doi
stringlengths 28
28
| title
stringlengths 19
311
| abstract
stringlengths 217
5.08k
| plain language summary
stringlengths 115
4.83k
| article
stringlengths 3.87k
161k
|
|---|---|---|---|---|
10.1371/journal.pgen.1007730 | The telomere bouquet is a hub where meiotic double-strand breaks, synapsis, and stable homolog juxtaposition are coordinated in the zebrafish, Danio rerio | Meiosis is a cellular program that generates haploid gametes for sexual reproduction. While chromosome events that contribute to reducing ploidy (homologous chromosome pairing, synapsis, and recombination) are well conserved, their execution varies across species and even between sexes of the same species. The telomere bouquet is a conserved feature of meiosis that was first described nearly a century ago, yet its role is still debated. Here we took advantage of the prominent telomere bouquet in zebrafish, Danio rerio, and super-resolution microscopy to show that axis morphogenesis, synapsis, and the formation of double-strand breaks (DSBs) all take place within the immediate vicinity of telomeres. We established a coherent timeline of events and tested the dependence of each event on the formation of Spo11-induced DSBs. First, we found that the axis protein Sycp3 loads adjacent to telomeres and extends inward, suggesting a specific feature common to all telomeres seeds the development of the axis. Second, we found that newly formed axes near telomeres engage in presynaptic co-alignment by a mechanism that depends on DSBs, even when stable juxtaposition of homologous chromosomes at interstitial regions is not yet evident. Third, we were surprised to discover that ~30% of telomeres in early prophase I engage in associations between two or more chromosome ends and these interactions decrease in later stages. Finally, while pairing and synapsis were disrupted in both spo11 males and females, their reproductive phenotypes were starkly different; spo11 mutant males failed to produce sperm while females produced offspring with severe developmental defects. Our results support zebrafish as an important vertebrate model for meiosis with implications for differences in fertility and genetically derived birth defects in males and females.
| Inherent to reproduction is the transmission of genetic information from one generation to the next. In sexually reproducing organisms, each parent contributes an equal amount of genetic information, packaged in chromosomes, to the offspring. Diploid organisms, like humans, have two copies of every chromosome, while their haploid gametes (e.g. eggs and sperm) have only one. This reduction in ploidy depends on the segregation of chromosomes during meiosis, resulting in gametes with one copy of each chromosome. Missegregation of the chromosomes in the parents leads to abnormal chromosome numbers in the offspring, which is usually lethal or has detrimental developmental effects. While it has been known for over a century that homologous chromosomes pair and recombine to facilitate proper segregation, how homologs find their partners has remained elusive. A structure that has been central to the discussion of homolog pairing is the bouquet, or the dynamic clustering of telomeres during early stages of meiosis. Here we use zebrafish to show that the telomere bouquet is the site where key events leading to homologous chromosome pairing are coordinated. Furthermore, we show that deletion of spo11, a gene required for proper recombination in most studied organisms, resulted in very different effects in males and females where males were sterile while females produced deformed progeny.
| Meiosis is a process that generates haploid gametes via one round of DNA replication and two rounds of chromosome segregation. Typically, homologous chromosomes (homologs) separate during meiosis I, and sister chromatids separate at meiosis II. Errors at either stage can lead to the production of aneuploid gametes, which is a major contributor to miscarriage and birth defects in humans [1]. During meiosis I prophase, homologs undergo pairing and crossing over, which is essential for their proper segregation in nearly every organism studied to date [2, 3]. Crossovers are created through a process termed recombination, where programmed DNA double-strand breaks (DSBs) are repaired using the homolog as a template, resulting in the exchange of homologous chromosome arms [2, 3]. Meiotic DSBs are formed by Spo11, a conserved topoisomerase-like enzyme [4]. They are then processed to reveal 3’ single stranded stretches of DNA that bind the strand exchange proteins, Dmc1 and Rad51, and engage in homology search [4–6]. In meiosis, the repair template is preferentially skewed towards the homologous chromosome rather than the sister chromatid, thus facilitating the pairing process [2]. While hundreds of DSBs might form in a single meiotic cell, only a subset go on to form crossovers, with the others being repaired via a noncrossover pathway [2, 3, 7].
DSB formation and crossing over occur in the context of the chromosome axis, a proteinaceous structure to which a linear array of chromosome loops is attached; both sister chromatids of a homolog are attached to a single axis and held together by cohesins [7, 8]. In many organisms, including mouse, budding yeast, and several plants, recombination initiation and repair intermediates are necessary for formation of the synaptonemal complex (SC), a tripartite proteinaceous structure that forms between the two homolog axes and holds them together along their lengths [2, 3, 7–9](S1 Fig). However, DSBs are dispensable for SC formation (synapsis) in some organisms such as C. elegans, Drosophila, and the planarian Schmidtea mediterranea [10–13]. In C. elegans and Drosophila, synapsis is initiated at “pairing centers” and ensues without the need for recombination [12]. In most organisms studied to date, recombination and the SC are used to link the homologous chromosomes. Notable exceptions are Drosophila males [14, 15] and Lepidopteran females [16] which do not form crossovers, and Tetrahymena [17] and fission yeast [18] which do not form the SC.
SC formation initiates primarily near telomeres in many organisms, including human males [19, 20], cattle males [21], the silkworm Bombyx mori [22, 23], the planarian Schmidtea mediterranea [13], and some plants such as tomato [24] and barley [25, 26]. In mouse males, while synapsis initiates interstitially as well as near the telomeres, there is a skew toward initiation at chromosome ends [27]. By contrast, synapsis in mouse and human females initiates primarily in interstitial regions [20, 28], while synapsis in female cattle initiates both near telomere ends and interstitially [21]. In many organisms, SC is nucleated preferentially at crossover fated sites [2]. Correspondingly, in mouse, human, and cattle, there is a skew toward crossovers in the distal regions of chromosomes in males but not in females [20, 29, 30].
During meiosis, telomeres are tethered to the nuclear envelope and their movement is directed by cellular cytoskeleton components [31–37]. One type of motion that is prominent in many species is the movement of chromosomes into and out of the bouquet, a conserved arrangement of chromosomes where telomeres are clustered together to one side of the nucleus. The bouquet has been hypothesized to restrict the chromosomes to one region of the nucleus thereby facilitating homolog recognition and pairing, possibly by limiting the homology search area or by active chromosome motion to disrupt weak non-specific interactions [2, 38]. However, in some organisms the bouquet does not exist (e.g. C. elegans and Drosophila) or does not form until after homologs are already co-aligned (budding yeast, Sordaria, mouse, and some plants) [2, 39], in which case it may play additional roles such as removing interlocks that form between two synapsed chromosomes [40]. Previous studies have shown that telomeres can form end-to-end associations in mammalian spermatids [41, 42] as well as in somatic and pachytene cells of some plants such as the dandelion-like smooth hawksbeard, Crepis capillaris [43, 44], pachytene cells of the cricket, Gryllus argentinus [45], and human spermatocytes [46]. The number and timing of these associations during meiotic prophase is poorly understood.
Our understanding of meiosis has been facilitated by the breadth of model organisms that have been studied, with each contributing new insight into the process. Budding and fission yeasts, C. elegans, Drosophila, mouse, and several plants have been instrumental to the study of the chromosome events of meiosis [9, 47]. While the basic features of meiosis are well conserved, the order of events and their functional dependencies vary significantly across species [2, 3, 48]. Indeed, the analysis of species-specific features has greatly informed our understanding of the seemingly fluid relationships between double-strand breaks (DSBs), the telomere bouquet, homolog pairing, synapsis, and recombination over evolutionary time scales [47]. There are remarkable similarities relating pairing, synapsis and recombination across phyla, even though difference in genome sizes can vary by orders of magnitude, especially when comparing mouse and humans to yeast, Drosophila, and C. elegans. Understanding how the chromosome events of meiosis are accommodated by larger genomes (and vice versa) necessitates the inclusion of additional model genetic organisms.
Zebrafish has many advantages as a vertebrate model for meiosis. Zebrafish can produce hundreds of offspring from a single cross, and external development allows for early detection of developmental abnormalities, including those caused by aneuploidy [49–51]. Unlike mammalian females, zebrafish females produce new oocytes throughout adulthood [52, 53], simplifying the characterization of female meiosis, which occurs in the fetal ovary in mammals. In addition, transparent gonads allow for observation of multiple stages of meiosis in a whole mount. Several cytological studies have provided insights into some aspects of zebrafish meiosis. For example, it has been shown in males that DSBs and the initial loading of the chromosome axis protein, Sycp3, and the SC transverse filament protein, Sycp1, are polarized to one side of the nucleus near the bouquet [6, 54–56]. Mlh1 foci, indicating sites of crossovers, have been shown to be distally skewed in males but not females [57]. These data are consistent with the genetic map where recombination is skewed to the telomeres in males yet more evenly distributed in females [58, 59]. A major gap of knowledge is that the order of events and the relationship between chromosome structure and recombination are not known. For example, it is not known if Spo11, a protein required for the formation of meiotic DSBs, is necessary for synapsis and homolog pairing in the zebrafish. This is an important relationship to determine as it sets the stage for further analyses of chromosome dynamics.
In this study we set out to establish the relationship between DSBs, synapsis initiation, and the establishment of close, stable homolog juxtaposition (which we refer to here as pairing) in the zebrafish males and females. We analyzed the progression of chromosome synapsis and pairing, telomere interactions, and double strand break localization at the super-resolution level. We created a knockout mutation in the spo11 gene and found that both pairing and synapsis in the zebrafish are Spo11-dependent. We found dramatic sex-specific outcomes from disrupting Spo11: although synapsis and pairing defects were similar between spo11 mutant males and females, males were completely sterile while females were able to produce offspring, though with severe developmental defects. Our results establish zebrafish as a tractable vertebrate model for understanding the chromosome events of meiosis I prophase from an evolutionary vantage and opening new lines of research with implications for human fertility and genetically derived birth defects.
To better understand the relationship between the bouquet, Sycp3 loading, and synapsis in zebrafish, we set out to find when and where these events occur relative to one another in the prophase I nucleus. We stained spermatocyte nuclear spread preparations using a fluorescently tagged PNA probe by in situ hybridization to mark repeated telomere sequences, an Sycp3 antibody to mark the chromosome axis, and an Sycp1 antibody to mark the SC. The images were collected using structured illumination microscopy (Fig 1, S1 Fig).
A general overview emerged. The telomere bouquet was prominent at the leptotene and zygotene stages (Fig 1, panels 1D-5D). Early Sycp3 loading occurred adjacent to the telomere probe and extended toward the middle of the chromosomes as meiotic progression advanced (see magnified regions in Fig 1, rows E, F, and G). When the average Sycp3 length reached about 1 μm from the end of the chromosome, Sycp1 lines appeared close to the telomeres and then extended inward; as Sycp3 lines extended, Sycp1 lines closely followed, yet lagging somewhat behind (Fig 2A). Interestingly, some chromosomes appeared to have more than one synapsis initiation site, albeit they were still in close proximity to the telomere (Fig 1, panels 3F and 4F).
To facilitate comparisons between nuclei, we established staging criteria based on the total length of Sycp1 in 30 pre-pachytene images (Fig 2B). We divided the nuclei into five classes: leptotene (L; Sycp1 = 0 μm), leptotene to early zygotene transition (L/EZ; Sycp1 = 1–10 μm, with few short stretches of SC near the telomeres), early to mid-zygotene (EZ/MZ; Sycp1 = 10–50 μm), mid- to late zygotene (MZ/LZ; Sycp1 = 50–100 μm), and late zygotene (LZ; Sycp1 > 100 μm). Spreads that contained more than three fully synapsed chromosomes were not measured as they were considered to be transitioning to pachytene and were referred to as “pre-pachytene”.
In wild type zebrafish, we observed frequent end-to-end associations between telomeres, either as doublets or in higher order structures at all stages of prophase I (Fig 1, panels 1E and 1G-6G, Fig 2C). We assessed the extent of these associations by counting the total number of engaged telomere ends, defined by an Sycp3 line and its telomere associating with another end. One association could involve two or more engaged telomere ends. For a detailed description of our criteria see the methods. We observed the highest numbers of associations at the L to L/EZ stages, with an average of ~ 31% of telomeres engaged in associations with each other (Fig 2C). The associations subsequently decreased to an average of ~ 6% but were still detected throughout the later stages including pachytene. In some cases, a short stretch of Sycp3 could be seen bridging the telomeres of unrelated chromosomes (Fig 1, panels 5G and 6G). The nature of these bridges is not known, however, Sycp3 protein can both bind dsDNA and form self-assemblies which are consistent with what we see [60, 61]. Associations involving more than two chromosomes indicated that at least a subset of telomere associations occurred between nonhomologous chromosomes or at the opposite end of the same chromosome. This was especially evident at the zygotene stage where telomeres of synapsing chromosomes were seen forming associations with a non-partner telomere, and at the pre-pachytene and pachytene stages where telomeres of synapsed chromosomes associated with nonhomologous chromosomes (Fig 1, panel 6G). Our data suggest that telomere associations are a normal part of zebrafish meiosis I and not a pathological occurrence such as fusions caused by nonhomologous end-joining.
In nearly every organism studied to date, axes of homologous chromosomes undergo some degree of co-alignment, or pairing, prior to synapsis, in which the distance between co-aligned axes is typically about 0.4 μm [2, 62]. Analysis of co-alignments seen in several species suggest the chromosomes are held in close proximity by DNA intermediates of the homologous recombination pathway [2]. We inspected images of spermatocyte nuclear spreads for evidence of axis co-alignment in the absence of detectable SC, marked by Sycp1. A detailed description of co-alignment assessment is provided in the methods. In brief, chromosome regions were considered co-aligned when Sycp3 lines were closely juxtaposed with the narrowest region between them at a distance of < 0.5 μm (Fig 1, Panels 1F and 2F, Fig 2D). We found that presynaptic co-alignment of chromosomes occurred near the telomeres and adopted two main types of configurations: funnel and pinch. In the funnel configuration, co-alignment occurred directly adjacent to the telomeres to form the stem of the funnel while the other ends of the Sycp3 lines diverged away from the stem. The diverging lines could either be long or short (Fig 1, panels 1F and 2F; Fig 2D), cross each other, or even fold backward toward the stem. In the pinch configuration, the narrowest region between the Sycp3 lines did not occur directly adjacent to the telomere, but a short distance away (Fig 2D).
We found a total of 23 funnel configurations and 10 pinch configurations among 24 wild-type spermatocyte cells from the L to the MZ/LZ stage. During leptotene, co-alignments were rare, indicating that chromosomes were not stably juxtaposed with their partners at this stage (Fig 2E). We found the highest number of co-alignments during the L/EZ and EZ/MZ stages, when the chromosomes begin to actively engage with each other to initiate synapsis. By the MZ/LZ and LZ stages, we found almost no co-alignments since most or all of the chromosomes had already engaged in telomere proximal synapsis (Fig 2E). As individual nuclei with presynaptic co-alignment usually have less than a total of five co-alignments (Fig 2F), we believe that this is a transient stage mediated by homologous recombination that quickly progresses to synapsis initiation for any given chromosome pair. The funnel and pinch configurations likely represent recombination events that initiate very close to the telomere or slightly inward. It is likely that the pinch and funnel axis shapes are precursors to synapsis initiation since we saw similar configurations with short stretches of SC (Fig 1, panels 2E and 2F; Fig 2D). Moreover, unpaired telomeres in the pinch configurations suggest that stable telomere associations are not the primary driver of homolog pairing and that initial homology recognition can occur in sub-telomeric regions.
While we observed pre-synaptic co-alignment near the telomeres, we found no evidence for similar co-alignment at interstitial regions of the axes. For chromosomes where synapsis had initiated, the distances between the diverging edges of the two Sycp3 lines were often much greater than 0.4 μm and were frequently bent in non-parallel orientations with respect to each other (Fig 1, panels 2E, 3E, 3F, 5F, 3G, 4G). This suggested that axes were not stably co-aligned at interstitial regions prior to synapsis. To determine when interstitial sites become stably co-aligned relative to SC formation, we performed fluorescence in situ hybridization (FISH) on cells at different stages of meiotic prophase I using a 68 kilobase bacterial artificial chromosome (BAC) probe located ~10 Mb from the end of chromosome V (total 72.5 Mb) (Fig 3, S2 Fig). We found that the BAC signals were far apart in the early prophase I stages but paired up as synapsis progressed (EZ/MZ to LZ). In a few instances, we found cells where the BAC probe localized to forked regions of Sycp3 just ahead of synapsis (Fig 3C). These data suggest that stable juxtaposition between homologs does not occur until they are synapsed at that region.
Prior to pairing, the BAC signal presented as an amorphous shape, while at synapsed regions the shape was elongated and appeared to lie perpendicular to the axis (Fig 3, S2 Fig, LZ and Pachytene panels). The two prominent features of a meiotic chromosome are the axes that run along the chromosome length and the DNA loops that attach to the axis. It is possible that the change in shape of the BAC signal reflects a change in chromosome architecture, for example, from a fractal globule, characteristic of interphase chromosomes [63], to a looped region that characterizes the DNA component of meiotic chromosomes [62].
Several lines of evidence point to differences between female and male meiosis in zebrafish. Females have longer chromosome axes and exhibit a more even distribution and higher numbers of crossovers [57–59, 64]. In order to determine if early prophase events in females differed from males, we stained ovary nuclear surface spreads with the telomere probe and antibodies against Sycp3 and Sycp1. We staged nuclei based on the overall appearance of axis extension and synapsis progression. We found that the progression of prophase I in females was similar to that in males: Sycp3 loading and synapsis initiated near both telomere ends in the bouquet and elongated toward the center of the chromosome, with synapsis lagging behind Sycp3; chromosomes appeared to become stably juxtaposed as synapsis progressed (Fig 4).
Chromosome interlocks are common during the zygotene stage in several organisms including the silkworm, Bombyx mori, and maize, Zea mays, but most interlocks are resolved by pachytene [22, 62, 65, 66]. In the zebrafish spermatocytes, we regularly found chromosomes that were intertwined or sometimes trapped between another set of homologs at the pre-pachytene stages when most of the chromosomes were already synapsed (i.e. 16–24 fully synapsed chromosomes; Fig 5). These configurations closely approximated interlock structures seen at late zygotene in other species [2]. Interlocks occur when one chromosome or a pair of chromosomes becomes entrapped between the space of two synapsing homologs. Thus, from a first approximation, the structures we see are likely interlocks. Interestingly, nuclei at this stage frequently also had individual pairs of chromosomes with extensive or complete de-synapsis, sometimes with another homolog appearing to be entrapped in the desynapsed region (Fig 5, Panels 1A-1C, 2A-3D, 5A-6D). Of 8 cells at this pre-pachytene stage, only 2 showed no anomalies. Of the remaining cells, 3 had both interlocks and de-synapsis, 1 had just an interlock, and 2 had just de-synapsis. The de-synapsis is unlikely to be due to the cells transitioning to diplotene, as some de-synapsed chromosomes were completely separated with no evident crossover connections, and some were entangled around other chromosomes. Although interlocks are not a common feature of pachytene cells, it is also possible that there is a subset of cells where interlocks persist through pachytene and the chromosome-wide de-synapsis we see are chromosomes in diplotene. Interestingly though, we never saw a spread nucleus showing a classic diplotene state as seen in other organisms where a complete set of de-synapsed bivalents were held together by one or more chiasmata, suggesting this state in males is transient or full-length Sycp3 axes start to degrade at this stage.
Since the sites of crossovers in zebrafish are skewed toward the ends of chromosomes in males [57], we suspected that co-alignment and synapsis near telomeres might be initiated by local DSBs. We first tested if γH2AX, a biomarker for DSBs, co-localizes with telomeres in sectioned testes and found a sharp polarization of γH2AX staining to one side of the nucleus when chromosomes were in the bouquet and then a more dispersed signal in cells where chromosomes had exited the bouquet (Fig 6A). These results are consistent with a study that showed γH2AX signal clustered with initiation of axis formation marked by Sycp3 [55].
To evaluate the distribution of DSBs at the super-resolution level, we probed spermatocyte nuclear surface spreads with telomere probes and antibodies to Sycp1 and the DSB repair protein Rad51. Previous work at lower resolution showed that Rad51 foci were primarily found near sites where Sycp3 loading had initiated [6]. Consistent with this finding, and with the γH2AX localization, we found that Rad51 foci were interspersed with the telomere foci in the bouquet cluster (Fig 6B). If DSBs are required for initiating synapsis, then we expected to find that most SC stretches would be associated with a Rad51 focus. This was not the case, however, since there were many instances of SC stretches with no associated Rad51 foci (Fig 6B, panels 1C-3C). We found that 39% (n = 269) of synapsed ends in early zygotene had no associated Rad51 focus. This was surprising given that we also found that synapsis requires Spo11-dependent DSBs (below). Three possible reasons could account for this observation: 1) some synapsis may occur independent of DSBs, 2) synapsis is initiated at Rad51-associated DSBs but Rad51 signal has been lost due to repair prior to imaging, or 3) some synapsis is initiated at Dmc1-associated DSBs that do not co-localize with the Rad51-associated DSBs. The latter is supported by data from Arabidopsis thaliana where Dmc1 and Rad51 DSBs do not colocalize [67].
We found that some cells at pachytene had Rad51 foci and they were located both near the telomeres and in interstitial regions. There are two ways we can envision the interstitial foci could arise: 1) All meiotic DSBs form at the same time when the cells are in the bouquet stage, in which case DSBs at interstitial locations would be recruited to the bouquet, or 2) breaks continue to form throughout prophase I. Further studies are required to distinguish between these models. Combined, our results show that DSBs are primarily clustered near the telomeres but are also found at interstitial regions during pachytene, which reflects the crossover pattern in zebrafish males [57, 58].
In order to determine whether synapsis can occur in the absence of Spo11, which is required for the formation of meiotic DSBs, we created a spo11-/- mutant. The spo11 gene in zebrafish consists of 13 exons encoding a 383-amino acid protein product (GenBank: AAI65825.1) with the predicted TP6A_N superfamily domain at 96–157 aa and the predicted TOPRIM superfamily domain at 205–367 aa (NCBI BLASTP 2.8.0). We used TALENs targeted to the second exon to introduce an indel mutation by error prone repair. Sequencing of genomic DNA isolated from offspring of founder backcrosses identified an 11 bp deletion resulting in a frameshift mutation in the coding region that predicts a truncated protein of 57 aa lacking both the TP6A_N and TOPRIM domains (Fig 7A). To confirm disruption of Spo11 function in the mutant, we probed whole mount testes of spo11-/- males with antibodies to γH2AX and the germ-cell specific Vasa protein and found that γH2AX clusters were absent in the germ cells, showing that the mutant is deficient for DSB formation (Fig 7B).
We next examined evidence of synapsis and pairing in nuclear surface spreads from spo11 mutants. Although Sycp3 loading initiated near the telomeres and elongated inward as in wild type, Sycp1 loading did not follow (Fig 8A). We divided spo11 mutant spermatocytes into the L—L/EZ-like, EZ-LZ-like, and Post-LZ-like categories based on the overall resemblance of Sycp3 loading in the nucleus to equivalent wild-type stages. Post-LZ included pre-pachytene-like or pachytene-like stages. In spo11 mutant spermatocytes, 30 out of 40 cells had no synapsis and the remaining cells had between 1 and 4 short fragments of Sycp1, which appeared either between two axes, on one axis, or as a lone filament (Fig 8A, panels 3C-4C). These Sycp1 stretches may have been due to self-assembly of Sycp1 filaments [68]. In addition, the bouquet was also maintained. Telomere-proximal co-alignment between chromosomes, however, was disrupted (Fig 2F).
Intriguingly, we found an average of ~ 42% of telomeres engaged in associations in the L–L/EZ-like mutant cells as compared to the ~ 31% we see in equivalent stages of wild type (Fig 8A, panels 1C-5C, Fig 8B; p = 0.0389). Unlike in wild type, the telomere associations were maintained at high levels in the mutant throughout prophase I. Several possibilities could account for the loss of telomere associations in wild-type cells. Pairing and synapsis between the ends of homologous chromosomes could physically displace weak associations between non-related chromosome ends. Alternatively, a regulatory feature associated with the transition from leptotene to zygotene could signal loss of a subset of associations, or the reduction of entanglements later in meiosis could allow associations to be disrupted by the physical force of spreading. Any one of these possibilities could account for the persistence of associations in the spo11 mutant.
Wild-type cells that were in the EZ/MZ to LZ stages gave a distribution of inter-BAC distances that were overall shorter than those in the L or L/EZ stage (0.24–3.13 μm vs. 9.3–25.3 μm, p = 0.0004, Fig 8C). In the L to L/EZ stages the average length of Sycp3 lines was less than 2 μm while in the later stages, the average length of Sycp3 lines was greater than 2 μm. The sharp decrease in inter-BAC probe distance measurements suggests that pairing at the probed locus occurs shortly after synapsis is initiated. For the spo11 mutant, we staged the nuclei based on Sycp3 axis length since the SC was absent. BAC foci in the spo11 mutants remained at approximately the same distance from each other when the axes were short (< 2 μm) or long (> 2 μm) (Fig 8C, S3 Fig). In mutant females, synapsis and pairing were also disrupted as was seen in males (Fig 9, S3 Fig). Together, these data indicate that Spo11 is required for the initiation and/or stabilization of synapsis and homolog juxtaposition in males and females.
We found that spo11 mutant males could induce spawning in females but failed to fertilize eggs (S4 Fig), indicating that they were either unable to produce or release their sperm. We inspected spo11 mutant testes using light microscopy and found they appeared more translucent compared to wild type (Fig 10A, panels 1A and 2A), a phenotype that suggested a defect in sperm production. To confirm this, we isolated and stained whole testes with an antibody to the Vasa protein (Fig 10A, panels 3A-4C). In wild-type zebrafish, Vasa is highly expressed in early germ cell clusters but diminishes as the spermatocytes progress in maturity, and is absent in mature spermatozoa clusters which can be identified by their tightly compacted nuclei [69]. We found that the spo11 mutant males lacked sperm, and correspondingly, Vasa was expressed in all cell clusters, though it did diminish compared to the early germ cells.
Surprisingly, spo11 mutant females produced similar numbers of fertile eggs as wild type, however, the vast majority of their embryos died before 5 days post fertilization (dpf) and displayed a spectrum of abnormalities (Fig 10B, 10C and 10D). We expect that the severe developmental defects displayed among the progeny of spo11 mutant females were a result of aneuploidy since it is unlikely that the chromosomes would be able to segregate properly with gross synapsis and pairing defects. The offspring that were normal at 5 dpf continued to grow into adults that developed as males. A similar offspring phenotype was seen in mlh1 mutants in the zebrafish, where the offspring were shown to be aneuploid, and the ones surviving to adulthood developed as males that were found to be triploid [49]. Together, our data show that despite similar synapsis and pairing defects, males and females display dramatically different reproductive outcomes. This suggests a difference in checkpoint response between the sexes in the zebrafish.
Super-resolution analysis of homologous chromosome synapsis and pairing in the zebrafish revealed a coherent timeline of events (Fig 11). 1) Assembly of the chromosome axis protein, Sycp3, initiates almost exclusively at both ends of chromosomes and elongates inward. 2) DSBs cluster near the telomere region. 3) Co-alignments form between telomere-proximal chromosome axes in funnel or pinch configurations. 4) The synaptonemal complex protein, Sycp1, loads between peri-telomeric axes and elongates slightly behind Sycp3 assembly. 5) Stable homolog juxtaposition at interstitial loci is not evident until the synaptonemal complex spreads across the region. As meiosis progresses, interlocks between chromosomes can be observed. Throughout the leptotene to pachytene stages telomere associations are present.
One of the most striking findings of our analysis was that the key events of meiotic chromosome metabolism, including axis morphogenesis, DSB formation, stable homolog juxtaposition, and synapsis all occurred within the limited region of the nucleus defined by the bouquet. Moreover, the focus of these events was specifically limited to the ends of chromosomes. From a first approximation, the general lack of close, stable homolog juxtaposition at interstitial sites suggests that the two ends of the same chromosome are not distinguished as such. Since zebrafish have 25 pairs of chromosomes, any given end is thus challenged to find its homologous partner among 99 possible choices within the bouquet prior to zygotene. Processes that promote the efficiency of pairing could include time intervals that favor collisions by diffusion [70, 71], rapid prophase movement to increase the rate of collisions via attachment of telomeres to cytoskeletal motor proteins outside the nucleus [36, 40, 72, 73] and/or through one or more DSB-independent pairing interactions [2, 27].
For organisms that initiate synapsis at sites of DSBs, homolog juxtaposition along the lengths of chromosomes can often be detected prior to synapsis (e.g. Sordaria)[2]. However, in these organisms a dramatic polarization of DSBs toward the telomere region, like that seen in zebrafish, is not evident. By contrast, DSBs and synapsis initiation in human males and the planarian Schmidtea mediterranea show a polarization similar to zebrafish [13, 74]. In the planarian, synapsis was shown to drive homolog pairing. It is possible that zebrafish homologs are paired by synapsis as well. A notable difference between the planarian and zebrafish, however, is that the planarian does not require Spo11 for synapsis whereas the zebrafish does, suggesting distinct SC nucleation methods between the two species. Our study does not answer the question of whether interstitial regions are physically juxtaposed by “zippering up” as SC spreads, or if a wave of DSBs creates new synapsis initiation sites and/or stabilizes SC.
Interestingly, in zebrafish, the bouquet is also the organizing center of the Balbiani body (Bb), a collection of embryonic patterning factors, mitochondria, and organelles which defines the animal-vegetal axis of the oocyte and is found in a wide variety of organisms including Drosophila, Xenopus, and mouse [75, 76]. In zebrafish, disruption of the bouquet ex vivo by the addition of the microtubule inhibitor nocodazole also disrupts Bb precursors showing the two structures are mechanistically linked [77]. It will be interesting to test if other meiotic chromosome features are also linked to the Bb, or if the Bb contributes to meiotic progression.
Our work uncovered several features of zebrafish biology that can stimulate new lines of enquiry to understand meiotic chromosome dynamics. First, nonhomologous telomere associations were prominent throughout meiosis, yet the nature of these associations is not well understood. One possibility is that they represent associations between heterochromatic regions like those seen in crickets [45], or telomere-bound protein interactions, as has been proposed for Trf1 [46]. We also do not know if they represent interactions between the same chromosomes from cell to cell. Associations could represent an early phase of the pairing process where the bouquet facilitates interactions between all telomeres, and rapid chromosome movements act to disrupt weak nonhomologous interactions to favor stronger DSB-dependent homologous interactions [32, 33, 36]. In addition, it is unknown what structure at or near the telomere seeds the initial loading of Sycp3, or the significance of the Sycp3 “bridges” sometimes seen between nonhomologous telomeres.
Second, our results show that Spo11 is required for the co-alignment of axes and SC formation. We attribute this effect to the formation of DSBs by Spo11 since the earliest occurrences of Rad51 and γH2AX signals are skewed toward telomeres where co-alignment and SC first appear. Observing Rad51 foci and γH2AX staining first near telomeres and later at interstitial locations suggests that DSBs may form in a wave, where initial breaks near telomeres bring homologs together to initiate local synapsis, while subsequent breaks form as Sycp3 is progressively loaded to initiate and/or to stabilize SC elongation. Consistent with the latter model, we occasionally see more than one synapsis initiation site between two chromosomes, albeit close to the telomere. A previous study showed that RPA foci, known to mark intermediates of DNA replication and DSB repair, form lines along the elongating axis in the zebrafish [56]. It is not known, however, if these are a result of DNA replication or Spo11-dependent recombination intermediates. Another possibility is that one or a few DSBs near the telomere are sufficient to promote synapsis along the length of a chromosome. The kinetic relationship between Sycp3 loading and Spo11-dependent SC initiation and elongation points to a possible regulatory mechanism that couples these processes. A study in the medaka fish, Oryzias latipes, has shown that loading of Sycp3 and Sycp1 is polarized to one side of the leptotene nucleus, with Sycp3 lines appearing to slightly anticipate Sycp1 [78]. This suggests that the mechanism of Sycp3 loading and synapsis in zebrafish may be common to other fishes.
Third, a transient interlock stage suggests a robust resolution mechanism. Interestingly, in cells where interlocks are observed, we also see pairs of chromosomes separated by long stretches of de-synapsed regions, sometimes disjoining two chromosomes completely. One possibility is that zebrafish employ long-range chromosome de-synapsis to resolve chromosome interlocks, as has previously been suggested in other organisms [62, 66]. It is possible that late-forming interstitial DSBs may play a role in re-establishing homologous synapsis at pachytene following this method of interlock resolution. In budding yeast, mouse, and C. elegans, SC components are involved in downregulation of DSB formation [7], thus it seems possible that local de-synapsis could activate new DSB formation.
Fourth, our analysis shows that synapsis initiates near the telomeres and progresses inward in both males and females, despite the differences in Mlh1 distribution and the recombination landscape between the two sexes [57, 58]. In mammals, the SC nucleation and crossover distribution landscapes correlate and are sex-specific (mouse, human, cattle) [20, 27–30]. In zebrafish, it is possible that the relationship between SC nucleation and crossover designation differs between the sexes, given that in males the SC nucleation pattern resembles the crossover pattern whereas in the females it does not appear to.
Zebrafish is not only an excellent model to study the events of meiosis per se, but also to study sexually dimorphic responses to meiotic perturbations. Zebrafish has previously been proposed as a model for germ cell aneuploidy [50, 79]. We show here that despite exhibiting similar defects in synapsis and pairing, spo11 mutant males and females show vastly different outcomes in reproduction. The males are unable to produce sperm, while females produce eggs that result in severely deformed offspring. This is in line with previous studies that show sexually dimorphic outcomes: Disruptions of Mlh1 [49, 51], and Mps1, a kinase required for the spindle assembly checkpoint [50], show a tendency for females to produce aneuploid offspring. Unlike in spo11 mutants, where males are sterile, and mlh1 mutants, where males are predominantly sterile, both male and female mps1 mutants produce aneuploid offspring, although the rate is higher in females than males (~46% vs. ~26% respectively). This suggests complex mechanisms underlying causes of increased aneuploidy in zebrafish females.
Sex specific differences are also seen in spo11 mutant mice; spermatocytes die by early pachytene whereas oocytes survive until diplotene/dictyate stage [80, 81]. The arrest seen in mouse males, however, is likely different than the arrest seen in zebrafish. Among organisms with heterogametic sex determination, mechanisms have evolved to specifically accommodate unpaired chromosomes in the heterogametic sex, including meiotic sex chromosome inactivation (MSCI) [82, 83]. As such, mutations that disrupt pairing might be expected to have a weaker effect in the homogametic sex, where the MSCI checkpoint may not be as robust [84]. In domesticated zebrafish, sex determination is polygenic, with no universal structural differences between chromosome sets of sexes in lab strains [58, 85]. Consistent with these findings, we did not observe any chromosomal regions that remained unpaired during pachytene. Thus, the pronounced effect of the spo11 mutation in males is likely not due to the activation of the MSCI checkpoint. Instead, the failure to produce sperm may depend on another checkpoint, such as the synapsis or the spindle assembly checkpoints, that operate in other model systems [1, 86–89].
Our findings highlight the importance of studying multiple model systems. While homolog pairing and recombination are considered universal features of meiosis, the means to getting there is quite varied among species. Interestingly, some meiotic prophase events in zebrafish resemble the corresponding events in human spermatogenesis, including the tendency of DSBs to skew near the ends of chromosomes and the initiation of synapsis at telomeres followed by inward synaptic progression [74, 90]. Telomere-proximal synapsis initiation while Sycp3 loading is not yet complete has also been reported in human spermatocytes [19, 20]. However, in humans the Sycp3 loading appears more extensive than in the fish by the time that synapsis ensues. Understanding spermatogenesis is important since sperm concentration and total sperm count has declined 50–60% between 1973 and 2011 among men in western countries [91], and the causes behind male infertility remain unknown in about 40% of patients [92]. In addition, human females are more prone to generating aneuploidy as compared to males [1, 93–95], which resembles the situation in zebrafish. While the causes of aneuploidy and reduced fertility in humans are complex and the contributions are manifold, zebrafish could provide valuable insights into environmental, genetic, and sex-specific effects on adverse meiotic outcomes.
The UC Davis Institutional Animal Care and Use Committee (IACUC) has approved of this work under the protocol #20199; For noninvasive procedures (e.g. fin clips for genotyping), zebrafish were anesthetized using tricaine. Invasive surgical methods were performed on fish euthanized by submerging fish in ice water.
The wild type AB strain was used in the production of spo11 mutants. Wild type data presented in Figs 1–5 and 10 are from tank mates of spo11 mutants. AB strain fish were used for Fig 6A, and NHGRI strain fish were used for Fig 6B. NHGRI strain fish were used for test crosses in one of the two pooled data sets in Fig 10. Other test crosses were done with AB strain fish. Fish were maintained as previously described [96].
Spo11 mutants were generated using TALENs targeting the second exon of spo11. The TALENs were assembled and injected as previously described [97]. The TALEN sequences were: HD-NG-NI-NI-NI-NN-NN-NG-NN-NI-NI-NN-HD-NI-HD-half repeat HD, and NG-HD-HD-NI-NN-HD-NI-NN-NN-NI-NG-HD-NG-NI-NG-half repeat NG. Injected founder fish were raised to adulthood and outcrossed to wild type fish; the resulting offspring were screened for mutations in spo11 via high resolution melt (HRM) analysis and subsequent sequencing. HRM primer sequences are: fwd TCACAGCCAGGATGTTTTGA, and rev CACCTGACATTGTTCCAGCA. The HRM analysis was performed with either Light Scanner Master Mix (BioFire Defence, Murray, UT, Catalog# HRLS-ASY-0003), 10X LCGreen Plus+ Melting Dye (Biofire Defence, Catalog# BCHM-ASY-0005), or 20X Eva Green dye (VWR, Radnor, PA, Catalog# 89138–982) using a CFX-96 real-time PCR machine and Precision Melt Analysis software (BioRad, Hercules, CA). The data presented in this paper is from individuals of a population with an 11 bp deletion mutation in exon 2 that has been outcrossed 2–3 times.
All our conclusions are based on experiments that were performed at least two times. All data sets comparing WT and spo11 mutants were collected from tank mates processed in parallel on the same days, including the spreads and staining the slides. The antibodies, the BAC probe and the Telomere PNA probes were tested multiple times on spreads and/or whole-mount gonads prepared on different days. The Student t-test was used for statistical analysis. All numerical data used for each plot is tabulated in S1 Table. Raw SIM data for all cells are available upon request. Images shown in each figure will be deposited at the Dryad Digital Repository (https://datadryad.org/).
About 15–20 gonads were freshly dissected in 1X Phosphate Buffered Saline (PBS). The gonads were placed in 2 ml Dulbecco’s Modified Eagle Medium (DMEM) in a 5 ml Eppendorf tube on ice. 4 mg of collagenase (Sigma-Aldrich Chemical Co Inc, St. Louis, MO, Catalog# C0130-500MG) dissolved in 200 μl DMEM were added and the gonads were gently shaken horizontally at 32°C for 50 minutes to an hour, until the liquid was cloudy, and the gonads were in small chunks. The tube was inverted rapidly several times every 10 minutes to facilitate dissociation. The collagenase was then washed out: DMEM was added up to 5 ml and the gonads were pelleted at 218g for 3 minutes. Then 3 ml of the supernatant were removed to reduce the liquid down to 2 ml. This was repeated 2 additional times for a total of 3 DMEM washes with the supernatant reduced to 1 ml after the last wash (the pellet was not resuspended between the washes).
DMEM was added up to 2 ml total, and 1.4 mg trypsin (Worthington Biochemical Corporation, Lakewood, NJ, Catalog# LS003708) dissolved in 200 μl DMEM and 20 μl of 400 μg/ml DNaseI (Roche Diagnostics, Pleasanton, CA, Catalog# 10104159001) were added for cell dissociation. The tube was gently shaken horizontally at 32°C for 5–15 minutes until the solution contained few clumps. The tube was inverted rapidly several times every 5 minutes to facilitate dissociation. 10 mg of trypsin inhibitor powder (VWR, Catalog# IC100612.5) dissolved in 500 μl DMEM and 50 μl of 400 μg/ml DNase I solution were then added. The tube was briefly spun down, and the cell suspension was pipetted repeatedly up and down with Pasteur pipettes for 2 minutes to facilitate dissociation of any remaining clumps.
The cell suspension was put through a 100 μm nylon Falcon filter (Fisher Scientific, Waltham, MA, Catalog# 08-771-19) and transferred to a fresh 5 ml tube. DMEM was added to 5 ml total volume and the cells were pelleted at 218g for 5 minutes. The supernatant was removed and 5 μl of the DNase I solution was added directly to the pellet which was then resuspended by scraping the bottom of the tube on an empty tube rack. DMEM was added up to 5 ml and the cells were pelleted at 218g for 2 minutes. The DNase I treatment was repeated a total of 2–4 times until the resuspended pellet did not clump upon addition of DMEM. After the last treatment, the pellet was resuspended in 1–2 ml of 1X PBS and pelleted again at 218g for 5 minutes. The supernatant was removed and the pellet resuspended with a pipette tip (~3 mm cut off from tip to widen the aperture) in 80–100 μl of 37°C 0.1M pH ~8 sucrose, and allowed to sit at room temperature for 3–5 minutes.
Slides (Fisher Scientific Premium Superfrost, Catalog# 12-544-7) were coated with 100 μl of 1% Paraformaldehyde (PFA; Acros Organics, Catalog# 30525-89-4) with 0.15% Triton X-100 (Fisher BioReagents, Catalog# 9002-93-1) and then ~20 μl of cell suspension was added directly to the center of the slide in a straight line. The slide was tilted to facilitate spreading. The slides were placed in a slightly cracked open flat humid chamber. The chamber was placed in a dark drawer and allowed to sit overnight. It was then opened, and the slides allowed to completely dry. The slides were rinsed for 5 minutes in H2O and then twice for 5 minutes in 1:250 Photo-Flo 200 (Electron Microscopy Sciences, Catalog# 74257) in Coplin jars. The slides were dried and stored at -20°C until they were stained.
About 6–10 gonads of females aged 60–80 dpf were dissected in 1X PBS. The gonads were placed in 2 ml DMEM in a 5-ml tube and passed through an 18-gauge needle and then a 20-gauge needle 15 times each. The cells were briefly spun down and then pipetted up and down with Pasteur pipettes for 2 minutes. The cell suspension was put through a 100 μm nylon Falcon filter and transferred to a clean 5 ml tube. The cells were then pelleted at 218g for 5 min. The pellet was composed of two layers, a bottom whitish layer and a top yellowish layer. The top layer was carefully removed with pipette and the remaining bottom layer was resuspended in 2 ml 1X PBS. The cells were then pelleted at 218g for 5 min, the supernatant was removed, and the pellet was resuspended with cut pipette tip in 80–100 μl 37°C 0.1M pH ~8 sucrose. The suspension was allowed to sit at room temperature for 3–5 minutes. The slides were prepared as in the “Adult testes chromosome spreads” protocol.
PNA telomere probes TelC-Alexa647 and TelC-Cy3 were acquired from PNA Bio Inc, Thousand Oaks, CA (Catalog# F1013 and F1002 respectively); 50 μM stocks were prepared in formamide as per manufacturer’s instructions and stored at -80°C. The hybridization solution was prepared to a final concentration of 0.2 μM PNA telomere probe and 1.33 mg/ml bovine serum albumin (Fisher Scientific) in pre-hybridization solution. The pre-hybridization solution was composed of 50% formamide (Fisher Scientific, Catalog# BP228-100), 5X Saline-Sodium Citrate (SSC; 20X stock: 3M NaCl and 0.3M Sodium Citrate), 50 μg/ml Heparin sodium salt from porcine intestinal mucosa (Sigma-Aldrich Chemical Co Inc, Catalog# H3393-100KU), 500 μg/ml transfer RNA from wheat germ (Sigma-Aldrich Chemical Co Inc, Catalog# R7876-2.5KU), 0.1% Tween 20 (Bio-rad, Catalog# 170–6531), and 1M Citric acid to bring the solution to pH ~6. The pre-hybridization and hybridization solutions were stored in -20°C in the dark.
BAC clone CH211-31P3 (https://zfin.org/ZDB-BAC-050218-850) was obtained from the BACPAC Resources Center (BPRC). The BAC was purified via Midiprep as previously described [98]. Purified BAC quality was assessed by running a sample on a 1% agarose gel, and the BAC’s identity was confirmed by PCR amplification of a segment of the nanos2 gene using the following primers: fwd ATGCAGTCCGAGAGTCAGCAGAG, and rev ATAACGGACACACGTAGCTCCTCAG.
The Cot-1 preparation was adapted from [98]. Salmon testes DNA (Sigma-Aldrich, Catalog# D1626-1G) was prepared at 10 mg/ml in H2O by dissolving overnight at 55°C. 300 μl aliquots of the testes DNA were sonicated in Diagenode tubes (Fisher Scientific, Catalog# NC0065146) in a Diagenode Bioruptor UCD-300 for ~60 15-second cycles or until the average fragment size was ~400–500 bp. The fragment size was checked on a 1% agarose gel. 500 μl of sonicated salmon testes DNA was denatured at ~100°C for 15 minutes and then incubated at 65°C for 4 minutes. 250 μl of 1M NaCl (pre-heated to 65°C) was added and the mix was incubated at 65°C for the duration of time needed for the Cot-1 fraction to re-anneal (equation: 5.92/DNA concentration in mg/ml = time (in minutes)). Then 1 unit of S1 nuclease (Thermo Fisher Scientific, Catalog# EN0321) per 1 μg of DNA was added together with 5X S1 nuclease reaction buffer. The mixture was incubated at 37°C for 30 minutes.
The Cot-1 solution was transferred to a 15 ml conical tube, mixed with 10 ml of pH 8 Phenol:Chloroform:Isoamyl alcohol 25:24:1 (Fisher Scientific, Catalog# BP1752-100) and centrifuged at 1500g for 5 minutes. The aqueous phase was transferred to a new 15 ml tube and mixed thoroughly with 0.1X volume of 3M sodium acetate. 1X volume of 100% isopropanol was added, the solution mixed gently to precipitate the DNA, and then centrifuged at 3000g for 10 minutes at 4°C. The supernatant was removed, and the pellet was allowed to air dry with the tube inverted at an angle. The pellet was re-hydrated in 30 μl H2O and the concentration was determined by nanodrop. The pellet was further cleaned with 1 ml Phenol:Chloroform:Isoamyl alcohol 25:24:1 followed by 70% EtOH. The final pellet was dried and resuspended in 30 μl H2O and the concentration was determined by nanodrop.
The BAC probe labeling and preparation was adapted from [98]. The probe was labelled with Green dUTP (Abbott Molecular, Abbott Park, IL, Catalog# 02N32-050) using the Nick Translation Kit (Abbott Molecular, Catalog# 07J00-001). 14 μl of the purified BAC was mixed with 23.4 μl of 0.1 mM dNTP mix (1:2:2:2 of dTTP:dATP:dCTP:dGTP), 10 μl of 10X Nick translation buffer, 10 μl of the Nick translation enzyme mix, 12 μl of 0.2 mM Green dUTP, and H2O to bring up the volume to 100 μl. The reaction was incubated in a thermocycler at 15°C for 16 hours, heated to 70°C for 10 minutes, and then held at 4°C. The labeled BAC was purified using DNA Clean & Concentrator-5 (Zymo Research, Irvine, CA, Catalog# D4013) in 50 μl batches and eluted in 10 μl of the elution buffer. 25 μg of salmon sperm Cot-1 was added per batch and the batches were mixed together. The mixture was vacuum dried, and the pellet was resuspended in 10 μl of LSI buffer (LSI/WCP Hybridization Buffer, Abbott Molecular, Catalog# 06J67-011) to make the stock BAC probe mix. The stock was stored in the dark at -20°C. For staining, the stock was further diluted in LSI buffer at a 1:19 stock:LSI ratio.
The BAC probe staining procedure was adapted from [98]. Chromosome spread slides were placed in 3:1 MeOH:HAc at -20°C for 15 minutes. The slides were then washed 2 times in 1X PBS for a minimum of 2 minutes each and treated with 0.5 mg/ml Protease II (Abbott Molecular Inc., Catalog# 06J93-001) at 37°C for 5 minutes. The slides were washed 2 times in 1X PBS for a minimum of 5 minutes each, and then progressively dehydrated in 2-minute washes with 70%, 85%, and 100% EtOH. The slides were allowed to air dry completely and used immediately for staining.
Prior to BAC probe staining, PNA telomere probe staining was performed as described in the “PNA telomere probe staining” section, and the slides were allowed to air dry completely after the final 1X PBS wash. At this point, 10 μl of the BAC probe (1:19 dilution in LSI buffer) was added per slide, covered with a 24 x 50 coverslip, and sealed with rubber cement (Elmer’s, Atlanta, GA, Catalog# E904). The slides were heated in a hybridization oven at ~70–71°C for 3 minutes and then the oven temperature was allowed to drop to ~50°C after which the slides were transferred to a flat, humid chamber and incubated at 37°C overnight in the dark. The coverslip was peeled off and the slides were washed in coplin jars in 1) 50% formamide in 2X SSC at 45°C, 2 times for 5 minutes each, 2) 2X SSC at 45°C, 2 times for 5 minutes each, 3) 4X SSC + 0.05% Tween 20 for 8 minutes, 4) 1:1 2X SSC:PBSTw (1X PBS + 0.1% Tween 20) at room temperature (RT) for 5 minutes, and 5) PBSTw at RT, 3 times for 5 minutes. Excess PBSTw was removed from the slides by tapping their sides on a paper towel. The antibody staining and slide mounting were performed as described in the “Primary antibody staining” and “Secondary antibody staining” sections, with PBSTw used instead of PBT.
All images were collected at the Department of Molecular and Cellular Biology Light Microscopy Imaging Facility at UC Davis. Chromosome spreads were imaged using the Nikon N-SIM Super-Resolution microscope in 3D-SIM imaging mode with Apo TIRF 100X oil lens. The images were collected and reconstructed using the NIS-Elements Imaging Software. Sections and fluorescent whole mounts were imaged using the Olympus FV1000 laser scanning confocal microscope. Images were processed using the Fiji ImageJ software. Only linear modifications to brightness and contrast of whole image were applied. All raw image files are available upon request.
To analyze fertility, individual mutant fish were crossed to wild type fish to assess their ability to generate offspring. Offspring that were produced were tracked daily for up to 5 days to assess morbidity and mortality.
A cDNA fragment of zebrafish ddx4/vasa encoding the COOH-terminal amino acids 479–651 (based on accession number BC129275) was cloned into pET100 using the following primers: fwd 5’- CACCATGTTCATAGCAACATTTCTCTGTCAAG-3’ (ATG initiation codon added); rev 5’- TAACAGGTGTGAGGCCAGTTATTCC-3’. The His-tagged protein was expressed in E. coli, purified using standard procedures and used to immunize chicken hens (91-day protocol, Pocono Rabbit Farm & Laboratory Inc. Canadensis, PA). Polyclonal IgY from crude serum was used at 1:500.
An N-terminal fragment of Sycp1 cDNA was amplified with Phusion DNA polymerase (Thermo Fisher Scientific, Catalog#: M0530L) using the following primers: Fwd 5’-aactttaagaaggagatataccATGCAAAAAGCATTCAACTT-3’, and Rev 5’-tctcagtggtggtggtggtggtgctcGGTAACTTCTATTTCTGCATtt-3’. The Sycp1 PCR product (1272 bp) was then cloned into pET28b using NEBuilder HiFi DNA Assembly Master Mix (NEB, Ipswich, MA, Catalog#: E5520S). BL21 (DE3) cells containing pRARE and Sycp1 overexpression construct were grown in 2.6 L of LB with kanamycin and chloramphenicol until an OD600 = 1 and induced with a final concentration of 1 mM IPTG at room temperature for six hours. The Sycp1 peptide was purified under denaturing conditions using Novagen NiNTA purification resins (Sigma, Catalog#: 70666) according to the manufacturer’s instructions. The Sycp1 peptide was concentrated to a final concentration of 1mg/ml in PBS using a 10kDa centrifugal filter (Sigma, Catalog# UFC901008). The Sycp1 peptide was injected into two chickens by Pocono Rabbit Farm and Laboratory following the 91-day polyclonal antibody production protocol.
|
10.1371/journal.pcbi.1006424 | Mechanisms of hysteresis in human brain networks during transitions of consciousness and unconsciousness: Theoretical principles and empirical evidence | Hysteresis, the discrepancy in forward and reverse pathways of state transitions, is observed during changing levels of consciousness. Identifying the underlying mechanism of hysteresis phenomena in the brain will enhance the ability to understand, monitor, and control state transitions related to consciousness. We hypothesized that hysteresis in brain networks shares the same underlying mechanism of hysteresis as other biological and non-biological networks. In particular, we hypothesized that the principle of explosive synchronization, which can mediate abrupt state transitions, would be critical to explaining hysteresis in the brain during conscious state transitions. We analyzed high-density electroencephalogram (EEG) that was acquired in healthy human volunteers during conscious state transitions induced by the general anesthetics sevoflurane or ketamine. We developed a novel method to monitor the temporal evolution of EEG networks in a parameter space, which consists of the strength and topography of EEG-based networks. Furthermore, we studied conditions of explosive synchronization in anatomically informed human brain network models. We identified hysteresis in the trajectory of functional brain networks during state transitions. The model study and empirical data analysis explained various hysteresis phenomena during the loss and recovery of consciousness in a principled way: (1) more potent anesthetics induce a larger hysteresis; (2) a larger range of EEG frequencies facilitates transitions into unconsciousness and impedes the return of consciousness; (3) hysteresis of connectivity is larger than that of EEG power; and (4) the structure and strength of functional brain networks reconfigure differently during the loss vs. recovery of consciousness. We conclude that the hysteresis phenomena observed during the loss and recovery of consciousness are generic network features. Furthermore, the state transitions are grounded in the same principle as state transitions in complex non-biological networks, especially during perturbation. These findings suggest the possibility of predicting and modulating hysteresis of conscious state transitions in large-scale brain networks.
| Hysteresis, characterized by distinct forward and reverse phase transitions, is ubiquitous in nature. For example, there are distinct temperatures for water freezing and ice melting. Similarly, it has been found that state transitions related to consciousness exhibit hysteresis. In particular, the concentration of general anesthetics required to achieve loss of consciousness is significantly higher than the concentration at which consciousness is regained. However, it is unknown whether this is trivially reducible to the pharmacology of these drugs or if it is something related to brain function itself. In this study, we took a novel, network-based approach and hypothesized that the hysteresis observed during anesthetic state transitions shares the same underlying mechanism as that observed in non-biological networks. Our computational modeling, analytic study, and high-density human EEG analysis suggest that various hysteresis phenomena during loss and recovery of consciousness can be explained in principled ways by generic network features. Identifying these network mechanisms of hysteresis in the brain also provides a unified framework for understanding the radically different conscious state transitions associated with sleep, anesthesia, and disorders of consciousness.
| Hysteresis, the differential pathway of forward and reverse state transitions, is a universal phenomenon observed in nature and has been investigated in various fields such as physics, engineering, biology, and economics[1–12]. Hysteresis has also been observed during state transitions in the brain, such as sleep[6,7] and general anesthesia[8–12]. It has been found in both drosophila and murine models that the concentration of general anesthetic required to induce unconsciousness is higher than the concentration at which consciousness is regained. Thus, identification of the mechanism of hysteresis will be essential to the complete understanding of conscious state transitions, with clinical applications to fields such as anesthesiology and neurology.
Traditionally, it has been assumed that the induction of and emergence from general anesthesia are mirror images of one another. However, the asymmetric anesthetic concentrations associated with induction and emergence have long been recognized and explained by pharmacokinetics[13–17]. Recently, it has become clear that the neural circuits mediating loss of consciousness do not entirely overlap with the neural circuits mediating recovery of consciousness. Max Kelz proposed ‘neural inertia’ as a fundamental and biologically conserved principle by which neural circuits of the central nervous system resist behavioral state transitions, such as those between consciousness and unconsciousness[8]. Neural inertia suggests that hysteresis is not a pharmacokinetic attribute, but a fundamental neurobiological process that stabilizes states of consciousness and creates resistance to rapid and potentially catastrophic transitions[8]. Accumulating evidence indicates that the relevant sites mediating the induction of and emergence from anesthesia are distributed globally rather than localized in the brain[13]. If hysteresis is a large-scale network phenomenon in the brain then, like many other biological and physical systems, the state transitions may be governed by the same network mechanism, which is referred to as ‘explosive synchronization.’
Explosive synchronization is a discontinuous transition between incoherent and synchronized states of a network[18,19]. We previously demonstrated that healthy humans undergoing frequent loss and recovery of consciousness in a lightly anesthetized state demonstrate conditions of explosive synchronization in networks reconstructed from high-density EEG[20]. We also conducted a modeling study in human brain networks that suggests varying patterns of explosive synchronization as a mechanism of diverse state transition patterns[21]. Based on our work and that of others, we hypothesized that the hysteresis observed during the loss and recovery of consciousness might be mediated by patterns of explosive synchronization, as occurs in many other physical and biological systems.
To test our hypothesis, we analyzed high-density EEG data from 22 healthy human volunteers during consciousness, anesthetic-induced unconsciousness (ketamine, n = 15; sevoflurane, n = 7), and recovery as well as the transitions between them. To identify empirical evidence of hysteresis based on EEG, we examined the EEG networks during the loss and recovery of consciousness induced by anesthesia. Assuming that characteristic features of EEG networks (i.e., altered connectivity, high modularity, and reconfigured hub structures) observed during general anesthesia reflect brain states that influence behavioral responses[22], we examined the EEG network and its trajectories in a 2-dimensional parameter space consisting of topographic similarities and connection strengths (Table 1). We also developed a neuroanatomically informed brain network model to identify the onset points of the state transitions and the conditions for hysteresis during loss and recovery of consciousness. Finally, with empirical data and analytic study, we tested the control parameters that were identified in the model for their ability to modulate hysteresis patterns across conscious state transitions.
This study was conducted at the University of Michigan Medical School and approved by the Institutional Board Review (HUM00061087); after careful discussion, written informed consent was obtained from all participants.
We used high-density EEG data from two independent studies using ketamine and sevoflurane; detailed methodology can be found in the S1 Text and the previous studies by Vlisides et al[23] (ketamine) and Blain-Moraes et al[24] (sevoflurane). For this study, we selected four states (baseline, induction, unconscious, and emergence) from the state transitions of each data set. The hypotheses and related analyses of the current study were completely distinct from that of the original studies.
Fifteen human volunteers with 128-channel EEG were studied in the ketamine experiment, which included a subanesthetic dose (0.5 mg/kg ketamine administered over 40 minutes) and general anesthesia (induced by a single bolus dose of 1.5 mg/kg). Seven volunteers with 64-channel EEG were studied in the sevoflurane experiment during anesthetic concentrations gradually increasing from 0.4% to 0.6% to 0.8% (the average concentration at which unconsciousness was induced in the original study) or beyond, then decreased from 0.8% to 0.6% in high-flow oxygen (8 L/min). The EEG in both experiments was recorded with eyes closed. The loss and recovery of consciousness were defined as the loss and recovery of response to the verbal command ‘squeeze your left [or right] hand twice,’ on a recording loop every 30 seconds, with right/left hand commands randomized.
The EEG includes four different levels of consciousness: baseline, induction, unconscious, and emergence. Since the lengths of induction and unconsciousness were variable across participants, we chose one EEG epoch for each state, which also enabled us to calculate the hysteresis areas in a parameter space. We determined that the selection of an EEG epoch for each state does not change the results qualitatively.
The states analyzed in this study are defined as:
The average reference was used for referencing and the windowed sinc-FIR filter (in the MATLAB toolbox from EEGLAB) was used to avoid a possible shifting of the signal phases in both analyses. We analyzed 2-minutes-long EEG epochs with 10-seconds-long moving windows for each state.
The following procedure, illustrated in Fig 1, was implemented:
Phase Lag Index (PLI), a measure of phase locking between two EEG signals, was used to define the functional connectivity in the EEG network[25]. We chose a Hilbert transform to extract the instantaneous phase of the electroencephalogram from each channel and calculate the phase difference Δθij(t) between channels i and j, where Δθij(t) = θi(t) − θj(t), t = 1,2,…,n, and n is the number of samples within one epoch. PLIij between two nodes i and j is then calculated using Eq (1):
PLIij=|<sign(Δθij(t))>|,0≤PLIij≤1.
(1)
Here, the sign() function yields: 1 if Δθij(t) > 0; 0 if Δθij(t) = 0; and -1 if Δθij(t) < 0. The mean < > is taken over all t = 1,2,…,n. If the instantaneous phase of one signal is consistently ahead of the other signal, the phases are considered locked and PLIij ≈ 1. However, if the signals randomly alternate between a phase lead and a phase lag relationship, there is no phase locking and PLIij ≈ 0.
To control for spurious connectivity of EEG, 20 surrogate data sets were generated with a random shuffling method, in which a time point is randomly chosen in each EEG channel; the EEG epochs are then shuffled before and after the time point. The shuffled data have the same amplitude distribution and power spectrum of the original EEG but there are disruptions of the original connectivity between two EEG signals. The non-zero PLI from the shuffled data is regarded as spurious connectivity.
We expected that different EEG frequency bands and different states would have different levels of spurious connectivity[26]. Thus, after subtracting the median PLI of 20 surrogate data sets, if the remaining PLI was larger than 0.1 then the connectivity of two EEG signals was set as 1; otherwise, it was set as 0. The threshold (0.1) was chosen to avoid isolated nodes in the EEG network in the baseline states (S1 Fig). The basic EEG network properties were examined across states during the two anesthetic experiments. The node degree of an EEG channel was defined as the number of links in the network.
Temporal coordination of neural activity is a necessary condition for neural communication in the brain[27]. An EEG network constructed with phase lag index (phase synchronization) and its topography of node degrees may reflect a coarse-grained structure facilitating neural communication across brain regions. We previously demonstrated that the topography of EEG networks can differentiate various states of consciousness during general anesthesia[28]. Thus, in this study, we defined the brain state by the topography and the average node degree of an EEG network, which reflect the structure and strength of neural communication, respectively, across the brain regions associated with each EEG electrode. To quantify hysteresis, we constructed a two-dimensional parameter space that consists of the average node degree and topographic similarity of EEG networks, and then examined the trajectories of EEG networks in the parameter space (Fig 1). Topographic similarity measures how far the EEG networks under general anesthesia diverge from that of the baseline. Considering the large variability, we used a relatively long epoch length and small moving window size to represent the brain state. We segmented the 2-minutes-long EEG epochs with 10-seconds-long moving windows. The topographic similarity and the average node degree of EEG networks were calculated for each 10-seconds-long moving windows, and the median value of every epoch in the same state was used to represent a state in the parameter space. The hysteresis size for a given subject and anesthetic was calculated with the four median topographic similarities (representing four states) and average node degrees in the parameter space (Fig 2A and 2D).
Topographic similarity was defined by the Pearson correlation coefficient between the average node degrees of the baseline and the node degrees of each epoch from the four states (baseline, induction, unconscious, and emergence). Average node degree was defined by the average node degree of every node in each epoch.
Topographicsimilarity=Corr(B¯i,j,Ai,j(k)),i,j=1,⋯,Nandk=1,⋯,T.
(2)
Averagenodedegrees=1N∑i[Ai,j(k)],i,j=1,⋯,N.
(3)
where Ai,j(k) is a binary connection matrix of each epoch and B¯i,j is the averaged connection matrix over every epoch of the baseline. Corr (⋯) is the Pearson correlation coefficient function. i and j are the indices of every node (N) and k is the index of each epoch.
For all selected periods within each subject, spectral power was computed based on the short-time Fourier transform using the ‘spectrogram.m’ function in the MATLAB Signal Processing Toolbox (time window: 3s hamming window, overlap: 50%). The median absolute power (μV2/Hz) was then computed for each experimental period at selected frequency bands, for all channels.
We performed one-way ANOVA (“anova1.m”, MATLAB toolbox) with Tukey-Kramer correction (“multcompare.m” with alpha = 0.05 and ctype = “tukey-kramer” in MATLAB) for the comparison of the hysteresis areas among various frequency bands. The statistical tests were carried out for each experiment separately. The adjusted P-values of 0.05 or lower (*P < 0.05, **P < 0.01, and ***P < 0.001) were considered to be statistically significant (S1 Table and S2 Table).
The main goal of large-scale brain network modelling with simple oscillatory models is to identify general computational principles rather than to achieve biological realism. Our goal of the model study was to identify an underlying network mechanism of hysteresis phenomena and to test our hypothesis that the property of explosive synchronization identified in generic network models also holds for state transitions of the brain network during general anesthesia. Many recent studies have successfully applied Kuramoto/Stuart-Landau models to the brain in order to understand the organizational principles of multiscale brain function, surrogates of information flow, and complex dynamics at the whole brain network level[29–31]. Similarly, we believe that the application of simple oscillatory models to anatomically informed brain network structure can capture the essence of the hysteresis phenomena.
To model the hysteresis phenomena that are empirically observed during anesthetic state transitions, we used a large-scale brain network model that implements a general coupled oscillator model on the scaffold of an anatomically informed human brain network structure. The human brain network consists of 78 parcels of the cerebral cortex constructed from diffusion tensor imaging (DTI) of 80 young adults[32]. Previous studies with a Kuramoto model demonstrated that the fraction of the nodes adaptively controlled by local order parameters (average phase synchronization of a node with its linked nodes) modulates the hysteresis during state transitions between incoherent and synchronized states[19,33,34]. The result was robust for various network configurations. Here, we extended the Kuramoto model with a modified Stuart-Landau model that includes an adaptive feedback term, i.e., a recursive interaction process between a node and the other nodes. The modified Stuart-Landau model shows how an adaptive feedback term modulates the hysteresis in both the phase and amplitude dynamics.
Here rj(t) is the amplitude of oscillator j at time t. λj is a parameter governing the amplitude of each oscillator. S is the coupling strength between oscillators and Ajk denotes the anatomical connections between oscillator j and k, yielding 1 if a connection exists and 0 otherwise. τjk is the time delay between node j and k. θj(t) is the phase of oscillator j at time t. ωj is the intrinsic frequency of oscillator j. We modified the original Stuart-Landau model by adding the RjZ term. We define Rj as synchrony of node j: Rj≡|1/2(eiθj+1/N∑k=1Neiθk)|, to measure the extent to which the node j is synchronized with the other nodes. In the setting of a heterogeneous network like the brain, Rj avoids a bias of the local order parameter. For instance, the local order parameter of a node whose node degree is 1 is determined by the phase synchronization of only the linked node and its initial frequency, randomly arranged. Rj is defined to be confined between 0 and 1, where 0 signifies complete incoherence and 1 means complete synchronization. Z is a scale term for the feedback process between a node j and its linked nodes. The adaptive feedback process incorporates the memory of a given state of synchronization into the dynamics, simultaneously enhancing heterogeneity in the phase and amplitude dynamics of the brain network. Notably, the adaptive feedback term was multiplied by the coupling strength of the phase equation, not by the amplitude equation. However, the phase and amplitude dynamics mutually interact with each other in Eqs (4) and (5). Direct implementation of a feedback term in the amplitude Eq (4) causes a divergence.
We analytically and computationally showed that hysteresis naturally occurs during state transitions between incoherent and synchronized states with mean field approximation[19,31,35], and identified the control parameters that modulate the hysteresis. In this work, the degree of synchrony will be measured by global order parameter R≡|1/N∑k=1Neiθk|. The analytic derivation is shown in the S1 Text. We extended the previous studies in several ways. First, with the modified Stuart-Landau model, we can study the hysteresis in phase dynamics and amplitude dynamics independently, which enabled us to interpret the hysteresis in the phase-based connectivity and EEG power during state transitions independently. Second, the newly defined node synchrony considers heterogeneous local network connectivity and Z in the adaptive feedback terms as associated with the steepness of dose-response slope of anesthetics. Third, from the model studies, we can infer the range of frequency Δω as another control parameter to modulate hysteresis during the state transition. Fourth, we analytically derived the specific feedback term, RjZS, from a pharmacokinetic equation (i.e., Hill coefficient equation), suggesting the role of the feedback term as the anesthetic effect on brain network synchronization (the analytic derivation is in S1 Text).
All parameters for the models were set to simulate alpha oscillations in the brain, because the hysteresis areas were highest in the alpha peak (9-11Hz) in the experimental results. The natural frequencies of the oscillators in our simulation were given as a Gaussian distribution around 10 Hz with standard deviation of 1 Hz. Time delay was given proportional to the physical distances between edges with a propagation speed of 8.6m/s[36]. The coupling strength between the oscillators was continuously increased from 0 to 50 and decreased from 50 to 0 for continuous change between a fully synchronized and fully unsynchronized network. The amplitude parameter λj was given identically for all oscillators with a value of 1. For all simulations, we also added a Gaussian white noise variable ξj(t) with a mean and standard deviation of 2 Hz. We changed the power of the adaptive feedback term Z from 2 to 8, and the distribution range of natural frequencies Δω from 0.1 to 4, in order to investigate the relationship between the parameters and hysteresis size. In each parameter set, 200 configurations were simulated and the results were averaged over all configurations.
To find empirical evidence for hysteresis phenomena at the network level, we measured average node degree and topographic similarity of node degree. These variables represent the connection strength and structure of the functional network, respectively, and have been demonstrated to help differentiate levels of consciousness in the context of anesthetic state transitions[28]. The two measures were used to construct a trajectory of functional brain networks during the loss and recovery of consciousness.
For the trajectory, we constructed the functional networks with 2-minutes-long EEG epochs and 10-seconds-long moving windows. Phase Lag Index (PLI) [25], a simple phase locking measure between two EEG signals, was used to define the edges in the functional network. For each epoch, we computed average node degree and topographic similarity. Topographic similarity measures the correlation between the node degrees of the baseline state and the node degrees of each epoch for every node in experiment 1 (ketamine) and experiment 2 (sevoflurane). The change of topographic similarity indicates how the connectivity structure changes during general anesthesia compared to the baseline state.
Fig 2A and 2D show the trajectories of the EEG network for different frequency bands during the state transitions in both experiments (from baseline (circle) through induction (down-pointing triangle), unconscious (square), and emergence (up-pointing triangle)). The point representing each state is the median value of all individual subjects. The states of individual subjects were calculated with an average of every epoch. Each EEG frequency band shows a distinctive hysteresis pattern. Fig 2B and 2E present the hysteresis sizes of different frequency bands (2Hz frequency bins for 0.1-31Hz frequency range). The hysteresis size was defined by the area encompassed by the forward and reverse pathways. The result shows that the alpha band (9-11Hz) has the largest hysteresis compared to the other bands (about 5-fold larger than the smallest hysteresis). Fig 2C and 2F shows the trajectory of the alpha band. Each marker represents the average node degree and average topographic similarity over the 15 (ketamine) and 7 (sevoflurane) subjects for each epoch. Filled stars represent the first epoch within each state. The brain topographic map for each state represents the node degree pattern averaged over all subjects. Notably, in the two experiments, the trajectories turn clock-wise, which implies that the topographic similarities during emergence are always higher than during the induction path at the same connection strength. In other words, the connection structure responds more sensitively to the induction and emergence than the connection strength. There is evidence that general anesthetics might disproportionately affect hub structures, which occupy a relatively small portion of the brain network but play a significant role in information transmission. Hub disruptions (leading to faster or slower dynamics) change functional connection structure at an earlier stage of anesthesia, while the average connection strength is still intact. A significant change in average connection strength follows later with overall damage of functional networks. Such differential responses to a perturbation of connection structure and strength is a generic feature for a heterogeneous network that is perturbed.
In order to compare the simulation and empirical data, we assumed brain networks with nodes that are synchronized correspond to conscious states and networks with incoherent nodes correspond to the unconscious state induced by general anesthetics. The empirical data support this assumption; deep anesthesia reduces the global synchronization level of the brain network. In Fig 3A and 3B, we simulated the typical state transition pattern (i.e., the clock-wise turn of trajectories) observed in the empirical data in Fig 2C and 2F. The topographic similarity during the emergence period is always higher than the induction period for the same average node degree. When changing the coupling strength S in the brain network model, the order parameter R shows a hysteresis in both forward and reverse pathways (Fig 3B). When S increases from 0, the transition to a synchronous state occurs at a critical coupling strength Sinc. However, when the coupling strength decreases, the transition to the incoherent state begins at a different critical coupling strength Sdec. Analytic and simulation studies show that the transition points of Sinc and Sdec are not equal, indeed, Sinc < Sdec, showing that, at each value of coupling strength S, the state of the system is not defined uniquely. In other words, if the synchronization level of a network is changed with an adaptive feedback process, then it naturally produces a path-dependence during state transition. The details of the analytic derivation for the different transition points, Sinc and Sdec, will be explained in the last section. Consequently, when we convert the typical path-dependence in the order parameter R (Fig 3B) into the trajectories of the topographic similarity and the connection strength (PLI), it manifests as a clock-wise turn (Fig 3A).
Furthermore, hysteresis is observed in the brain network connectivity but not the power. Here, we defined average power as the average over the powers of all nodes and the topographic similarity of power as the correlation of the power topographies between baseline and all epochs. Fig 3C presents the trajectory in the 2-dimensional space of the averaged power and the topographic similarity of power. For both the model data (Fig 3C) and empirical data (Fig 3D and 3E), no significant hysteresis was found. This result suggests that the hysteresis in the human brain is mainly due to global interactions across brain regions, rather than activities within brain regions.
According to the model, we tested two control parameters, (1) the strength of adaptive feedback process, i.e., the power coefficient Z in the model, and (2) the variance of peak frequencies among nodes Δω in the brain network. We studied how the two control parameters modulate the onset points of state transitions between incoherent and synchronized states. As expected, Fig 4A and 4B show that a larger Z produces larger hysteresis (Z = 4 and 8 for Fig 4A and 4B, respectively). The Z value reflects the strength of an adaptive and recursive feedback process. The locally different feedback process with RjZ promotes the heterogeneity of the network dynamics and modulates the onsets of state transitions. To determine the onset of the state transition, we defined the synchronized state as R = 0.5 and the critical coupling strength S as R = 0.5. The simulation was repeated 200 times with random initial frequencies and the results were robust with any other thresholds between 0.1 and 0.9. The mean of the frequencies was set as 10Hz with variance between 0.1 and 4 Hz to emulate the EEG of the alpha band. Fig 4D demonstrates the critical coupling strengths of both pathways with respect to Z. The critical coupling strengths increase along with Z. Notably, the critical coupling strength more steeply increases during emergence compared to induction. This implies that a stronger adaptive feedback process delays the onset of emergence from the incoherent state and has less influence on the onset of induction.
Second, the variance of frequency Δω is another control parameter that is measurable from EEG. The variance of initial frequency Δω in the model positively correlates with the critical coupling strengths in both forward and reverse pathways. Fig 4A and 4C present small and large variance of frequencies, Δω = 1 and 2.5, respectively. In the comparison, the larger Δω has larger critical coupling strengths than the smaller Δω. Fig 4E shows that the critical coupling strengths linearly increase with Δω. According to the relationship, if we assume that the coupling strength of the system is increased and decreased continuously between 0 and 6, we can predict that a network with a larger Δω would transition more easily at induction but resist emergence. Conversely, a network with a smaller Δω may have resistance to induction but be more permissive for emergence. For instance, the network of Δω = 4 easily reaches its critical coupling strength during the induction (from 6 to 5 in Fig 4E), whereas in the network of Δω = 1 it is relatively hard to reach the critical coupling strength (from 6 to 1.5). Conversely, it is difficult for the network of Δω = 4 to reach its critical coupling strength during the emergence (from 0 to 6 in Fig 4E), but relatively easy for Δω = 1 (from 0 to 2). To empirically test this model prediction, we investigated the correlation between the Δω of EEG during anesthesia and the recovery time, that is, the duration between the loss and recovery of consciousness. The test was carried out only with the data from the ketamine experiment, in which we could determine the recovery time precisely after administering the anesthetic, because ketamine was given as a single bolus as opposed to the downward titration of sevoflurane. Fig 4F demonstrates significant correlation between the recovery time and the Δω of the EEG (0.73, p<0.01, Spearman coefficient). The clear relationship between recovery time and the frequency range of EEG matches the model prediction.
A strong pharmacological perturbation such as general anesthesia induces a significant decrease of long-range network synchronization, which has been proposed as a neural mechanism of the loss of consciousness; the converse situation for the recovery of consciousness has been proposed [37–39]. The analytic derivation with mean-field approximation explains why the critical coupling strengths in both pathways are different as the synchronization level varies [33]. In the model, the adaptive feedback process plays a key role in the hysteresis. As S decreases, the inherent frequencies at which the oscillators join a synchronized cluster is smaller than certain value: ωj < Ωdec where Ωdec ≈ S R˜2kj. By contrast, as S increases, the permissive condition to join a synchronized cluster is ωj < Ωinc where Ωinc ≈ S R˜kj. Here, R˜ is an arbitrary normalized synchronization measure (0 to 1) that is used as the adaptive feedback term in the model and kj is the degree of node j. In the solutions, if the R˜ is near 1, the difference between Ωdec and Ωinc becomes negligible, but if R is near 0, the difference between Ωdec and Ωinc becomes larger, and Ωdec< Ωinc. Therefore, when the system transitions from an incoherent to a synchronized state (from R˜≈0 to R˜≈1), the possible number of oscillators to join a synchronized cluster is smaller than in the case of the opposite transition (from R˜≈1 to R˜≈0). In other words, as a network transitions from an incoherent to synchronized state, it is harder to reach the synchronized state (defined by the threshold), but in the opposite path it is easier to get to the desynchronized state. Such asymmetry in the possible number of nodes that can join a synchronization cluster along forward and reverse pathways is the key network mechanism of the hysteresis. As a result, hysteresis naturally appears when a network contains a feedback process.
Hysteresis is a universal phenomenon observed in many biological and physical systems[1–12]. In the brain, the hysteresis phenomenon is observed in various state transitions that can be spontaneous (sleep-wake cycle) or pharmacologically induced (general anesthesia) [6–12]. Hysteresis has been reported at various scales of observation in the brain [8–12,40,41]. At the molecular level, models of general anesthesia suggest that hysteresis may be caused by heterogeneous site effects resulting in pharmacokinetic/dynamic delays[13–17]. At the neuronal circuit level, Steyn-Ross el al. proposed that hysteresis naturally occurs as a first-order phase transition (i.e., discontinuous state transition) in the population average membrane voltage of cortical neurons[40]. A recent study in rodents reported distinctive sequences of thalamic and medial prefrontal cortex activity at loss and recovery of consciousness[42]. At the behavioral level, it has been demonstrated that different anesthetic concentrations are associated with loss and recovery of consciousness in diverse species such as mouse and fruit fly. As a result of this and other work, Kelz suggested a fundamental neurobiological process referred to as ‘neural inertia,’ which helps to maintain aroused and anesthetized states and creates resistance to state transitions[8]. Proekt and Hudson, using a mathematical model, tested the hypothesis that neural inertia is a consequence of the stochastic switching between the waking and anesthetized state. They showed that properties of a bistable system can account for phenomena related to neural inertia, supporting the hypothesis that emergence from anesthesia is independent of pharmacokinetic factors[43,44].
In this study, we hypothesized that the hysteresis observed during state transitions induced by general anesthesia is a generic network feature. Furthermore, we hypothesized that the state transition between consciousness and unconsciousness in the human brain may be governed by the same principle that is operational in non-biological complex networks responding to perturbations. Even though the molecular mechanisms of ketamine and sevoflurane are significantly different, we found that the trajectories of both state transitions were similar to each other, showing a clock-wise turn in the 2-dimensional parameter space composed of average node degree and topographic similarity. The consistent trajectory pattern of functional brain networks suggests a common mechanism of state transition during the loss and recovery of consciousness. We discovered in the model study that the larger topographic similarity of functional brain networks during the emergence process compared to that of the induction process is a spontaneous phenomenon, while the synchronization level decreases and increases in a network with an implemented feedback process.
In the EEG analysis, we also found that the lower frequency bands (<12Hz) showed significantly larger hysteresis, whereas the higher frequency bands (12- 30Hz) have a smaller hysteresis that approaches randomness. Notably, the alpha frequency band (8-12Hz) showed the largest hysteresis (about 5-fold larger than the smallest hysteresis in the high frequency band). The role of the alpha band during anesthetic state transitions is still unclear but the anteriorization process during propofol anesthesia and surgical levels of sevoflurane anesthesia has been well described[45]. Furthermore, a prominent response of the diverse functional connectivity measures in the alpha band has been reported during the loss of consciousness[24,46–49].
The large hysteresis of the alpha band can be explained by the fact that alpha oscillations are associated with global brain connectivity. For example, Zhang et al. demonstrated that the travelling waves of the alpha band propagate across brain regions[50]. The direction and the speed of traveling waves are determined by a spatial gradient of local frequencies and connections that also correlate with the performance of a memory task. Independently, we identified a general relationship between brain network structure and direction of connectivity[30,31]. Among the frequency bands, only the alpha bandwidth of the human and monkey brains matched the model predictions. These two studies demonstrate the association of alpha oscillations with global connectivity in the brain. Although anesthetics can affect all EEG frequencies, alterations of other frequency bands are relatively small because their original connectivity is significantly lower than the alpha band[23,48]. Since we defined hysteresis based on an EEG network derived with a particular kind of functional connectivity measure, the alpha band predictably shows the largest alteration of EEG network, which would allow the largest hysteresis to be observed relative to other frequency bands.
Interestingly, the hysteresis was observed only in the functional connectivity, not in the power of EEG. The empirical data in this study thus suggest that hysteresis is a function of global interactions, not regional brain activities. These empirical findings motivated us to develop the phase-based human brain network model and to investigate the generic features of hysteresis phenomena.
In the field of statistical physics, especially in condensed matter physics, the hysteresis phenomenon during state transitions has been studied extensively[1,2]. Recently, studies of hysteresis have extended to network science, including the underlying network mechanism of explosive synchronization[18,19,35,51]. Filatrella et al. and Zhang et al. suggested that the suppression of giant synchronization cluster formation in a network is a general mechanism of explosive synchronization[33,35]. Under the conditions of such a synchronization suppression process, a network will exhibit explosive synchronization, i.e., a discontinuous state transition between incoherent and synchronized states, during which a hysteresis can occur. Incorporating previous studies, we developed a global brain network model of hysteresis. Our model generalizes the previous phase dynamics models[33,35], including the amplitude dynamics as well as time delays between oscillators. The model of amplitude dynamics was able to recapitulate the empirical finding. In particular, we empirically demonstrated that there was no hysteresis in the power of EEG, lending credibility to the model. The adaptive feedback term in the model plays an essential role in producing asymmetric suppression during destruction and reformation of giant clusters during the synchronization process. The feedback primarily acts on the phase synchronization rather than the amplitude of the oscillators. Our analytic solution for the model quantitatively explains the asymmetric pathways of a hysteretic state transition. When a network changes its overall level of synchronization, the threshold required to join a synchronized cluster is different in the forward and reverse pathways (Ωdec ≈ S Ωdec ≈ S R˜2kj for decreasing synchronization; Ωinc ≈ S Ωdec ≈ S R˜kj for increasing synchronization). For instance, when the network transitions from an incoherent to synchronized state (from R˜≈0 to R˜≈1), the possible number of oscillators to join a synchronized cluster is smaller than that of the opposite transition (from R˜≈1 to R˜≈0). This discrepancy is greatly amplified at small R˜ (<<1) with Ωdec (R˜2) << Ωinc (R˜). Considering the fact that a consistent effect of general anesthetics is to decrease the temporal coordination in large-scale brain networks, we expect that the network mechanism of hysteresis may be more influential during the state transition around a lower level of consciousness.
The brain requires a higher anesthetic concentration to lose consciousness than to maintain unconsciousness. That is, despite the same behavioral response ratio, there can be different anesthetic concentrations depending on whether the pathway is moving in the forward or reverse direction[8]. In the same way, the brain network requires a higher coupling strength to restore synchronization (i.e., crossing over the threshold of state transition, R = 0.5) than to lose synchronization. Applying this to empirical observations of anesthetic state transitions, a lower anesthetic concentration during emergence would be more potent because the network now has higher thresholds to cross in order to resynchronize. However, the state transitions empirically observed during the loss and recovery of consciousness are not linear and monotonic but rather nonlinear and complex; here we simplified the problem as much as possible to identify a general network feature in the empirical data. In Fig 5, we present a conceptual schema to correlate the hysteresis pattern of the brain network with the hysteresis pattern of the behavioral response.
Another benefit of our model study is that it identified the two control parameters for modulating the hysteresis size. The Z value in the model reflects the strength of the adaptive feedback process in the brain, which relates to the dose-response slope of anesthetics. The simulation result demonstrated that a larger adaptive feedback process gives rise to a larger hysteresis. The dose-response curve for each anesthetic has a unique slope (H; Hill coefficient in Hill equation determines the dose-response curve). For instance, the H of halothane is 10–30 and that of isoflurane is 3–4. The hysteresis of halothane is larger than isoflurane[8]. The Hill equation determines the degree of cooperativity of the ligand binding to the enzyme or receptor. The Hill coefficient H quantifies the degree of interaction between ligand binding sites and measures the sensitivity of the response curve[52,53]. Analytically, the power coefficient Z of RZ is equivalent to the Hill coefficient H in the dose-response curve for the anesthetic (S1 Text). However, to biologically link the adaptive feedback term Z in our model and the Hill coefficient H in the dose response curve will require further investigation.
Our model also suggests the variance of frequencies as another control parameter that modulates hysteresis in functional brain networks. We found a strong correlation between the variance of EEG peak frequencies during anesthesia and the onset of recovering consciousness in the empirical analysis of the ketamine experiment. Larger variance of the peak frequencies of channels was associated with a more delayed onset of recovering consciousness. This is consistent with our model prediction, i.e., larger variance of frequencies lead to larger hysteresis with relatively delayed onset of a state transition from the incoherent to synchronized state. Our model prediction also holds for the opposite pathway. Chennu et al. demonstrated that subjects who have lower EEG coherence (weaker phase-based connectivity in the alpha band) at baseline were more likely to become unresponsive during sedation[49]. The model with a larger variance of frequencies demonstrated that the brain network is less synchronized at a higher coupling strength, i.e., a relatively small perturbation of coupling strength induces a transition into an incoherent state. This may explain why individuals with lower functional brain connectivity, which can be caused by a large variance of EEG frequencies, lose consciousness more easily.
Conventionally, the dose-response slope of the anesthetic is described with the Hill coefficient of the dose-response equation. However, the steepness of dose-response slope of the anesthetic and its Hill coefficient have not been linked to a brain network model that can explain anesthetic effects at the network level. In the analytic study, we proposed for the first time the relationship between the dose-response slope of the anesthetic, drug concentration, and coupling strength of the brain network. Two assumptions were made to link a Hill-type equation and a brain network model. The first assumption was that the anesthetic reduces coupling strengths between brain regions (in S1 Text equation S2). Empirical evidence supports this assumption with disintegrated brain functions at the network level under general anesthesia[22]. The second assumption was that drug concentration is inversely proportional to the brain network synchronization (in S1 Text equation S4) (that is, deep anesthesia induces an incoherent brain network). If these two assumptions are satisfied, then the dose-response slope of the anesthetic, drug concentration, and coupling strength of the brain network are related (in S1 Text equation S5). Interestingly, the effective coupling strength under anesthesia, Seff, is analytically equivalent to the adaptive feedback term, SRz, that has been used in the previous network models to facilitate explosive synchronization. Thus, in our brain network model under general anesthesia, the power of the feedback term, z, represents the anesthetic effect on the brain network (in S1 Text equation S11).
This study has several limitations. First, we did not perform a slow up- and down-titration in the ketamine experiment. As such, there was pharmacokinetic asymmetry that is independent of neurobiology or principles of hysteresis. However, the results at the network level during ketamine state transitions were similar to that of sevoflurane, which was up- and down-titrated, as well as the model. Second, we acknowledge that consciousness and unconsciousness cannot be trivially reduced to, respectively, synchronized and incoherent networks. It is temporal coordination rather than synchrony, per se, that is critical for consciousness. However, it is well known that functional brain networks become more modular during general anesthesia, as coordinated and synchronized interactions across the cortex break down. Thus, for the purposes of large-scale modeling, we consider it reasonable to assume that the conscious brain will have, in aggregate, more synchronized interactions compared to the unconscious brain. Third, our model focused on the coarse-grained features of brain network state transitions to identify the fundamental factors that produce and modulate hysteresis patterns. State transitions induced by general anesthetics are complex and the relationship between anesthetic concentration and behavioral response is nonlinear [54,55]. More realistic neural population models may extend the interpretation of our study to such nonlinear relationships. Fourth, there is still a large gap between the adaptive feedback term in the brain network model and the Hill coefficient in the dose-response curve of the anesthetic. However, precisely defining this relationship was beyond the scope of the current study.
We characterized hysteresis phenomena in functional brain networks during anesthetic-induced state transitions. Our brain network model and its analytic derivation suggest that the asymmetry of synchronization suppression is the key mechanism of the hysteresis observed during loss and recovery of consciousness. Furthermore, we propose variance of frequencies and strength of the adaptive feedback process as the control parameters for modulating the onsets of state transitions in the human brain.
|
10.1371/journal.pntd.0006278 | Multiplex serology for impact evaluation of bed net distribution on burden of lymphatic filariasis and four species of human malaria in northern Mozambique | Universal coverage with long-lasting insecticidal nets (LLINs) is a primary control strategy against Plasmodium falciparum malaria. However, its impact on the three other main species of human malaria and lymphatic filariasis (LF), which share the same vectors in many co-endemic areas, is not as well characterized. The recent development of multiplex antibody detection provides the opportunity for simultaneous evaluation of the impact of control measures on the burden of multiple diseases.
Two cross-sectional household surveys at baseline and one year after a LLIN distribution campaign were implemented in Mecubúri and Nacala-a-Velha Districts in Nampula Province, Mozambique. Both districts were known to be endemic for LF; both received mass drug administration (MDA) with antifilarial drugs during the evaluation period. Access to and use of LLINs was recorded, and household members were tested with P. falciparum rapid diagnostic tests (RDTs). Dried blood spots were collected and analyzed for presence of antibodies to three P. falciparum antigens, P. vivax MSP-119, P. ovale MSP-119, P. malariae MSP-119, and three LF antigens. Seroconversion rates were calculated and the association between LLIN use and post-campaign seropositivity was estimated using multivariate regression. The campaign covered 68% (95% CI: 58–77) of the population in Nacala-a-Velha and 46% (37–56) in Mecubúri. There was no statistically significant change in P. falciparum RDT positivity between the two surveys. Population seropositivity at baseline ranged from 31–81% for the P. falciparum antigens, 3–4% for P. vivax MSP-119, 41–43% for P. ovale MSP-119, 46–56% for P. malariae MSP-119, and 37–76% for the LF antigens. The seroconversion rate to the LF Bm33 antigen decreased significantly in both districts. The seroconversion rate to P. malariae MSP-119 and the LF Wb123 and Bm14 antigens each decreased significantly in one of the two districts. Community LLIN use was associated with a decreased risk of P. falciparum RDT positivity, P. falciparum LSA-1 seropositivity, and P. malariae MSP-119 seropositivity, but not LF antigen seropositivity.
The study area noted significant declines in LF seropositivity, but these were not associated with LLIN use. The MDA could have masked any impact of the LLINs on population LF seropositivity. The LLIN campaign did not reach adequately high coverage to decrease P. falciparum RDT positivity, the most common measure of P. falciparum burden. However, the significant decreases in the seroconversion rate to the P. malariae antigen, coupled with an association between community LLIN use and individual-level decreases in seropositivity to P. falciparum and P. malariae antigens show evidence of impact of the LLIN campaign and highlight the utility of using multiantigenic serological approaches for measuring intervention impact.
| Plasmodium falciparum malaria is the principal cause of illness and death in Mozambique. However, the same mosquitoes that transmit P. falciparum parasites also transmit three other species of malaria (P. malariae, P. ovale, and P. vivax) and the worm that causes lymphatic filariasis. To date, we do not know how much transmission of the three other species of malaria occurs. We also do not know if control interventions such as the distribution of bed nets reduce the transmission of lymphatic filariasis and non-falciparum malaria. To address this question, we sampled community members immediately following and one year after a bed net distribution campaign in Mozambique. We analyzed their blood for the presence of antibodies to four species of malaria and lymphatic filariasis. We found that a substantial proportion of individuals had antibodies to P. falciparum, P. malariae, P. ovale, and the worms causing lymphatic filariasis. We found much lower rates of seropositivity to P. vivax. Individuals reporting using bed nets had a lower risk of testing positive for P. falciparum and P. malariae antibodies. The proportion of the population with access to and using bed nets was too low to cause a population-wide decrease in malaria transmission. There was a significant decline in lymphatic filariasis seropositivity between the two surveys, but we could not attribute it to the bed net distribution campaign. Measuring antibody levels for multiple diseases simultaneously has utility in assessing intervention impact.
| Northern Mozambique has one of the highest rates of Plasmodium falciparum transmission and disease burden in the world [1]. As in the rest of Mozambique, infection with P. falciparum malaria parasites is among the principal causes of outpatient visits, hospitalizations, and deaths. Malaria transmission occurs year-round and P. falciparum prevalence in children under 5 was measured to reach up to 65% in the central and northern provinces in 2015 [2].
Although P. falciparum parasites are the most deadly and common agent transmitted by Anopheles vectors in Mozambique, anopheline mosquitoes are also responsible for transmission of three other species of human malaria–P. malariae, P. ovale, and P. vivax–as well as the lymphatic filariasis (LF) parasite Wuchereria bancrofti. In contrast to P. falciparum, the distribution and burden of non-falciparum malaria and LF in Mozambique, as in most of sub-Saharan Africa, has not been well characterized. Reasons for this include poor diagnostic capability, less severe manifestations of disease, and less attention and funding from ministries of health and international donors. In general, P. vivax has historically been thought to be largely absent from sub-Saharan Africa due to the lack of the Duffy coat receptor in populations originating in West Africa [3]. Although P. ovale and P. malariae are thought to be in circulation in Mozambique, estimating their incidence of infection has been difficult, particularly since they are typically present at low parasite densities and are difficult to detect through slide microscopy, especially in the presence of a concomitant P. falciparum infection. Lymphatic filariasis has been mapped to be most prevalent in northern and central Mozambique [4], but the highly focal nature of LF transmission and the delay between infection and development of disease makes surveillance of this major cause of disability difficult.
Currently, the most effective strategies for reducing the burden of P. falciparum infection focus on vector control [5]. In Mozambique, Anopheles mosquitoes are targeted by indoor residual spraying with insecticides and the distribution and use of long-lasting insecticidal nets (LLINs), which, in addition to protecting the user with a physical barrier, in practice also function as human-baited insecticidal mosquito traps and can significantly reduce mosquito populations [6]. The Mozambican National Malaria Control Program (NMCP) adopted a strategy of universal coverage with LLINs throughout the country in 2010–2011, aiming to cover each sleeping space with an LLIN. Implementation began with a series of sub-provincial mass distribution campaigns, some of which were implemented by the Ministry of Health and local authorities, and some by non-governmental partner organizations. Although primarily used to prevent P. falciparum infections, LLINs are known to reduce transmission of P. vivax [6] and have been postulated to also reduce transmission of P. ovale and P. malariae [7] and the LF parasites through their effect on the common vector [8]. Although there is some evidence of the impact of LLINs on LF transmission [8, 9], nets have not been widely adopted as a primary intervention against LF, with current strategies largely limited to mass drug administrations (MDAs) of the antifilarial drugs ivermectin or diethylcarbamazine in combination with albendazole [10].
As part of its monitoring and evaluation program, the Mozambican NMCP periodically evaluates LLIN distribution campaigns. The objectives are to both monitor the operational performance of the campaigns, as assessed through coverage and usage indicators, as well as to measure the impact of the campaigns on malaria prevalence and estimates of transmission. The latter is particularly important amid the rise of insecticide resistance and the potential for diminishing effectiveness of LLINs in controlling malaria transmission. In 2013, the Mozambican NMCP chose to evaluate a LLIN distribution campaign in the northern province of Nampula. Besides measuring coverage and falciparum malaria prevalence, additional components were added to the evaluation to measure the prevalences and assess the impact of the campaign on non-falciparum Plasmodium species. A further component was included to assess the additional impact of the LLIN campaign on LF; this was complex as both districts received MDAs for LF during the evaluation period. These additional objectives were made possible by the recent development of multiplex serology methods that allow detection of antibodies to multiple antigens simultaneously [11]. It was hypothesized that this laboratory technique could detect changes in the levels of antibodies to various malaria and LF antigens following the campaign, indicating changes in exposure to, and hence transmission of, malaria and LF parasites as a result of the LLIN campaign.
Two consecutive cross-sectional household surveys were carried out one year apart, with the first, baseline survey occurring two weeks following the mass LLIN distribution campaign in 2013. Achieved coverage with LLINs was assessed during the first survey, LLIN usage was assessed during the follow-up survey one year later in 2014, and impact of the campaign was assessed by comparison of biological markers of infection from the first and second surveys.
A LLIN distribution campaign encompassing six districts of the northern province of Nampula was implemented in the second half of 2013. Of the six districts, two were purposefully chosen to be included in the survey: the coastal district of Nacala-a-Velha, and Mecubúri District in the interior (S1 Fig). Both districts are predominantly rural. Nacala-a-Velha is in close proximity to the port city of Nacala-Porto, and Mecubúri, although close to the provincial capital of Nampula, is particularly difficult to access. Both districts were classified as endemic for LF as of 2013, and began undergoing annual MDAs of albendazole and ivermectin starting 2012 in Nacala-a-Velha and starting 2013 in Mecubúri. Administrative coverage for the MDAs in Nacala-a-Velha and Mecubúri was 70% and 114%, respectively, in 2013 and 76% and 82% in 2014.
For the survey, twenty enumeration areas (survey clusters) in each district were chosen randomly with probability proportional to size from the full list of census enumeration areas for each district. Four clusters in Mecubúri were inaccessible and were substituted with four randomly chosen replacement clusters. Immediately prior to the start of the survey, trained enumerators visited each selected cluster and compiled a full list of households, recording the name of head of household and the latitude/longitude coordinates of the household. For each cluster, the list of households was randomly sorted, and in the first survey in 2013, survey teams visited households according to the order of the list, continuing until 16 households were visited per cluster. In the second survey in 2014, survey teams revisited households from 2013, matching households based on the name of head of household and coordinates, and no replacement of households was allowed. In each household, all household members present were invited to participate in the survey. The target sample size was 1,320 individuals and 367 households per district, designed to provide 80% power to detect a 10% change between the two surveys in the proportion of the population testing positive for P. falciparum parasites, assuming a baseline prevalence of 43% and a design effect of 3.
Trained teams, each composed of a national-level supervisor and a district surveyor, visited the selected households and, after obtaining consent, administered a household questionnaire. Surveyors collected socioeconomic data, including occupation and education of the head of household and household ownership of goods; generated a roster of household members; enumerated all sleeping spaces and recorded who slept in which sleeping space; visually inspected each sleeping space and recorded the presence and location (hanging or stored) of the bed net designated for that sleeping space, denoting whether or not the bed net bore the marking specific to the distribution campaign; and asked the interviewee about how often on average each bed net was used during the wet and dry seasons. In both years, each household member, regardless of age, present during the visit and providing written consent was administered a P. falciparum HRP2-specific rapid diagnostic test (RDT) (SD Bioline, Yongin, Republic of Korea), and had up to six 10 mcL spots of capillary blood collected on filter paper (TropBio, Cellabs, Sydney, Australia). Individuals testing positive for malaria were treated by survey teams in accordance with national treatment guidelines [12]. In 2014, the same questionnaires and procedures were followed as in 2013, with an additional module where household members were cross-linked to those in the first round based on name and age.
The surveys were carried out in September 2013 and October 2014 in Nacala-a-Velha and December 2013 and November 2014 in Mecubúri, with data collection lasting three weeks for each survey in each district.
After collection in the field, the filter paper was dried overnight, placed into individual plastic bags with desiccant sachets, and then refrigerated prior to shipment to central laboratories in Maputo and later to CDC laboratories in Atlanta. Blood spots were eluted overnight at 4°C at a serum dilution of 1:40 (assuming 50% hematocrit) and further diluted in casein-containing dilution buffer as previously described for a final dilution of 1:400 of serum [13, 14].
A multiplex bead platform [11] was used to measure immunoglobulin G (IgG) antibody response to ten antigens: six malaria, three LF, and one control antigen (Strongyloides stercoralis) (S1 Table). The 19-kDa subunit of the merozoite surface protein 1 (MSP-119) from each of the four main human malaria species was cloned and expressed as recombinant Schistosoma japonicum glutathione-S-transferase (GST) fusion proteins [14–16]. A P. vivax MSP-119 expression clone that included the carboxy-terminal hydrophobic tail sequence was used [16]. A (NANP)5 peptide corresponding to the carboxy-terminus of the P. falciparum circumsporozoite protein (CSP) was cross-linked to GST and then coupled to a SeroMap bead as previously described [14]. The Pl1043 epitope from P. falciparum Liver Stage Antigen 1 (LSA-1) [17] was synthesized and coupled to beads at a concentration of 60ug/mL at pH 5.0. The Strongyloides stercoralis NIE antigen-GST fusion protein, GST fusion partner with no inserted sequence, and the Brugia malayi Bm14- and Bm33-GST fusion proteins were cloned and expressed as previously described [18–21]. W. bancrofti Wb123 antigen expressed as a GST fusion protein was a gift of T. Nutman (NIH, Bethesda, MD). Although only W. bancrofti occurs in Africa, the serological test for the B. malayi antigens cross-reacts with W. bancrofti.
With the exception of the P. vivax GST/MSP-119, all other antigens were coupled to SeroMap beads (Luminex Corp., Austin, TX) using the buffers and protein amounts previously described [16]. The P. vivax antigen was coupled to a BioPlex COOH bead (BioRad, Hercules, CA) using the protein amount and buffer previously specified [16]. Total IgG multiplex bead assays were performed using the biotin-streptavidin system previously described [19, 20]. Assays included beads coated with the 10 proteins described above and an additional 31 antigen-coated beads representing viral, bacterial, and parasitic diseases. Each assay plate included a buffer-only blank and 6 control sera to ensure consistent assay performance throughout the study. Assays were run in duplicate, and results were reported as the average of the two median fluorescent intensity values minus the buffer-only blank value (MFI-bg). Samples (N = 8) that had discordant result between the two runs (coefficient of variation >15%) for >4 positive antigen responses were repeated.
For malaria and LF antigens, cutoff values were determined using a panel of 81 presumed negative sera from adult US citizens who had no history of foreign travel. Values greater than the mean plus 3 standard deviations of these negative control values were considered positive. For the S. stercoralis NIE assay, a cutoff value determined by a receiver-operator characteristic curve analysis using a different lot of coupled beads was translated to the current study bead set using a 2-fold serial dilution curve of a strong positive control sample as an inter-assay standard.
A subset of samples (20%) did not have full blood spots available; for these, partly filled blood spots were analyzed. Comparison of average MFI-bg values from full and partial blood spots revealed an average difference of <10%; thus, all samples were included in the final statistical analysis.
Demographic characteristics for the heads of household were tabulated, and a socioeconomic status (SES) index was constructed and calculated for each household using a previously described methodology [22]. Households were divided into equal quintiles based on the SES index score. Key ownership, access, and usage indicators [23] were calculated separately for each district. The proportion of the population testing positive for P. falciparum infection by RDT was calculated for each district and year, stratifying by age. Estimation of coverage indicators and malaria positivity was adjusted taking into account the complex sample design using the R survey package [24]. Data were weighted by the inverse of the probability of selection, calculated as the product of the probability of selection for the cluster, the household, and, where applicable, the individual.
For each of the antigens the mean seropositivity, defined as the percent of individuals with MFI-bg above the predetermined threshold, was calculated for the total population and also stratifying by ten age categories. A reversible catalytic model was fit to the seropositivity by age data for each antigen, and the estimates for the serological conversion rate (SCR) and serological reversion rate (SRR) per year were directly calculated from the likelihood model [25]. The SRR was assumed to be constant for both districts and both years, but the SCR was separately calculated for each district and each year of the survey. For each antigen, the average of the population log MFI-bg value was calculated for each district and year, and the 95% confidence intervals were calculated assuming a normal distribution of the log MFI-bg values. The Strongyloides NIE antigen was included as a control antigen to aid in discriminating between the effects of the LLIN and MDA campaigns, as Strongyloides transmission was presumed to be unaffected by the LLIN campaigns but sensitive to the MDA campaigns.
Poisson regression, which allows direct estimation of the relative risk, [26] was used to model the association between individual- and cluster-level LLIN use and ten binary biological outcomes in individuals sampled in the second survey: RDT positivity and seropositivity for the six malaria and three LF antigens. Each model was fit adjusting for age, sex, and household SES. Use of LLIN was quantified on a 0 to 1 scale by a principal components analysis of data on LLIN location and reported use during the wet and dry season, as previously described [23]. The individual and community effects were jointly estimated by normalizing the individual-level variable by subtracting the average cluster-level value from the individual-level variable [27].
All statistical analyses were performed in R version 3.3.2 (R Foundation for Statistical Computing, Vienna, Austria).
The study was approved by the National Bioethics Committee in Mozambique. Adult participants provided written consent prior to enrollment in the study, and also provided written consent on behalf of child participants. CDC investigators provided technical assistance and were not considered to be engaged in the research.
A total of 282 households in Nacala-a-Velha and 300 households in Mecubúri were visited in the first survey in 2013 (Table 1). The total number of people living in these households was 1,172 in Nacala-a-Velha and 1,443 in Mecubúri, and of these, 539 (46%) household members in Nacala-a-Velha and 598 (41%) in Mecubúri were present and consented to have blood drawn during the survey. In the follow up survey in 2014, 81% (228/282) of the households in Nacala-a-Velha and 72% (217/300) in Mecubúri were revisited; a total of 578 household members in Nacala-a-Velha and 704 in Mecubúri were sampled in the second survey.
The coverage attained by the LLIN distribution campaign was low (Table 2). The campaign reached 80% (95% CI: 72–86) of households in Nacala-a-Velha and 54% (44–65) in Mecubúri, but only 58% (48–67) of households in Nacala-a-Velha and 36% (29–43) in Mecubúri received at least one LLIN per sleeping space. The proportion of the population sleeping in spaces with an available LLIN was 68% (58–77) in Nacala-a-Velha and 46% (37–56) in Mecubúri. Usage of any LLIN in the year following the distribution campaign was also low, with only 40% (27–55) of the population in Nacala-a-Velha and 23% (17–30) in Mecubúri reporting having used LLINs more than 4 times per week during the wet season, falling to 21% (14–30) and 17% (13–22), respectively, in the dry season.
There was no statistically significant change in P. falciparum RDT positivity from 2013 in 2014 in either district with overlapping 95% confidence intervals, although the point estimates for RDT positivity were higher in 2014 versus 2013 (Table 3). One year after the campaign, RDT positivity in the key <5 year age group was 61% (95% CI: 44–76) in Nacala-a-Velha and 87% (76–94) in Mecubúri. Overall, RDT positivity was significantly higher in Mecubúri than in Nacala-a-Velha.
The serological data confirm high P. falciparum transmission in both districts. Virtually all sampled individuals were positive for P. falciparum MSP-119 antibodies, with very high antibody responses (S2 Fig) even in infants, indicating that individuals’ first P. falciparum infection likely occurs early in infancy. Seropositivity for P. falciparum CSP also eventually reached saturation, with close to 100% of the older age categories testing positive, but the slope of the seropositivity by age curve was more gradual (Fig 1). In contrast to the other two P. falciparum antigens, seropositivity to P. falciparum LSA-1 in general did not surpass 40% for any age group.
For both P. falciparum CSP and P. falciparum LSA-1 antigens, the SCR was higher in Mecubúri than Nacala-a-Velha, consistent with the difference in RDT positivity by district. Generally, there was no statistically significant difference in SCR between the two years for these antigens (t-test p-values ranging from 0.07 to 0.35) (Table 4). The only difference approaching statistical significance was a slightly lower SCR for P. falciparum CSP in 2014 versus 2013 in Nacala-a-Velha, which fell by 15% (t-test p-value 0.07).
Robust antibody responses to the three other Plasmodium species were detected in both young and old individuals in both districts (Fig 2, S3 Fig), an indication of ongoing transmission of all three other species. After P. falciparum, the species registering the highest seropositivity was P. malariae, with 46% (95%CI: 42–50) of the population in 2013 in Nacala-a-Velha and 56% (52–60) in Mecubúri seropositive for antibodies against P. malariae MSP-119, with a corresponding SCR of 0.068 (0.054–0.081) in Nacala-a-Velha and 0.095 (0.077–0.11) in Mecubúri. Similar levels of exposure were observed for P. ovale, with the proportion of the population with antibodies to P. ovale MSP-119 in 2013 ranging from 41% (37–45) in Nacala-a-Velha to 43% (39–47) in Mecubúri, and an SCR estimated to be 0.12 (0.087–0.15) in Nacala-a-Velha and 0.13 (0.093–0.16) in Mecubúri. Much lower levels of antibody positivity to P. vivax were observed, with only 2.7% (1.5–4.5) of the population in Nacala-a-Velha and 3.9% (2.5–5.9) in Mecubúri with detectable antibodies to P. vivax MSP-119. There was a statistically significant difference in SCR for P. malariae between the two surveys in Nacala-a-Velha, with the estimate for SCR for 2014 24% lower than in 2013 (t-test p-value 0.03). There was a similar reduction in the SCR for P. ovale in Nacala-a-Velha, with a 22% reduction, but this was not statistically significant (t-test p-value 0.096). The SCRs for P. malariae and P. ovale in Mecubúri did not show a similar reduction, falling by only 1% and 3%, respectively, with neither antigen showing a statistically significant difference in SCR between the two surveys (t-test p-values ranging from 0.43 to 0.46).
The absolute levels of antibody response and seropositivity by age curves for the three LF antigens were consistent with the geographic distraction of LF (Fig 3, S4 Fig). The highest rates of seropositivity at baseline were to the Bm33 antigen, with 67% (95%CI: 56–65) of the population in 2013 in Nacala-a-Velha and 76% (72–79) in Mecubúri seropositive, compared to 37% (33–41) in Nacala-a-Velha and 45% (41–49) in Mecubúri seropositive for antibodies against Bm14, and 50% (46–54) in Nacala-a-Velha and 41% (37–45) in Mecubúri seropositive for antibodies against Wb123 (Table 4). There were significant reductions in SCR for the LF antigens between the two surveys in Nacala-a-Velha, with the SCR for Wb123 declining by 27% (chi-square test p-value 0.011), the SCR for Bm14 declining by 17% (chi-square test p-value 0.084), and the SCR for Bm33 declining by 22% (chi-square test p-value 0.026). In Mecubúri, the SCR fell by 19% for Wb123 (t-test p-value 0.052), 31% for Bm14 (t-test p-value <0.001), and 59% for Bm33 (t-test p-value <0.001).
A substantial proportion of the population in Nacala-a-Velha (58%, 95%CI: 53–62) and Mecubúri (65%, 95%CI: 61–69) had antibodies against the control Strongyloides NIE antigen in 2013. In Nacala-a-Velha, there was no statistically significant difference in SCR (t-test p-value 0.26) or seropositivity (chi-square p-value 0.53) to the NIE antigen between the two surveys, whereas Mecubúri witnessed statistically significant declines of 35% for the NIE SCR (t-test p-value <0.01) and 17% for NIE seropositivity (chi-square p-value <0.01) from 2013 to 2014.
The achieved sample size was lower than the target sample size, due to a lower than expected number of people tested per household (2.4 vs 3.6). Although there was no detected change in overall P. falciparum RDT positivity from 2013 to 2014, both individual- and cluster-level LLIN use was associated with lower risk for P. falciparum RDT positivity in the second year after adjusting for age, sex, and SES (Table 5). The relative risk for testing RDT positive in LLIN users compared to non-users was 0.81 (95%CI: 0.64–1.0). Moreover, the relative risk of RDT positivity for the individuals living in clusters where everyone would be sleeping under an LLIN was estimated to be 0.43 (0.24–0.77) compared to individuals living in clusters with no LLIN use, demonstrating the indirect effects of LLINs, independent of individual use of LLINs. Cluster-level LLIN use was also associated with lower risk for seropositivity to P. falciparum LSA-1 and P. malariae MSP-119 antigens. In contrast, no significant protective effect of individual- or community-level LLIN use on LF, P. vivax MSP-119, and P. ovale MSP-119 antibody positivity was observed. Generally, RDT positivity and seropositivity to all antigens decreased with increasing SES. P. falciparum RDT positivity was negatively associated with age, in contrast to antibody seropositivity, which increased with age for all antigens.
The high rates of P. falciparum RDT positivity observed in these surveys, coupled with the high rates and age distribution of antibody levels to P. falciparum antigens, confirm holoendemic transmission of P. falciparum in the survey area. The levels of RDT positivity are particularly striking given that the surveys were conducted at the end of the dry season, when transmission would be expected to be lowest. Areas of large malaria burden would benefit most from a mass LLIN campaign. However, the coverage indicators provide evidence that the LLIN campaign evaluated here was far from reaching its intended target of universal coverage. Reported usage, ranging from 17% to 40%, was far removed from the 65% threshold postulated to be necessary in providing a demonstrable community reduction in malaria incidence [28]. Nevertheless, those individuals using LLINs and living in clusters with high overall usage of LLINs did have significantly lower risk for testing positive for P. falciparum infection by RDT, mirroring results from previous studies of LLIN effectiveness in Mozambique [22] and confirming the continued effectiveness of LLINs as a malaria prevention strategy in Mozambique.
This study highlights the added benefit of simultaneously measuring seropositivity to multiple antigens to estimate P. falciparum transmission intensity. The results show that not all antigens are consistently informative in this setting. For example, the P. falciparum MSP-119 antigen, which to date has been one of the standard antigens used to assess population-level P. falciparum exposure [25, 29] provides little information on changes in P. falciparum intensity in a setting of such high transmission as northern Mozambique. The transmission intensity is such that virtually all sampled individuals regardless of age had antibodies to MSP-119, evidence that the first P. falciparum infection likely occurs in early infancy. However, the two other P. falciparum antigens included in the assay, CSP and LSA-1, generated seroprevalence curves with increasing likelihood of transitioning to seropositive with increasing age. Both antigens are thought to be less immunogenic than MSP-119, and the data presented here suggest that repeated P. falciparum infections throughout life are needed to generate a consistently high antibody level to each antigen. Due to the slower acquisition of antibodies to these two antigens, the seroprevalence by age data were informative, allowing differentiation between the higher transmission in Mecubúri versus Nacala-a-Velha, an observation also seen in the RDT positivity data. Additionally, P. falciparum LSA-1 seropositivity was lower in LLIN users, further evidence of the protective effect of LLINs against P. falciparum infection.
In addition to the two districts being a setting of very high P. falciparum transmission, the populations in both districts showed substantial serological responses to P. ovale and P. malariae antigens. A small but non-zero proportion of the population showed evidence of exposure to P. vivax, consistent with the results of a 2015 household survey which showed a national 0.2% P. vivax prevalence in children under 5 years of age [2]. MSP-119 antigen competition studies have not indicated antibody cross-reactivity in most individuals [14], and species-specific MSP-119 antibody responses were common even among patients who had high responses to multiple malaria MSP-119 antigens. Although the MSP-119 antigens from P. vivax and P. falciparum share 51% identity at the amino acid level, some of the conserved residues are cysteines and other hydrophobic amino acids that are unlikely to be exposed to the immune response [30]. Bousema et al. used the two MSP-119 antigens in ELISA studies of sera from a population living in a region endemic for both parasites and did not observe any correlation between the P. vivax and P. falciparum antibody responses [31]. In a separate study, 79% of women who were positive for antibodies to malaria reacted with the MSP-119 antigen from only one species [16]. Thus, it is unlikely that the observed antibody responses to P. malariae and P. ovale antigens can be solely attributed to assay cross-reactivity.
Although subject to many limitations, the SCR at its most basic definition is a measure of incidence, the annual rate at which individuals acquire antibodies to a certain antigen [32]. Taken at face value, the SCRs estimated for P. ovale MSP-119 and P. malariae MSP-119 in these two districts in 2013 suggest an annual incidence of P. ovale infection of 125/1000 in Nacala-a-Velha and 133/1000 in Mecubúri, and an annual incidence of P. malariae infection of 70/1000 in Nacala-a-Velha and 100/1000 in Mecubúri. This magnitude of incidence would elevate P. ovale and P. malariae as major contributors to malaria burden in northern Mozambique.
In contrast to P. falciparum, there was a statistically significant population-level decrease in P. malariae seropositivity and a borderline significant decrease in P. ovale seropositivity from 2013 to 2014 in one of the districts. Since LLIN use was associated with lower post-campaign risk of testing positive for P. malariae antibodies, there is evidence that the LLIN campaign, at least in Nacala-a-Velha where coverage was higher, might have had an impact on decreased P. malariae transmission. The fact that there were detectable changes in seroprevalence and distribution of P. ovale and P. malariae markers and no change in P. falciparum might be due to the differences in magnitude of transmission intensity. One hypothesis is that with such high levels of P. falciparum transmission, there would need to be a much larger decrease in vectorial capacity to result in a detectable change in incidence, whereas non-falciparum transmission might be low enough to be sensitive to smaller changes in vectorial capacity.
In both districts there was evidence of significant declines in LF transmission between 2013 and 2014, as seen by decreases in both the SCR and overall proportion of the population seropositive for the LF antigens. The results are robust as they hold across all three different LF antigens included in the assay. However, attributing this change to the LLIN distribution campaign is hampered by the concurrent MDAs of antiparasitic drugs in both districts. Ivermectin is effective against LF microfilariae, in addition to a postulated killing effect on mosquitos feeding on individuals treated with ivermectin, and albendazole kills adult worms. Together, the two-drug combination could be expected to influence antibody levels through its effect on transmission and worm load. In Mecubúri, the substantial declines in SCR and proportion of the population seropositive for the NIE Strongyloides antigen, which should be influenced by the MDAs but not by the LLIN campaign, suggest that there was high enough coverage from the antiparasitic MDAs to reduce LF transmission. In Nacala-a-Velha, however, there was no significant difference in NIE Strongyloides SCR and seropositivity between the two surveys, and thus the population-level declines in SCR and seropositivity to LF antigens could be due to the LLIN distribution campaign. Overall, there was no association between individual or community LLIN use and seropositivity to LF antigens. Given the confounding due to the concurrent use of MDAs in the survey districts, this result cannot be interpreted as evidence of no effect, as the effect might have been masked by the antifilarial MDAs.
Several aspects of the study’s design prevent direct inference of a causal relationship between the LLIN distribution campaign and the observed changes in malaria positivity and serological outcomes for malaria and LF. The lack of a control group and the MDA campaigns in the intervening year hamper direct estimation of the impact of the LLIN campaign. In addition, the low coverage and usage resulting from the LLIN campaign and the lower-than-expected sample size limited the ability of the study to assess the impact of LLINs on malaria and LF transmission. Additionally, as IgG against some antigens is known to persist for years following infection [33], more elapsed time may be needed to detect a substantial change in serological metrics following a successful intervention. The extraordinarily high rates of P. falciparum transmission on the backdrop of low LLIN coverage argue for follow-up campaigns in these two districts, both of which will take part in the nationwide universal coverage campaign in Mozambique launched in 2016. Similar evaluations are recommended to evaluate coverage and usage of future campaigns. Finally, the results presented here provide evidence for the enhanced utility of the multi-antigenic and multi-disease assay for quantifying baseline exposure to the non-falciparum malarias and LF, and evaluating the impact of vector control intervention campaigns on these diseases.
|
10.1371/journal.pgen.1005802 | Functional Investigation of a Non-coding Variant Associated with Adolescent Idiopathic Scoliosis in Zebrafish: Elevated Expression of the Ladybird Homeobox Gene Causes Body Axis Deformation | Previously, we identified an adolescent idiopathic scoliosis susceptibility locus near human ladybird homeobox 1 (LBX1) and FLJ41350 by a genome-wide association study. Here, we characterized the associated non-coding variant and investigated the function of these genes. A chromosome conformation capture assay revealed that the genome region with the most significantly associated single nucleotide polymorphism (rs11190870) physically interacted with the promoter region of LBX1-FLJ41350. The promoter in the direction of LBX1, combined with a 590-bp region including rs11190870, had higher transcriptional activity with the risk allele than that with the non-risk allele in HEK 293T cells. The ubiquitous overexpression of human LBX1 or either of the zebrafish lbx genes (lbx1a, lbx1b, and lbx2), but not FLJ41350, in zebrafish embryos caused body curvature followed by death prior to vertebral column formation. Such body axis deformation was not observed in transcription activator-like effector nucleases mediated knockout zebrafish of lbx1b or lbx2. Mosaic expression of lbx1b driven by the GATA2 minimal promoter and the lbx1b enhancer in zebrafish significantly alleviated the embryonic lethal phenotype to allow observation of the later onset of the spinal curvature with or without vertebral malformation. Deformation of the embryonic body axis by lbx1b overexpression was associated with defects in convergent extension, which is a component of the main axis-elongation machinery in gastrulating embryos. In embryos overexpressing lbx1b, wnt5b, a ligand of the non-canonical Wnt/planar cell polarity (PCP) pathway, was significantly downregulated. Injection of mRNA for wnt5b or RhoA, a key downstream effector of Wnt/PCP signaling, rescued the defective convergent extension phenotype and attenuated the lbx1b-induced curvature of the body axis. Thus, our study presents a novel pathological feature of LBX1 and its zebrafish homologs in body axis deformation at various stages of embryonic and subsequent growth in zebrafish.
| Scoliosis is the most common type of spinal deformity with a lateral spinal curvature of at least 10 degrees, affecting 2–4% of the population. Scoliosis caused by a primary problem related to the spine itself is classified into congenital scoliosis (CS) and idiopathic scoliosis (IS). Among these, adolescent idiopathic scoliosis (AIS), the most common form of scoliosis, is known as a common polygenic disease. Severe curving of the spine in scoliosis leads to profound psychological and social impacts, but etiology-based therapies have not been established since the precise pathological mechanisms of both IS and CS remain undefined. Previously, we identified an AIS susceptibility locus near human ladybird homeobox 1 (LBX1) by a genome-wide association study. Here, we report the functional characterization of the most significantly associated single nucleotide polymorphism (SNP), rs11190870 and LBX1 as well as its zebrafish homologues. Overexpression of LBX1 and zebrafish lbx genes caused lateral body curvature in association with the impairment of non-canonical Wnt/planar cell polarity signaling. Thus, our study presents a novel pathological feature of LBX1 in body axis deformation.
| Scoliosis is defined as lateral curvature of the spine with a Cobb angle greater than 10 degrees [1]. It is categorized into congenital, idiopathic, and secondary scoliosis [2]. Congenital scoliosis (CS) is caused by embryonic vertebral malformation that results in deviation of the normal spinal alignment [3]. Idiopathic scoliosis (IS) is a twisting condition of the spine characterized by rotation of the vertebrae without their malformation and is further categorized into infantile, juvenile, and adolescent type by age of onset. Among these forms, adolescent IS (AIS) accounts for 80% of all human scoliosis and develops in 2–4% of children aged between 10 and 16 years across all racial groups [1, 4]. Secondary scoliosis is attributed to a wide variety of causes such as cerebral palsy, paralysis, Duchenne muscular dystrophy, Marfan syndrome, and Ehlers-Danlos syndrome [5–7]. In contrast, the precise disease mechanisms of both IS and CS are understood poorly [8].
Axial skeletal development occurs through a sequential and coordinated series of events regulated by various growth/differentiation factors [2, 9]. During gastrulation, the vertebrate embryo elongates along the anterior-posterior axis through a process called convergent extension [10]. The notochord is then formed ventral to the neural tube as the transient embryonic backbone prior to vertebral bone formation in vertebrates [11]. Following somite segmentation in the paraxial mesoderm, which is formed in a well-defined order along the head to tail axis, the sclerotome derived from the ventral part of the somite eventually gives rise to the vertebrae, the annulus fibrosus of the intervertebral discs, and the rib cage [2]. Any anomalies in these processes are considered to result in the development of both CS and IS.
The role of hereditary or genetic factors especially in the development of AIS has been widely accepted [8]. AIS is a complex polygenic disease influenced by more than one allele at different loci [12]. Indeed, genome-wide association studies identified several novel susceptibility loci including ladybird homeobox 1 (LBX1), G protein-coupled receptor 126, zinc finger protein basonuclin 2, and paired box 1 (PAX1) [13–16]. Among them, a single nucleotide polymorphism (SNP), rs11190870 in the 3′-flanking region of LBX1, has been replicated consistently in independent studies using Chinese [17–19] and Caucasian populations [20].
Human LBX1 was first identified as a gene with homology to the ladybird late (lbl) gene in Drosophila [21]. The ladybird protein is a member of the homeobox transcription factor family with an engrailed repressor domain [22]. In vertebrates, Lbx genes are expressed in the dorsal spinal cord and hindbrain [23], a subpopulation of cardiac neural crest cells [24], muscle precursor cells, and satellite cells of regenerating adult skeletal muscle [25, 26]. Ectopic expression of LBX1 in chicken somites and limbs activates myogenic markers such as myogenin and myod, owing to the expansion of the myoblastic cell population [27]. Previous in vivo studies using Lbx1 knockout mice and lbx gene knockdown morphants in zebrafish or Xenopus did not reveal phenotypes associated with scoliosis [25, 28–30]. To our knowledge, the pathological features of Lbx1 and lbx genes in body axis deformation have not been explored.
Scoliosis has long been considered to be exclusive to bipedal vertebrates [31]. It has been proposed that the unique human upright posture alters spinal conditions toward the eventual development of scoliosis [32]. Naturally occurring scoliosis is quite rare in quadrupedal vertebrates such as rats and mice [31]. The lack of good animal models in vivo has been a major challenge for studying the etiology of scoliosis. Previously, the experimental animal model available for scoliosis research was the young melatonin-deficient chicken, which develops a three-dimensional spinal deformity consisting of lateral curvature after pinealectomy [33]. Recently, it has become clear that several types of fish including zebrafish are suitable for exploring human scoliosis [14, 34–38]. AIS-like scoliosis develops in loss-of-function mutants of protein tyrosine kinase 7 (ptk7) [37] and kinesin family member 6 in zebrafish [34]. Sharma et al. reported that the PAX1 enhancer locus in humans is associated with susceptibility to IS in females and its enhancer activity is disrupted by IS-associated SNPs [14]. Loss of collagen type VIII alpha 1 function also reportedly causes CS-like vertebral malformations [38].
In this study, we characterized the most significantly associated SNP, rs11190870, using chromosome conformation capture (3C), electrophoretic mobility shift assays (EMSAs), and dual luciferase assays, and then examined the effects of the misregulated expression of LBX1 on axial skeletal development using zebrafish as an animal model by both gain-of-function and loss-of-function approaches. We demonstrate that the elevated expression of human LBX1 or zebrafish lbx1 homologs in zebrafish causes axial developmental defects including defective convergent extension movement and body curvature, which could be attributed to the impairment of non-canonical Wnt/planar cell polarity (PCP) signaling. Some zebrafish transiently overexpressing lbx1b survived to develop mild body axis deformation including spinal curvature during larval or juvenile stage. Taken together, our study demonstrated the pathological contribution of lbx genes to body axis deformation in zebrafish.
Human LBX1 and FLJ41350 are located approximately 0.6 kb apart in a head-to-head arrangement on human chromosome 10, and rs11190870 lies 7.5 kb downstream of LBX1 (Fig 1A). FLJ41350 is a hypothetical gene that is found only in the human genome, and its function is uncharacterized [15]. We characterized FLJ41350 through exon connection and 5′-rapid amplification of cDNA ends. We confirmed that FLJ41350 is composed of 3 exons, with the predicted translational start site located at exon 1 followed by an open reading frame of 120 amino acids with no known motif (S1 Fig). No orthologs of FLJ41350 are found in any other species except humans.
To investigate the functional impact of rs11190870, we performed EMSAs and found that some nuclear proteins bound specifically to the genome sequences around rs11190870 with higher affinity to the risk allele than the non-risk allele (Fig 1B). We also analyzed the physical interaction between the genome sequence surrounding rs11190870 and its nearby genome regions using the 3C assay [39] with A172 human glioblastoma cells (A172 cells) (Fig 1C). The 3C assay is a powerful technique for analyzing chromatin organization to reveal the physical interaction between two distal genomic elements [40]. Digestion of cross-linked chromatin with a restriction enzyme and subsequent intra-molecular ligation produces novel junctions between restriction fragments in proximity in the nucleus, which can be detected by PCR. We confirmed that the specific band with primers F4 and R1 was of the expected length (Fig 1C) and corresponded to each primer region by sequencing. This result indicates that the F4 and R1 primer regions are adjacent to each other and that the genome sequence surrounding rs11190870 physically interacts with the promoter region of LBX1-FLJ41350.
We then cloned approximately 1 kb of the LBX1 promoter region (-917 to +153) and evaluated its promoter activity by luciferase assay. In A172 cells, the region had relatively high promoter activity in the direction of LBX1, but not in that of FLJ41350 (Fig 1D). Moreover, the LBX1 promoter, combined with a 590-bp sequence around rs11190870 that is highly conserved across species, had higher transcriptional activity with the risk allele than with the non-risk allele in HEK 293T cells (Fig 1E). These results suggest that rs11190870 confers AIS susceptibility by upregulating LBX1 transcription.
To investigate the effect of the elevated expression of LBX1 on body axis formation, we performed a series of gain-of-function experiments using zebrafish. We overexpressed zebrafish lbx1a, lbx1b, lbx2, and their mutated genes without the homeodomain or the engrailed homology domain by mRNA injection (Fig 2A and 2B). By 48 hours post-fertilization (hpf), the larvae developed body curvature by the ubiquitous overexpression of any one of these lbx genes, but not by that of the mutated genes without the functional domains (Fig 2B). The incidence of body curvature was highest with lbx1b overexpression and increased in a dose-dependent manner (Fig 2B and 2C). Notably, some lbx1b-overexpressing larvae exhibited notochord deformity and a displaced dorsal melanophore stripe (S2 Fig). In addition, a reduction or complete deletion of the forebrain and eyes was observed in many larvae (S2 Fig). Injection of human LBX1 mRNA caused body curvature in embryos, but injection of human FLJ41350 mRNA failed to induce any obvious phenotype related to body axis morphology (Fig 2B).
We also examined the loss-of-function effect on axial development in transcription activator-like effector nuclease (TALEN)-mediated knockout zebrafish (S3 and S4 Figs). Unlike overexpression of lbx genes, lbx1b-/- and lbx2-/- mutant larvae displayed a straight trunk comparable to wild-type larvae (Fig 2D), suggesting the involvement of gain-of-function but not loss-of-function of lbx1b in the body curvature phenotype that might be related to scoliosis susceptibility.
To confirm defective axial development in the established line with uniform expression of lbx1 in an inducible manner, we employed a Gal4/UAS-based bidirectional expression system for the stable overexpression of lbx1b in zebrafish [41]. The F0 driver transgenic carriers were crossed with the F0 responder transgenic carriers to produce Tg(hsp:Gal-VP;EGFP:UAS:lbx1b) F1 progeny with different copies of the Tol2 insertion. Responding to heat shock, embryos with both driver and responder transgenes expressed EGFP and lbx1b driven by the E1b promoter (Fig 3A). The positive correlation between the levels of EGFP and lbx1b expression was examined in Tg(hsp:Gal-VP;EGFP:UAS:mcherry) (S5 Fig). Overexpression of lbx1b in Tg(hsp:Gal-VP;EGFP:UAS:lbx1b) after heat shock was confirmed by western blotting (Fig 3B). By 48 hpf, body curvature became evident in Tg(hsp:Gal-VP;EGFP:UAS:lbx1b) embryos exposed to heat shock at 4 hpf (Fig 3C and 3D). All larvae with lbx1b overexpression died within 7 days post-fertilization (dpf). The severity of body curvature was related to the fluorescence intensity of EGFP in a dose-dependent manner (Fig 3C and 3E). Taken together, we conclude that overexpression of lbx genes, especially lbx1b, induces body curvature in zebrafish embryos.
To elucidate the mechanism by which lbx1 overexpression causes embryonic body curvature, we traced back to the pregastrulation stages. Convergent extension movement during gastrulation (5.25–10.33 hpf) shapes the body axis, narrowing all germ layers in the mediolateral direction and extending them along the anterioposterior axis (Figs 4A and 5A). Embryos overexpressing lbx1b in the gastrulation stage showed mediolateral elongation of somites (Fig 4B and 4C), suggesting some perturbations occur in the formation of the body axis due to abnormal convergent extension. Embryos exposed to heat shock at 4 hpf exhibited a more profound convergent extension defect and more severe body curvature than those at 12 hpf (Fig 4B–4E), demonstrating a positive correlation between the extent of defective convergent extension with the severity of body curvature and the presence of a critical time window for lbx1b overexpression. In situ hybridization for the characteristic markers for the ectoderm or mesoderm revealed a marked delay of convergent movement in embryos overexpressing lbx1b (Fig 5B and 5C). Compared with sibling controls, lbx1b-overexpressing embryos showed a wider neural ectoderm border (dlx3b), broader paraxial mesoderm (papc), and mediolateral elongation (uncx4) of somites (Fig 5B). We also found a significant delay of extension movement (Fig 5D and 5E), which elongated the embryo from head to tail. By contrast, the expression pattern of a dorsal marker, chordin (chd), and a ventral marker, ventral homeobox (vox), in early gastrula was not significantly altered in lbx1b-overexpressing embryos (S6 Fig), indicating that dorsoventral patterning was not affected in these embryos. These results indicate that defective convergent extension resulting from elevated lbx1b expression during gastrulation provokes impaired body axis formation.
In vertebrates, non-canonical Wnt/PCP signals are mainly involved in the regulation of convergent extension [10]. Loss of function of wnt5b or wnt11, the ligand for the non-canonical Wnt/PCP signaling pathway, leads to severely defective convergent extension movement in zebrafish [42, 43]. Our in situ hybridization study revealed that wnt5b was downregulated in gastrulation embryos upon lbx1b overexpression (Fig 6A and 6B and S7 Fig). In contrast, no significant change was observed in wnt11 expression (Fig 6A and 6B and S7 Fig). We also confirmed by quantitative RT-PCR that wnt5b expression was significantly downregulated at the gastrulation stage in lbx1b-overexpressing embryos (Fig 6C).
We then performed an in vivo luciferase assay in zebrafish embryos to analyze the in vivo effects of lbx1b overexpression on the transcriptional activity of two potential promoter regions of wnt5b. Two transcripts encoding 363 (MN1309037) or 380 amino acids were found on the Ensembl website. We tested the sequences upstream of wnt5b including these promoters (P1 and P2). In 90% epiboly embryos, the P2 promoter had much stronger transcriptional activity (about 50-fold induction) than the P1 promoter. Co-injection of lbx1b mRNA repressed the transcriptional activity of P2 by 66.7% (Fig 6D). Thus, lbx1b overexpression during gastrulation downregulated the expression of wnt5b largely through repression of the P2 promoter. These results suggest that defective convergent extension caused by the overexpression of lbx1b in embryos could be attributed to impairment of non-canonical Wnt/PCP signaling.
To evaluate further the effect of misregulation of non-canonical Wnt/PCP signaling in defective convergent extension caused by lbx1b overexpression, we performed a rescue experiment by overexpressing wnt5b, a ligand of the Wnt/PCP pathway. We optimized the amount of wnt5b mRNA injection to avoid defects caused by its overexpression in embryos. Defective migration of dlx3b-positive cells in embryos injected with lbx1b mRNA was rescued by co-injection of lbx1b and wnt5b mRNA (Fig 7A and 7B). Wnt5b mRNA injection mostly rescued the body curvature phenotype in Tg(hsp:Gal-VP;EGFP:UAS:lbx1b) with heat shock at 4hpf (Fig 7C and 7D). We further examined whether defects caused by lbx1b overexpression can be rescued by overexpressing RhoA or Rac1 small GTPases, both of which are downstream effectors of the Wnt/PCP pathway. RhoA rescued both the defective convergent extension and body curvature phenotype (Fig 7A and 7B), whereas Rac1 failed to rescue the convergent extension defects and body curvature (S8 Fig). Interestingly, RhoA overexpression was not effective in larvae with heat shock at 12 hpf (S9 Fig). These results demonstrate that impairment of non-canonical Wnt/PCP signaling, especially the wnt5b/RhoA pathway, caused by lbx1b overexpression, contributes to defective convergent extension and curvature of the body axis.
To investigate the effects of lbx1b overexpression on endogenous expression domains during axial development, we forced lbx1b expression under the control of the previously characterized lbx1b enhancer [44] and the GATA2 minimal promoter by microinjecting a GATA2-1b:lbx1b plasmid (Fig 8A). We confirmed that reporter expression driven by the regulatory elements in Tg(GATA2-1b:EGFP) generally recapitulated the endogenous expression of lbx1b, lbx1a, or lbx2 at different developmental stages (S10 Fig). Similarly to embryos injected with mRNAs, many 48 hpf embryos injected with GATA2-1b:lbx1b developed severe body curvature (S11 Fig), abnormalities in somite morphology (Fig 8B), notochord deformity (S12A Fig), and a displaced dorsal melanophore stripe (Fig 8C). Some of the larvae with a displaced dorsal melanophore stripe had no apparent notochord deformity (S12B Fig). The majority of Tg(GATA2-1b:lbx1b) F1 embryos presented with a severe malformation and died within 24 hpf (S13 Fig). Some were alive at 48 hpf, developing serious axial body curvature, but died within 72 hpf (S13 Fig).
Unlike the F1 generation of Tg(GATA2-1b:lbx1b), which is embryonic lethal, some founder Tg(GATA2-1b:lbx1b) with almost a straight trunk could survive to adulthood, thus allowing our observation of the later developmental stages in this model. We monitored embryos with a mild notochord deformity induced by injection of GATA2-1b:lbx1b (S12A Fig) (n = 41) until adulthood, together with wild-type siblings as controls (n = 45). Thirteen Tg(GATA2-1b:lbx1b) and two control zebrafish died within 21 days. The deformed notochord (S12A Fig red arrow) gradually ossified to form a spine, leading to vertebral malformations (n = 27, p < 0.01) such as hemivertebrae (Fig 8D, white arrow) and block vertebra (Fig 8D, yellow arrow) at the location of the notochord deformity. Eventually, these zebrafish showed scoliosis with vertebral malformations mimicking CS (Fig 8E and S1 video). No apparent spinal deformity was identified in the control (Fig 8B–8E). Thus, local notochord deformity in founder Tg(GATA2-1b:lbx1b) develops into CS-like scoliosis with vertebral malformations.
To investigate the possibility of AIS-mimicking scoliosis in Tg(GATA2-1b:lbx1b) during the period corresponding to human adolescence (Fig 8F), we kept transgenic larvae that had a displaced dorsal melanophore stripe without an apparent notochord deformity (n = 45), together with their wild-type siblings (n = 60). Eight Tg(GATA2-1b:lbx1b) and two control zebrafish died within 30 days. In 19 of the 37 surviving Tg(GATA2-1b:lbx1b) (p < 0.01), significant scoliosis, with rotation of the longitudinal body axis but without visible vertebral malformations, was observed by 55 dpf and then developed progressively until 90 dpf (Fig 8F, 8G and S2 video). Additionally, there was a significant female bias (16/19) for the prevalence of scoliosis (p < 0.01). No apparent spinal deformity was identified in the control group (0/58). These results indicate that mild body axis deformation resulting from the increased expression of lbx1b could cause irregular trunk development such as notochord deformity and a displaced dorsal melanophore stripe, further leading to the later development of CS- or AIS-like scoliosis.
We demonstrate here that the most significantly associated SNP, rs11190870 [15] could confer AIS susceptibility by activating LBX1 transcription. Our gain-of-function approaches using the zebrafish model revealed that the elevated expression of human LBX1 or any of the zebrafish genes lbx1a, lbx1b, and lbx2 causes body axis deformation at various stages of embryonic and subsequent growth in zebrafish. Embryonic body curvature prior to vertebral column formation is associated with defective convergent extension involving the downregulation of wnt5b during gastrulation to disrupt axial development. Defective convergent extension and embryonic body curvature phenotypes were mostly rescued by the overexpression of wnt5b and RhoA, key molecules in the Wnt/PCP signaling pathway. An embryonic lethal phenotype could be alleviated by chimeric expression of lbx1b under the control of the GATA2 minimal promoter and the lbx1b enhancer in larvae, thus allowing observation of the later onset of the spinal curvature with or without vertebral malformation in zebrafish. Thus, as a step towards better understanding of the genetic pathophysiology of scoliosis, our study provide a new evidence for a pathological role of LBX1 and its zebrafish homologs in body axis deformation.
The most significant SNP associated with AIS (rs11190870) is located in the intergenic region [15]. The nearest gens are LBX1 and FLJ41350, which are 7.5 kb upstream and 8.1 kb downstream of rs11190870, respectively. Using 3C assays, we found that the genome sequence surrounding rs11190870 physically interacts with the LBX1 and FLJ41350 promoters. In luciferase assays, significantly higher promoter activity was detected in the direction toward LBX1, but not toward FLJ41350. EMSAs revealed that some nuclear proteins bound specifically to the genome sequences around rs11190870 with higher affinity to the risk allele. Given that risk variants could disrupt or create a binding site for a transcription factor, any change of LBX1 expression driven by the variants, including downregulation, upregulation, and alteration in temporospatial distribution, would be possible. Expression quantitative trait loci (eQTL) data are available only for peripheral blood cells, which showed no association between the LBX1 expression level and the rs11190870 genotype (Human genetic variation database. (http://www.genome.med.kyoto-u.ac.jp/SnpDB/index.html). However, further studies on eQTL are hampered by a lack of information on which types of tissues or cells are relevant to AIS pathogenesis.
So far, phenotypes associated with CS and AIS have not been reported in Lbx1 null mice and lbx gene knockdown morphants in zebrafish or Xenopus [25, 28–30]. Our lbx1b or lbx2 single knockout zebrafish mutants generated by targeting the first exon using Platinum TALENs [45] also did not exhibit embryonic axial body curvature or scoliosis. The database of Zebrafish Mutation Project also shows that normal development is observed in lbx1a or lbx2 nonsense mutants (https://www.sanger.ac.uk/sanger/Zebrafish_Zmpsearch/lbx1), although information on the associated phenotype of double or triple knockout zebrafish is not available. Previous studies demonstrated a dominant-negative effect by the removal of the engrailed domain from Xenopus Lbx1 that normally functions as a repressor [22, 46]. Injection of lbx1aΔeh, lbx1bΔeh, or lbx2Δeh mRNA did not cause any body curvature as shown in our study. Thus, the current data do not support the possibility that loss-of-function of LBX1 is involved in susceptibility to scoliosis. In contrast, we found a significant increase of promoter activity in the presence of the genomic region with rs11190870 found in the risk allele. Considering that rs11190870 could confer AIS susceptibility by activating LBX1 transcription, it would be reasonable to assume that upregulation of human LBX1 may contribute to some aspects of the pathogenic mechanism in scoliosis.
The ladybird protein is a member of the homeobox transcription factor family with an engrailed repressor domain at the N-terminus [22]. Overexpression of LBX1 and any one of lbx1a, lbx1b, or lbx2 caused defective convergent extension movements that led to curvature of the body axis. Upon overexpression of the lbx genes without the engrailed repressor domain, body curvature was not observed in the embryos, suggesting that Lbx genes negatively regulate their target genes as repressors. Indeed, our in vivo luciferase assays revealed that lbx1b significantly represses the promoter activity of wnt5b. Hence, lbx1b downregulates wnt5b expression during gastrulation at the transcriptional level, thereby causing defective convergent extension followed by deformation of the body axis.
Both canonical and non-canonical Wnt signaling pathways are involved in convergent extension movements during gastrulation. A shortened-curled tail was reported in a Wnt-5 mutant (ppt−/−) with defective convergent extension [47]. AIS- and CS-like scoliosis are also observed in zebrafish mutants of ptk7, which regulates both canonical and non-canonical Wnt signaling activity [37, 48]. The same group identified a novel sequence variant within a single IS patient that disrupted PTK7. In this study, we found that the elevated expression of lbx1 in zebrafish evokes wnt5b downregulation, suggesting that aberrant Wnt/PCP signaling causes defective convergent extension in our experimental model. Interestingly enough, our approach investigating the etiology of scoliosis from the opposite direction also led to the hypothesis that a dysregulated Wnt signaling pathway is involved in both CS and IS pathogenesis.
Non-canonical Wnt/PCP signaling is involved in a variety of events independently of β-catenin [49]. During axis formation in vertebrates, the Wnt/PCP pathway regulates cell polarity and cell motility by modulating the activity of Rho family small GTPases. Especially, RhoA-ROCK signaling mainly acts downstream of wnt5 and wnt11 in zebrafish embryos [50]. Co-injection of mRNA for wnt5b or RhoA mRNA with lbx1b mRNA rescued defective convergent extension leading to embryonic body curvature. These findings strongly support our hypothesis that misregulation of Wnt/PCP signaling induced by lbx1b overexpression is responsible for defective convergent extension followed by body axis deformation.
To date, CS and IS have been considered not to be etiologically relevant, but it has been reported previously that a family history of IS was observed in 17.3% of 237 families with CS [51]. Another study of 31 CS cases also reported that three (10%) had first-degree relatives with IS [52]. The overlapping familial aggregates of CS and IS suggested the possibility of a common cause for these clinically distinct diseases. The uniform overexpression of lbx1b either ubiquitously or in the endogenous expression domain results in severe defective convergent extension leading to morphological defects in both mesoderm and ectoderm patterning followed by early death prior to notochord mineralization to form the spine. In contrast, mosaic expression of lbx1b under the control of the lbx1b enhancer in larvae alleviated the embryonic lethal phenotype with body curvature and thereby allowed the later onset of scoliosis with or without vertebral malformation in zebrafish. The F1 embryos generated by the AIS- and CS-like mosaic transgenic founders presented with severe body axis deformation including convergent extension defects and body curvature. Thus, our observations that the elevated expression of lbx1b causes both AIS- and CS-like scoliosis may provide a new perspective for the shared genetic basis of AIS and CS.
Some of the founder Tg(GATA2-1b:lbx1b) with a displaced dorsal melanophore stripe without apparent notochord deformity developed scoliosis with rotation around the longitudinal axis of the body, but without visible vertebral malformations. These results suggest that subtle deformities in the early body axis may be later accentuated during the growth spurt. In fact, AIS patients appear to be quite normal until adolescence. It is reasonable to postulate that an early event such as defective axial development resulting from the upregulation of LBX1 may be too mild to be detected in potential AIS patients until the growth spurt. In a late-onset polygenic disease such as AIS, even such subtle abnormalities may be sufficient to accumulate growth irregularities and greatly aggravate biomechanical instability during adolescence in association with additional genetic or environmental factors. However, at present, considering that the GATA2 minimal promoter and an lbx1b enhancer could drive lbx1b expression in neural tissue later in development [44], we cannot exclude the possibility that lbx1b expression after convergent extension causes idiopathic scoliosis. Thus, we need to determine carefully the mechanism by which lbx1b causes the AIS-like phenotype in the mosaic transgenic founders.
Polygenic diseases including AIS are triggered by the combination of a number of susceptibility genes whose individual contribution may be relatively small. It is also considered that these diseases could occur when a threshold of quantitatively-varying risk or liability influenced genetically and environmentally is exceeded [53]. Unlike a monogenic disease caused by a mutation in one gene, it appears that the cumulative effects combined with additional factors for a relatively long time lead to the onset of clinical manifestations of AIS, even though the contribution of each individual gene is small. Our study provides a new evidence for the possible involvement of LBX1-induced mild defects during embryonic axial development in AIS susceptibility. As the faithful recapitulation of the late-onset polygenic disease in the animal model has not been generally established yet, our current experimental approaches are still fraught with limitations. Further studies are necessary for establishment of a genetic animal model recapitulating the expression of LBX1 in an analogous way to that in AIS patients.
All of the animal experimental procedures used in this study were approved by the Animal Care Committee of the Institute for Frontier Medical Sciences, Kyoto University and conformed to institutional guidelines for the study of vertebrates
Rhabdomyosarcoma cells and A172 human glioblastoma cells (A172 cells) from ATCC and HeLa cells were obtained from the Japanese Collection of Research Bioresources Cell Bank (Osaka, Japan). The cells were maintained at 37°C under 5% CO2 in Dulbecco’s modified Eagle’s medium-high glucose supplemented with penicillin (50 U/mL), streptomycin (50 g/mL), and 10% fetal bovine serum. The cells were crosslinked with 37% formaldehyde solution at a final concentration of 1% in a 37°C dry incubator for 10 min, followed by an additional incubation at 4°C for 2 h. The crosslinked protein-chromatin material was purified by 8 M urea ultracentrifugation and digested with Sau3AI as described previously. A 2-g aliquot of chromatin was diluted in a ligation buffer and ligated with T4 DNA ligase (Fermentas) for 4 h. After reversing the crosslinks, the ligated DNA was amplified by PCR with various combinations of primers using GoTaq Hot Start Master Mix (Promega).
We prepared nuclear extracts from rhabdomyosarcoma and A172 cells as described previously [54]. We prepared probes for the risk (R) and non-risk (N) alleles of rs11190870 by annealing 17-bp complementary oligonucleotides and labeling with digoxigenin (DIG)-11-ddUTP (Roche). For competition experiments, nuclear extracts were pre-incubated with excess unlabeled probes. We detected DNA-protein complexes using a DIG gel shift kit according to the manufacturer’s instructions (Roche).
We amplified the LBX1 promoter fragment (-917 to +153) in both directions by PCR and cloned them into the pGL4.10 promoter-less luciferase reporter vector (Promega). The constructs were co-transfected with the pGL4.73 Renilla luciferase vector (hRluc/SV40) as an internal control. Transfection of each construct was performed using TransIT-LT1 (Mirus Bio LLC). HEK 293T cells were maintained at 37°C under 5% CO2 in Dulbecco’s modified Eagle’s medium-high glucose supplemented with 10% fetal bovine serum. Transfection was performed with Lipofectamine LTX and PLUS reagent (Life Technologies). After 24 h of transfection, the cells were harvested and luciferase activity was measured using a Pick&gene dual luciferase detection kit (Toyo B-Net Co.).
The RIKEN Wako (RW) strain and AB strain were obtained from the Zebrafish National BioResource Center of Japan (http://www.shigen.nig.ac.jp/zebra/) and Kondoh ERATO Laboratory, respectively. Adult fish were maintained under a 14 h light–10 h dark cycle at 28°C. Embryos were kept at 28°C and staged by hpf or dpf [55]. The RW strain was subjected to micro-injection and whole-mount in situ hybridization. The AB strain was used for the preparation of total RNA. The established line Tg(UAS:EGFP) [41] was generously provided by Dr. Koichi Kawakami (National Institute of Genetics).
Specific primers for zebrafish lbx1a (NM_001025532), lbx1b (NM_001163312), and lbx2 (NM_001007134), and human LBX1 (NM_006562.4), FLJ41350 (NR_029380), and RhoA (NM_001664.2) were designed based on the nucleotide sequences from GenBank (S1 Table). The cDNAs were amplified by PCR from a cDNA library and cloned into the pCS2(+) vector. Deletion constructs of the engrailed domain and homeodomain of lbx1 were generated by inverse PCR. Capped mRNAs were synthesized using an SP6 RNA polymerase in vitro transcription kit (Life Technologies) and purified using a MEGAclear Kit (Life Technologies) according to the manufacturer’s instructions. A mixture containing 50/100/150 pg mRNA for lbx1a, lbx1b, lbx2, LBX1, and FLJ41350, 15 pg mRNA for RhoA, and 40 pg mRNA for wnt5b and RAC1 was injected into the cytoplasm of one-cell-stage embryos.
Highly active Platinum TALENs were constructed using two-step Golden Gate assembly method as described previously with a slight modification [45]. DNA-binding modules were assembled with the two-step Golden Gate method using the Platinum Gate TALEN Kit (Addgene, Kit #1000000043). pCS2-based vectors were used as destination vectors. The target sequence was 5’-TAAACCCCCTGGACCACcttccaccacccgcgAGCTCCAACAA GCCCTTA-3’, where uppercase and lowercase letters indicate lbx1b TALEN recognition sequence and spacer sequence, respectively. We found polymorphism of RW WT in the left TALE-binding sequence; TAAACCCCCTGGACCAC and TGAATCCCCTGGACCAC, both of which are silent mutations. The target sequence was 5’-TTGCAGTCCAGCGGCGAG gagaggcggcggggtCCCTTGGACCAACTCCCA-3’, where uppercase and lowercase letters indicate lbx2 TALEN recognition sequence and spacer sequence, respectively.
Genomic DNA was extracted from the caudal fins of lbx1b TALENs mRNA-injected zebrafish. For sequencing, PCR products were amplified from the genomic DNA and phosphorylated by T4PNK (TAKARA BIO INC.), and then subcloned into EcoRV site of pBluescript II SK(+) vector. We identified F0 fishes carrying multiple mutations in the target site, and then generated and screened a F1 fish with a nonsense mutation by crossing the F0 and wild type fishes. lbx1b+/- and lbx2+/- mutant were generated by crossing F1 and wild type fishes. lbx1b-/- and lbx2-/- were further generated by intercrossing lbx1b+/- and lbx2+/-, respectively.
The Tol2 transposon/transposase system [41, 56–59] was employed for the establishment of transgenic zebrafish. The coding sequence of lbx1b was cloned into pME-MCS to generate pME-lbx1b. The complementary sequence of the E1b promoter, EGFP, and polyA were cloned into p5E-UAS-E1b to generate p5E-polyA-EGFP-E1b-UAS-E1b. The driver construct (Fig 3A) was generated by recombining p5E-hsp70I, pME-Gal4VP16, p3E-polyA, and pDestTol2CG2 with Gateway LR Clonase II Enzyme mix (Life Technologies). Similarly, the responder construct (Fig 3A) was generated by recombining p5E-polyA-EGFP-E1b-UAS-E1b, pME-lbx1b, p3E-polyA, and pDestTol2CG2. Capped mRNA of medaka Tol2 transposase was prepared by in vitro translation as described above. A mixture containing 50 pg transposase mRNA and 40 pg Tol2 transgenic plasmid was injected into the cytoplasm of one-cell-stage embryos. F1 fish were acquired by outcrossing EGFP-positive F0 with RW fish, and screened by cardiac fluorescence. F2 lines were then generated by outcrossing F1 and RW fish. All kept F2 lines yielded about 50% EGFP-positive progeny when mated to RW fish, which suggested there was a single Tol2 insertion site. The lbx1b enhancer located from +1316 to +2383 bp downstream of the lbx1b transcription start site [44] was cloned into pME-MCS. The enhancer activity in vivo was confirmed using zebrafish enhancer detection (ZED) [60]. To generate p5E-lbx1b enhancer-GATA2, the 2.3-kb BamHI fragment from the ZED-lbx1b enhancer, was cloned into the BamHI site of p5E-MCS. The constructs GATA2-1b:lbx1b or GATA2-1b:MCS were generated by recombining p5E-lbx1b enhance-GATA2, pME-lbx1b or pME-MCS, p3E-polyA, and pDestTol2CG2.
Embryos at 4 hpf or 12 hpf in E3 buffer were placed on block incubator and heated up to 38°C gradually, and then maintained at 38°C for 30 min. After heat shock treatment, they were gradually cooled to 28.5°C. The heat shock treatment causes neither an anomaly nor a decrease in viability.
Larvae were homogenized with a Dounce tissue grinder and lysed with lysis buffer (50 mM Tris-HCl [pH 7.5], 150 mM NaCl, 1% NP-40, 0.1% SDS) containing protease inhibitor cocktail (Roche). Lysates were mixed with 5× Laemmli sampling buffer containing 100 mM DTT and boiled at 95°C for 3 min. Proteins were separated by SDS-PAGE and transferred onto PVDF membranes (Merck Millipore). After blocking with BLOCKING ONE (Nacalai Tesque), the membranes were incubated with primary antibodies in phosphate-buffered saline containing 0.1% Tween 20 and 10% BLOCKING ONE, followed by incubation with horseradish peroxidase-conjugated secondary antibodies. Signals were detected with SuperSignal West Pico Chemiluminescent Substrate (Thermo Scientific) and images were captured by ImageQuanta LAS 4000 (GE Healthcare Bio-Sciences).
Whole-mount in situ hybridization was performed as described previously [61]. Sense and antisense riboprobes for hgg1, dlx3b, ntl, papc, uncx4, wnt5b, wnt11 [42], chd and vox [48] were generated by in vitro translation using a digoxigenin (DIG) RNA labeling kit with T7 or T3 RNA polymerase (Roche). Hybridization signals were detected with an alkaline phosphatase-conjugated anti-DIG antibody (Roche) according to the manufacturer’s instructions. For quantification, the image colors of in situ hybridization were inverted, and Area, Integrated Density, and Mean Gray Value were measured by ImageJ. The corrected Gray Value = Integrated Density − (Area of the selected embryos × Mean Gray Value of background readings).
Total RNA was extracted from zebrafish embryos injected with lbx1b mRNA using an RNeasy Plus Mini kit (QIAGEN). Two hundred nanograms of total RNA were used to synthesize cDNA with a PrimeScript RT reagent Kit (Takara Bio). Quantitative RT-PCR was performed using SYBR Premix Ex Taq II (Takara Bio) on a StepOne instrument (Life Technologies). Relative mRNA expression was normalized to ef-1α and calculated using the 2−ΔΔCT method. Specific primers for quantitative RT—PCR are listed in S1 Table.
A 569-bp insulator of chicken β-globin (BGI) and firefly luciferase (luc) were amplified from the ZED vector and pGL3-basic vector (Promega), respectively. The resultant amplification products were cloned into pME-MCS to construct the promoter-less pME-BGI-luc plasmid. Two fragments of approximately 2 kb upstream of each transcription start site of wnt5b were amplified from zebrafish genome DNA with the primers listed in S1 Table. These were cloned into pME-BGI-luc, pME-BGI-P1-luc plasmid (P1), and pME-BGI-P2-luc (P2) by the In-Fusion technique (Clontech). A mixture containing fluorescein isothiocyanate (FITC)-dextran (SIGMA) and luciferase plasmids with or without lbx1b mRNA was injected into one-cell-stage embryos. FITC fluorescence intensity was quantified using a fluorescence microscope (Leica MZ 16 FA) and ImageJ software. Embryos were then lysed individually and luciferase activity was measured as described previously [62]. The measured activity was normalized by the FITC fluorescence intensity of an individual embryo.
Bones in fixed larvae were stained with alizarin red (Wako).
Vertebral bone morphology of adult zebrafish was analyzed by micro-computed tomography scans with inspeXio SMX-90CT (SHIMADZU). Three-dimensional reconstruction and videos were generated with ImageJ software.
Embryos were examined and scored for relevant phenotypes. Statistical analysis (SPSS 16.0) was performed by chi-square analysis for enumeration data and independent-samples t test or one-way ANOVA for measurement data to calculate p values under various conditions. Spearman’s correlation between relative fluorescence intensity and body curvature severity was calculated. A linear regression equation was calculated with SPSS.
|
10.1371/journal.ppat.1006367 | Mycobacterium tuberculosis subverts negative regulatory pathways in human macrophages to drive immunopathology | Tuberculosis remains a global pandemic and drives lung matrix destruction to transmit. Whilst pathways driving inflammatory responses in macrophages have been relatively well described, negative regulatory pathways are less well defined. We hypothesised that Mycobacterium tuberculosis (Mtb) specifically targets negative regulatory pathways to augment immunopathology. Inhibition of signalling through the PI3K/AKT/mTORC1 pathway increased matrix metalloproteinase-1 (MMP-1) gene expression and secretion, a collagenase central to TB pathogenesis, and multiple pro-inflammatory cytokines. In patients with confirmed pulmonary TB, PI3Kδ expression was absent within granulomas. Furthermore, Mtb infection suppressed PI3Kδ gene expression in macrophages. Interestingly, inhibition of the MNK pathway, downstream of pro-inflammatory p38 and ERK MAPKs, also increased MMP-1 secretion, whilst suppressing secretion of TH1 cytokines. Cross-talk between the PI3K and MNK pathways was demonstrated at the level of eIF4E phosphorylation. Mtb globally suppressed the MMP-inhibitory pathways in macrophages, reducing levels of mRNAs encoding PI3Kδ, mTORC-1 and MNK-1 via upregulation of miRNAs. Therefore, Mtb disrupts negative regulatory pathways at multiple levels in macrophages to drive a tissue-destructive phenotype that facilitates transmission.
| The mechanisms whereby Mycobacterium tuberculosis (Mtb) evades host immunity are well described, but Mtb must also engage the host immune response to drive tissue destruction, cavitation and transmission as part of its life cycle. We identify negative regulatory pathways that suppress pathogenic matrix metalloproteinase-1 expression in primary human macrophages, including a previously unidentified role of the MAP kinase-interacting kinase (MNK) pathway in inhibiting protease secretion. Furthermore, these pathways are suppressed in granulomas of patients with culture-proven pulmonary tuberculosis and in infected macrophages in vitro. Stability of mRNA encoding negative regulatory molecules is reduced, and Mtb upregulates multiple microRNAs predicted to target their 3’UTR. Together, these findings demonstrate that Mtb skews the macrophage phenotype towards tissue destruction by disrupting negative regulatory pathways.
| Tuberculosis (TB) is a global pandemic, killing more than any other infectious disease [1], and ongoing transmission in high incidence settings impedes control measures [2]. Mycobacterium tuberculosis (Mtb), the causative organism, must cause lung destruction to create highly infectious individuals with pulmonary cavities who drive the pandemic [3]. Cavitation results from tissue-destructive matrix metalloproteinases (MMPs) [4], in particular MMP-1 from macrophages [5–7]. The signalling pathways driving MMP-1 expression have been described [5, 8], but relatively little is known about regulatory pathways that limit immunopathology in tuberculosis [9].
Pro-inflammatory signalling in LPS-stimulated dendritic cells is negatively regulated by the phosphoinositol-3 kinase (PI3K) signalling pathway, which inhibits IL-12 secretion and TLR signalling [10, 11]. Intracellular signalling is highly complex, with cross-talk between cascades such as the mitogen-activated protein kinase (MAPK), PI3K and MAP kinase-interacting kinase (MNK) pathways [12]. MNKs (MNK1/2) are protein kinases which phosphorylate the translation initiation factor eIF4E and are therefore thought to regulate mRNA translation [13], but have not previously been studied in TB. The precise role of MNK-mediated eIF4E phosphorylation is unclear, but is considered to differentially affect the translation of multiple mRNAs [12–14]. Expression of these intracellular signalling molecules can be regulated by microRNAs [15], and Mtb infection of macrophages can modulate this microRNA profile [16–18].
We hypothesised that negative regulatory pathways in macrophages limit excessive immunopathology in TB, and that the pathogen specifically targets them to exacerbate tissue destruction and consequently transmission. In the present study, we have identified for the first time regulatory pathways which limit MMP-1 production in human macrophages, including PI3K, AKT and mTORC1, and show that PI3K expression is reduced in pulmonary granulomas of patients with TB. Intriguingly, MNK inhibition also increases MMP-1 secretion by as yet an undescribed signalling pathway. Furthermore, Mtb infection suppresses mRNA levels of multiple regulatory pathways in macrophages via augmenting expression of several key micro-RNAs (miRNA). Therefore, the pathogen skews the macrophage response to promote tissue destruction.
As previously demonstrated, Mtb infection of primary human macrophages significantly increased MMP-1 secretion and expression (Fig 1A and 1B). However, global inhibition of PI3K signalling with the pan-PI3K inhibitor LY294002 further augmented Mtb-induced MMP-1 secretion (Fig 1A) and gene expression (Fig 1B). The PI3K pathway has multiple subunits, and we studied the PI3Kδ subunit which is specifically expressed in cells derived from the blood by specific inhibition with IC87114 (PI3Kδ, PI3Kγ and PI3Kβ = IC50 0.5, 29 and 75μM, respectively). Consistent with the global inhibition, PI3Kδ inhibition significantly upregulated MMP-1 secretion (Fig 1C) and MMP-1 gene expression (Fig 1D). LY294002 did not suppress expression of PIK3CD, the gene encoding PI3Kδ, in macrophages after 24h of incubation (S1 Fig). We subsequently evaluated AKT phosphorylation, immediately downstream of PI3K within the signalling cascade. Mtb infection induced AKT phosphorylation at Ser 473 within 30 minutes, peaking at 60 minutes and was still evident at 120mins. This downstream signalling was completely inhibited by LY294002 (Fig 1E and S2 Fig). To address the question of potential off-target effects of the inhibitors, we evaluated an additional PI3K/PDK1 inhibitor, NVP-BAG956, and again demonstrated increased MMP-1 secretion after inhibition (Fig 1F). To investigate whether MMP-1 upregulation was augmented by intercellular networks, we stimulated macrophages with conditioned media from Mtb-infected monocytes. MMP-1 secretion was increased after stimulation with media from infected macrophages, demonstrating that intercellular networks can augment MMP-1 driven by Mtb infection (S3 Fig).
To determine whether the PI3K regulation of MMP-1 was specific or part of a more widespread phenomenon, we profiled secretion of MMPs, cytokines, chemokines and growth factors by uninfected and infected macrophages using a Luminex array. Mtb infection upregulated numerous mediators, and inhibition of PI3K signalling had an additive effect, causing further upregulation of secretion of multiple MMPs, TH1 and TH2 cytokines, chemokines and growth factors (Fig 2). In the context of PI3K pathway inhibition, the majority of pro-inflammatory mediators were upregulated.
Next, we investigated PI3Kδ signalling in lung lesions of patients with confirmed pulmonary TB. Initially, we performed immunohistochemistry for phosphorylated PI3K to determine activation in vivo, but were unable to demonstrate immunoreactivity. Therefore, we analyzed total PI3Kδ, but again were unable to detect expression. In normal lung tissue, alveolar macrophages expressed both CD68 and PI3Kδ (Fig 3A and 3B). Within tuberculosis granulomas, CD68 expression is widespread (Fig 3C) and multinucleate giant cells also express CD68 (Fig 3E). In contrast, expression of PI3Kδ is globally absent throughout the granuloma (Fig 3D) and also absent in macrophages and multinucleate giant cells (Fig 3F). Positive controls excluded a technical cause for the absent PI3Kδ staining, and analysis of further granulomas confirmed the absence of PI3Kδ despite widespread CD68 expression (S4 Fig).
Therefore, we examined the effect of Mtb infection on macrophage PI3Kδ expression. Mtb upregulated expression of MMP-1 gene expression at 24h (Fig 3G), but in the same cells significantly suppressed expression of PIK3CD, the gene endcoding PI3Kδ (Fig 3H). Therefore, the increase in MMP-1 expression is accompanied by suppression of PI3Kδ in both patients and primary human macrophages. Stimulation of macrophages with conditioned media from Mtb-infected cells did not suppress PIK3CD, suggesting that intercellular networks were not the primary driver of PIK3CD suppression.
We then studied the signalling pathway directly downstream of PI3Kδ, which includes AKT and mTORC1. Inhibition of AKT signalling, which is phosphorylated following PI3Kδ activation, similarly upregulated MMP-1 secretion from macrophages (Fig 4A). Similarly, mTORC1 inhibition with rapamycin augmented Mtb-driven MMP-1 secretion at 72h (Fig 4B), and this associated with increased MMP-1 gene expression at 24h. Luminex profiling of MMPs and pro-inflammatory cytokines demonstrated a global effect of upregulation of MMPs and cytokines after mTORC1 inhibition (S5 Fig), as was observed for PI3K inhibition (Fig 2).
We then investigated whether these inhibitory pathways increased MMP-1 via crosstalk with the pro-inflammatory MAPK pathways, which regulate MMP-1 secretion in macrophages [8]. PI3K inhibition did not increase p38 MAPK phosphorylation at either 30 or 240 minutes (Fig 4D and S2 Fig), nor ERK MAPK phosphorylation at these time points (Fig 4E). We measured cyclo-oxygenase II (COX-II) accumulation, which regulates MMP-1 downstream of p38, but PI3K inhibition did not alter COX-II levels within infected macrophages (Fig 4F). Finally, we investigated whether PI3K linked with NFκB signalling. Mtb increased the nuclear translocation of p65 in macrophages, but this was not altered by LY294002, suggesting that crosstalk was not occurring at this level (Fig 4G). Therefore, the PI3K pathway suppresses MMP-1 expression independent of regulating the MAPK, COX-II or NFκB axes.
Next, we investigated the MNK pathway, which is downstream of p38 MAPK and regulates mRNA translation. We hypothesised that MNK inhibition would suppress MMP-1 production, but surprisingly MNK inhibition significantly augmented Mtb-driven MMP-1 secretion (Fig 5A). The increased secretion was secondary to increased MMP-1 gene expression (Fig 5B). To confirm that this is indeed an effect of disabling MNK function, we studied bone marrow-derived macrophages from mice in which the genes encoding MNK1 and/or MNK2 (termed Mknk1 and Mknk2) have been disrupted [19]. Mice lack an orthologue of MMP-1 [20], and so we analyzed MMP-3 secretion, which is regulated in a very similar manner to MMP-1 [21]. Mtb infection increased MMP-3 secretion by murine wild type macrophages, and MMP-3 secretion was markedly higher in the MNK double knock-out cells, confirming that the effect of the MNK inhibitor does indeed reflect a negative input from the MNKs to MMP-3 secretion (Fig 5C). To confirm the efficacy of the MNK-I1 inhibitor [22], we performed Western blotting for phosphorylated eIF4E, a component of the eIF4F translation complex which is specifically phosphorylated by the MNKs and the only validated in vivo substrate. MNK inhibition suppressed eIF4E phosphorylation as expected (Fig 5D and S6 Fig). To determine if this effect involved the p90RSK pathway, which is downstream of ERK MAPK, we specifically inhibited this with BI-D1870 and found no change in MMP-1 secretion (Fig 5E), thereby demonstrating the crosstalk was not via this pathway.
To determine whether the negative regulatory effect of the MNKs was global, as we had observed for the PI3K/AKT/mTORC1 axis, or more specific, we performed luminex profiling of MMPs and cytokines. We demonstrated that the MNK effect was relatively specific to MMP-1, 3 and 10, and only significantly upregulated MCP-1 and EGF amongst the other inflammatory mediators studied (Fig 6). Blockade of the MNK pathway suppressed secretion of the majority of TH1 and TH2 cytokines, whereas they had been augmented by PI3K/mTORC1 pathway inhibition.
We sought to identify a point where the negative regulatory pathways mediated by PI3K and MNK converge. First, we studied the effect of inhibition of the eIF4F translation initiation complex, which is required for 5’-cap-dependent protein synthesis. Inhibition of the translation complex, by disrupting the eIF4E and eIF4G interaction with 4EGI-1 [23], significantly suppressed MMP-1 secretion (Fig 7A), demonstrating that disrupting the translation factor complex suppressed synthesis as expected. We therefore then studied eIF4E phosphorylation. The positive control, insulin, an agonist of the PI3K pathway, increased eIF4E phosphorylation (Fig 7B, lane one and S6 Fig), whereas inhibition with LY294002, in the absence of any stimuli, inhibited eIF4E phosphorylation (lane two). MNK inhibition both without and with Mtb completely suppressed eIF4E phosphorylation as expected. In the context of Mtb stimulation, specific inhibition of PI3K/PDK1 reduced eIF4E phosphorylation (lane 8), providing evidence of convergence between MNK and PI3Kδ signalling at this level.
Finally, to investigate the mechanism whereby Mtb suppresses PI3Kδ expression, we studied mRNA stability within macrophages. First, we analyzed total cellular mRNA and demonstrated that Mtb not only suppressed PIK3CD in macrophages (Fig 8A), but also significantly suppressed expression of MLST8, a key subunit of mTORC1, and the MNK1 gene, MKNK1 (Fig 8B and 8C). To determine whether this suppression was specific to Mtb, we studied different microbial stimuli. Mtb, TLR-2 stimulation and zymosan, a fungal wall component, each suppressed MKNK1 gene expression in macrophages, while LPS and purified mycolic acid did not (S7 Fig).
To investigate the underlying mechanism, we performed mRNA pulldown experiments to characterise newly transcribed gene expression following Mtb infection. 4-Thio uridine was used to label mRNA and determine the ratio of newly-transcribed mRNA with total mRNA levels, and thereby demonstrate whether changes were due to altered synthesis or stability. Newly-transcribed PIK3CD mRNA expression was increased in infected cells, and MLST8 and MKNK1 were not suppressed (Fig 8D), suggesting the significantly reduced total mRNA levels were secondary to reduced mRNA stability. This suggested a role for post-transcriptional regulation via microRNAs increasing mRNA degradation, and we therefore analyzed microRNAs predicted by bioinformatic approaches to target these three mRNAs, using miRBase v21. Mtb infection upregulated multiple microRNAs, including miR27a, miR125b and miR199, all of which are predicted suppressors of the PIK3CD, MLST8 and MKNK1 mRNAs (Fig 8E). To confirm that PIK3CD could be targeted by these miR’s, we generated reporter constructs comprising the 3’-UTR fused to Renilla-luciferase. Transfection of HeLa cells with a plasmid expressing miR7, which is predicted to target PIK3CD, suppressed luminescence, whilst a site-directed mutant within the predicted binding region did not, confirming that miR-7 may directly bind to PIK3CD to reduce mRNA or protein levels (Fig 8F). Therefore, the reduced stability of mRNA is most likely secondary to Mtb-induced upregulation of microRNAs that target these transcripts.
Mtb must cause pathology to be transmitted to new hosts, and patients with pulmonary cavities are the most infectious [3, 24]. Whilst the mechanisms whereby Mtb evades the host immune responses have been extensively investigated [25], relatively little is known about how Mtb engages the immune response to drive tissue destruction, cavitation and transmission [26]. Immune evasion will only lead to latent TB without onward transmission, and so the initiation of immunopathology is an essential event in the Mtb life cycle [27]. We have identified negative regulatory pathways that limit pathogenic MMP-1 secretion by primary human macrophages and observed the absence of PI3Kδ in TB granulomas in patients. We demonstrated that Mtb disables the PI3Kδ/AKT/mTORC1 and MNK regulatory pathways to drive a pro-tissue destructive phenotype, thereby uncovering a previously unidentified role of the MNK pathway as inhibitor of MMP-1 expression (Fig 9).
Direct macrophage infection by Mtb was required in vitro for suppression of PIK3CD, while widespread suppression was observed in lung granulomas, where mycobacteria are relatively sparse [28]. This suggests that either accumulation of Mtb antigens within granulomas may suppress PIK3CD expression in non-infected cells, or that a more complex regulation occurs during the long host-pathogen interaction in patients relative to the short-term cellular experiments possible in vitro. Since intercellular networks can upregulate MMP-1 secretion without suppressing PIK3CD, we propose that the down-regulation of the negative regulatory pathways serves to further augment a tissue-destructive proteolytic pathway caused by both direct infection and intercellular signalling to facilitate cavitation and transmission. Interferon-γ, a key cytokine in the host immune response to TB, also targets mTORC1 and MNK signalling [29], demonstrating the complex interplay between pathways and consistent with the emerging hypothesis that either an insufficient or excessive host immune response may be deleterious [30].
Pathology in TB results from dysregulation of inflammation [31] and MMPs are emerging as key pathological mediators [4]. p38 and ERK MAPKs are positive regulators of MMP expression, and p38 is phosphorylated in patients [8], while PI3K is a negative regulator in stromal cells [32]. We demonstrate in primary human macrophages that PI3K, AKT, mTORC1 and MNKs are all negative regulatory pathways, and are suppressed by Mtb infection. Consistent with our findings, PI3K limits IL-12 secretion and TLR signalling in LPS-stimulated dendritic cells [10, 11], suggesting that these early signalling events that occur immediately after receptor activation having a broad regulatory effect to diverse stimuli and lead to increased cytokine secretion. A similar role for PI3Kγ in controlling a macrophage switch between immune stimulation and suppression in cancer has very recently been described in murine macrophages [33] In that study, the signalling was via NFκB, but we were unable to demonstrate cross-talk at this level. Similar to our results in primary macrophages, in human PBMCs, Mtb has been shown to phosphorylate AKT and mTORC1, and rapamycin increases secretion of TNF-α, IL-1β and IL-6 [34]. The skewing of intracellular signalling in macrophages by Mtb may negate the effect of TH2 cytokines, such as IL-4 and IL-10, and regulatory T cells that are thought to limit immune-mediated tissue destruction in TB [30, 35].
We have identified a novel role for MNK signalling in limiting tissue-destructive MMPs and found that this was relatively specific, whereas the effect of PI3K/AKT/mTORC signalling on inflammatory mediators was much more widespread. The effect of MNK inhibition was striking and more pronounced that phenotypes observed with stimuli other than Mtb. MNK is downstream of p38 and ERK MAPK signalling [13], and is considered to regulates protein synthesis [36]. Therefore, we had predicted that MNK inhibition would suppress MMP-1. MMP-1 downregulation was observed when we directly suppressed eIF4F complex formation, whereas MNK inhibition increased MMP secretion while suppressing cytokines. We confirmed that the MNK inhibitor, MNK-I1, was acting specifically by blocking MNK function by studying MNK-deficient mouse cells. The majority of cytokines were suppressed by MNK inhibition, and so the overall effect may be to increase destruction of the extracellular matrix in the absence of an increased host inflammatory response that might favour bacterial killing. We demonstrated that other, but not all, infectious stimuli suppress MKNK gene expression, suggesting that this response may occur to a range of pathogen-derived molecules, and is usurped by Mtb to increase matrix destruction. This phenomenon requires further systematic dissection for full characterisation. MNK is thought to increase translation [12, 37], and therefore the phenotype of increased MMP-1 after MNK inhibition may be due to reduced inhibitory transcription factor production, as has been demonstrated for type I interferons in viral infection [38]. Therefore, our findings suggest that modulating MNK will exert a more nuanced effect within the cellular machinery than purely inhibiting protein synthesis.
Our RNA analysis suggested Mtb infection reduced stability of mRNAs encoding negative regulatory proteins, and we identified Mtb-driven increases in microRNAs that target the signalling cascade at multiple levels. We demonstrated that miR-7 binds to the 3’UTR of PIK3CD, as predicted bioinformatically. miRNAs are responsible for fine tuning transcription and usually multiple miRNAs combine to reduce the levels of a given mRNA with an additive effect [15]. We identified that multiple miRNAs predicted to target each negative regulatory pathway [39] were increased in Mtb infection, suggesting an overall effect skewing the transcriptional response to a matrix-degradative phenotype. However, systematic characterisation of each miR interaction with each regulatory pathway will be required to fully confirm this hypothesis. We demonstrated that intercellular networks could augment MMP-1 secretion, but did not suppress PIK3CD gene expression, indicating that dual regulatory effects are likely to be operant in vivo, with infected cells being particularly predisposed to excessive protease secretion. Cytokine networks may augment MMP secretion by bystander cells in the absence of PIK3CD suppression. Since mice do not express MMP-1 [40], further dissection of this regulatory pathway will likely require gene editing in advanced human cell culture model systems.
In summary, we identify a novel strategy employed by Mtb to drive immunopathology by disrupting negative regulatory pathways, both in primary macrophages and in patients. We demonstrate a previously unrecognised role of the MNK pathway as a negative regulator of MMP secretion. Inhibitors of PI3Kδ are already in clinical use, and other drugs are in development, and therefore it is possible that they may increase the risk of active TB in the same way as anti-TNF-α agents [41]. Furthermore, novel host-directed therapies that target intracellular signalling pathways to enhance Mtb killing [42] must not inadvertently suppress the negative regulatory pathways and thereby augment immunopathology.
Samples used in this study were sourced from the Southampton Research Biorepository, University Hospital Southampton NHS Foundation Trust and University of Southampton, Mailpoint 218, Tremona Road, Southampton, SO16 6YD. Lung biopsy tissue was taken as part of routine clinical care and tissue blocks excess to diagnostic testing were analyzed in this study. The project was approved by the Institutional Review Board (Reference 12/NW/0794 SRB04_14). The ethics committee approved the analysis of this tissue without individual informed consent since it was surplus archived tissue taken as part of routine care. For analysis of blood from healthy donors, this work was approved by the National Research Ethics Service committee South Central—Southampton A (ref 13 SC 0043) and all donors gave written informed consent.
Standard laboratory reagents were from Sigma Aldrich. Chemical inhibitors were pan-PI3K: LY294002; PI3Kδ: IC87114; mTORC1: Rapamycin (Merck Millipore); AKT: MK-2206 (Insight Biotechnology); PI3K/PDK-1: NVP-BAG956; eIF4E/eIF4G interaction: 4EGI-1 (Merck Chemicals); P90RSK: BI-D1870 (Selleckchem). MNK-I1 was kindly synthesised by Professor Jiang Tao and her colleagues at the Ocean University of China, Qingdao, China. Infectious stimuli were: Zymosan (Sigma, 100μg/ml), LPS (Sigma, 1μg/ml), TLR-2 agonist Pam Cys-Ser-(Lys) [Pam3Cys] (MerkMillipore, 100ng/ml), Mycolic acid (Sigma, 10 μg/ml).
PBMCs were isolated from single donor leukocyte cones (National Health Service Blood and Transfusion, Southampton, UK) or fresh blood from healthy donors by density gradient centrifugation over Ficoll-Paque (GE Healthcare Life Sciences). Monocytes were plated at 250,000cm2, adhered for 1 hour and then washed 3 times to remove non-adherent cells. Monocytes were matured to macrophages for 4 days in complete RPMI with 10% human serum with 100ng M-CSF, then rested for 1 day in complete medium without growth factors, and then the experiment was started (Day 0).
M. tuberculosis H37Rv (Mtb) was cultured in Middlebrook 7H9 medium (supplemented with 10% ADC, 0.2% glycerol and 0.02% Tween 80) (BD Biosciences, Oxford) with agitation. Cultures at 1x 108 CFU/ml Mtb (OD = 0.6) was used for all experiments.
Macrophages were infected with Mtb at MOI of 1. Cells were washed one hour after infection to remove non-phagocytized Mtb. Experimental duration was between hours for analysis of protein phosphorylation and 3 days for secretion. For experiments involving chemical inhibitors, cells were pre-incubated with inhibitor for 2h prior to infection with Mtb.
MMP-1 concentrations were analyzed by ELISA assay according to manufacturer’s protocol (R & D Systems). For multianalyte profiling, MMP and cytokine concentrations in cell culture supernatants harvested at 72h were analyzed on a Bioplex 200 platform (Bio-Rad, Hemel Hempstead, U.K.). MMP concentrations were analyzed by the MMP Fluorokine multianalyte profiling (R&D Systems, Abingdon, U.K) and cytokine concentrations were measured using the Cytokine Human 30-Plex Panel for the Luminex platform (Invitrogen, Paisley, UK) according to manufacturers’ protocol.
MDMs were infected with Mtb and lysed with 200μl SDS sample buffer (62.5mM Tris pH 6.8, 2% SDS, 10% glycerol, 50mM DTT, 0.01% Bromophenol blue) at defined time points. Samples were filtered through 0.2μM Anopore filter and frozen at -80°C. 20μl aliquots were heat denatured and run on a 10% acrylamide gel at 200V (Running buffer 25mM Tris base, 192mM glycine, 0.1% SDS) for 3h. Gels were electro-transferred to a nitrocellulose membrane (Amersham) and blocked for 1h with 5% milk protein / 0.1% Tween-20. The membrane was incubated with primary antibody (AKT or phospho-AKT Ser 473, p38 or phospho-p38, ERK or phospho-ERK, Cell Signalling Technology; eIF4E and phosphor eIF4E, Merck Millipore) in 5% BSA / 0.1% Tween at 4°C overnight. Blots were then washed three times and incubated with HRP-linked anti-rabbit secondary antibody (Cell Signalling Technology, 1/2000 dilution in 5% milk protein / 0.1% Tween) for 1h. Luminescence was detected with the ECL system (Amersham) according to manufacturer’s protocol. Immunoblotting for total p38, ERK or AKT confirmed equal loading between samples.
MDMs were lysed with TRI-reagent (Sigma-Aldrich), isolated with chloroform phase separation and precipitated with isopropanol. RNA was quantified by NanoDrop and retro-transcribed to complementary DNA by the High Capacity cDNA Reverse Transcription kit (Life Technologies, Paisley, UK). cDNAs obtained were then used for micro RNA and gene expression quantification assays by RT-qPCR for MMP-1 (Hs00899658_m1), PIK3CD (Hs00192399_m1), MLST8 (Hs00909882_g1), MKNK1 (Hs00374375_m1) and GAPDH (Hs02758991_g1) following manufacturer's instruction (Applied Biosystems, USA). For microRNA’s, miRNA primers were from TaqMan OpenArray Human MicroRNA Plate (ThermoFisher Scientific, UK), with the following catalogue numbers: miR-7 TM000268, miR-27a TM000408, miR-30a TM000417, miR-221 RT524, miR-125b TM449, miR-199 TM498, miR-22 TM000398, miR-135a TM000480. Taqman Universal master mix and primers specific for the gene of study and GAPDH as house-keeping gene were used. Each RT-qPCR experiment was performed in duplicates and results were analyzed using SDS version 2.3 sequence detection systems (Applied Biosystems, USA). Comparative CT method was employed to analyze all RT-qPCR data.
Paraffin-embedded lung tissues were mounted at 4μm thin onto APS coated glass slides and dried. Sections were dewaxed and 30% hydrogen peroxide used to block endogenous peroxidase. Sections were washed three times in TBS buffer and heat induced-epitope retrieval employed by boiling the slides in 1mM EDTA (pH 8.0) in distilled water for 25 minutes. Slides were incubated in Avidin solution for 20 minutes, followed by 3 washes. This was followed by incubation in biotin solution for 20 minutes followed by another wash step. The slide was blocked with Dulbecco’s Modified Eagle Medium (DMEM) containing with 10% FCS and 2% BSA for 20 minutes. Slides were incubated at 4°C overnight in appropriately diluted primary antibody; 1:1000 dilution of anti-PIK3δ (LifeSpan BioScience,Inc) or ready–to-use monoclonal mouse CD68 antibody, clone PG-M1 (Dako, IR613). Sections were washed and incubated in 1:400 of biotinylated rabbit anti mouse (Dako) secondary antibody for 30 minutes. After a second wash, sections were incubated for 30 minutes in streptavidin biotin-peroxidase complexes (Elite vectastain ABC kit, Vector laboratories). Sections were washed and incubated in DAB (2-component DAB pack, BioGenex) substrate for 5 minutes. Counter staining was performed in Mayer’s haematoxylin for 20 seconds. Dehydration of slides was performed at 1 minute in graded alcohols and mounted in pertex. Images were captured on Olympus CC12 (dotSlide) microscope.
MDMs were infected with Mtb, in the presence or absence of inhibitors, as above. Cells were fixed in 4% paraformaldehyde for 30 minutes at room temperature then lifted mechanically. Anti-COX II FITC labelled intracellular staining was performed according to manufacturer’s instructions (Cayman Chemicals). Briefly, cells were washed and permeabilized with 0.5% BSA, 0.1% Na azide, 0.1% Saponin solution. After a further wash step, cells were incubated with primary antibody (1/10 dilution) or IgG control (1/200 dilution, mouse IgG1 FITC, Serotec) for 30 minutes at room temperature in the dark. Cells were re-suspended in PBS and analyzed by flow cytometry on a Becton Dickinson FACS Calibur.
MDMs were infected with Mtb, in the presence or absence of inhibitors and cytoplasmic and nuclear extracts were prepared using the NE-PER kit (Pierce Biotechnology, Perbio) according manufacturer’s instructions. Briefly, cells were mechanically lifted into ice cold PBS, microcentrifuged and the supernatants removed leaving the pellet as dry as possible. Halt Protease Inhibitor Cocktail (Pierce, Perbio) was added to the stock solutions and the appropriate volume of ice cold CERI was added to the pellet. After re-suspension by vortexing, samples were incubated on ice for 10 minutes and CERII solution was added. Samples were vortexed and after centrifugation at 4°C for 5 minutes the supernatant containing the cytoplasmic extract was immediately collected, and filtered through a 0.2μM Durapore filter (Millipore). Ice cold NER was added to the remaining pellet containing nuclei, and vortexed for 15 seconds every 10 minutes for a total of 40 minutes on ice. After centrifugation the supernatant was harvested, filtered through a 0.2μM Durapore filter (Millipore) and all extracts stored at -80°C.
Transcription factor activation in nuclear extracts was determined by TransAm ELISA based assay kits (Active Motif, UK) according to manufacturer’s instructions. Total protein concentration was measured by Bradford assay (Biorad) and 5μg of nuclear extract was used for each sample. The lower level of sensitivity was <0.5μg of nuclear or whole cell extract.
Bone Marrow Derived Macrophages (BMDM) from MNK1, MNK2 or MNK Double mutant mice were acquired from South Australian Health and Medical Research Institute, Australia. After defrosting, macrophages were cultured in complete macrophage medium, which is composed of L929 cell conditioned medium, 20% FCS and DMEM with 10ng/ml of M-CSF overnight, and were infected with Mtb H37Rv at MOI of 1 for 24 h. Supernatants were harvested for analysis and sterile filtered. Mouse MMP-3 Fluorokine beads (R&D Systems) were used to measure concentrations of MMP-3 in supernatant from BMDM on the Luminex 200 platform (Bio-Rad, Hertfordshire, UK). The lower limit of detection of the assay was 2 pg/ml. Assays were performed per manufacturer’s instructions.
To analyze newly synthesised mRNA, 100μM of 4-Thio uridine was added to infected macrophages for at least 24h, lysed in Tri-reagent and total RNA extracted. 1mg/ml of EZ-Link biotin-HPDP (Pierce, Thermo Fisher Scientific UK Ltd.) solution was prepared in Dimethylformamide (DMF). 1μl of 1mg/ml of EZ-Link biotin-HPDP solution was used per μg of total 4-TU labelled RNA. Master mix was prepared from 1M Tris (pH 7.4) and 0.5M EDTA to a final concentration of 10mM Tris and 1mM EDTA, in RNase free water. Isopropanol at volume equal to the final reaction volume and 5M NaCl (at 1/10th final reaction volume) were added to the biotinylated 4-TU labelled RNA. The mixture was vortexed and incubated at room temperature for 5 minutes before being centrifuged at 13000rpm for 20 minutes. The RNA pellet was washed in 250μl 75% Ethanol, followed by 10 minutes centrifugation. The pellet was allowed to air-dry until semi-transparent, and RNA was re-suspended in 20–50μl of RNase free water. The re-suspended RNA was purified immediately or stored at -80°C for future pulldown analysis.
To isolate and purify the 4-TU labelled RNA, Magnetic Porous Glass (MPG) streptavidin beads (Pure Biotech LLC) was used to bind and pull down the biotinylated 4-TU labelled RNA. The beads were incubated with tRNA (1μg per 5μl of beads) and rotated at room temperature for 20 minutes. Tubes were placed in magnetic stand beads for 1 minute. This was followed by three washes in 300μl of MPG buffer (1M NaCl, 10mM EDTA, 100mM Tris-HCL at pH 7.4 in RNase free water). Beads were re-suspended in MPG buffer equal to the original volume of beads. Volume of RNA and beads were adjusted to be equal, to allow 1:1 combination ratio. The biotinylated 4-TU RNA was added to the beads and incubated at room temperature with rotation for 1 h. Beads were collected in a magnetic stand for 1 minute. The supernatant was collected and kept as unbound, non-4TU RNA. Beads were washed two times in 250μl of room temperature MPG buffer, one wash in 65°C MPG buffer, and a final wash in 50μl MPG buffer. The supernatant from the last wash was kept as ‘wash RNA’ to check for flow through RNA.
To elute the bead-bound, 4-TU labelled RNA, freshly prepared 5% β-mercaptoethanol was added at volume equal to original bead volume and incubated at room temperature with rotation for 20 minutes. The tubes were centrifuged, and beads collected in a magnetic stand for 1 minute. The supernatant was kept as bound, 4-TU labelled RNA. RNA was precipitated as described above by adding 5M NaCl at 1/10th the RNA volume, isopropanol at the same volume as the RNA, and 1μg glycogen. The mixture was incubated at room temperature for 5 minutes and centrifuged for 20 minutes. The RNA pellet was washed in 250μl 75% Ethanol followed by 10 minutes centrifugation. The pellet was allowed to air-dry until semi-transparent, and RNA was re-suspended in 20–50μl of RNase free water. The re-suspended RNA was placed in the magnetic stand to collect any residual beads. The supernatant was collected in newly labelled nuclease free tubes as purified, newly synthesised RNA samples.
Purified 4-TU labelled RNA was treated with rDNase (DNA-free Kit, Life Technologies) to remove potential genomic DNA contamination. SuperScript III Reverse Transcriptase (Invitrogen), Oligo (dt)12–18 primer (Invitrogen) and random hexamer (Component of high capacity cDNA kit, Life Technologies Ltd.) were used to retro transcribe 5μg of the pulled down RNA to cDNA in a 20μl reaction volume according to the manufacturer’s protocol. The mixture was heated at 65°C for 5 minutes, after which tubes were incubated on ice for a further 1 minute. 4μl, 1μl and 2μl of 5x first strand buffer, 0.1MDTT, SuperScript III Reverse Transcriptase respectively (all from Invitrogen) and 1μl of RNaseOUT (Invitrogen) were added. Reverse transcription was performed by incubating tubes at 25°C for 5 minutes, 55°C for 40 minutes and at 70°C for 15 minutes. The cDNA was stored in -20°C for RT-qPCR amplification as above.
To quantify expression of the micro RNAs of interest, cellular RNA was converted to cDNA as above. For micro-RT, 10ng/μl of RNA concentration (diluted using RNase-free water) and stem-loop primers that are specific for each of the microRNA of interest were used. RT-qPCR was performed using TaqMan Gene Expression Assays (Thermo Fisher Scientific).
Regions of interest were first amplified from genomic DNA by PCR using GoTaq G2 Polymerase (Promega) with standard additions to the PCR mastermix. The genomic region encompassing miR-7-3 was amplified using the following forward; CTCGAGGGGTCTCAGACATGGGGCAGAGGG and reverse; AAGCTTCCACTGGCCAGCCCATTGAAGGCG primers with XHOI and HINDIII restriction sites. For the PIK3CD-3’UTR, forward; TCTAGACAAGCACATTGGTCCTAAAGGGGC and reverse; GCGGCCGCAAGGCATCCTGTCGGACAGTAGGC primers with XBAI & NOTI sites were used to amplify a 362nt fragment of the PIK3CD-3’UTR containing the 8mer miR-7-5p binding site. Each product was cloned separately into pCR 2.1-TOPO (Invitrogen) and amplified in plasmid DNA using the TOPO TA cloning kit method (Invitrogen) and chemically competent E. Coli. The amplified sequences of interest were removed from pCR 2.1 TOPO and inserted into pCDNA 3.1 (miR-7) or pRLTK (PIK3CD-3’UTR) using XHOI/HINDIII or XBAI/NOTI restriction sites, respectively.
To assess the specificity of the miR-7-5p putative binding site in the PIK3CD-3’UTR fragment, a 4nt substitution removing sequence complementarity to the miR-7-5p seed sequence was performed by QuickChange Site Directed Mutagenesis (Stratagene), using a previously outlined method [43]. Primer sequences for the PIK3CD-3’UTR mutant were forward; GGATTGTCACCCCAAGGATCCCAGCTGGTGGATCTG and reverse; CAGATCCACCAGCTGGGATCCTTGGGGTGACAATCC.
To determine direct targeting of PIK3CD by miR-7, the pCDNA 3.1_miR-7 construct was co-transfected in to HeLa cells with either pRL-TK_PIK3CD-3’UTR or pRL-TK_PIK3CD-3’UTR_MUTANT vectors, using SuperFect (Qiagen) and manufacturers recommendations. Empty pRL-TK vector and a pCDNA 3.1 construct containing a non-related insert (PU.1 3’UTR) were used as control vectors to elucidate miR-7 activity against PIK3CD-3’UTR. The pGL3 Luciferase reporter (Promega) was used as a normalising vector to assess transfection efficiency. Luminometry was performed using the Dual-Glo Luciferase assay system (Promega) following manufacturer’s instructions. Experimental conditions performed in duplicate were averaged, and 4 independent experiments were performed.
Analysis was performed using Graphpad Prism v6.0. Multiple intervention experiments were compared with the One Way ANOVA followed by Tukey’s multiple comparison. A p value of <0.05 was taken as statistically significant. For secretion data, experiments were performed in triplicate on a minimum of 2 occasions, while for RNA analysis representative data from at least 3 independent experiments is shown.
Uniprot accession numbers for the principal proteins discussed are; MMP-1 P03956, PI3Kδ O00329, mTORC1 P42345, MNK Q9BUB5
|
10.1371/journal.pgen.1007689 | A metabolic checkpoint protein GlmR is important for diverting carbon into peptidoglycan biosynthesis in Bacillus subtilis | The Bacillus subtilis GlmR (formerly YvcK) protein is essential for growth on gluconeogenic carbon sources. Mutants lacking GlmR display a variety of phenotypes suggestive of impaired cell wall synthesis including antibiotic sensitivity, aberrant cell morphology and lysis. To define the role of GlmR, we selected suppressor mutations that ameliorate the sensitivity of a glmR null mutant to the beta-lactam antibiotic cefuroxime or restore growth on gluconeogenic carbon sources. Several of the resulting suppressors increase the expression of the GlmS and GlmM proteins that catalyze the first two committed steps in the diversion of carbon from central carbon metabolism into peptidoglycan biosynthesis. Chemical complementation studies indicate that the absence of GlmR can be overcome by provision of cells with N-acetylglucosamine (GlcNAc), even under conditions where GlcNAc cannot re-enter central metabolism and serve as a carbon source for growth. Our results indicate that GlmR facilitates the diversion of carbon from the central metabolite fructose-6-phosphate, which is limiting in cells growing on gluconeogenic carbon sources, into peptidoglycan biosynthesis. Our data suggest that GlmR stimulates GlmS activity, and we propose that this activation is antagonized by the known GlmR ligand and peptidoglycan intermediate UDP-GlcNAc. Thus, GlmR presides over a new mechanism for the regulation of carbon partitioning between central metabolism and peptidoglycan biosynthesis.
| Bacterial cells are surrounded by a peptidoglycan cell wall that is, under most conditions, required for viability. Synthesis of the cell wall requires a considerable diversion of resources from central carbon metabolism into a lipid-linked precursor (lipid II) that is exported from the cell for wall assembly. Here, we propose that GlmR presides over a new mechanism for the regulation of carbon partitioning between central metabolism and peptidoglycan biosynthesis: GlmR activates the GlmS-dependent diversion of carbon from the glycolytic pathway into peptidoglycan synthesis. This effect is particularly important during gluconeogenesis since the GlmS substrate fructose 6-phosphate is present at a reduced level under these conditions.
| Bacillus subtilis provides a powerful model system for understanding cell wall homeostasis in Gram positive bacteria. Disruption of pathways for the synthesis of peptidoglycan (PG) and other cell envelope components elicits complex adaptive responses often controlled by alternative σ factors or two-component systems [1, 2]. The ECF σ factor σM regulates numerous operons involved in PG synthesis and mutants are sensitive to PG synthesis inhibitors [3]. Previously, we found that mutation of gdpP, which encodes a cyclic-di-adenosine monophosphate (c-di-AMP) hydrolase, can suppress the sensitivity of B. subtilis sigM null mutants towards beta-lactam antibiotics [4]. This suggests that c-di-AMP plays some role in PG homeostasis. Mutations in the yvcK gene (herein renamed glmR) also exhibit cell envelope defects, as evidenced by cell bulging and lysis when inoculated into non-glycolytic carbon sources [5]. Moreover, a yqfF::Tn insertion suppressed the inability of a glmR mutant to grow on gluconeogenic media [5]. Although unknown at the time, yqfF is now known to encode a second c-di-AMP hydrolase renamed PgpH [6, 7]. These observations encouraged us to investigate possible connections between GlmR, c-di-AMP, and cell envelope homeostasis.
In B. subtilis, GlmR (formerly YvcK) is essential for growth on amino acids and intermediates of the tricarboxylic acid cycle and pentose phosphate pathway, but dispensable for growth on glucose and other glycolytic carbon sources [5]. Previous genetic studies revealed that mutations in genes affecting central carbon metabolism (CCM), including zwf and cggR, allow a glmR null mutant to grow on gluconeogenic carbon sources [5]. These observations suggest that GlmR has a yet undefined role in regulating metabolism. In the absence of GlmR, cells display cell envelope defects and lyse under gluconeogenic growth conditions.
The function of GlmR in CCM, and how this relates to cell envelope integrity, is not yet clear. One model suggests that GlmR may function as a cytoskeletal filament protein analogous to MreB to help coordinate cell wall synthesis [8]. MreB, an actin-like cytoskeletal protein, is important for maintaining a rod shape in B. subtilis and deletion of mreB leads to severe morphological defects and eventual cell lysis, effects attributed to mislocalization of penicillin binding protein 1 (PBP1) [9]. B. subtilis GlmR localizes to the membrane in a helical fashion, and overexpression of GlmR rescues the cell defects seen in an mreB deletion mutant and restores proper localization of PBP1. Conversely, overexpression of MreB rescues the morphological defects of a glmR null mutant when grown on gluconeogenic carbon sources [8].
Recently, GlmR was found to possess a ligand binding site for UDP sugars such as UDP-glucose and UDP-N-acetylglucosamine (UDP-GlcNAc) [10]. Since UDP-GlcNAc is a precursor of PG synthesis, this suggests that GlmR may sense this intermediate to somehow modulate CCM or cell envelope homeostasis. Mutations altering the UDP-sugar binding site did not affect growth on gluconeogenic media in B. subtilis, but did lead to increased sensitivity to bacitracin [10].
Although the biochemical details are unclear, the role of GlmR in metabolism and cell wall homeostasis seems to be widely conserved. Homologs of GlmR are present diverse bacteria and a glmR mutant can be complemented by expression of the Escherichia coli homolog, YbhK [5]. Mutation of glmR homologs in the intracellular pathogens Mycobacterium tuberculosis (cuvA) and Listeria monocytogenes (yvcK) leads to alterations in cell morphology and sensitivity to cell wall acting antibiotics, as well as defects in carbon source utilization and establishment of infection in the host cell [11, 12]. Although these diverse phenotypes, biochemical properties and cell localization studies are all intriguing, a unifying model to account for the role of GlmR in the cell has been elusive.
Here, we show that a B. subtilis strain lacking glmR is susceptible to peptidoglycan (PG) biosynthesis inhibitors such as beta-lactams, vancomycin and moenomycin. Characterization of glmR suppressor mutations indicates that increased expression of genes involved in UDP-GlcNAc biosynthesis is sufficient to increase beta-lactam resistance and restore growth on gluconeogenic carbon sources. Moreover, supplementation with GlcNAc can bypass the requirement for GlmR even in strains where GlcNAc cannot enter into CCM. Our results support a model in which GlmR functions to help divert carbon to PG biosynthesis, likely through direct interaction with GlmS. We propose that this effect is particularly important during gluconeogenesis since the GlmS substrate fructose 6-phosphate is present at a reduced level under these conditions [13].
To test the role of GlmR in the connection between CCM and PG biosynthesis (Fig 1), we generated a B. subtilis strain with an in-frame, unmarked deletion of glmR (ΔglmR) and characterized its growth properties and sensitivity to cell wall antibiotics. Mueller-Hinton (MH) is a gluconeogenic medium containing amino acids as primary carbon source and is commonly used for antibiotic sensitivity experiments. However, ΔglmR is unable to grow on MH. This phenotype can be complemented by an ectopic, inducible copy of glmR (Fig 2A) or addition of glucose (S1A and S1B Fig), consistent with prior results [5].
To monitor the impact of the ΔglmR mutation on antibiotic sensitivity we performed zone-of-inhibition assays using LB (lysogeny broth) medium, a complex medium containing a variety of mono- and disaccharides (a total carbohydrate concentration of ~0.16%; [14]) and abundant amino acids. The ΔglmR mutant is much more sensitive to the beta-lactam antibiotic cefuroxime (CEF) (Fig 2B) as well as to other beta-lactam antibiotics (oxacillin and cefixime), moenomycin, and vancomycin (S2A–S2D Fig), all of which act by affecting the assembly and cross-linking of the peptidoglycan sacculus. However, we did not observe any significant difference in susceptibility between wild-type (WT) and ΔglmR to fosfomycin, bacitracin or nisin (S2E–S2G Fig). The lack of significant effect with these compounds may be due to the presence of inducible resistance mechanisms that might mask the effects of the ΔglmR mutation [15–18].
We selected CEF for further study due to the significantly higher sensitivity of the ΔglmR strain. Induction of an ectopic, IPTG-inducible glmR gene partially complements ΔglmR cefuroxime sensitivity (Fig 2B). Incomplete complementation may indicate that GlmR levels from this construct, while sufficient to restore growth (Fig 2A), are insufficient for robust cell wall synthesis. Consistent with this idea, induction of an N-terminally 3X-FLAG-tagged glmR allele with an optimized ribosome-binding site (AGGAGG-seven base pairs upstream from start codon), complemented CEF resistance to WT levels (S3A Fig). Mutations affecting PG synthesis can often be suppressed by high concentrations of Mg2+ [19, 20]. Indeed, Mg2+ suppresses the growth defect of a glmR deletion mutant on non-glycolytic carbon sources (S1A Fig), as shown previously [5], and also partially suppresses CEF sensitivity (S3B Fig). These results suggest that a ΔglmR strain is impaired in PG synthesis, and therefore more susceptible to antibiotics that interfere directly with PG assembly such as beta-lactams.
Both the ΔglmR and ΔsigM mutants are CEF sensitive, and in both cases mutations known to increase c-di-AMP levels suppress this sensitivity (see below). This suggests that GlmR and σM may function in the same pathway. However, a ΔglmR ΔsigM double mutant is much more sensitive than either single mutant (Fig 2C), suggesting that these are two independent (and additive) pathways for intrinsic CEF resistance.
The CEF sensitivity of the ΔglmR strain is suggestive of a defect in PG synthesis. GlmR is also known to be modified on Thr304 by the penicillin binding protein and serine/threonine associated (PASTA) kinase PrkC and phosphatase PrpC [21]. PrkC is activated by muropeptides during spore germination [22] and is regulated by interaction with the cell division protein GpsB during growth [23]. PrkC-dependent phosphorylation of GlmR has been linked to its role in morphogenesis and to resistance to bacitracin, but appears not to be required for growth on gluconeogenic carbon sources [21]. Similarly, this post-translational modification is not required for suppression of CEF sensitivity: both the phosphomimetic GlmRT304E and phosphoablative GlmRT304A mutant proteins complement the null mutant as well as wild-type (Fig 2B).
To gain insight into the role of GlmR in B. subtilis, we characterized suppressors (both spontaneous and transposon-generated) that either increased CEF resistance or restored the ability of ΔglmR to grow on MH medium. We isolated CEF resistant ΔglmR suppressors from CEF zone-of-inhibition assays or as colonies on MH medium (S1B Fig). We identified the causative mutations using whole-genome resequencing (spontaneous mutations) or by sequencing of junction fragments (transposon insertions) followed by linkage analysis and/or genetic reconstruction and complementation (Table 1). In general, the selected mutations suppressed both phenotypes associated with ΔglmR. Those suppressors selected for increased CEF resistance also recovered an ability to grow on MH medium. Conversely, for those selected for growth on MH medium, nearly all displayed at least a partial increase in CEF resistance relative to the ΔglmR starting strain (Table 1). In general, in this and previous studies, we find that CEF sensitivity is an excellent reporter for defects in cell wall synthesis. Often, suppressor mutations that fully restore growth may only partially rescue intrinsic CEF resistance. Here, we will focus on those suppressor mutations in the cdaA-cdaR-glmM-glmS region of the chromosome, which encodes the two initial enzymes in the peptidoglycan biosynthesis pathway, a major cyclic-di-AMP synthase (CdaA) and a regulator of CdaA (CdaR). We also recovered mutations in other genes in carbon metabolism, including pgcA and zwf, consistent with prior genetic studies of glmR function [5]. The possible mechanisms of suppression for these and other mutations are considered in the Discussion.
Many of the ΔglmR suppressors (Table 1) contained changes in a chromosomal region around two neighboring operons: sigW-rsiW and cdaA-cdaR-glmM-glmS (Fig 3A). These included a transposon insertion immediately after the rsiW stop codon (rsiW3) and point mutations in the glmS ribozyme (glmS1; 200068A>T), in the penultimate codon of rsiW (rsiW1; 196049G>A), and downstream of rsiW (rsiW2; 196071C>T). Note that the identical glmS mutation (glmS1) was recovered independently in both selection conditions.
Since most of the suppressor mutations did not fully restore CEF resistance to WT levels (Table 1), we selected several with intermediate levels of resistance as a starting point for selection of further increased CEF resistance. The most frequent secondary mutations were in rho (S1 Table). A rho deletion mutant has been associated with beta-lactam resistance in B. subtilis previously [24]. Interestingly, a ΔglmR Δrho double mutant is actually more sensitive to CEF than ΔglmR (S4 Fig), and it is only when a primary suppressor mutation (such as glmS1) is present in ΔglmR that rho mutations confers significant CEF resistance (S4 Fig and S1 Table).
GlmS is an amidotransferase that catalyzes the first step in PG synthesis (Fig 1) by conversion of the glycolysis intermediate fructose-6-phosphate (F6P) into glucosamine-6-phosphate (GlcN6P) using glutamine as an amino group donor [25]. Expression of GlmS is under negative feedback control mediated by a ribozyme structure encoded in the 5'-untranslated region (5’-UTR) of the glmS mRNA. Upon binding to the GlmS product, GlcN6P, the ribozyme promotes site specific self-cleavage of glmS mRNA and consequently reduces glmS expression [26].
The glmS1 suppressor mutation is a base change in the catalytic domain of the glmS ribozyme (Fig 3B) [27]. After moving the glms1 mutation into a ΔglmR strain, the reconstructed ΔglmR glmS1 strain regains the ability to grow on gluconeogenic carbon sources (Fig 4A) and has increased resistance to CEF (Fig 4B). We hypothesized that glmS1 might interfere with the catalytic activity of the glmS ribozyme. Consistent with this idea, the glmS1 mutation caused a >50-fold increase in glmS mRNA compared to WT (Fig 4C) and a corresponding increase in GlmS protein levels (Fig 4D). We did not see any significant difference in glmS mRNA level between WT and ΔglmR.
Reconstruction of ΔglmR strains with mutations rsiW1 or rsiW2 confirmed that these changes allow growth of ΔglmR on gluconeogenic growth medium (Fig 5A) as well as increased resistance to CEF (Fig 5B). The rsiW1 mutation is silent with respect to the sequence of RsiW and rsiW2 is downstream of the rsiW coding region (Fig 3A). We hypothesized that these point mutations might affect the intrinsic transcription terminator of the sigW-rsiW operon. In silico analysis indicated that each mutation generates a mismatch in the stem of the transcription terminator that is predicted to decrease stability and therefore increase readthrough from the sigW-rsiW operon into the downstream cdaA-cdaR-glmS-glmM operon (S5 Fig). Indeed, the rsiW1 or rsiW2 suppressor mutations led to a >10-fold increase in the mRNA level for the first gene of this operon, cdaA (Fig 5C).
Expression of the sigW-rsiW operon is dependent on an autoregulatory σW-dependent promoter. An in-frame deletion mutation of sigW abolished the ability of the rsiW1 and rsiW2 mutations to suppress the ΔglmR phenotype (Fig 5B). However, in a strain with a sigW::erm disruption mutation the rsiW1 and rsiW2 mutations still conferred increased CEF resistance since the erm σA promoter now reads into the cdaA operon (S5B Fig). These observations support our hypothesis that rsiW1 and rsiW2 increase expression of cdaA-cdaR-glmM-glmS. A similar increase in transcription may explain the phenotype of the rsiW3 Tn insertion (Table 1).
We reasoned that the rsiW1, rsiW2 and rsiW3 mutations likely lead to elevated expression of the cdaA-cdaR-glmM-glmS operon. The first two genes encode the major synthase (CdaA) for c-di-AMP and an activator protein (CdaR) [6, 7]. The final two genes encode enzymes for the initial steps of PG biosynthesis that (together with GlmU; also known as GcaD; [28]) convert F6P to UDP-GlcNAc (Fig 1). To determine which gene(s) in this operon are involved in suppression of the ΔglmR phenotypes we integrated IPTG-inducible copies of various portions of this operon (including cdaA, cdaA-cdaR, cdaA-cdaR-glmM, cdaA-cdaR-glmM-glmS, glmM-glmS) at the amyE locus in the ΔglmR strain. These strains were tested for CEF sensitivity and growth on MH medium. Overexpression of cdaA or cdaA-cdaR was not sufficient to increase CEF resistance of ΔglmR (Fig 6A), although we did note an increased frequency of spontaneous suppressors. Overexpression of cdaA-cdaR-glmM or glmM-glmS partially restored CEF resistance (Fig 6A). However, when the whole operon (cdaA-cdaR-glmM-glmS) was induced CEF resistance was restored to essentially WT levels (Fig 6A). Increased expression of cdaA-cdaR-glmM or cdaA-cdaR-glmM-glmS also suppressed the essentiality of ΔglmR on gluconeogenic MH medium (Fig 6B). In contrast, induction of cdaA-cdaR alone has a comparatively weak and variable effect on growth, which may reflect the rapid emergence of suppressors in this strain (Fig 6B). From these results we conclude that the key factor in increased fitness of the ΔglmR strain is elevated expression of GlmS and/or GlmM, but that c-di-AMP may also play a role.
An increase of c-di-AMP has been previously associated with CEF resistance since mutations in gdpP, encoding the major c-di-AMP hydrolase, suppress the CEF sensitivity of a sigM mutant [4]. Moreover, a yqfF::Tn insertion, affecting a second c-di-AMP hydrolase renamed PgpH [6, 7], suppresses the inability of a glmR(yvcK) mutant to grow on gluconeogenic media [5]. We have confirmed these findings and here demonstrate that inactivation of gdpP increases CEF resistance of ΔglmR, although pgpH does not have a significant effect under our conditions (S6A and S6B Fig). It is interesting to note that a gdpP pgpH double mutant, which has greatly elevated c-di-AMP levels and is growth impaired [7], is also highly sensitive to CEF. This effect is not additive with ΔglmR, suggesting that excess c-di-AMP may affect the same pathway as GlmR (S6A and S6B Fig). Consistently, the ability of CdaA and CdaR to increase CEF resistance in a ΔglmR mutant seems to be contingent on the additional expression of GlmM and GlmS, as noted above (Fig 6A). CdaA forms a complex with both CdaR and GlmM [7, 29], suggesting that c-di-AMP may modulate GlmM activity.
We next considered whether a ΔglmR strain might be phenotypically suppressed by over-expression of other individual enzymes upstream and downstream of UDP-GlcNAc. Induction of glmS, glmM or glmU (Fig 1), partially restored CEF resistance (Fig 7A) and restored the ability of ΔglmR to grow on gluconeogenic medium (Fig 7B). We suggest that these enzymes increase the forward reaction catalyzed by GlmS by consumption of the product, GlcN6P. GlcN6P is potent inhibitor of GlmS (product inhibition) [30], a property shared with the human ortholog [31].
A portion of cellular UDP-GlcNAc is converted to UDP-MurNAc, the second building block of PG, by MurA and MurB (Fig 1). B. subtilis has two MurA paralogs, MurAA and MurAB, but only MurA is essential. UDP-MurNAc is then modified by addition of a pentapeptide side-chain and transferred to the undecaprenylphosphate carrier lipid to ultimately generate lipid II (Fig 1), a lipid-linked GlcNAc-MurNAc-pentapeptide that is the substrate for extracellular PG synthesis [32]. Overexpression of murAA or murB increased the sensitivity of the ΔglmR strain to CEF (Fig 7C), and neither rescued the growth defect of ΔglmR on MH medium (Fig 7D).
We reasoned that the effects of MurAA and MurB overproduction might be relieved in cells that have increased capacity to synthesize UDP-GlcNAc. To test this hypothesis, we introduced the glmS1 mutation (which abolishes negative feedback regulation of glmS) into the ΔglmR amyE::Pspac(hy) murAA and ΔglmR amyE::Pspac(hy) murB strains. In these glmS1 strains, induction of murAA or murB no longer increases sensitivity to CEF (Fig 7C). Based on these observations we hypothesize that B. subtilis lacking GlmR is impaired specifically in UDP-GlcNAc biosynthesis. The resulting inability to efficiently synthesize PG is a likely reason for the essentiality of glmR on gluconeogenic media.
GlmR was recently found to bind UDP-sugars such as UDP-glucose and UDP-GlcNAc [10]. UDP-GlcNAc bound with five times higher affinity that UDP-Glc, suggesting that the former may be a regulatory ligand for GlmR. We used CRISPR-gene editing to introduce single amino acid substitutions in the UDP-GlcNAc binding site of GlmR that were previously shown to abolish ligand binding (Y265A, R301A and R301E). Consistent with prior results [10], none of these three mutations affected the ability of GlmR to support growth on gluconeogenic MH medium (Fig 8A), nor did they have a significant impact on CEF resistance (Fig 8B). We therefore suggest that ligand binding serves as a feedback mechanism to down-regulate GlmR activity when UDP-GlcNAc levels are high. Under gluconeogenic conditions, when GlmR is required for redirecting carbon from CCM into PG synthesis, this binding site would be vacant, and therefore these mutations would not affect the stimulatory function of GlmR (Fig 1).
Since ΔglmR suppressor mutations lead to increased glmS expression (Fig 4C and 4D), we reasoned that the ΔglmR strain may be specifically defective in GlmS activity. If this is the case, we hypothesized that provision of cells with GlcNAc would chemically complement the ΔglmR growth defect. Indeed, when a disc containing GlcNAc was placed on a MH medium plate strong growth of the ΔglmR strain was observed (Fig 9A).
GlcNAc is taken up by the GlcNAc-specific phosphoenolpyruvate phosphotransferase system (PTS) protein NagP and enters the cell as GlcNAc-6-phosphate [33]. Deacetylation by NagA then generates GlcN6P (Fig 1), which is also the product generated by GlmS [34]. GlcN6P can either feed into peptidoglycan biosynthesis (GlmM and GlmU) or feed CCM by conversion to F6P by either of two inducible deaminases (NagB and GamA) [33, 35] (Fig 1). The ability of GlcNAc to support growth of the ΔglmR strain requires NagA, but is independent of the GamA and NagB deaminases (Fig 9B). This indicates that the limiting step in metabolism during growth of the ΔglmR strain on largely gluconeogenic carbon sources is the GlmS-catalyzed conversion of F6P to GlcN6P. This limitation can be by-passed by up-regulation of GlmS (e.g. by overexpression, Fig 7B, or in the glmS1 mutant strain, Fig 4) or by provision of cells with GlcNAc. The ability of overproduced GlmM or GlmU to support growth (Fig 7B) may therefore seem surprising, but may be explained by more rapid consumption of GlcN6P, which would prevent product inhibition of GlmS and also increase translation of GlmS by inhibiting glmS ribozyme cleavage.
To test if GlcNAc addition also suppresses the increased CEF sensitivity, we tested WT and ΔglmR strains on LB agar supplemented with 0.5% and 1% GlcNAc. Addition of GlcNAc partially suppressed the CEF sensitivity of ΔglmR, but had no significant effect on a strain in which GlmS was up-regulated by the glmS1 suppressor mutation (Fig 9C). In a ΔglmR ΔnagB ΔgamA strain in which added GlcNAc cannot re-enter CCM, CEF resistance is restored to near WT levels (Fig 9D). The greater suppression seen in this strain may result from the inability of this strain to catabolize incoming GlcNAc, which thereby further increases the flux into PG synthesis. This supports the notion that a major contributor to CEF sensitivity is a metabolic defect that limits the ability of the cell to synthesize PG, apparently due to a limitation in the ability of GlmS to redirect carbon from CCM to cell wall synthesis. We hypothesize that GlmR may directly stimulate GlmS enzyme activity. This is supported by evidence of a GlmR-GlmS protein interaction in bacterial two-hybrid assays (Fig 10). The observed interaction is robust, as compared to the positive control, and GlmR did not interact with other proteins tested including CdaA, GlmM or CdaR (Fig 10).
Here we present a forward genetic analysis that indicates that GlmR regulates the redirection of carbon from CCM into PG biosynthesis, likely by stimulation of GlmS activity. The regulation of CCM as cells adapt to nutrient availability is exceptionally complex and involves numerous transcriptional regulators and post-transcriptional regulatory mechanisms [36, 37]. The carbon catabolite control protein CcpA plays a central role in this process and represses genes for the utilization of non-preferred carbon sources when glucose is available [38], as well as the operon encoding glmR: yvcI-yvcJ-glmR-yvcL-crh-yvcN [39]. As a result, GlmR should be most abundant when CcpA activity is low. CcpA repressor activity is indirectly stimulated by elevated levels of fructose-1,6-bisphosphate present during growth on preferred carbon sources [40–42]. During growth on non-preferred, gluconeogenic carbon sources GlmR will be more abundant, consistent with its role in diverting carbon to PG synthesis under these conditions.
The GlmR (formerly YvcK) protein is conditionally essential and plays a poorly defined role in cell morphology and antibiotic resistance [10–12, 21]. Homologs in M. tuberculosis (CuvA) and L. monocytogenes (YvcK) appear to also play a role in helping maintain cell shape [11, 12]. GlmR was suggested to lead to a dysregulation of carbon metabolism since mutations affecting metabolic enzymes (e.g. Zwf) and CCM regulatory proteins (e.g. CggR) suppress the glmR null mutant and allow growth on gluconeogenic carbon sources [5]. Cytological evidence suggests that GlmR and CuvA localize to membrane sites associated with PG synthesis, and it has been noted that GlmR and MreB appear to functionally substitute for one another, perhaps in coordinating the assembly of PG biosynthetic complexes [8, 11]. Despite intensive study, the connection between these disparate phenotypes has been elusive. Here, we propose that several of these phenotypes can be explained by GlmR-dependent stimulation of the key branchpoint enzyme, GlmS.
It remains possible that, in addition to stimulation of GlmS activity, GlmR may have other functions. This is suggested by the observation that the role of GlmR in intrinsic CEF resistance is independent of protein phosphorylation as judged by the analysis of phosphomimetic and phosphablative mutants (Fig 2). In contrast, phosphorylation of GlmR was shown to affect bacitracin sensitivity and cell morphogenesis in an mreB mutant background [21]. Although the M. tuberculosis GlmR ortholog CuvA is also modified by phosphorylation by Ser/Thr PASTA kinases, this modification is not important for complementation of carbon source specific growth defects or for localization to sites of PG synthesis [11, 12], and perhaps regulates other functions. Analysis of phosphosite mutants of the L. monocytogenes GlmR ortholog suggests that a phosphomimetic variant is unaffected in metabolism and cell wall homeostasis, but is impaired in virulence [11, 12]. Further studies are needed to clarify how GlmR phosphorylation affects some, but not all, activities of this protein.
Our genetic analysis supports a model in which GlmR activates GlmS, and we suggest that this activity is inhibited when GlmR is bound to the downstream metabolite, UDP-GlcNAc (Fig 1). This model is supported by several key observations. First, overproduction of GlmS, in either the glmS1 mutant or by induction from an ectopic glmS gene, is sufficient to restore growth of the glmR null mutant on MH medium (Figs 4 and 7). Second, a glmR mutant can be chemically complemented by GlcNAc, even under conditions where GlcNAc cannot be routed into CCM (Fig 9). Since metabolism of GlcNAc generates GlcN6P, this addition specifically bypasses the GlmS reaction (Fig 1). Therefore, we suggest that GlmS (rather than GlmM or GlmU) is limiting the flux of carbon into PG in the ΔglmR strain. Third, GlmR and GlmS interact in vivo as judged by a bacterial two-hybrid assay (Fig 10). Fourth, previous metabolomics measurements indicate that F6P levels are ~16-fold lower during growth on gluconeogenic carbon sources when compared to glucose [13], consistent with the requirement for GlmR under these conditions (Fig 1). Fifth, GlmR was recently found to bind UDP-GlcNAc [10]. However, mutations that abolish binding do not affect the ability of GlmR to stimulate growth under gluconeogenic conditions [10] or to provide intrinsic CEF resistance (Fig 8), as predicted by the hypothesis that UDP-GlcNAc antagonizes GlmR function (Fig 1). GlmS is recognized as the key branch-point enzyme in bacteria for diverting carbon from CCM into PG synthesis, and in eukaryotes the GlmS ortholog diverts carbon into hexosamine synthesis. Both classes of enzyme are in some cases feedback regulated by UDP-GlcNAc [43–47]. Here, UDP-GlcNAc binding is proposed to antagonize GlmR function, and therefore reduce stimulation of GlmS.
In addition to GlmS, we also demonstrate that overproduction of either GlmM or GlmU, but not by enzymes downstream of the key intermediate UDP-GlcNAc, can suppress the glmR growth defect under gluconeogenic conditions. GlmS catalyzes a reversible reaction, and its product (GlcN6P) is a potent inhibitor of the forward reaction [30]. Moreover, GlcN6P binds to the glmS ribozyme to cleave the mRNA and suppress translation [26]. Therefore, we suggest that increasing the level of GlmM and/or GlmU likely helps pull the reaction in the forward direction and may also stimulate GlmS translation.
With a defined model in hand, we can revisit the other suppressor mutations recovered both in our selection conditions (Table 1) and the studies of Görke et al. [5]. As noted previously, many of the mutations that suppress glmR affect CCM. We recovered a frameshift mutation in zwf, a gene also recovered in the previous transposon-based selection for glmR suppressors [5]. Normally, Zwf diverts a substantial fraction of glucose-6-phosphate from glycolysis into the pentose phosphate pathway [48]. We speculate that in the absence of Zwf there is increased flux leading to F6P, the GlmS substrate. We also recovered a mutation in pgcA, which encodes another branch point enzyme that uses glucose-6-phosphate. Previously, it was reported that a mutation in cggR, encoding the central glycolytic genes regulator, also suppresses glmR [5]. Since a cggR null mutant will have increased levels of several key enzymes that function in both glycolysis and gluconeogenesis [49], we speculate that this mutation alleviates the metabolic restriction in the glmR strain by increasing gluconeogenesis and therefore F6P levels.
A second class of mutations that increase the fitness of the ΔglmR strain are those that lead to elevated c-di-AMP levels. This was foreshadowed by the finding that a pgpH (formerly yqfF) mutation suppresses glmR [5]. In our studies, we find that gdpP suppresses glmR both for growth on MH medium and for CEF resistance, whereas pgpH has a lesser effect (S6 Fig). CdaA is regulated by interaction with the CdaR protein and also forms a complex with GlmM [7, 29]. Indeed, the cdaA-cdaR-glmM genes are co-transcribed in a wide variety of species, suggesting a functional connection. This has led to the suggestion that GlmM may regulate c-di-AMP synthesis [7, 29]. Conversely, CdaA may regulate GlmM. In this scenario, conditions that lead to elevated c-di-AMP may alter the CdaA-CdaR complex to favor a stimulatory interaction of CdaA with GlmM. Indeed, it is striking that induction of the entire cdaARglmMS operon fully restores CEF resistance to a glmR mutant (Fig 6), whereas this is not the case for the glmR glmS1 strain (Fig 4) or for induction of glmS alone (Fig 7). Alternatively, c-di-AMP is also known to regulate potassium homeostasis by interaction with both protein and RNA (riboswitch) targets [50–53]. This c-di-AMP dependent osmolyte transport is important for maintaining turgor pressure in the cell and it has been proposed that perturbations of c-di-AMP metabolism can affect cell envelope integrity by increasing resistance against osmotic stresses [54].
A third class of suppressor mutations is in genes important for energy generation by the electron transport chain. These include mutations in qoxB, encoding cytochrome aa3 quinol oxidase, and yqiD(ispA), encoding a geranyltransferase that is involved in synthesis of isoprenoid compounds including menaquinone, an electron carrier important for respiration (Table 1). Mutations in both of these loci have been previously associated with an increased ability of cells to survive the transition to L-forms that lack a peptidoglycan cell wall [55]. This observation led to a model in which a lethal consequence of cell wall defects is oxidative damage triggered by increased flux through the electron transport chain when carbon flux into peptidoglycan is eliminated [55]. Regardless of the precise mechanism, it is intriguing that mutations in these same genes were recovered as suppressors of ΔglmR.
Finally, we recovered one strain containing a missense mutation in yvcJ (Table 1), the gene immediately upstream of glmR. The role of YvcJ is unknown, but it has GTPase activity, affects phosphorylation of an uncharacterized cell component, and has an apparent role in natural competence [56, 57]. Since this strain contained an additional mutation in sigA (Table 1), further work is needed to determine the effect of the yvcJ mutation on CEF resistance. Curiously, mutants of the E. coli YvcJ homolog (RapZ; formerly YhbJ) lead to overproduction of GlmS [58]. RapZ appears to sense GlcN6P and regulates the processing and stability of a small RNA, GlmZ, that activates GlmS synthesis [46, 58, 59]. It is presently unknown whether YvcJ plays a related role in B. subtilis, perhaps by interacting either with GlmR or the glmS ribozyme.
In conclusion, the results presented here highlight the importance of the GlmS branch point in regulating the flow of carbon from CCM into PG synthesis. In eukaryotes, GlmS orthologs serve as the initiating enzyme for hexosamine biosynthesis, and are sensitive to both GlcN6P product inhibition [31] and feedback regulation by UDP-GlcNAc, which binds to the isomerase domain [43, 44]. In bacteria, GlmS is also subject to complex regulation at the level of both synthesis and activity [45–47]. In B. subtilis, GlmS is feedback inhibited by its immediate product, GlcN6P [30], which also activates the glmS ribozyme [26]. GlmR provides another layer of regulation. Our results support a model in which GlmR stimulates GlmS activity, and we propose that binding of UDP-GlcNAc may attenuate this stimulation.
B. subtilis strains used are derived from strain 168 (trpC2) (S2 Table). E. coli strain DH5α was used for cloning and strain BTH101 [60] for bacterial two hybrid experiments. Bacteria were cultured in LB broth. Strains with a glmR deletion mutation were cultured on LB with 20 mM MgSO4 unless specified otherwise. Antibiotics were added to growth media when required at the following concentrations: 100 μg/ml ampicillin, 30 μg/ml chloramphenicol for E. coli, 10 μg/ml kanamycin, 10 μg/ml chloramphenicol, 5 μg/ml tetracycline, 100 μg/ml spectinomycin and 1 μg/ml erythromycin with 25 μg/ml lincomycin (erm; macrolide-lincomycin-streptogramin B resistance).
For cloning procedures, restriction digestion and ligation with T4 ligase was done as per manufacturer's instructions (NEB, USA). Plasmids were then transformed into competent DH5α cells [61]. Cloning was confirmed by polymerase chain reaction (PCR) followed by Sanger sequencing. B. subtilis transformation was carried out in minimal competence media with 12 mM MgSO4. DNA was added when cells reached OD600 of ~0.7–0.8. Generation of B. subtilis strains overexpressing gene(s) at amyE was achieved using pPL82 [62] carrying gene(s) of interest followed by transformation into the indicated B. subtilis recipient strain.
Bacillus knockout erythromycin (BKE) strains with various gene deletion mutations of B. subtilis were obtained from the Bacillus Genetic Stock Center (BGSC) [63]. Chromosomal DNA from each BKE strain was transformed into our lab strain B. subtilis 168. The erythromycin resistance cassette was removed using pDR244 [63], which produces Cre recombinase at the permissive temperature of 30°C, to generate in-frame deletions. pDR244 was transformed into B. subtilis strain at 30°C and plated on LB plates with spectinomycin. Colonies were picked after two overnight incubations and patched three successive times on LB plates incubated at the non-permissive temperature 42°C overnight. Strains were then patched on spectinomycin- and erythromycin-containing plates to confirm the absence of both markers. All the deletion mutants used in study are markerless deletions except Δrho (rho::erm).
Single nucleotide mutations glmS1, rsiW1 and rsiW2 were reconstructed using the integration vector pMutin4 that has an erm resistance marker and lacZ [64]. A fragment of DNA with the mutation of interest was cloned into pMutin4 and confirmed with PCR and Sanger sequencing. The vector was transformed into B. subtilis where it integrated at locus by single crossover homologous recombination. Transformants were selected on plates with Erm and 40 μg/ml X-gal. After overnight incubation, a few blue color colonies were picked. Since pMutin4 integration is unstable, cells were grown without antibiotic selection three consecutive times with each time adding 1:100 dilution of cells from previous culture. Cells were then plated on LB plates with X-gal and white colonies were picked and sequenced to find those strains that retained the single nucleotide mutation of interest.
Mariner transposon mutagenesis procedure was carried out in ΔglmR as described previously [65]. In brief, ΔglmR was transformed with the pMarA vector. The strain with pMarA was grown in 5 ml LB broth until mid-exponential phase and various dilutions of cells were plated on selection medium. In independent experiments CEF resistance and ability to grow on MH media were used as a selection.
Spontaneous suppressors of ΔglmR were picked from the clear zone of CEF disc diffusion plates and independently from MH plates after overnight incubation at 37°C. Chromosomal DNA extracted from these suppressors was sequenced using an Illumina machine. The sequencing data were analyzed using CLC genomics workbench.
Antibiotic sensitivity was tested using disc diffusion assays, which were carried out on LB medium. Strains to be tested were grown in 5 ml LB broth at 37°C with vigorous shaking to an OD600 of ~0.4. 100 μl of cells were added to 4 ml top LB agar (0.7% agar) kept at 50°C. 1 mM IPTG was added to top agar when indicated. Top agar with cells was poured over 15 ml LB bottom agar (1.5%) plate. A Whatman paper disc (7mm dia) with 6 μg CEF was put on the plate unless specified otherwise. Plates were incubated at 37°C overnight and the clear zone of inhibition was measured the next day. Values for CEF resistance (Table 1) report the diameter of the zone of growth inhibition. For all histograms, the values shown have the diameter of the filter disk (7 mm) subtracted from the average diameter.
To test the ability of B. subtilis mutants to grow under gluconeogenic conditions we used MH medium (Sigma-Aldrich, USA) prepared per the manufacturer's instruction. Growth was monitored using a Bioscreen growth analyzer with 200 μl of MH broth in 100 well Bioscreen plates inoculated with 2 μl of B. subtilis strains pre-grown in LB broth at 37°C to an OD600 of ~0.4. When required, glucose, MgSO4 and IPTG were added to the final concentrations of 1%, 20 mM and 1 mM respectively.
Strains of interest were grown to an OD600 of ~0.5. 1.5 ml of culture was used for RNA extraction. RNA isolation (Qiagen, USA) and cDNA preparation (Thermofisher, USA) was carried out as suggested by the manufacturer. qRT-PCR was carried out using a Bio-Rad iTaq universal SYBR green super mix. 23S rRNA was used to normalize the cycle threshold (Ct) value.
For GlmS measurements, ΔglmR and ΔglmR glmS1 strains were grown in LB medium to an OD600 of ~0.3 at 37°C with shaking. 30 ml of culture was withdrawn and centrifuged at 5000 rpm for 10 minutes. Cell pellets were frozen at -20°C. Pellets were washed once with 1X phosphate buffer saline (pH 7.4). 150 μl of lysis buffer (20 mM tris-HCl, 100 mM NaCl, 1 mM EDTA, 1 mM DTT, 10% glycerol and protease inhibitor cocktail) was used to re-suspend the cell pellets. One tablet of protease inhibitor cocktail from Roche diagnostics was added to 10 ml of lysis buffer. Cells were lysed by sonication. After centrifugation cell lysates were transferred to fresh tubes. Protein concentration was measured by Bradford assay (Bio-Rad). 5 μg of protein was run on a 4–15% gradient gel from Bio-Rad. Protein was transferred onto a PVDF membrane using a Bio-Rad transblot turbo transfer system. The membrane was blocked with 5% milk powder for one hour followed by overnight incubation with primary anti-GlmS polyclonal antibodies [66] added to 1:3000 dilution in 1X tris buffer saline with 0.1% tween 20 and 0.5% milk powder. After three washes, the membrane was incubated with a 1:3000 dilution of HRP conjugated anti-Rabbit antibodies (Sigma). Bands were visualized on a Bio-Rad Chemidoc MP imaging system.
Strains of interest were grown in 5 ml LB medium to an OD600 of ~0.4. 100 μl of cells were added to 4 ml top MH agar (0.7% agar) preheated at 50°C and was laid on a 15 ml MH agar (1.5%) plate. A disc with 0.5 mg GlcNAc (Sigma, USA) was put on the plate. After overnight incubation at 37°C, the zone of growth surrounding the disc was measured.
DNA changed encoding single amino acid substitutions (GlmRY255A, GlmRR301A and GlmRR301E) were generated at the native glmR locus using CRISPR editing as described [67]. In brief, oligonucleotides encoding a 20 nucleotide gRNA with flanking BsaI sites and a repair fragment carrying mutations of interest with flanking SfiI restrictions sites were cloned sequentially into vector pJOE8999 followed by transformation into E. coli DH5α cells. The resultant plasmid was transformed into recipient B. subtilis strain and cells were plated on 15 μg/ml kanamycin plates with 0.2% mannose. Transformation was carried out at 30°C as pJOE8999 cannot replicate at higher temperatures. The transformants were patched on LB agar plates and incubated at the non-permissive temperature of 42°C. The loss of vector was confirmed by the inability of selected isolates on kanamycin plates. The presence of the desired mutations was confirmed by Sanger sequencing.
Vectors pT18 and pT25 and strains for bacterial two hybrid were prepared as described [60]. E. coli BTH101 strains carrying pT18 and pT25 with genes of interest were grown in LB broth overnight at 30°C with 100 μg/ml ampicillin, 50 μg/ml chloramphenicol and 0.5 mM IPTG. 10 μl of cells were spotted on LB plate with 100 μg/ml ampicillin, 50 μg/ml chloramphenicol, 0.5 mM IPTG and 40 μg/ml X-gal. Plates were incubated overnight at 30°C.
In silico analysis was carried out using NUPACK web application [68].
|
10.1371/journal.pntd.0000568 | Human Probing Behavior of Aedes aegypti when Infected with a Life-Shortening Strain of Wolbachia | Mosquitoes are vectors of many serious pathogens in tropical and sub-tropical countries. Current control strategies almost entirely rely upon insecticides, which increasingly face the problems of high cost, increasing mosquito resistance and negative effects on non-target organisms. Alternative strategies include the proposed use of inherited life-shortening agents, such as the Wolbachia bacterium. By shortening mosquito vector lifespan, Wolbachia could potentially reduce the vectorial capacity of mosquito populations. We have recently been able to stably transinfect Aedes aegypti mosquitoes with the life-shortening Wolbachia strain wMelPop, and are assessing various aspects of its interaction with the mosquito host to determine its likely impact on pathogen transmission as well as its potential ability to invade A. aegypti populations.
Here we have examined the probing behavior of Wolbachia-infected mosquitoes in an attempt to understand both the broader impact of Wolbachia infection on mosquito biology and, in particular, vectorial capacity. The probing behavior of wMelPop-infected mosquitoes at four adult ages was examined and compared to uninfected controls during video-recorded feeding trials on a human hand. Wolbachia-positive insects, from 15 days of age, showed a drastic increase in the time spent pre-probing and probing relative to uninfected controls. Two other important features for blood feeding, saliva volume and apyrase content of saliva, were also studied.
As A. aegypti infected with wMelPop age, they show increasing difficulty in completing the process of blood feeding effectively and efficiently. Wolbachia-infected mosquitoes on average produced smaller volumes of saliva that still contained the same amount of apyrase activity as uninfected mosquitoes. These effects on blood feeding behavior may reduce vectorial capacity and point to underlying physiological changes in Wolbachia-infected mosquitoes.
| Mosquitoes transmit diseases when they are actively searching for a source of blood. This so called probing behavior comprises the “searching” time, the beginning of the feeding process until the first sign of blood can be seen within the insect body. The manipulation of this behavior can have important consequences for the mosquito's ability to transmit pathogens, such as dengue virus or Plasmodium. In this study we examined the probing behavior of the main vector of dengue viruses, Aedes aegypti, when infected with an intracellular bacterium, Wolbachia pipientis. This bacterium alters the probing behavior of older mosquitoes such that they take longer to find a feeding site and longer to imbibe blood, which may make them more susceptible to human defense responses. The bacterium appears to reduce mosquito feeding success by preventing the mosquito from successfully inserting its stylet into human skin. The old age onset of reduced mosquito feeding success due to Wolbachia could selectively promote a reduction in dengue transmission.
| Insect transmitted diseases such as malaria and dengue occur in more than 100 countries worldwide, placing at risk around half the world's population. The disease burden is high with more than 500 million cases each year. Despite various vector control measures, there continues to be emergence and resurgence of these diseases [1]. Aedes aegypti is the main vector of dengue fever causing millions of cases and thousands of deaths each year. Vector control is the only method for dengue and dengue hemorrhagic fever (DHF) prevention, however current strategies are failing to prevent the increasing global incidence of dengue fever [2]. The development of practical alternative strategies to control dengue, which could be used in conjunction with current measures, is much needed.
Recently our group reported the successful stable infection of A. aegypti with the wMelPop Wolbachia strain that reduces insect lifespan [3]. Wolbachia is an inherited bacterium able to manipulate the insect host's reproductive biology [4] in a manner that promotes its rapid spread through insect populations [5]. Releasing Wolbachia-infected mosquitoes that could initiate an invasion of Wolbachia into a wild mosquito population [6] and that resulted in reduced lifespan of wild mosquitoes could theoretically greatly reduce transmission of dengue virus. This is because only old mosquitoes transmit the virus [3],[7],[8]. In addition to lifespan reduction, we have recently shown that the wMelPop infection substantially reduces dengue load in A. aegypti mosquitoes [9] and reduces their ability to successfully obtain blood meals as they age [10].
Mosquitoes rely on chemical and physical cues (e.g. carbon dioxide, body odors, air movement and heat) to locate suitable hosts for feeding [11],[12] and this complex set of activities is known as host-seeking behavior [13],[14]. Once the host is located, the mosquito must quickly obtain blood to avoid any host defensive behavior [15],[16]. Behavior during feeding can then be divided into the stages of pre-probing (foraging) [17] and probing (feeding) activities [18]. After finding a suitable blood vessel and thrusting its stylet into the host skin [19], the saliva plays an important role in preventing blood clotting, through the anti-platelet aggregation activity of the enzyme apyrase [20],[21] and other antihemostatic and anti-inflammatory compounds [19]. Insect vector competence for the transmission of viruses and parasites is dependent upon the successful execution of these upstream steps in the process.
Here we report the results of an examination of the effects of wMelPop infection on pre-probing behavior, probing behavior, saliva production, and apyrase content of saliva of A. aegypti. The goals of the study were to identify possible mechanisms for the insect's reduced ability to obtain a blood meal with age and further evaluate the capacity for Wolbachia infection to reduce vector competence.
This study was conducted according to the principles expressed in the Declaration of Helsinki. The study was approved by the Medical Research Ethics Committee at the University of Queensland (Project #2007001379). Volunteers were made aware of the risks of bloodfeeding (allergy and discomfort) and the plans to analyze and publish all data prior to providing written consent to participate in the study.
Aedes aegypti mosquitoes, wMelPop infected (PGYP1) and its Tetracycline-cured counterpart (PGYP1.tet) [3], were kept in a controlled environment insectary at 25 °C and 80% RH. Larvae were maintained with fish food pellets (Tetramin, Tetra) and adults were offered 10% sucrose solution, ad libitum. Adult females were fed on human blood (UQ human ethics approval 2007001379) for egg production and eggs were dried for at least 96 h prior to hatching.
Fertilized and non-blood fed females of different ages (5, 15, 26 and 35 days old) were used in all behavior experiments. Sucrose solutions were removed from cages on the night before the experiments. Forty females were used per age and per infectious status. Single mosquitoes were transferred to a transparent Perspex cage (25 cm3) and filmed through a digital camera with 6mm Microlens (IEEE-1394, Point Grey Research) mounted on a tripod. Mosquitoes were given about five minutes to settle within the cage before a human gloved-hand was inserted into the cage. A window of about 15 cm2 was cut of the upper part of the latex glove in order to delineate the probing field.
Movies were recorded (QuickTime Player) for a maximum of 10 minutes or until blood was seen within the mosquito midgut and sub-sequentially watched for time calculations. Two electronic timers were used, one for recording pre-probing time and the second for probing time. Pre-probing time was defined and the time since the mosquito has landed on the bare hand area until the insertion of mouthparts into the human skin. Probing time is defined as the initial insertion of insect mouthparts until blood can be seen within the mosquito midgut through the abdominal pleura [21]. Timing stopped when mosquitoes left the bare hand area or withdrew their mouthparts before taking blood and began again when the mosquito came back or after subsequent stylet penetration. If blood was not found by the end of 10 minutes, we defined this case as unsuccessful probing and it was measured as a proportion. Movies were also used to visualize additional abnormal phenotypes as the jittering action of mosquito body while landed on top of the human hand, and named “shaky”. Furthermore, the inability of mosquitoes to insert they mouthparts due to a bendy proboscis was also analyzed. The bendy phenotype was recently described by Turley et al [10].
Mosquitoes of different ages (5, 26 and 35-days-old) and infectious status were starved overnight (without sucrose solution or water). On the following morning mosquitoes were briefly anesthetized with CO2 and placed on a glass plate over ice. Wings and legs were removed with forceps and their proboscis introduced into a 1cm piece of polypropylene tubing (0.61×0.28mm, Microtube Extrusions, NSW, Australia) (modified from [21]). Females were allowed to salivate for 5 minutes and then the diameter of the saliva droplets was measured through an ocular micrometer at 40× magnification. Volumes were calculated via the sphere formula [22]. Saliva was then collected into 20 µL of 0.05mM Tris-HCl pH 7.5 by attaching the needle of a 10 µL Hamilton syringe and rinsing the tubing content a few times. Samples were centrifuged at 14,000g for 2 minutes and kept frozen (−80°C) in 20 µL of 0.05mM Tris-HCl, pH 7.5 for enzymatic assay (see below).
Saliva samples (8 µL) were transferred, in duplicates, into individual wells of a plastic 96-well ELISA plate (NUNC Maxisorp). For the blank, 8 µL of the 0.05mM Tris buffer was added to the wells. To each well, 100 µL was added of a mixture containing 100mM NaCl, 50mM Tris–HCl (pH 8.95), 5mM CaCl2, 2mM ATP and 20 mM B-Mercapthanol. The plate was incubated at 37°C for 10 min and then the reaction was immediately stopped, by adding 25 uL of acid molybdate solution (1.25% ammonium molybdate in 2.5mM H2SO4). Immediately after termination of the reaction, 2 µL of a reducing solution (0.11mM NaHSO3, 0.09mM Na2SO3 and 8mM 1-amino-2-naphthol-4-sulphonic acid) was added to each well and the plate was incubated at 37°C for 20 min [22]. Plates were read at a FLUOstar OPTIMA ELISA plate reader (BMG Technologies) at 660nm. Readings were quantified by comparison with an inorganic phosphate standard curve (1, 0.5, 0.25, 0.125, 0.06125, 0.03125, 0.015625 mM of sodium phosphate).
Wolbachia infection was confirmed through PCR to detect both mosquito (apyrase gene: ApyF: 5′-TTTCGACGGAAGAGCTGAAT-3′ and ApyR: 5′-TCCGTTGGTATCCTCGTTTC-3′) and Wolbachia (IS5-F: 5′-CTGAAATTTTAGTACGGGGTAAAG-3′ and IS5-R: 5′- CAAGCATATTCCCTCTTTAAC-3′) sequences. Saliva screening to check the presence of Wolbachia was also done via PCR (with IS5 primers) using saliva samples of infected and non-infected mosquitoes. Mosquito sequences in this case were detected with primers for the ribosomal protein gene RpS17 [23].
In all cases, general linear models were employed to examine the effects of the variables age and infection status and their interaction with one another. Models demonstrating significance for the variable infection status were then followed by individual t-tests examining the differences between infected and uninfected mosquitoes for each age class. The proportion of infected and uninfected mosquitoes that obtained blood meals were examined using Mann-Whitney U tests instead of linear models, given the deviation of the data from normality, and were based on four populations of mosquitoes. Chi-square 2×2 contingency tests were employed to examine the relationship between observed behavioral traits and lack of feeding success. The correlation between these traits was quantified using a cox-proportional hazards model for age, with the behavioral traits and lack of blood meal success covariates. All statistical analyses were carried out in STATISTICA v8 (StatSoft, Tulsa, OK).
We measured the time mosquitoes spent from first contact with a human volunteer until the insertion of the insect's mouthparts as a measure of pre-probing time. All feeding trials were carried out with individual mosquitoes, which had been starved prior to the assay, at four adult ages (5, 15, 26 and 35-days-old). Mosquitoes that never successfully achieved a blood meal were excluded from this analysis. Overall both age (df = 3, F = 13.73, P<0.0001) and infection status (df = 1, F = 23.18, P<0.0001) had a significant effect on the length of pre-probing time. On average infected mosquitoes spend more time pre-probing especially as they age (Fig. 1). This change with age is clearly exhibited by a significant interaction between the variables age and infection status (df = 3, F = 8.11, P<0.0001). At five days of age infected and uninfected mosquitoes do not differ in their pre-probing time (df = 78, t = 0.63, P = 0.52), which lasted on average 11 seconds. Uninfected mosquitoes maintained the same foraging time as they aged, while wMelPop insects exhibited a steady and significant increase (15d: df = 75, t = −3.37, P = 0.0012; 26d: df = 63, t = −4.17, P = 0.014; 35d: df = 48, t = −2.25, P = 0.0034), reaching a mean length of 45 sec by 35 days of age (Fig. 1).
In the same feeding trials described above, the length of time between insertion of mouthparts and the first visible sign of blood in the abdominal pleura [21] was recorded as probing time for the mosquitoes. As with pre-probing time, the variables of age (df = 3, F = 11.36, P<0.0001), infection status (df = 1, F = 29.46, P<0.0001) and the interaction (df = 3, F = 10.56, P<0.0001) between these two variables were highly significant. Infected and uninfected mosquitoes did not differ in their probing time (∼33 sec) at 5 (df = 78, t = −0.46, P = 0.64) and 15 (df = 75, t = 1.43, P = 0.15) days of age (Fig. 2). In contrast, infected mosquitoes at 26 (df = 63, t = −3.76, P<0.001) and 35 (df = 48, t = −4.06, P<0.001) days of age took significantly longer during probing, exhibiting up to a seven-fold increase in their probing time relative to uninfected mosquitoes (Fig. 2).
In the assays detailed above we then compared the ability of infected and uninfected mosquitoes to obtain blood meals (Fig. 3) using Mann-Whitney U tests. At 5 (Z = 0, P = 1) and 15 (Z = 0, P = 1) days of age infected and uninfected mosquitoes did not differ in their success. At 26 (Z = −2.39, P = 0.020) and 35 (Z = −2.39, P = 0.020) days of age infected mosquitoes were less successful at obtaining blood meals in comparison to their uninfected counterparts. Only Wolbachia infected mosquitoes refused to land on the hand during the assay or landed but did not take a blood meal. Percentages of individuals that did not land on the human were: 2.5% at 15 d; 0% at 26 d and 7.5% at 35d. Percentages that landed but that did not feed were 5% at 15 d; 37.5% at 26d and 67.5% at 35 d.
It is important to note that as infected mosquitoes aged, the frequency of events where they pierced the skin did not increase despite failed attempts at feeding (Fig. 4). A general linear model revealed that age (df = 3, F = 20.47, P<0.0001), infection (df = 3, F = 29.12, P<0.0001) and age X infection (df = 3, F = 27.18, P<0.0001) were significant determinants of the number of probings. Subsequent t-tests comparing the number of probings between infected and uninfected mosquitoes at each of the age points (data not shown), however, demonstrated that only 35 day old (df = 1, t = −8.44, P<0.0001) mosquitoes differed. In this case, uninfected females probed more on average per session (1.05±0.05) than wMelPop infected mosquitoes (0.3±0.073). This is due to other behaviors, which impaired the infected mosquitoes to feed (see below).
We recently have reported the appearance of a “bendy” proboscis in association with wMelPop, which was the inability of the mosquito to properly orient its mouthparts and insert the stylet into the skin [10]. Here we quantified the occurrence of this trait. The bendy proboscis was never observed in any of the uninfected mosquitoes in this study, nor was it present in a small cohort of very old mosquitoes (∼90 days) we examined in a second pilot study. The phenotype was also not present in 5 day-old infected mosquitoes. The trait first appeared at a low level (2.5%) in 15 day-old wMelPop infected mosquitoes and rose to a frequency of 65% by 35 days of age (Fig. 5).
Another phenotype observed, although in lower frequencies, was the jittering action of the insect body (named here as “shaky”) when the mosquito was sitting on top of the human hand (Fig. 5, Video S1). The association between each of these traits and lack of success in blood meal acquisition was explored using 2×2 contingency tests in each of the age classes where the trait was expressed. There was a significant association between the failure to obtain a blood meal and both the bendy phenotype (26d: df = 1, χ2 = 14.1, P = 0.0002; 35d: df = 1, χ2 = 11.8, P = 0.0006) and the shaky phenotype (35d: df = 1, χ2 = 4.2, P = 0.038). Using survival analysis we obtained estimates of the correlation between lack of feeding success and the bendy phenotype (0.63) and the shaky phenotype (0.19). These correlations reveal the presence of a relationship between the traits and success in feeding, but do not completely explain lack of success. There are mosquitoes in the older age classes that fail to feed and that are not shaky or bendy. To discard any possibility that this other abnormal phenotypes were due to the lack of blood feeding, which could have physiologically compromised the mosquitoes we also blood fed females of both groups when they were 3 to 5-days-old and then after 38 days evaluated their feeding behavior. None of the wMelPop mosquitoes were able to feed and all presented the bendy proboscis, although all the tetracycline-treated mosquitoes successfully imbibed blood (data not shown).
To check whether the probing behavior and the additional phenotypes we observed were due to differences in saliva volume and salivary gland apyrase activity we measured both traits in infected and uninfected mosquitoes at three adult ages. Apyrase activity (Fig. 6A) did not differ in infected and uninfected mosquitoes regardless of age (df = 1, F = 0.44, P = 0.51). Infection status (df = 1, F = 11.99, P<0.01) and age (df = 2, F = 14.54, P<0.0001), however, were determinants of saliva volume (Fig. 6B) and on average infected mosquitoes produced less saliva. When saliva volumes of infected and uninfected mosquitoes were compared to each other for each age class, only the 26 days old mosquitoes were significantly different (df = 1, t = −2.9, P<0.01). During collection, 80–100% of the mosquitoes produced saliva droplets. Infection status was not a predictor of whether saliva was produced or not (data not shown).
To begin to dissect the functional role of Wolbachia in these feeding effects we sought to determine the presence of the bacterium in both the saliva and salivary glands. PCR amplification of the Wolbachia wsp gene or mosquito apyrase has shown only the presence of Wolbachia in salivary glands, but not in saliva (see Figure S1). The transposable element IS5, present in at least 13 copies within the bacteria genome [24], was also used in extra samples as a very sensitive PCR target (N = 16 of each group) but no amplification was obtained (data not shown). These results are supported by the size of the intracellular Wolbachia (around 1µm in diameter) [4] and the diameter of mosquito salivary ducts (also about 1 µm) [25], which indicate that even if Wolbachia was able to be present in the secreted salivary fluid it would be unlikely to be able to freely move through the salivary ducts.
The results presented here indicate that infection with wMelPop alters the blood-feeding behavior of female A. aegypti mosquitoes in a manner that intensifies with increasing mosquito age. Older infected mosquitoes spend more time pre-probing and probing, commonly exhibit evidence of shaking and a bendy proboscis, produce less saliva and overall are less successful in obtaining blood meals. None of our experiments revealed any evidence of Wolbachia in the saliva or Wolbachia-associated changes in salivary apyrase activity.
Female mosquitoes must balance the feeding time required to obtain a sufficient blood meal against the risk of death from human defensive behavior upon their detection [26]. On the other hand, parasites or viruses that are transmitted through the insect saliva benefit from feeding times that enhance transmission but possibly cause a higher risk for the mosquito. Parasite manipulation of insect feeding behavior often includes increases in the insect's probing time or number of probings. This has been seen in mosquitoes infected with malaria parasites [27] and viruses [28], and in trypanosome-infected triatomine bugs [29]. As Wolbachia is not transmitted via saliva, the increased length of time infected mosquitoes are taking to feed could be advantageous only if blood-meal sizes acquired were greater and subsequent fecundity rates higher. Our group recently demonstrated evidence to the contrary, with infected females taking smaller blood meals on average [10]. This in conjunction with the data reported here, with overall reduced rates of feeding in combination with physiological changes like the observed “bendy” and “shaky” traits, suggest that the feeding biology is not likely to be a Wolbachia parasite “adaptation.”
In the absence of evidence for a specific Wolbachia manipulation of the host, blood-feeding effects are likely explained as a by-product of infection or as the direct result of pathogenesis on host cells or tissues. Wolbachia are found in many different insect tissues, especially in nervous tissue, muscles and reproductive systems [9],[30]. While in Drosophila the deleterious effects of wMelPop are related to age and bacterial density [4], the relationship between bacterial density and pathogenicity has yet to be empirically determined for wMelPop-infected A. aegypti. The decreasing ability to feed with increasing age in combination with the increased prevalence of shaking behavior and the bendy proboscis could be explained by direct damage on host cells or tissues, resulting from Wolbachia over-replication. Interestingly, failure to obtain a blood meal cannot be completely explained by these two traits. Alternatively, Wolbachia infection may be having direct and targeted effects on neuro-regulators, which would influence their ability to forage for blood and actually feed. The bendy proboscis and shaky phenotypes may represent very extreme endpoints in the process.
The wMelPop strain of Wolbachia was initially targeted for biocontrol development given its life-shortening capacity [7]. More recently, the presence of wMelPop has been shown to block the accumulation of dengue virus in A. aegypti [9]. The pattern of the reduced blood-feeding success seen here, like life-shortening leaves the young breeding individuals unaffected while targeting older members of the population. While the dengue transmission rates of wMelPop-infected mosquitoes are yet to be estimated, initial concerns that infected mosquitoes, failing to successfully feed, might actually bite more and hence possibly enhance viral transmission appear unfounded. While probing frequency decreased with age in Wolbachia infected individuals probing time increased. This was compensated by a reduction in saliva produced by Wolbachia infected individuals, which should counteract any potential increases to pathogen transmission risk in Wolbachia infected individuals. In conclusion, the presence of Wolbachia bacterium in A. aegypti mosquitoes significantly reduces their feeding success in an age dependent manner that is more likely a by-product of virulence than a direct manipulation of host behavior. These effects on feeding could substantially improve the efficacy of wMelPop biocontrol strategies in combination with the traits of life-shortening and viral blocking.
|
10.1371/journal.pgen.1003091 | MicroRNA–Mediated Repression of the Seed Maturation Program during Vegetative Development in Arabidopsis | The seed maturation program only occurs during late embryogenesis, and repression of the program is pivotal for seedling development. However, the mechanism through which this repression is achieved in vegetative tissues is poorly understood. Here we report a microRNA (miRNA)–mediated repression mechanism operating in leaves. To understand the repression of the embryonic program in seedlings, we have conducted a genetic screen using a seed maturation gene reporter transgenic line in Arabidopsis (Arabidopsis thaliana) for the isolation of mutants that ectopically express seed maturation genes in leaves. One of the mutants identified from the screen is a weak allele of ARGONAUTE1 (AGO1) that encodes an effector protein for small RNAs. We first show that it is the defect in the accumulation of miRNAs rather than other small RNAs that causes the ectopic seed gene expression in ago1. We then demonstrate that overexpression of miR166 suppresses the derepression of the seed gene reporter in ago1 and that, conversely, the specific loss of miR166 causes ectopic expression of seed maturation genes. Further, we show that ectopic expression of miR166 targets, type III homeodomain-leucine zipper (HD-ZIPIII) genes PHABULOSA (PHB) and PHAVOLUTA (PHV), is sufficient to activate seed maturation genes in vegetative tissues. Lastly, we show that PHB binds the promoter of LEAFY COTYLEDON2 (LEC2), which encodes a master regulator of seed maturation. Therefore, this study establishes a core module composed of a miRNA, its target genes (PHB and PHV), and the direct target of PHB (LEC2) as an underlying mechanism that keeps the seed maturation program off during vegetative development.
| Seed development can be conceptually divided into two phases: namely the morphogenesis phase, in which cell division is active and all the major organs are formed, and the maturation phase, in which cells enlarge and storage reserves are synthesized and accumulated. Expression of the seed maturation program is tightly controlled such that it only occurs during the late phase of seed development. To uncover the molecular mechanisms underlying the repression of seed genes during vegetative development, we performed a reporter-assisted genetic screen, and one mutant identified is a weak allele of ARGONAUTE1 (AGO1) that displays ectopic seed gene expression. We then performed a series of transgenic and genetic analyses to search for the molecular mechanisms underlying the mutant phenotype. We first demonstrate that the decrease in miR166 in ago1 is a major cause of the mutant phenotype. Further, we show that the targets of miR166, type III HD-ZIP transcription factors PHB and PHV, are sufficient for derepressing seed maturation genes in seedlings, likely by binding directly to the promoter of a master regulator gene of maturation. Thus, this work establishes a miRNA–mediated pathway that represses the embryonic program and also establishes PHB/PHV as direct activators of the maturation program.
| Seed maturation is a highly coordinated developmental phase in which storage reserves, including seed storage proteins (SSPs), are synthesized and accumulated to high levels. The maturation genes need to be repressed, however, in order to allow seedling development to occur. Indeed, these genes are not expressed in vegetative organs of the plant [1]. Research in the past decade with the model plant Arabidopsis has led to the identification of repressors of seed maturation genes in vegetative organs (reviewed in [2]), including chromatin-remodelling ATPases [3]–[5], polycomb group (PcG) proteins [6]–[9], histone deacetylases [10], and DNA-binding transcription factors [11]–[13]. However, our understanding of the molecular mechanisms that repress the seed maturation program during vegetative development remains fragmented, and thus continued efforts are needed to identify additional factors involved and, more importantly, the molecular and functional links between the various components.
In Arabidopsis, ABA-INSENSITIVE3 (ABI3), FUSCA3 (FUS3), LEC1 and LEC2 are master regulators of seed maturation [14]–[17], and they regulate one another [18], [19]. ABI3, FUS3 and LEC2 are closely-related members of a plant-specific B3-domain transcription factor family. LEC1 encodes a novel homolog of the CCAAT-binding factor HAP3 subunit. Loss-of-function mutations in ABI3, FUS3, and LEC1 give rise to pleiotropic seed phenotypes including a strong reduction of SSPs. These regulatory genes are predominantly expressed in the seed. When misexpressed in vegetative tissues, they induce ectopic expression of the SSP genes and even the formation of somatic embryos [15], [17], [20]–[22]. It remains poorly understood, however, how the expression and activity of these master regulators are in turn regulated.
Small RNAs of 20–30 nucleotides (nt) have emerged as key sequence-specific regulators of gene expression that influence almost all aspects of plant biology (reviewed in [23]–[26]). There are two major types of small RNAs in plants, microRNA (miRNA) and small interfering RNA (siRNA). Plant miRNAs are generated from longer precursors arising from defined genomic loci – the MIRNA genes. The biogenesis of miRNAs involves several evolutionarily conserved families of proteins, including DICER-LIKE (DCL), ARGONAUTE (AGO), HUA ENHANCER 1 (HEN1), and HASTY (HST). Plant miRNAs regulate target mRNAs temporally and spatially through transcript cleavage and/or translational inhibition. Conserved miRNAs tend to target transcription factor genes that play crucial roles in almost all aspects of plant development. Plants are rich in endogenous siRNAs, which can be classified into several types, such as trans-acting siRNAs (ta-siRNAs), natural cis-antisense transcripts-associated siRNAs, and heterochromatic siRNAs.
Here, we show that mutations in AGO1 resulted in the ectopic expression of seed maturation genes in seedlings. Taking advantage of the weak ago1 allele identified in this work, we were able to identify the miRNA species (miR166) responsible for the repression of seed genes. We demonstrated that targets of miR166, the class III homeodomain leucine zipper (HD-ZIPIII) family of transcription factor genes, PHB and PHV, are positive regulators of seed genes. Further, we provided evidence to suggest that PHB acts directly at LEC2. This work thus uncovered an important role of miR166 in the repression of seed genes during seedling development.
We have recently conducted a genetic screen in Arabidopsis to isolate mutants exhibiting ectopic expression of a soybean β-conglycinin gene promoter:GUS transgene (βCG:GUS), which is normally expressed only in the seed [5], [27], [28]. Here, we describe the characterization of one of the mutants identified from the screen, initially named essp5 (ectopic expression of seed storage proteins 5). The essp5 mutant plants exhibited strong ectopic GUS activity in leaves, but not in other organs (Figure 1A–1D). In addition, the mutant plants had pleiotropic developmental defects, such as late flowering, narrow and dark green leaves, shorter siliques and fewer seeds (Figure 1B and Figure S1).
The mutation segregated as a recessive allele and was mapped to the AT1G48410 gene (Figure S2), which encodes AGO1, the major effector protein that associates with small RNAs [29]. A single missense mutation in this gene was identified that would lead to the conversion of a leucine residue at position 740 to a phenylalanine. The leucine 740 is a highly conserved residue in the PIWI domains of AGOs from diverse species (Figure 1E). A number of mutant ago1 alleles have been described previously and their pleiotropic morphological phenotypes have been documented [29]–[32]. The morphological phenotypes of essp5 resemble those documented for weak ago1 alleles. We obtained T-DNA insertion mutants of ago1, including ago1-36 (SALK_087076), ago1-39 (SALK_089073), ago1-40 (SALK_076199), ago1-41 (SALK_096625), and ago1-42 (SALK_116845) (Figure 1F), crossed them with βCG:GUS and examined GUS expression in F2 progeny seedlings. The T-DNA insertion lines, regardless of their insertion sites, all displayed similar morphological phenotypes: long rod-shape cotyledons, delayed emergence of true leaves, and premature death with only a couple of small true leaves. As shown in Figure 1G–1J, ectopic GUS activity was clearly observed in several T-DNA alleles with insertion sites located throughout the gene. Furthermore, we performed an allelism test to provide additional evidence that essp5 is a weak ago1 allele. A weak ago1 allele (ago1-25) that exhibits similar morphological phenotype [33] was crossed with essp5 and the F1 progeny were examined for GUS activity. As shown in Figure S3, the F1 seedlings displayed ectopic GUS expression, indicating that the essp5 GUS phenotype cannot be complemented by a weak ago1 allele. Together, these observations suggest that essp5 is allelic to AGO1 and thus designated as ago1-100.
To find out whether the endogenous seed maturation genes are indeed ectopically expressed in ago1 mutant seedlings, we performed northern blot analysis to profile the expression of both the 2S and 12S storage protein genes. As shown in Figure 2A, the transcripts of the storage protein genes are highly accumulated in the two strong alleles, ago1-41 and ago1-42, both with T-DNA insertion sites located in the 5′ end of the gene but barely detectable in ago1-100/essp5 and other weak alleles with insertion sites located in the middle and 3′ end of the AGO1 gene. We further examined the transcript levels of the four master regulators by quantitative real-time RT-PCR (qRT-PCR). In line with the ectopic expression of the storage protein genes, all the master regulators are expressed to varying levels in the mutants, especially ago1-42 (Figure 2B). In addition, we also profiled the temporal expression pattern of the maturation genes, using 2S2 as a marker. ago1-41 seedlings from 5 d to 19 d after germination were examined. As shown in Figure 2C, the 2S2 transcript peaked in abundance at around 13 d, but was clearly detected throughout the time course. These expression analyses clearly demonstrate the involvement of AGO1 in the repression of seed maturation genes.
AGO1 associates with miRNAs and some endogenous siRNAs to mediate their activities [34]. AGO1 association also stabilizes the small RNAs such that ago1 mutants show a reduction in the steady-state levels of miRNAs and siRNAs [35]–[36]. To confirm that the ectopic expression of seed maturation genes in ago1 mutants is due to defects in small RNA biogenesis or activity, we examined seed gene expression in seedlings of loss-of-function alleles of genes commonly involved in small RNA biogenesis. For this purpose, we obtained mutant alleles of HEN1, hen1-5 (SALK_049197) and hen1-6 (SALK_090960), and of HST, hst-1 [37], hst-15 (SALK_079290) and hst-16 (SALK_056352). We introduced these mutations, individually, into the βCG:GUS background and examined GUS expression in the F2 generation. We were able to detect clear ectopic GUS activity in the hen1 backgrounds (hen1-5; Figure 3A and 3B), but not in the hst alleles. We then generated double mutants between ago1, hen1, and hst. As shown in Figure 3C–3J, both ago1-41 hst-16 and hen1-5 hst-16 double mutants exhibited a high level of expression of the storage protein genes and the master regulators. The ago1-41 hen1-5 plants were very small, which precluded seed gene expression analysis. These results indicate synergistic genetic interactions among AGO1, HEN1, and HST in repressing seed genes during seedling development and, more importantly, the involvement of a small RNA pathway(s) in this repression process.
Since AGO1, HEN1 and HST are essential players in small RNA biogenesis and are involved in several small RNA pathways [38], it was necessary to determine which pathway underlies the mutant phenotype. To this end, we took advantage of pathway-specific components to define the specific pathway responsible for the ago1 mutant phenotype. Specifically, RDR2 is an essential component of the heterochromatin pathway and RDR6 is required for the biogenesis of trans-acting siRNAs. We obtained and introduced the rdr2-1 (SAIL_1277_H08) and rdr6-11 [39] mutations into the βCG:GUS background by genetic crosses and examined GUS expression in the F2 progeny. A large number of F2 seedlings were stained for GUS and no ectopic GUS activity was observed in either population. This genetic evidence suggests that it is unlikely that the trans-acting siRNA or the hc-siRNA pathway is involved in the repression of seed genes in seedlings. Since AGO1, HEN1, and HST all act in miRNA biogenesis, a miRNA(s) is thought to be a strong candidate for the repression of seed genes during vegetative development.
To provide evidence that a miRNA pathway is indeed underlying the mutant phenotype, we examined the steady-state levels of a number of conserved miRNAs in ago1 and other mutant backgrounds. In ago1 mutants, it was documented previously that the accumulations of a number of conserved miRNAs decline markedly and their target gene transcripts are concomitantly elevated [35]–[36]. Here, we performed a miRNA northern blot analysis to examine and compare the accumulation of conserved miRNA species in ago1, hen1, hst, and the two double mutants, ago1 hst and hen1 hst. As shown in Figure 4, we confirmed the published observation for ago1 in that all the miRNAs examined were clearly reduced. More importantly, we observed further reduced accumulation of most examined miRNAs in ago1 hst and hen1 hst double mutants compared to the single mutants (Figure 4). These findings are consistent with a genetic model for explaining the ectopic seed gene expression in ago1and other mutants: the steady-state level of a specific miRNA was reduced below a threshold to lead to the ectopic expression of its target gene, which encodes a positive regulator of seed maturation genes leading to the ectopic expression of seed genes in leaves.
Post-germination repression of seed genes is critical in order for the seedling to develop normally. We thus reasoned that such a fundamental developmental program should be controlled by a conserved miRNA(s). Therefore, to find out which miRNA was involved in conferring the essp5/ago1-100 GUS phenotype, we over-expressed each of the 15 conserved miRNA species, as listed in [23] and Table S1, in the essp5/ago1-100 background and examined GUS expression of the resulting transgenic plants. The transgenic plants overexpressing different miRNAs displayed unique morphological phenotypes, which are consistent with previously published observations (reviewed in [23]).
Analysis of leaf GUS expression was conducted in the T2 generation. For each miRNA transgene, multiple independent transgenic lines were analyzed (in most cases 10 lines); and for each line, at least 20 T2 progeny homozygous for essp5 were stained for GUS activity. We only observed loss of leaf GUS activity in miR166 and miR156 overexpressing lines. In this study, we have focused on the characterization of miR166. In total, we only obtained four miR166 transgenic lines, miR166ox-1-4, of which two showed clear loss of leaf GUS activity (miR166ox-1-2) while the other two (miR166ox-3-4) did not show as obvious a change compared with essp5/ago1-100 seedlings (Figure 5A–5D). The extremely low rate of positive transgenic plants for miR166 is likely due to the fact that some transgenic seedlings failed to develop the shoot apical meristem and could not survive in soil, as observed by others [40]. To confirm that the loss of leaf GUS activity in the transgenic lines was indeed due to the elevated accumulation of miR166, a northern blot analysis was performed. As shown in Figure 5E, there were clearly higher levels of miR166 in lines miR166ox-1-2 than lines miR166ox-3-4. In addition, we observed the formation of aberrant structures on leaves of miRNA166ox-1-2 (Figure 5F). Similar aberrant structures were observed by Zhou et al in miRNA166 overexpressors [40]. These observations suggest that the reduction of miRNA166 and the concomitant accumulation of its target gene transcripts are likely the cause underlying the ectopic GUS phenotype of essp5/ago1-100 seedlings.
To demonstrate that the specific loss of miR166 can cause the ectopic expression of seed maturation genes, we obtained the recently developed transgenic lines that exhibit a dramatic reduction in miR165/166 accumulation achieved by the expression of a short tandem target mimic (STTM165/166) [41]. RNA blot analysis was performed to examine the expression of seed storage protein genes in these transgenic lines, using 2S2 as a probe. As shown in Figure 5G, the 2S2 gene is clearly expressed in the strongest line (STTM165/166-48), but not detectable in a weaker line (STTM165/166-31). This observation indicates that miR166 plays an important role in repressing seed genes in seedlings.
It has been well established that the miR165/166 family miRNAs target the transcripts of the HD-ZIPIII genes, controlling their expression level and domain, to fulfill their roles in plant development including leaf polarity determination [42]–[45]. The HD-ZIPIII family consists of five transcription factors (REV, PHB, PHV, AtHB8, and AtHB15), and they play both redundant and unique roles in diverse plant developmental processes [46]. In this context, it is worth noting that the transcript level of PHB was found to be decreased in miR166 overexpressors (Figure S4). To investigate whether the HD-ZIPIII proteins are responsible for conferring the ectopic GUS phenotype of essp5/ago1-100, we introduced loss-of-function mutations in PHB and PHV genes into essp5/ago1-100 by genetic crosses and examined leaf GUS expression in F2 and F3 seedlings. A large number of F2/F3 seedlings were examined and no clear loss of leaf GUS activity was observed in phb essp5 or phv essp5. We further introduced phb phv double mutations into the essp5/ago1-100 background, but still saw no detectable loss of leaf GUS activity. Obviously, the potential redundancy among the five HD-ZIPIII genes could be confounding the genetic analyses above.
Next, taking advantage of the previously identified gain-of-function mutations in HD-ZIPIII family genes, we investigated whether these proteins are sufficient to cause the ectopic expression of seed genes. These gain-of-function alleles have mutations in the miR166 target regions to cause a mismatch between the miRNA and the target mRNA and thus render the transcripts resistant to miRNA-mediated degradation and consequently the ectopic accumulation of HD-ZIPIII transcripts. First, the gain-of-function mutations phb-1d [47] and phv-1d [44] were introduced into the βCG:GUS background by genetic crosses and ectopic GUS activity was examined in F2 seedlings (Figure 6A–6I). Meanwhile, another gain-of-function phb allele driven by the CaMV 35S promoter, 35S:PHB G202G [43], was also introduced into the βCG:GUS background by Agrobacterium-mediated transformation and GUS activity was examined for each independent T1 plant (Figure 6A and 6J–6M). As shown in Figure 6A–6M, ectopic GUS activity was clearly observed for phb-1d, phv-1d, and 35S:PHB G202G. Further, we performed northern blot and qRT-PCR analyses to examine the ectopic expression of endogenous seed maturation genes in the gain-of-function mutant plants (Figure 6N and 6O). Two representative maturation genes 2S2 and 2S3 were clearly expressed in the mutant seedlings (Figure 6N). Similarly, the four master regulators were all elevated to varying levels (Figure 6O). In addition, a gain-of-function REV mutant, rev-10d [42], was also analyzed in the βCG:GUS background but no ectopic GUS activity was detected. In summary, our gain-of-function genetic evidence indicates that the HD-ZIPIII proteins PHB and PHV are each sufficient for ectopic expression of seed genes.
HD-ZIP proteins are plant-specific transcription factors and named for the combination of homeodomain and leucine zipper domains at their N termini [48]. They bind a palindromic DNA sequence in vitro as dimers [49]. To determine whether PHB acts directly at maturation gene loci, we performed chromatin immunoprecipation (ChIP) experiments to examine PHB occupancy at the promoters of these genes.
For the ChIP assays, we generated Arabidopsis plants transgenic for a YFP-tagged gain-of-function allele of PHB under the control of the PHB native promoter (PHB:PHB G202G-YFP). Morphologically, the transgenic plants resemble the phb-1d mutant (Figure 7A–7D). When the transgene was introduced into the βCG:GUS background, it resulted in ectopic GUS activity (Figure 7E–7H). Expression of the master regulators of seed maturation in the transgenic seedlings was also examined by qRT-PCR. As shown in Figure 7I, these genes were ectopically expressed to similar levels compared to those of phb-1d, and the expression levels in homozygous seedlings were clearly higher than those in the hemizygous siblings. These observations demonstrate that the PHB:PHB G202G-YFP plants resemble phb-1d. In addition, we observed, at a low frequency, disorganized growth and/or formation of somatic embryo-like structures in the transgenic plants (Figure 7J–7L). Some parts of these plants could be stained by the neutral lipid dye fat red (Figure 7M–7O), indicating a high level accumulation of seed storage-specific triacylglycerols in these plants (fat red/sudan red stains only seed storage-specific lipids).
ChIP was performed with anti-GFP antibodies and an Arabidopsis line transgenic for GFP driven by the CaMV 35S promoter (35S:GFP) was used as a negative control. The ChIP DNAs were analyzed by qPCR to examine the enrichment of promoter region genomic DNAs of the four master regulator genes. One region in the LEC2 promoter was highly enriched relative to the 35S:GFP and no antibody controls (Figure 7P), but no enrichment was found for the promoter regions of other maturation genes examined (Figure S5). The enrichment level in homozygous plants was about double that in the hemizygous siblings, consistent with the ectopic expression level of seed genes in these plants (Figure 7I). Interestingly, a partial palindromic sequence, aaAATCATTAC, was found in the vicinity of the enriched genomic region in the LEC2 promoter, but not at other maturation loci. This sequence is very similar to the HD-ZIPIII binding consensus sequence, GTAAT(G/C)ATTAC, derived from an in vitro binding site selection experiment [49]. These observations suggest that PHB, when ectopically expressed, binds to the LEC2 promoter and activates the expression of the gene. LEC2 can, in turn, activate a network of maturation-related genes including ABI3, FUS3, LEC1, and the SSP genes (Figure 8).
In this work, we first identified a weak EMS ago1 allele, which exhibited ectopic expression of a GUS reporter driven by a seed gene promoter. Taking advantage of the weak ago1 allele and its GUS phenotype, we then performed a series of transgenic and genetic analyses to search for the molecular mechanisms underlying the mutant phenotype. We first demonstrated that miR166 reduction is a major cause of the mutant phenotype and further showed that the targets of miR166, PHB and PHV, are sufficient for derepressing seed maturation genes in seedlings. Finally, our ChIP assay using a tagged PHB transgenic line suggests that PHB may act directly at the LEC2 promoter (summarized in Figure 8). However, in addition to LEC2, PHB may also regulate other factors that in turn regulate seed maturation genes directly or indirectly. Future studies, such as ChIP-seq, are needed to address this issue. Therefore, this work has added miR165/166 to the documented repertoire of postgermination repressors of the embryonic program (reviewed in [2]), and more importantly, established PHB, and possibly PHV, as direct positive regulators of the master regulator of seed maturation LEC2. A major future challenge in the field is to find out the genetic and molecular relationships amongst the various players, including transcription factors, chromatin remodelers and modifiers, and the newly added miRNA, and build an integrated genetic network.
Given the well-established expression patterns and roles of miR166 and its targets in leaf polarity determination (reviewed in [50], [51]), an obvious outstanding question is why the normal expression of the PHB and PHV genes in the adaxial domain of leaf primordia in wild type plants is not sufficient to cause the ectopic expression of seed maturation genes. miR165/166 is concentrated in the abaxial domain to restrict the expression of the HD-ZIPIII transcription factor genes to the adaxial domain in the lateral organs in Arabidopsis [42]–[44] and maize [45]. In phb and phv gain-of-function mutants, the expression of PHB and PHV is not restricted to the adaxial domain but extends into the entire primordium. We observed ectopic expression of seed maturation genes only in these gain-of-function mutants, indicating that the normal, adaxial expression of the HD-ZIP III genes is not sufficient to activate the seed maturation program. There could be at least two underlying reasons. First, the ectopic expression of the seed maturation genes in the phb and phv gain-of-function mutants only occurs in the abaxial domain. In this scenario, the lack of necessary co-factors or the presence of negative factors in the adaxial domain may prevent the HD-ZIPIII genes from activating the seed maturation genes. Alternatively, it might be a matter of thresholds – the adaxial domain normally does not have sufficient levels of HD-ZIPIII expression to trigger the seed maturation program, but when the miRNA is compromised, the expression level is high enough to trigger the program. Our preliminary observation is in support of the first scenario. GUS expression along the adaxial-abaxial axis in essp5/ago1-100 was examined and GUS activity was found only on the abaxial side (Figure S6). In addition, interestingly, GUS was also observed in both the upper and lower epidermal cells (Figure S6).
The seed maturation program is a tightly regulated developmental process. Mechanisms are in place to not only ensure its repression during seedling development but also prevent its precocious induction during early embryogenesis [2], [52]. The induction of seed maturation is also referred to as the morphogenesis-to-maturation phase transition of embryogenesis. While our studies have established miR165/166 and implicated miR156 as players in the repression of the seed maturation program in vegetative development, two recent studies have also revealed important roles of miRNAs in regulating the morphogenesis-to-maturation phase transition [53], [54]. Of these, the work of Nodine and Bartel [53] demonstrated that miR156 and two of its target genes SPL10 and SPL11 play a major role in early embryo patterning and in preventing the precocious expression of maturation genes. An obvious question is whether miR165/166 also acts similarly in early embryogenesis to control the morphogenesis-to-maturation phase transition. Previous studies have shown that PHB and PHV promote embryonic development, and that the expression of these genes must be repressed by miR165/166 for embryonic development to proceed normally. For example, Grigg et al showed that serrate (se) mutants cause ectopic expression of PHB and PHV in the root pole of embryos, and that the embryonic lethal phenotype of se mutants can be rescued by loss-of-function mutations in PHB and PHV [55]. Smith and Long also showed that PHB and PHV promote shoot development during embryogenesis [56]. These studies focused on the roles of the miR165/166-PHB/PHV module in early embryo patterning. Our finding that this module plays an important role in repressing seed maturation genes during seedling development prompted us to test its role in the morphogenesis-to-maturation phase transition. We performed a ChIP analysis using a transgenic line expressing a tagged PHB driven by its endogenous promoter (PHB:PHB-YFP). Preliminary data suggests that PHB acts directly at LEC2 during embryogenesis (Figure S7). Future investigations are needed to sort out the contributions of each miRNA to the repression of the seed maturation program during the pre- and post-maturation stages.
Seeds of mutants including the T-DNA insertion mutants were obtained from the ABRC, unless otherwise indicated. Seeds were stratified at 4°C for 3-d. Then the seeds were sowed on soil or on agar plates containing 4.3 g/L Murashige and Skoog nutrient mix (Sigma-Aldrich), 1.5% sucrose, 0.5 g/L MES (pH 5.7), and 0.8% agar. Plants were grown under 16 h-light (22°C)/8 h-dark (20°C) cycles; except that the phb-1d/+ and phv-1d/+ mutants were grown at 17°C during reproductive development as described [47]. Homozygous T-DNA insertion mutants were identified by genotyping.
The mutant essp5 was isolated from the same genetic screen as essp1 and essp3 [5], [27]. For genetic mapping of the essp5 mutation, mutant plants were crossed with wild type plants of the Ler ecotype. A total of 644 homozygous essp5 mutants were collected from the F2 segregating population. Genomic DNA extracted from these seedlings was used for PCR-based mapping with simple sequence polymorphism markers, and the essp5 locus was mapped to a ∼127 kb genomic interval on BACs F11A17, T1N15 and F9P7 on chromosome one (17,852–17,979 kb). Sequencing of the genomic region revealed a mutation in At1g48410.
The modified GUS staining solution (0.5 mg/mL 5-bromo-4-chloro-3-indolyl-glucuronide, 20% methanol, 0.01 M Tris-HCl, pH 7.0) was used [5]. Seedlings immersed in GUS staining solution were placed under vacuum for 15 min, and then incubated at 37°C overnight. The staining solution was removed and samples were cleared by sequential incubation in 75% and 95% ethanol. Fat red staining was performed by incubating samples in a saturated solution of Sudan red 7B (Sigma) in 70% ethanol for 1 h at room temperature. Samples were then rinsed with 70% ethanol [57].
Plants grown on MS media were used for gene expression analyses. RT-PCR and RNA blot analyses were preformed as described previously [5]. Probes for detecting transcripts of the CRA1, CRB, and CRC genes were designed based on Pang et al [58]. Real-time PCR was conducted using the Bio-Rad CFX96 real-time PCR detection system and the SsoFast™ EvaGreen® Supermix kit (Bio-Rad Laboratories, Inc.). Data from three biological replicates were analyzed by the software Bio-Rad CFX96 Managertm V1.6.541.1028, using Actin8 as the internal reference. DNA oligonucleotides used as probes or in real-time PCR are listed in Table S1.
RNA isolation and hybridization for miRNA detection were performed as described [59], [60]. 5′-end-labeled 32P antisense DNAs or an LNA oligonucleotide (for miRNA166) were used to detect miRNAs from total RNAs (10 µg each sample). Oligonucleotide probes used are listed in Table S1.
Genome sequences surrounding the selected MIRNA genes (listed in Table S1) were amplified by PCR from genomic DNA isolated from wild-type Arabidopsis (Col). The amplified DNA was first cloned into the pDNR221 vector (Invitrogen), confirmed by sequencing, and then recombined into the pEarlyGate100 Gateway-compatible destination vector [61] where the MIRNA genes are under the control of the CaMV 35S promoter. The constructs were introduced into essp5 homozygous or heterozygous plants (essp5/+). PCR primers used for amplifying the MIRNA genes are listed in Table S1. Transgenic plants were selected on Basta, allowed to grow to maturity and seeds were collected, and GUS expression was analyzed in the next generation.
For the construction of the PHB:PHB G202G-YFP transgene plasmid, the PHB promoter was PCR amplified from Arabidopsis (Col-0) genomic DNA by Fusion DNA Polymerase (NEB, M0530) using primers EcoRI-PHBpr and PHBpr-NcoI, and inserted into the pBluscript SK vector. The plasmid was then fully digested by SpeI and partially digested by EcoRI. The full-length promoter fragment was purified and ligated with the pEARLEYGATE 104 vector [61] to generate the plasmid pEG104-PHBpro. The PHB G202G coding sequence was amplified from cDNAs made from 35S:PHB G202G transgenic plants [43] with primers PHBf and PHBr, cloned into the pENTR-D-topo vector (invitrogen), and subsequently cloned into the destination vector pEARLEYGATE104 by LR reaction. The generated plasmid pEG104-PHB G202G were digested by NcoI and SpeI, and the PHB G202G-YFP fragment was recovered and ligated with pEG104-PHBpro to obtain the pEG104-PHB:PHB G202G-YFP plasmid. The PHB:PHB-YFP transgene plasmid was constructed using a similar strategy. Primers are listed in Table S1.
Chromatin immunoprecipitation (ChIP) was carried out as described previously [62]. One gram of twenty-day-old Arabidopsis plants grown on MS agar was used for each ChIP. The sonicated chromatin was immunoprecipitated with 5 µL of anti-GFP antibody (ab290, Abcam). Quantitative ChIP PCR was performed with three technical replicates, and results were presented as percentage of input. ChIP experiments were performed at least two times. See Table S1 for primer sequences used for ChIP-PCR and construction of the PHB:PHB G202G-YFP transgene.
|
10.1371/journal.pcbi.1005470 | Quantifying the roles of host movement and vector dispersal in the transmission of vector-borne diseases of livestock | The role of host movement in the spread of vector-borne diseases of livestock has been little studied. Here we develop a mathematical framework that allows us to disentangle and quantify the roles of vector dispersal and livestock movement in transmission between farms. We apply this framework to outbreaks of bluetongue virus (BTV) and Schmallenberg virus (SBV) in Great Britain, both of which are spread by Culicoides biting midges and have recently emerged in northern Europe. For BTV we estimate parameters by fitting the model to outbreak data using approximate Bayesian computation, while for SBV we use previously derived estimates. We find that around 90% of transmission of BTV between farms is a result of vector dispersal, while for SBV this proportion is 98%. This difference is a consequence of higher vector competence and shorter duration of viraemia for SBV compared with BTV. For both viruses we estimate that the mean number of secondary infections per infected farm is greater than one for vector dispersal, but below one for livestock movements. Although livestock movements account for a small proportion of transmission and cannot sustain an outbreak on their own, they play an important role in establishing new foci of infection. However, the impact of restricting livestock movements on the spread of both viruses depends critically on assumptions made about the distances over which vector dispersal occurs. If vector dispersal occurs primarily at a local scale (99% of transmission occurs <25 km), movement restrictions are predicted to be effective at reducing spread, but if dispersal occurs frequently over longer distances (99% of transmission occurs <50 km) they are not.
| Diseases which are transmitted by the bites of insects can be spread to new locations through the movement of both infected insects and infected hosts. The importance of these routes has implications for disease control, because we can often restrict host movement, and so potentially reduce spread, but cannot easily restrict insect movements. Despite this, the importance of host movements has been little studied. Here we develop a mathematical model which allows us to disentangle and quantify transmission by insect dispersal and by host movement. We apply the model to two diseases of cattle and sheep transmitted by biting midges that have emerged in northern Europe in the past decade, bluetongue virus (BTV) and Schmallenberg virus (SBV). For both viruses, we show insect movements account for a majority of spread between farms. Although they cannot sustain an epidemic on their own, animal movements play an important role in introducing disease to new areas.
| The role of host movements in the transmission of vector-borne diseases has largely been ignored [1,2]. Recently, however, several studies have quantified the importance of human movements for the spread of vector-borne diseases. Mobile phone data were used to infer mobility patterns and, hence, the impact of large-scale movement patterns on the transmission of malaria [3] and dengue [4]. Alternatively, detailed social surveys were used to investigate the importance of house-to-house movements on dengue virus transmission [5,6].
Both approaches exemplify a central difficulty in studying the role of human movement in the transmission of vector-borne diseases: they must rely on detailed small-scale studies or proxy measures, because human movements are seldom recorded in detail [2]. By contrast, livestock movements, and those of cattle in particular, are well described in many countries. The role of livestock movements in the spread of infectious diseases has been widely studied for directly-transmitted infections, such as bovine tuberculosis [7–9] or foot-and-mouth disease [10–12]. However, their role in the spread of vector-borne diseases has not been explored in any great detail.
In this study we disentangle and quantify the relative importance of livestock movements and vector dispersal in the spread of two viral infections of cattle and sheep, both of which are transmitted by Culicoides biting midges and have recently emerged in northern Europe. In 2006 bluetongue virus (BTV) serotype 8 (BTV-8) appeared near Maastricht in the Netherlands and, subsequently, spread to much of northern Europe [13,14]. Schmallenberg virus (SBV), a novel orthobunyavirus, was first detected in Germany and the Netherlands in the summer of 2011 [15] and by spring of 2013 had been reported across much of Europe [16].
We develop a model describing the transmission of BTV and SBV within and between farms, which uses separate descriptions for transmission via dispersal of infected vectors and via movement of infected livestock. First, we apply the model to demographic and epidemiological data from the BTV-8 epidemic in Great Britain (GB) during 2007 in a Bayesian framework. This approach allows us to link (unobserved) infection with reported clinical disease. It also allows us to update previous estimates for epidemiological parameters related to the transmission of BTV within and between farms (cf. [17,18]). Next, the model is applied to SBV using previously derived estimates relating to transmission within a farm [19]. Finally, the practical implications of the results are explored by assessing the impact of controlling livestock movements on the spread of the two viruses.
The dynamics of BTV within a farm are described using a stochastic compartmental model that includes two ruminant host species (cattle and sheep) and the Culicoides vector [18].
The cattle and sheep populations are assumed to be constant (Hi), except for disease-associated mortality, and are subdivided into the number of susceptible (i.e. uninfected), infected and recovered animals, denoted by X(i), Y(i) and Z(i), respectively, where the superscript i indicates cattle (C) or sheep (S). To allow for a more general gamma distribution for the duration of viraemia, the infected host population, Y(i), is subdivided into a number of stages, with newly infected hosts entering the first stage and then passing through each successive stage. If the time spent in each stage follows an exponential distribution with mean 1/niri, the total length of time spent in the ni stages follows a gamma distribution, with mean 1/ri and variance 1/niri2 [22].
The vector population (N) is subdivided into the number of adult female midges that are susceptible (i.e. uninfected), latent (i.e. infected, but not infectious) and infectious, denoted by S, L and I, respectively. To allow for a more general gamma distribution for the extrinsic incubation (i.e. latent) period (EIP) [23], the latent class is subdivided into a number of stages in a similar approach to that described above for the duration of host viraemia. Vector mortality occurs at the same rate in all classes and is balanced by the recruitment of susceptible vectors, so that the total vector population (N) remains constant.
The force of infection for host species i, λi, is given by,
λi(t)=baϕimiθ(t)I(t)N,
(1)
where b is the probability of transmission from an infected vector to a host, a is the reciprocal of the time interval between blood meals for the vector (assumed to be equal to the biting rate), mi (= N/Hi) is the vector-to-host ratio and I/N is the proportion of bites which are from infectious vectors. The proportion of bites on cattle and sheep is given by
ϕC=HCHC+σHS, ϕS=1−ϕC,
(2)
respectively, where σ is the vector preference for sheep relative to cattle. The seasonal vector activity [24] on day t is given by
θ(t)∝exp(b11sin(2πt365)+b21cos(2πt365)+b12sin(4πt365)+b22cos(4πt365)),
(3)
normalised so the maximum value is one. The force of infection for vectors, λV, is
λV(t)=βaθ(t)(ϕCY(C)(t)HC+ϕSY(S)(t)HS),
(4)
where β is the probability of transmission from an infected host to a vector and Y(C) and Y(S) are the total number of infected cattle and sheep, respectively.
Infection on a farm was related to reported clinical disease by assuming there was a daily probability of a farm with infected cattle or sheep reporting clinical disease, ζC and ζS, respectively, where 0≤ζC,ζS≤1.
Parameters in the model are summarised in S1 Table. The reciprocal of the time interval between blood meals (a), the vector mortality rate (μ) and the reciprocal of the mean EIP (ν) were assumed to vary with the local temperature (see S1 Table for details).
Population sizes in the model take integer values, while transitions between compartments are stochastic processes (S2 Table). The number of transitions of each type during a small time interval δt was drawn from a binomial distribution with population size n and transition probability q (the appropriate per capita rate multiplied by δt) (S2 Table). However, binomial random variables are computationally expensive to simulate and an approximating distribution was used wherever possible. If: (i) nq(1-q)>25; (ii) nq(1-q)>5 and 0.1<q<0.9; or (iii) min(nq,n(1-q))>10, an approximating normal variate with mean nq and variance nq(1-q) was used, while if q<0.1 and nq<10, an approximating Poisson variate with mean nq was used [25].
To describe the spread of BTV between farms, a stochastic, spatially-explicit model with a daily time-step was used. Transmission between farms was assumed to occur via two routes: movement of infected animals or dispersal of infected vectors.
A total of 22 or 23 parameters were estimated by fitting the BTV model to the summary outbreak data for Great Britain in 2007: two or three related to vector dispersal (depending on the model used; Table 1); two related to under-ascertainment of infected farms (ζC and ζS); and 18 related to within-farm transmission (see S1 Table).
When applying the model framework to SBV, estimates for parameters related to the transmission of SBV within a farm were derived from an earlier study in which the within-farm component of the model was fitted to sero-prevalence data for Belgium and the Netherlands [19]. Parameters for transmission between farms were assumed to be the same as for BTV.
Unlike for BTV, the timing of incursions for SBV in GB has not been investigated in any great detail. Back-calculation from the dates of reported cases of malformed calves and lambs indicates that a plausible date for an incursion is 28 June 2011 [38], though it is not possible to rule out earlier dates or, indeed, multiple dates of incursion. For simplicity, each incursion was initialised on the 28 June by selecting a single farm at random from a county on the south-east coast of England (Suffolk, Essex, Kent, East Sussex, West Sussex, Hampshire and Isle of Wight), which were the earliest affected regions. The aim here is to compare the importance of transmission routes for SBV with those for BTV, rather than to reconstruct the SBV epidemic in GB. However, the sensitivity of the results for SBV to the timing of incursion was assessed by simulating incursions on five other incursion dates throughout the year (1 May, 1 June, 1 July, 1 August and 1 September).
To explore the impact of movement restrictions on the spread of BTV and SBV we assumed they were applied in a circular zone around known IPs. For BTV IPs were detected on the basis of reported clinical disease, while for SBV they were assumed to be detected when the first newly infected cattle or sheep occurred on the farm (adult animals show no or very mild clinical signs of disease). Farms within a specified radius of an IP became part of the movement restriction zone (MRZ) and were allowed to move animals to farms within the MRZ, but not to any farms outside the MRZ. For each radius, one hundred replicates of the model were run for five incursion dates (1 May, 1 June, 1 July, 1 August and 1 September) until 31 December. Each incursion was initialised by selecting a single farm at random from a county on the south-east coast of England (Suffolk, Essex, Kent, East Sussex or West Sussex). The model for transmission between farms via animal movements was parameterised using movement data for 2006 (i.e. a year in which there were no major outbreaks of disease in cattle or sheep). We explored the sensitivity of the model predictions for the effectiveness of movement controls to temperature by using data for two years: 2006 (a warmer year) and 2007 (a cooler year).
For the model in which vector dispersal is described as a diffusion process, the predicted number of newly infected holdings each week reaches its peak about seven weeks after the initial incursion (Fig 1A), preceding the peak in newly confirmed clinical farms by one or two weeks (Fig 1B). Moreover, there are substantially more infected farms than confirmed clinical farms, indicating a high level of under-ascertainment. The observed number of newly confirmed clinical farms each week lies close to the posterior mean for most weeks (Fig 1B), while the observed cumulative number of confirmed clinical farms in each county lies in the 95% prediction interval for all counties, except Essex (Fig 1C). In addition, the model predicts spread of BTV into areas where reported clinical farms were not confirmed in only a small proportion (2.7%) of simulations and, in each instance, only a single case (Fig 1C and 1F). The observed number of infected farms detected by pre-movement testing or by targeted surveillance lies within the 95% prediction intervals for the model (Fig 1D), while the posterior mean for the within-farm prevalence for cattle herds matches that observed (Fig 1E).
The predicted dynamics for the four kernel models are similar to those for the diffusion model (S1–S4 Figs; cf. Fig 1). However, all four kernel models predict more extensive spatial spread of BTV than the diffusion model. This has the consequence that they are less able to capture the number of infected holdings detected through targeted surveillance around the first two IPs, but are better able to account for the number of confirmed clinical farms in Essex. The kernel models also more frequently (>10% of simulations for each model) predict clinical cases in areas in which no cases were reported. In addition, the predictions of the kernel models are more variable than those for the diffusion model.
The marginal posterior distributions for the parameters in each model are plotted in S5 and S6 Figs and are summarised in S5 and S6 Tables. Summary statistics for the within-farm parameters are provided only for the model in which vector dispersal was described as a diffusion process (S6 Table). However, the posterior distributions did not differ greatly for most parameters (S6 Fig), except for the probability of transmission from host to vector (β), which was higher for the kernel models compared with the diffusion model (posterior mean: 0.05 vs 0.02).
All the models predict that transmission of BTV between farms occurs predominantly through dispersal of infected vectors. In simulated outbreaks, the median proportion of farms infected via dispersal of infected vectors was 86–91%, depending on the model for vector dispersal (Fig 2A). This compares with the median proportion of farms infected via movement of cattle and sheep of 5–7% and 3–6%, respectively (Fig 2A). A similar pattern is seen in the number of secondary infections via the three routes, in which most secondary infections are generated by vector dispersal (Fig 2C). In addition, the number of secondary infections per farm by vector dispersal is over-dispersed, with the majority of transmission attributable to a small number of farms (Fig 2C). The mean number of secondary infections for this route was around 1.3 for the diffusion model and around 0.9 for the kernel models, while the dispersion parameter was around 0.05 for all models (Table 2). By contrast, the mean number of secondary infections per farm arising by the movement of infected cattle or sheep was around 0.05 (Table 2).
The distance over which transmission occurred was strongly dependent on the transmission route. Transmission via movement of infected livestock occurred over considerably longer distances than via dispersal of infected vectors and this was independent of the model used for vector dispersal (Fig 2E). When transmission was via movement of infected livestock, the mean distance between source and recipient farms was around 50 km, with 99% of transmission occurring within around 150 km for both cattle and sheep. When transmission was via dispersal of infected vectors, the distances between source and recipient farms depended critically on the model for vector dispersal (Fig 2E). The median distance (distance within which 99% of transmission occurred) was 7.8 (25.2) km for the diffusion, 30.9 (49.6) km for the exponential kernel, 29.1 (49.5) km for the Gaussian kernel, 22.5 (49.3) km for the fat-tailed kernel and 19.9 (48.6) km for the stepped kernel.
The characteristics of each transmission route (frequent, but shorter range for vector dispersal; less frequent, but longer range for cattle and sheep movements) and the differences between models in vector dispersal distances are demonstrated in maps showing which farm infected which in the simulated outbreaks (Fig 3).
The sensitivity of the importance of the transmission routes to the time of incursion and temperature data was assessed for each model of vector dispersal (S7 and S8 Figs). The proportion of farms infected via each route was not substantially influenced by either the time of incursion or the temperature data used. Both the number of secondary infections per infected farm and the distance over which BTV spread via livestock movements were higher for incursions earlier in the year. Furthermore, the number of secondary infections per infected farm was higher when using the 2006 temperature data (a warmer year) compared with 2007 data (a cooler year).
The model predicts much larger outbreaks for SBV compared with BTV, in terms of both the number of infected farms and spatial spread (S9 Fig; cf. Fig 1). Furthermore, the proportion of transmission between farms via dispersal of infected vectors is higher for SBV than for BTV (Fig 2B). In simulated outbreaks the median proportion of farms infected via vector dispersal is 98% and this is independent of the model of vector dispersal. This compares with 1% each for transmission via movement of infected cattle and sheep. This difference in the importance of the transmission routes was reflected in the number of secondary infections per infected farm, which was higher for vector dispersal for SBV than for BTV, but which was lower for cattle and sheep movements (Fig 2D; cf. Fig 2C). For SBV, the mean number of secondary infections per infected farm was around 2.0 for the diffusion model and around 1.5 for the kernel models, while the dispersion parameter was around 0.07 for all models (Table 2). The mean number of secondary infections via livestock movements was 0.01 for both cattle and sheep (Table 2). The distance over which SBV spread via livestock movements was greater than for BTV (Fig 2F), but this is a consequence of the movement restrictions in place during the BTV outbreak. Finally, the importance of the transmission routes for SBV was not greatly sensitive to the time of incursion (S10 Fig).
Movement restrictions (in this case applied in a circular zone around known IPs) can potentially reduce the size of a BTV outbreak, but whether or not they are predicted to do so depends critically on assumptions about vector dispersal (Fig 4). When vector dispersal is described by a diffusion process, movement restrictions reduce the size of an outbreak and, furthermore, there is an optimal radius for the MRZ of approximately 20 km (Fig 4A). When vector dispersal is described by a fat-tailed kernel, there is also some evidence for an impact of movement restrictions on outbreak size, but in this case the optimal MRZ radius is around 35–40 km (Fig 4D). However, when vector dispersal is described by an exponential, Gaussian or stepped kernel, there is no evidence for an impact of movement restrictions (Fig 4B, 4C and 4E). In addition, movement restrictions do not substantially reduce outbreak size if the incursion occurs later in the year (August or September) and this conclusion is independent of assumptions about vector dispersal (Fig 4). Similar results are also obtained if temperature data for 2006 are used (S11 Fig) instead of for 2007 (Fig 4).
Whether or not movement restrictions are predicted to reduce the size of an SBV outbreak also depends critically on assumptions about vector dispersal (S12 Fig). Using the model describing vector dispersal as a diffusion process, movement restrictions were predicted to have a substantial impact on outbreak size and to a much greater extent than for BTV (S12 Fig; cf. Fig 4). By contrast, movement restrictions were predicted to have no impact on outbreak size for any of the kernel models (S12 Fig).
Initial modelling studies for BTV-8 in northern Europe used kernel- or wave-based approaches to explore spread, implicitly incorporating all modes of transmission in a single description [18,35,39]. Subsequently, models were developed which separate animal and vector movements [40–42], but these were not fitted to outbreak data nor did they quantify the relative importance of the two transmission routes.
Here we have developed a model framework that allows us to disentangle and quantify the roles played by livestock movements and vector dispersal in the transmission of two Culicoides-borne viruses. Our results show that dispersal of infected vectors accounts for the majority (around 90%) of spread of BTV between farms and an even higher proportion (98%) of spread of SBV between farms (Fig 2A and 2B). We are able to attribute spread to each route because of the detailed, independent data available to describe spread via movement of infected livestock. If this were not the case, it would be more challenging to estimate the relative contribution of each route, because a decrease in transmission due to one route could be compensated for by an increase in transmission due to another.
One previous study has quantified the role of vector dispersal in the spread of BTV-8 in northwest Europe in 2006 [31]. The authors could explain infection onset for 94% of reported BTV-infected farms based on wind and midge flight activity. As they did not consider livestock movements, this represents an upper bound on the proportion of farms infected via vector dispersal, but is consistent with our estimate for the BTV-8 outbreak in GB in 2007 (Fig 2A). Similar methods were subsequently applied to quantify the role of vector dispersal in the spread of SBV in northwest Europe in 2011 [32]. In contrast with BTV, the authors could explain infection onset for only 70% of reported SBV-infected farms based on midge flight activity, which is markedly lower than our estimate of 98% (Fig 2B). This discrepancy is likely to be a consequence of under-ascertainment of SBV-infected farms. This results in a greater distance between infected farms, making it more difficult for the vector-only approach to explain SBV transmission [32].
The number of secondary infections arising through each route also emphasises the major role played by vector dispersal in the transmission of BTV and SBV between farms compared with livestock movements (Fig 2C and 2D). In particular, the mean number of secondary infections via both cattle and sheep movements was estimated to be substantially below one for both viruses (Table 2), indicating that these routes alone cannot sustain transmission (cf. [41]). By contrast, the mean number of secondary infections via vector dispersal was above one for both BTV and SBV, indicating transmission can be sustained by this route. However, the number of secondary infections is over-dispersed, so that a small proportion of farms account for the majority of transmission: an example of the 80/20 rule [43].
In the model, it is the larger farms which account for most of the transmission via dispersal of infected vectors. This is primarily a consequence of our assumption that the number of vectors is proportional to the number of livestock on a farm (though the constant of proportionality does vary amongst farms). Few studies have investigated the relationship between vector abundance and host numbers. One recent study found that Culicoides abundance was higher at trap locations with a high density of cattle in the locality [24]. Another study suggested that catches in light traps increase linearly with sheep numbers, at least for small host numbers [44]. Although these results do not allow robust generalizations, the findings are compatible with the assumption of a constant vector-to-host ratio. In addition, the common alternative assumption is that the number of vectors is independent of host numbers, but this results in the conclusion that outbreaks are more likely on smaller farms because they will have higher vector-to-host ratios.
Both the proportion of transmission and the number of secondary infections via dispersal of infected vectors are independent of the model used for vector dispersal (Fig 2A–2D). This is not the case, however, for the distances between source and recipient farms, which differ markedly amongst the models (Fig 2E and 2F). The distances inferred in the present study of BTV-8 in GB using a diffusion model (median: 7.8 km; Fig 2E) are similar to those estimated for BTV-8 in northern Europe based on wind and midge flight activity (median: ~5 km; see [31], their Fig 3). These contrast with the considerably larger distances inferred using the kernel models (median for all models: ≥20 km; Fig 2E). Comparing the different models suggests that, while the diffusion model is able to capture the general pattern of local-scale spread, it is not able to capture the relatively infrequent longer-range dispersal events (see, e.g. [30]). By contrast, the kernel models can predict longer-range jumps, but at the expense of missing the detail of local-scale spread. However, given the under-ascertainment of infected holdings and the spatial resolution of the data, it will be difficult to infer a more robust model for vector dispersal from the 2007 GB outbreak. Moreover, the challenges associated with studying Culicoides biting midges in the field [45–47] make it difficult to estimate dispersal patterns empirically, especially over longer distances.
The lower proportion of transmission attributed to movement of infected livestock for SBV compared with BTV (2% vs 10%; Fig 2A and 2B) can be accounted for by two key differences between the viruses. First, vector competence (i.e. the probability of transmission from host to vector) is much higher for SBV (0.14; [19]) than for BTV (0.02; S6 Table). This means there is a higher prevalence of infectious vectors for SBV compared with BTV, increasing the importance of vector dispersal for SBV (see Eqs (5) and (6)). Second, the mean duration of viraemia is much longer for BTV (21 days; S6 Table) than for SBV (3–4 days; [19]). Consequently, there is a lower probability of moving an animal while it is infected for SBV than for BTV, reducing the importance of livestock movements for SBV.
Host movements may only account for a small proportion of transmission, but our results reinforce the important role that they play in the transmission of vector-borne diseases, particularly through the introduction of infection to new areas [3,4,48,49]. In the case of BTV and SBV, host (i.e. livestock) movements spread the virus over much longer distances than would be expected by vector dispersal (Fig 2E and 2F) and facilitate the establishment of new outbreaks away from existing foci (Fig 3).
While seldom practical for human movements, it is feasible to restrict livestock movements as part of disease control measures. The predicted impact of movement restrictions for both BTV and SBV depends critically on the model used for vector dispersal. In particular, there is a significant reduction in outbreak size (i.e. cumulative number of farms infected) only for the diffusion model (Fig 4; S11 and S12 Figs). Furthermore, the magnitude of the reduction resulting from movement restrictions is predicted to be much lower than alternative control measures, in particular, vaccination [50,51].
The difference in predicted effectiveness of movement restrictions amongst models reflects the distances over which dispersal of infected vectors occurs in each one (Fig 2E and 2F). When dispersal is primarily local, as is the case with the diffusion model, movement restrictions are effective because the virus is unlikely to escape the MRZ through vector dispersal into an area in which movements are allowed and, hence, can spread over longer distances. When vector dispersal occurs more frequently over longer distances, as is the case with the exponential and Gaussian kernels, infection is likely to escape any MRZ through vector dispersal alone. As a result, movement restrictions are predicted to be ineffective in this case.
When fitting the BTV model to outbreak data we have used summary or aggregated epidemiological measures, rather than more detailed data on location and timing of infected farms. For many outbreaks, including the 2007 BTV outbreak in GB, summary statistics are the only data available, which makes ABC a natural framework in which to implement epidemiological models [37,52]. Moreover, ABC methods facilitate integrating data from different surveillance sources (in the case of BTV, reported clinical farms, pre-movement testing and targeted surveillance), which helps overcome the limitations associated with individual sources (e.g. under-ascertainment of reported cases). This does, however, require a model relating disease occurrence and reporting to the underlying pattern of infection, which is not always straight-forward. Here, we have used a simple model (a fixed daily probability of reporting), which captures this acceptably for most areas in each of the models. Where there are discrepancies (e.g. for Essex in the diffusion model; Fig 1C), this could reflect differences in reporting behaviour between regions or changes in the probability of reporting over time. This is difficult to explore in detail, however, given the limited numbers of cases.
In this study we have demonstrated that both vector dispersal and host movements play important roles in transmission of vector-borne diseases of livestock, though for different reasons. Vector dispersal is the principal mode of spread between farms, while livestock movement is the principal means of introducing infection to new areas. However, the relative importance of the routes differs between viruses, even when they share the same vector species. This has practical implications for disease control and, in particular, movement restrictions, so that generic measures may not be effective.
|
10.1371/journal.pmed.1002673 | Patient-centered primary care and self-rated health in 6 Latin American and Caribbean countries: Analysis of a public opinion cross-sectional survey | Despite the substantial attention to primary care (PC), few studies have addressed the relationship between patients’ experience with PC and their health status in low-and middle-income countries. This study aimed to (1) test the association between overall patient-centered PC experience (OPCE) and self-rated health (SRH) and (2) identify specific features of patient-centered PC associated with better SRH (i.e., excellent or very good SRH) in 6 Latin American and Caribbean countries.
We conducted a secondary analysis of a 2013 public opinion cross-sectional survey on perceptions and experiences with healthcare systems in Brazil, Colombia, El Salvador, Jamaica, Mexico, and Panama; the data were nationally representative for urban populations. We analyzed 9 features of patient-centered PC. We calculated OPCE score as the arithmetic mean of the PC features. OPCE score ranged from 0 to 1, where 0 meant that the participant did not have any of the 9 patient-centered PC experiences, while 1 meant that he/she reported having all these experiences. After testing for interaction on the additive scale, we analyzed countries pooled for aim 1, with an interaction term for Mexico, and each country separately for aim 2. We used multiple Poisson regression models double-weighted by survey and inverse probability weights to deal with the survey design and missing data. The study included 6,100 participants. The percentage of participants with excellent or very good SRH ranged from 29.5% in Mexico to 52.4% in Jamaica. OPCE was associated with reporting excellent or very good SRH in all countries: adjusting for socio-demographic and health covariates, patients with an OPCE score of 1 in Brazil, Colombia, El Salvador, Jamaica, and Panama were more likely to report excellent or very good SRH than those with a score of 0 (adjusted prevalence ratio [aPR] 1.61, 95% CI 1.37–1.90, p < 0.001); in Mexico, this association was even stronger (aPR 4.27, 95% CI 2.34–7.81, p < 0.001). The specific features of patient-centered PC associated with better SRH differed by country. The perception that PC providers solve most health problems was associated with excellent or very good SRH in Colombia (aPR 1.38, 95% CI 1.01–1.91, p = 0.046) and Jamaica (aPR 1.21, 95% CI 1.02–1.43, p = 0.030). Having a provider who knows relevant medical history was positively associated with better SRH in Mexico (aPR 1.47, 95% CI 1.03–2.12, p = 0.036) but was negatively associated with better SRH in Brazil (aPR 0.71, 95% CI 0.56–0.89, p = 0.003). Finally, easy contact with PC facility (Mexico: aPR 1.35, 95% CI 1.04–1.74, p = 0.023), coordination of care (Mexico: aPR 1.53, 95% CI 1.19–1.98, p = 0.001), and opportunity to ask questions (Brazil: aPR 1.42, 95% CI 1.11–1.83, p = 0.006) were each associated with better SRH. The main study limitation consists in the analysis being of cross-sectional data, which does not allow making causal inferences or identifying the direction of the association between the variables.
Overall, a higher OPCE score was associated with better SRH in these 6 Latin American and Caribbean countries; associations between specific characteristics of patient-centered PC and SRH differed by country. The findings underscore the importance of high-quality, patient-centered PC as a path to improved population health.
| In the 40 years since the Declaration of Alma-Ata, empirical studies using ecological data have shown positive effects of access to primary care on population health outcomes, such as child mortality, adult overall mortality, and adult avoidable hospitalizations, both in high-income countries and in low- and middle-income countries (LMICs).
In the context of LMICs, the centrality of primary care has been questioned by findings of its poor quality and its limitations in adapting to urbanization and to the epidemiological transition, as well as the increasing population demand for responsive, high-quality services.
To our knowledge, no individual-level studies have examined the relationship between the attributes of patient-centered primary care and self-rated health (SRH) in the context of LMICs.
Using person-level survey data representative of the urban population of 6 countries in Latin America and the Caribbean, the current study shows that individuals who reported receiving more patient-centered primary care overall were more likely to report excellent or very good SRH status in all countries. Features of primary care associated with better SRH differed between countries.
The expansion of primary care in LMICs can be informed by better evidence on which of its features—including ease of communication, comprehensiveness of care, and support for coordination of care—are associated with better self-reported health status.
While individual-level administrative data are not yet widely available for health systems in LMICs, patient-reported survey data may serve as an instrument to assess healthcare services and to inform policy-makers in their efforts to increase the quality of primary care services.
| Primary care (PC) has been described as being uniquely positioned to promote health and well-being at the population level [1–3]. Its central role in providing adequate, efficient, and equitable access to preventive and curative healthcare was strongly emphasized by the Declaration of Alma-Ata 40 years ago [2]. However, PC’s centrality has been questioned by findings of its poor quality, especially in low- and middle-income countries (LMICs), and its limitations in adapting to urbanization and the epidemiological transition, as well as increasing population demand for responsive, high-quality services [4–6].
More recently, patient-centered healthcare has emerged as a person-oriented model of care aimed at meeting population needs, expectations, and preferences. Studies from the United States and the United Kingdom have shown the positive effect of patient-centered healthcare in improving the quality of the processes of care, reducing hospitalizations and emergency visits (and consequently healthcare costs), and improving users’ satisfaction and self-management [7–9]. Within PC, a number of patient-centered healthcare attributes have been shown to be associated with perception of good healthcare quality, such as the availability of a PC provider who “knows relevant information about a patient’s medical history,” “solves most of the health problems,” “spends enough time with the patient,” “coordinates healthcare,” and “is easy to communicate with” [10].
Previous work using ecological data has shown positive effects of PC on population health outcomes, such as child mortality, and avoidable hospitalizations, both in high-income countries and in LMICs [11,12]. However, despite the substantial attention and policy emphasis on PC, few studies have addressed the relationship between patient experience with PC and health in LMICs; none to our knowledge have done so with a cross-country perspective [13,14].
Self-rated health (SRH) is a broadly used measure: individuals evaluate their health status through a Likert scale or compare their health status with individuals of the same age [15]. Though SRH is a subjective indicator of health status, it has been found to be a robust predictor of mortality [15,16]; also, low SRH is associated with increased hospitalization and outpatient care in elderly populations [17]. Research studies addressing the relationship between health service characteristics and SRH have reported that in the US, individuals living in states with a higher ratio of PC physicians to population were more likely to report good SRH than those with a lower ratio of PC physicians [18]. Enhanced accessibility and continuity of PC in the US [19,20] and high total PC quality scores in South Korea were associated with better SRH of health service users [21].
The objectives of the present study were (1) to test the association between overall patient-centered PC experience (OPCE) and SRH in 6 Latin American and Caribbean (LAC) countries and (2) to identify specific features of patient-centered PC associated with better SRH. This work can help inform financing and policies at a moment of renewed global attention to PC.
We performed a secondary analysis of a recent (2013) public opinion survey focusing on perceptions and experiences with healthcare systems in 6 LAC countries: Brazil, Colombia, El Salvador, Jamaica, Mexico, and Panama. The detailed methodology of this survey was previously reported elsewhere [10,22,23]. In each country, the survey included a nationally representative urban sample of the population that comprised between 1,500 and 1,506 adults per country. According to the 2017 revision of World Population Prospects [24], the urban population constitutes the majority in these countries, ranging from 54% in Jamaica to 85% in Brazil. In total, 330 million individuals reside in urban areas in these countries.
During 2012 and 2013, Harris Interactive collected the data through telephone interviews. The sample frame for the survey consisted of random digit dialing listings of landline and mobile phone numbers in each country. The survey used an adapted version of the methodology and questionnaire that the Commonwealth Fund has been applying in Europe, Australia, Canada, and the US over the past 15 years [25]. The selection criteria considered any household member aged 18+ years. Only 1 adult per household was interviewed.
We did not have a formal prospective analysis plan. Prior to seeing the data, we identified the public opinion survey as a unique resource to test associations of healthcare quality and SRH in a representative, multi-country sample. We then reviewed the literature on patient-centered PC and identified 5 key domains (contact with clinic, time spent with provider, patient–provider communication, technical quality and solving problems, and healthcare coordination) relevant for patient-centered healthcare [10,23,25,26]; we mapped items from the survey to these domains and created single-item summaries as well as an overall score. We defined covariates based on relevance to health status and healthcare utilization. We planned to assess all countries in a pooled sample; on identifying substantial variation in the level of SRH between countries, we tested for interaction between patient-centered PC variables and country on the additive scale and report stratified models where evidence of interaction was found.
The dependent variable was “excellent or very good SRH,” obtained from the general SRH report and categorized as 1 = “excellent” or “very good” and 0 = “good,” “fair,” “poor,” or “not sure”.
The survey specified that PC is care provided by the doctors or other health professionals (i.e., nurses, social workers) at the family doctor’s practice or clinic. We selected items related to PC that fall into the domains of patient-centered healthcare identified in the literature [10,23,25,26] and organized them by domain:
The variable “preventive exams up to date” was defined as “yes” when the respondent reported having blood pressure measurement in the last year, serum cholesterol in the last 5 years, and, for women over 40 years, cervical cytology (Pap test) and mammography in the last 3 years.
All other PC variables were measured on a 5-options Likert scale and categorized as 1 = yes (“always” or “often”) and 0 = no (“sometimes,” “rarely or never,” and “not sure”). The decision to categorize the variables this way was based on previous studies [10,23,25,26]. We calculated OPCE as the arithmetic mean of these items, following the recommendation of previous research on the use of patient experience surveys to assess service provision [27]. OPCE score ranged from 0 to 1, where 0 meant that participant did not have any of the 9 patient-centered PC experiences, while 1 meant that he/she reported having all these experiences. We assumed that each component of OPCE score contributed equally to patients’ experiences and that a difference in patient experiences had a constant effect on SRH. We maintained the 9 binary items as individual components of patient-centered PC.
Several socio-demographic and health-service-related factors are associated with poor SRH. Individual factors linked to lower health status include unhealthy lifestyle [28–30] and chronic diseases that affect mental and physical health [31–34]. Although some aspects of the relationship between socio-demographic factors and SRH are still inconclusive, it has been reported that older age, low schooling, low socio-economic status, low social capital, and low health insurance (HI) coverage are associated with poor SRH [35–37].
Based on survey data availability, we included the following covariates: sex, age, education, chronic disease, and the type of HI. The variable education defines the level of education for participants who answered the survey in all countries except El Salvador, where it describes the education of the head of the household. We identified the participant as having a chronic disease if he/she reported that a doctor had told him/her of having arthritis, asthma or chronic lung disease, cancer, diabetes, heart disease, hypertension, or depression. The type of HI was categorized as: government HI (publicly subsidized insurance not related to job affiliation), social security HI (contributory insurance related to job affiliation), and private HI (voluntary private insurance; also, in Brazil and Jamaica, those who reported having private HI provided by workplace). Furthermore, respondents reporting both government and social-security-based HI (4.1% of participants in Colombia and 21.5% in Mexico) were grouped under social security HI. Participants who reported not having HI were placed in the government HI group, because, in all these countries, government HI is freely available for those without social security or private HI.
We used descriptive statistics to analyze the characteristics and PC experiences of the study participants. We performed a bivariable analysis including chi-squared tests between the dependent variable (SRH) and each independent variable (PC experience) or categorical covariate. We used Student t tests for comparison of the continuous variable OPCE score between people who reported excellent or very good SRH and those who reported good, fair, or poor (or not sure) SRH.
The survey asked the complete set of questions about PC experiences only to respondents who affirmed “having a regular doctor or regular place for primary care.” This skip pattern results in a high percentage of missing data, given that lack of access to a regular source of PC ranged from 16.3% in Jamaica to 43.1% in El Salvador; in addition, several PC variables also had missing information (S1 Table). In sum, in the 6 countries there were 6,100 participants with complete information, which represented 67.7% of the initial sample of 9,012. Thus, we applied a double-weighted strategy with the use of survey weights to account for the survey sample design and stabilized inverse probability (IP) weights to correct for potential selection bias [38]. This technique consists in assigning a weight to individuals with complete information so that they account for themselves as well as for those with similar characteristics who had missing information. It assumes that those with missing information are similar to those with complete information who share the same measured covariates [38]. In particular, to apply this technique to adjust for the missingness induced by not having a regular PC clinic or doctor, we assumed that the PC experience of individuals without a regular PC clinic or doctor can be represented by those with a regular PC clinic or doctor conditional on the specified covariates, i.e., that there are no unmeasured confounders that are a common cause of both having access to a regular PC clinic or doctor and SRH. We first compared the number of PC visits between those with and without a regular PC clinic or doctor. We found that the mean number of visits in the last 12 months in the group with a regular doctor was 3.07 and in the group without a regular doctor was 2.09 (p < 0.001). We then generated the denominator and numerator of the IP weights. The denominator for stabilized IP weights was the probability of having missing data conditional on the following covariates: sex, age, education, type of HI, and presence of chronic disease. The numerator was the probability of having missing data regardless of the covariates.
We used Poisson regression models with robust variance as recommended for cross-sectional studies with high-prevalence binary outcomes [39]. We initially fit pooled models across all 6 countries, and then calculated the relative excess risk due to interaction (RERI) between each country and PC experience as a measure of interaction on the additive scale [40]; additive interaction is more indicative of underlying causal interaction than interaction on the relative (ratio) scale. Where evidence of interaction was identified (RERI significant at p ≤ 0.05), we included interaction terms for country in the pooled multiple regression model, or stratified the model by country in the case of multiple interactions identified. Each multiple Poisson regression model included the dependent variable, independent variables, and conceptually relevant covariates. The pooled model included fixed effects for countries to control for country-level heterogeneity and to focus on the effect of the individual-level predictors [41,42].
Finally, we performed a sensitivity analysis in which the IP weights were calculated after the individuals without a regular PC clinic or doctor were dropped. The results were similar to those of the main analysis, suggesting that our findings were not distorted by including everyone when calculating the IP weights. All analyses were performed using the software Stata 14 and considering estimates with p ≤ 0.05 to be statistically significant.
The study consists of a secondary data analysis of a public opinion survey focusing on perceptions and experiences with healthcare systems in 6 LAC countries. The survey was commissioned by the Inter-American Development Bank, and the contracted surveying firm was responsible for obtaining all necessary regulatory approvals and verifying compliance with the ethical standards of the ICC/ESOMAR Code on Market, Opinion and Social Research. The survey data for this secondary data analysis were made available by, and their use approved by, the Inter-American Development Bank.
Tables 1 and 2 present the characteristics of study participants from 6 LAC countries (n = 6,100) double-weighted by survey and stabilized IP weights. Slightly more women than men participated in all countries (52.2% versus 47.8% in the full sample). Participants reported lower education levels in Brazil, Panama, and El Salvador (62.7%, 37.5%, and 33.4% with elementary school or less, respectively), while approximately half the sample had completed secondary school in Mexico, Colombia, and Jamaica. Government HI predominated in Brazil and Jamaica (76.5% and 61.5%, respectively), while social security HI was more common in Colombia, Panama, Mexico, and El Salvador (65.1%, 62.4%, 48.9%, and 47.4%, respectively). The proportion with private HI ranged from 9.8% in Colombia and El Salvador to 38.5% in Jamaica. Report prevalence of chronic conditions ranged from 31.5% in El Salvador to 52.2% in Jamaica. Finally, the percentage of participants with excellent or very good SRH was highest in Jamaica (52.4%), declining to a low of 29.5% in Mexico.
Tables 3 and 4 show the participants’ experience with PC services in the full sample and by country. The proportion of participants who reported that the PC facility was easy to contact by telephone during regular office hours ranged from 38.1% in El Salvador to 75.2% in Jamaica. Patients from Brazil reported less frequently that the PC provider spent enough time with them (31.8%), while in Colombia this figure reached 74.2%. Regarding patient–provider communication, the opportunity to ask questions and having the PC provider explain things in a way that was easy to understand were less frequent in Brazil (58.0% and 63.9%, respectively) and more frequent in Mexico (79.5% had the opportunity to ask questions) and Colombia (81.3% received explanations in a way that was easy to understand). Relating to the technical quality of care, only 40.9% in Brazil reported that the PC provider knew relevant information about their medical history, while this figure was 75.4% in Mexico. Only between 25.9% (in Jamaica) and 44.2% (in Panama) reported that the PC provider talked about healthy lifestyles, while between 25.8% and 26% (in Panama and El Salvador) and 40.7% and 40.2% (in Brazil and Mexico) had their preventive exams up to date. The percentage of participants who considered that the PC provider solved most of their health problems ranged from 54.2% in Brazil to 80.6% in Mexico, while only from 21.8% (in Brazil) to 45.4% (in Mexico) stated that the PC provider helped to coordinate healthcare. The average OPCE score ranged from 0.44 points in Brazil to 0.63 points in Mexico.
In bivariable analyses, the average OPCE score was significantly higher in participants with excellent or very good SRH in 4 out of 6 countries. For specific features of patient-centered PC, the proportion of respondents with excellent or very good SRH was significantly higher among those who had a PC facility that was easy to contact in Colombia, El Salvador, and Mexico; who reported that the PC provider spent enough time with them in Colombia and Mexico; who had the opportunity to ask questions in Brazil, Colombia, and Mexico; who had a PC provider who explained things in a way that was easy to understand in Brazil, Colombia, Jamaica, and Mexico; who perceived that the PC provider knew relevant information about their medical history in Colombia, Jamaica, and Mexico; who considered that PC provider solved most of their health problems in Brazil, Colombia, Jamaica, and Mexico; and who reported that PC provider coordinated care with other providers or sources of care in Colombia, El Salvador, Jamaica, and Mexico.
Table 5 shows the results of the pooled multiple Poisson regression model double-weighted by survey and stabilized IP weights to test the association of OPCE score with excellent or very good SRH. The coefficients represent prevalence ratios of the report of excellent or very good SRH; their interpretation is the same as for risk ratios. Assessment of interaction between countries and OPCE score identified a significant positive interaction in Mexico (RERI 0.55, 95% CI 0.09–1.02, p = 0.019) (S2 Table); we included an interaction term in the analytic model (Table 5). After adjustment for socio-demographic and health covariates, in all countries except Mexico, patients with an OPCE score of 1 were 1.6 times (95% CI 1.37–1.90, p < 0.001) as likely to report excellent or very good SRH as those with a score of 0. The association was significantly stronger in Mexico: incorporating the interaction term, patients with an OPCE score of 1 had a 4.27 (95% CI 2.34–7.81, p < 0.001) times higher probability of reporting excellent or very good SRH compared to those with an OPCE score of 0.
Tables 6 and 7 depict the association of specific PC patient experiences with excellent or very good SRH. We found evidence of multiple interactions between countries and specific features of patient-centered PC (S2 Table); we thus present results stratified by country (Tables 6 and 7). The analysis revealed differences among countries in patient experiences associated with a high probability of having excellent or very good SRH, when controlling for the study covariates. After adjustment for socio-demographic and health characteristics, the experience of easy contact with the PC facility by telephone during regular office hours was associated with excellent or very good SRH in Mexico (aPR 1.35, 95% CI 1.04–1.74, p = 0.023), the perception that the PC provider gives an opportunity to ask questions was associated with excellent or very good SRH in Brazil (aPR 1.42, 95% CI 1.11–1.83, p = 0.006), having a PC provider who knows relevant information about the patient’s medical history was associated with excellent or very good SRH in Mexico (aPR 1.47, 95% CI 1.03–2.12, p = 0.036) but was negatively associated with excellent or very good SRH in Brazil (aPR 0.71, 95% CI 0.56–0.89, p = 0.003), the perception that the PC provider solves most of the patient’s health problems was associated with excellent or very good SRH in Colombia (aPR 1.38, 95% CI 1.01–1.91, p = 0.046) and in Jamaica (aPR 1.21, 95% CI 1.02–1.43, p = 0.030), and coordination of care by the PC provider was associated with excellent or very good SRH in Mexico (aPR 1.53, 95% CI 1.19–1.98, p = 0.001). After adjustment for covariates, no individual features of patient-centered PC were associated with excellent or very good SRH in El Salvador or Panama.
This secondary analysis of a nationally representative survey of the urban population in 6 LAC countries found that higher OPCE was associated with excellent or very good SRH. At the same time, the specific features of patient-centered PC associated with excellent or very good SRH differed among countries, with features from the domains of contact with clinic, communication, technical quality, and coordination showing significant associations in at least 1 country. The findings underscore the importance of high-quality, patient-centered PC as a path to improved population health while identifying areas for future country-specific investigation.
Overall scores are considered a valid alternative to global ratings in patient experience surveys [27]. To our knowledge there were at least 2 previous studies in the US and South Korea that investigated the association between overall PC quality metrics and SRH [20,21]. The first study utilized summary metrics of accessibility, interpersonal relationships, and continuity, while the second included first contact, personalized care, coordination function, comprehensiveness, and family/community orientation. Both measures showed a positive association between better PC experience and better SRH. Consistent with these 2 studies, we found a significant association of the average OPCE score with excellent or very good SRH in the context of LMICs in the LAC region.
Interestingly, we found that in Mexico patients with an OPCE score of 1 had a 4.27 times higher probability of reporting excellent or very good SRH compared to those with an OPCE score of 0. Also, we found that the overall SRH in Mexico was substantially lower than in the other countries, and the factors that explain this difference might also help us understand why the relationship of SRH with patient-centered PC is stronger in Mexico. Further country-specific research would be needed to identify such factors.
While broad policy statements on the centrality of PC for achieving health for all are important [2], these often lack the specific guidance to policy-makers who intend to pursue health system reform and introduce PC orientation within their health systems. Thus, detailed knowledge on specific patient experiences associated with better SRH (i.e., excellent or very good SRH) is important to identify priority areas for improvement in the delivery of healthcare, together with further assessments with longitudinal or experimental data [43]. In our study, PC features associated with excellent or very good SRH varied among countries. Two countries (El Salvador and Panama) showed no significant associations for the individual features of patient-centered PC, suggesting that the totality of the experience was more salient than any component within it. In Mexico, having a facility easy to contact by telephone, having a provider who knows relevant information about the patient’s medical history, and having a provider who coordinates healthcare were associated with better SRH. Having a PC provider who gives an opportunity to ask questions was associated with better SRH in Brazil, and having a PC provider who solves most health problems was associated with better SRH in Colombia and in Jamaica. Taken as a whole, the results suggest that the domains of patient-centered PC are all important to patient-reported health, but that the individual components with greatest relevance vary across settings. These characteristics shape the definition, goals, and priorities of PC. The attainment of PC goals requires easy communication with the clinic or provider to guarantee timely access to care, coordination among healthcare providers to assure continuity of care, and the ability to solve health-related problems. Previous studies found the importance of these experiences to patients [44,45], yet, to the best of our knowledge, our study is the first to find the association of these characteristics with very good or excellent SRH. Effective patient-centered communication was associated with improved health outcomes in several studies [46,47]. In our study, the opportunity to ask questions was significant only in Brazil.
The study has several limitations. First, it is an observational analysis of a cross-sectional survey, which does not allow making causal inferences or identifying the direction of the association between the study variables. Bidirectional relationships could be possible between better SRH and some healthcare experiences. For instance, on the one hand, people with poorer health are less likely to give the clinician credit for solving issues, and on the other hand, worse health problems are harder to solve. Second, due to the high prevalence of missing data, the analysis included IP weighting; therefore, we had to assume that the population with a regular PC clinic or doctor was exchangeable, conditional on covariates, with the population without a regular PC clinic or doctor; if this assumption was violated, the results would not be generalizable to those without a regular PC clinic or doctor. Third, in cross-national comparisons of survey data, cultural differences may lead to different interpretations of the questions being asked of respondents. For this reason, questionnaires had to be adapted for the characteristics of each country. Rather than focusing on the specifics of service provision in each country, this study aimed at identifying the broader roles of PC that may affect patient experience. Fourth, the results of our study are generalizable only to the urban populations of the analyzed 6 countries, as the samples were designed to represent national urban populations in each country. The results do not represent experiences of rural populations. Fifth, our findings cannot be generalized to other LMICs because of the different characteristics of their healthcare systems. Finally, information on type of employment and income was not collected by the survey; however, information on level of education and HI was available and included in this study.
In the context of the 40th anniversary of the Declaration of Alma-Ata, there seems to be broad consensus that strengthening PC is an essential strategy to achieve universal health coverage and the Sustainable Development Goals. In parallel, there is a growing interest in the importance of patient-centered healthcare as a tool for improving outcomes. However, to date there is little empirical cross-country evidence from LMICs that tests whether the main attributes of patient-centered PC are associated with better individual health. This study contributes to closing this gap by showing specific characteristics of patient-centered PC, and an overall summary measure of patient-centered PC performance, that are associated with better SRH in a sample representative of nearly 330 million people in 6 LAC countries. While the current study focused on self-reported cross-sectional data, the expansion of PC coverage in LAC countries and the increasing availability of administrative and clinical data associated with the introduction of electronic health records should allow for more longitudinal analyses to be conducted in the future.
|
10.1371/journal.ppat.1000922 | The Early Stage of Bacterial Genome-Reductive Evolution in the Host | The equine-associated obligate pathogen Burkholderia mallei was developed by reductive evolution involving a substantial portion of the genome from Burkholderia pseudomallei, a free-living opportunistic pathogen. With its short history of divergence (∼3.5 myr), B. mallei provides an excellent resource to study the early steps in bacterial genome reductive evolution in the host. By examining 20 genomes of B. mallei and B. pseudomallei, we found that stepwise massive expansion of IS (insertion sequence) elements ISBma1, ISBma2, and IS407A occurred during the evolution of B. mallei. Each element proliferated through the sites where its target selection preference was met. Then, ISBma1 and ISBma2 contributed to the further spread of IS407A by providing secondary insertion sites. This spread increased genomic deletions and rearrangements, which were predominantly mediated by IS407A. There were also nucleotide-level disruptions in a large number of genes. However, no significant signs of erosion were yet noted in these genes. Intriguingly, all these genomic modifications did not seriously alter the gene expression patterns inherited from B. pseudomallei. This efficient and elaborate genomic transition was enabled largely through the formation of the highly flexible IS-blended genome and the guidance by selective forces in the host. The detailed IS intervention, unveiled for the first time in this study, may represent the key component of a general mechanism for early bacterial evolution in the host.
| It has been known for some time that bacteria undergo genome-reduction when they transition from a free-living state to a constantly host-restricted state. High levels of IS element expansion were also found in these bacteria, and the IS elements were suggested to play a role in genome reductive evolution. Here we provide evidence for stepwise IS actions as the exclusive mechanism that mediates bacterial genomic changes during the early stage of constant host-bacterial association, by unveiling the processes that resulted in the development of B. mallei genome. We show the details of the multi-level interplay of IS elements, which facilitate the wide spread of the IS copies, and the overall mechanics in genome reduction and rearrangement. These processes appeared to operate as chain reactions mediating elaborate genomic transition, without seriously affecting the original gene expression patterns. The absence of differential gene expression in the resulting genome suggests that changes in transcriptional regulation that are often observed in other old bacterial genomes may take place subsequent to the IS-mediated steps, along with gradual nucleotide-level changes.
| The genomes of host-adapted bacteria, including endosymbionts and obligatory intracellular pathogens, go through reductive evolution [1], [2], [3]. Such changes are partly due to a reduced pressure to maintain genes that are not essential for survival in the host. Similarly, decreased efficiency of purifying selection, resulting from the reduced population size from a restricted life, results in inactivated genes, including beneficial genes, through genetic drift [3]. During the early stage of the genome reduction process, the majority of genes are lost as large chromosomal fragments spanning multiple genes. Such genome reduction has been documented in diverse bacterial groups, including Firmicutes, Chlamydiae, Spirochetes, and γ-Proteobacteria [1], [3], [4], [5], [6], [7]. Most of these bacteria have large expansion of IS elements (insertion sequences), and thus it has been suggested that the IS elements may play an essential role during the genome reduction process [1], [3], [8], [9], [10].
Burkholderia pseudomallei and Burkholderia mallei belong to the ß-Proteobacteria family, and are the causative agents of melioidosis and glanders, respectively [11], [12], [13], [14], [15], [16], [17]. B. mallei has very recently (∼3.5 myr) evolved from a clone of B. pseudomallei through extensive genome reduction [18], [19], accounting for as much as 1.41 Mb or 20% of the genome, as estimated by the size difference between the genomes of B. mallei ATCC 23344 and B. pseudomallei K96243 [18], [20], [21]. Concomitant with this process, B. mallei became constantly associated with mammalian hosts, specifically equines [22], [23], while B. pseudomallei maintains an opportunistic pathogenic lifestyle [17]. Preliminary analyses of the two type strains, B. mallei ATCC 23344 and B. pseudomallei K96243, have suggested that genome reduction and rearrangement in B. mallei were mediated by IS elements that are widely spread throughout the genome [20], [21]. Genes that have been deleted from the B. mallei genome but are maintained in B. pseudomallei include genes that are required for environmental survival. Many of these genes encode metabolic functions for the synthesis of metabolites or the utilization of various sugars and amino acids, without which bacterial propagation in the environment could be significantly hindered [20].
While the genomic reduction during bacterial restriction to their hosts has been well documented [1], [8], [10], most of the stepwise processes have not yet been elucidated. The B. mallei genome has unique significance, as it is much younger than the other genomes in which the genome-reductive evolutionary processes have been most studied to date, including Buchnera (>150 million years) and other much older groups [1], [3], [4], [5], [6], [7]. The studies with these older genomes have been challenging due to the subsequent genomic- and nucleotide-level mutations that accumulated over a long evolutionary history. In this study, we dissected 10 genomes each of B. pseudomallei and B. mallei to understand the early-stage processes that drive genome-reductive evolution in host-associated bacteria.
It is well known that bacteria specialized to a (host) niche, often have a large number of IS elements compared to their free-living relatives [1], [3], [8], [9], [10]. Likewise, by comparing genome sequences, we found that three types of IS elements, ISBma1, ISBma2, and IS407A, were significantly increased in B. mallei compared to B. pseudomallei (Fig. 1A). By contrast, other types of IS, including IS1356, ISBma3, ISBma4, and ISBma5 were found in low copy number in both species of bacteria. These elements appeared to be mostly degenerate evolutionary remnants (i.e., part of the IS disrupted or deleted) of the Burkholderia lineage. ISBma1, ISBma2, and IS407A also had degenerate elements in each species; the ISBma1 elements had the highest levels of degeneration (44%), followed by ISBma2 (20%), and by IS407A (5%) (Fig. 1A).
Intriguingly, up to almost 90% of ISBma1, ISBma2, and IS407A (88.5%, 86.1%, and 89.6%, respectively) were found to be present at the corresponding loci in all 10 B. mallei strains, when examined after the rearranged genomic fragments in each strain were aligned against a reference genome of B. pseudomallei K96243 (Fig. 1B; for a scaled map with the IS insertion sites in all B. mallei and B. pseudomallei strains, see Fig. S1; for the patterns of genomic rearrangements in the strains of each species, see Fig. S2; for the actual comparative blast data, see Tables S1 and S2). In contrast to these “core” elements, those elements that were not present in all (singletons and those found in a few strains), collectively referred to as “accessory” elements, were much less common. That the core elements, expected to be associated with the speciation of B. mallei from B. pseudomallei, accounted for most of the elements clearly reflects the common origin of B. mallei strains from a clone of B. pseudomallei [18], [20]. More importantly, it also suggests that further transpositions were significantly slowed after subsequent geographical segregation of the bacteria. There are 13 core elements in B. mallei that have matching IS elements located at the same sites in B. pseudomallei strains (Table 1). These elements were found to be composed of elements of ISBma1 and ISBma2 but not of IS407A. This finding suggests that ISBma1 and ISBma2 have a longer history of association with B. pseudomallei than IS407A does.
Among the three largely expanded elements, we found that IS407A and ISBma2 were associated with almost all of the large genomic deletions and rearrangements in the B. mallei strains (Fig. 2; Figs. S1 and S2; Tables S1 and S2). The only exception to this was a large deletion found in the strain ATCC 23344 and its direct derivatives, FMH, JHU, and GB8 horse 4 [24], between the 43rd and the 44th elements in chromosome 2 (Fig. 2; Table S1). No genomic rearrangement was mediated by features other than the two IS elements. ISBma1, which was significantly increased in B. mallei, was not directly involved in any of the genomic deletions or rearrangements, however as many as 35% of it served as secondary entry points for IS407A. The majority of the core elements of IS407A, 71.8% and 63.3% in chromosomes 1 and 2, respectively, mediated rearrangements, deletions, or both (Fig. 3A). By contrast, accessory elements of IS407A contributed less, but were more active in chromosome 2 than in chromosome 1. By contrast, 50.4% and 53.2% of the core elements of ISBma2 in chromosomes 1 and 2, respectively, contributed to rearrangements and/or deletions, and the accessory elements in both chromosomes were very rarely involved (Figs. 2 and 3). We identified 59 and 28 genomic fragments in chromosomes 1 and 2, respectively, which were encompassed by core elements of IS407A or ISBma2; these core elements mediated genomic rearrangements in at least one strain (Figs. 2; Table S1). We referred to these genomic fragments as BRUs (basic rearrangement units), a set of basic units for genomic reduction and rearrangement in B. mallei. The BRUs formed various rearrangement patterns in the B. mallei strains (Fig. S2A). By contrast, B. pseudomallei strains had little variation in genome arrangement among one another due to low levels of IS elements- a few rearrangements were found but were around non-IS repeat sequences (Fig. S2B).
When the pattern of the IS insertions and their involvement in genome-reductive and rearrangement processes in strains were used to construct a phylogenetic tree, strains sharing a recent common ancestry (e.g., ATCC 23344 and its immediate derivative isolates, FMH, JHU, and GB8 horse 4) or common recent geographical origins (e.g., strains NCTC 10257, NCTC 10229, and 2002721280 from European countries) were grouped together (Fig. 3B). This phylogenic relationship supports the hypothesis that the accessory IS elements, which provided the major determinants for the tree rather than the common core elements, occurred following the speciation and geographical segregation of the B. mallei strains. By contrast, such patterns were not obvious among the B. pseudomallei strains which did not go through IS element expansions; Australian strains 1655 and 668 did not branch separately from the South Asian strains.
The deletions and rearrangements that were mediated by accessory elements were most frequently noted in strains SAVP1 and 2002721280, which lost virulence after successive passages in laboratory cultures [25] (Figs. 2 and S1). Most of the extra deletions in these strains were more prominent in chromosome 2 than in chromosome 1. In SAVP1, an IS407A-mediated deletion removed a major group of virulence genes encoding the animal-type type III secretion system in the BRU B22 (Figs. 2 and S1); this deletion may be a major cause of the avirulence of that strain. By contrast, there is no obvious deletion that may be responsible for the loss of virulence in strain 2002721280. That the strains SAVP1 and 2002721280 obtained deleterious mutations from in vitro culturing suggests that maintenance of the genomic contents in B. mallei requires selective pressure for survival in the host environment. By contrast, the fully virulent strain PRL-20 showed more frequent deletions and rearrangements mediated by accessory elements than other virulent strains. This strain may represent one of the more evolved (more genome-reduced) strains of B. mallei.
Although extra deletions and rearrangements were noted, the actual number of the accessory IS elements was not significantly increased in PRL-20, SAVP1 or 2002721280. Furthermore, none of the direct derivatives of the strain ATCC 23344 (i.e. FMH, JHU, and GB8 horse 4) had new IS insertions (Fig. 2 and Table S1). These ATCC 23344 derivatives also did not have genomic rearrangements; the only change found was a single IS407A-mediated deletion located within the BRU B17 in the strain JHU (Fig. 2 and Table S1). These lines of evidence suggest that B. mallei genomes are structurally flexible with regard to deletions, however perhaps not as much anymore for additional IS transpositions or genomic rearrangements.
IS407A elements are known to generate 4-bp target region duplications as direct repeats around them when they transpose [26]. We found that ISBma1 generates 8-bp target region duplications, and that ISBma2 generates longer repeats of various lengths (18–26 bp) (Table 2; for the entire data, see Table S3). In addition to the various lengths of duplications, these target regions of the three types of IS had different nucleotide compositions and patterns. Most notably, the sequences of ISBma1 contained homopolymers of A and/or T in up to 8-bp stretches of nucleotides (Fig. 4A). The target sequences of ISBma2 had a loose pattern in which the GC-rich central region was encompassed by strands of As and Ts on either side. Target sequences of IS407A had the least characteristic composition. It is intriguing to note that each IS element showed different levels of copy number expansion, ISBma1 with the lowest (3.3×), ISBma2 with an increased level (9.5×), and IS407A with the highest (16.7×) (Fig. 1A). Perhaps this difference, at least in part, resulted from the availability of genomic sites suited for insertion targets.
There were concordant patterns of disruption of the core elements of one type by another, in that ISBma1 and ISBma2 were intersected by transposed IS407A (Fig. 4B), while the reverse (IS407A disrupted by ISBma1 or ISBma2) was not found. A possible explanation for these insertion patterns may be that ISBma1 and ISBma2 could not transpose into IS407A due to the lack of sites suited for their rather uncommon target preferences, while IS407A did not have this problem. Consistent with this hypothesis, ISBma1 and ISBma2 also did not have self-disrupted elements, while there were several self-disrupted IS407A elements. The involvement of the three IS elements with different target sites increased the total number of IS insertions in the genome. Furthermore, this increase led to further spread of IS407A, because ISBma1 and ISBma2 provided neutral insertion points for the element. This in turn directly improved the efficiency of IS407A-mediated recombination in the genome, resulting in more sophisticated deletions and rearrangements.
We estimated that 83.7% of IS407A and 65.6% of ISBma2 elements in the B. mallei genomes lost their matching target duplicates, while all of the elements from intact ISBma1 elements were maintained (Table S3). Almost all of the IS407A (see Table S3 for details) and all of the ISBma2 elements that contained matching repeats were not involved in genomic rearrangements in B. mallei. This indicates that recombination among the elements were the major cause of the loss of the matching target duplicates.
B. mallei still has a high nucleotide-level identity (99%) to B. pseudomallei. Consistent with this, there was no AT-biased genome deviation in B. mallei, unlike that seen in many old symbionts or obligatory host-associated pathogens [1], [3]. Although the overall identity is still very high, significant nucleotide-level divergence exists, especially at the SSRs (simple sequence repeats), where there are intrinsically high mutation rates [27]. These SSRs were abundant in both B. mallei and B. pseudomallei at corresponding sites in the genomes. However, there were more genes that were disrupted by frameshift mutations in B. mallei compared to B. pseudomallei (Table S4). Most of these disrupted genes were commonly present in all B. mallei strains, reflecting the clonal origin of the strains. Some of these gene disruptions may have contributed to better adaptation of the bacteria (increased persistence) in the host environment or simply became obsolete [28]. One of the most characteristic loss of function or of surface structure in B. mallei is the loss of flagella [20]. A gene essential for flagellum biogenesis, fliP, [29] in the strain ATCC23344 was disrupted by a 65-kb fragment flanked by IS407A elements, and this mutation completely turned B. mallei flagella-less. This disruption in fliP is present in all B. mallei strains (Table S1; Fig. 2, between BRUs A2 and A3), implying the significance of losing flagella in the evolution of host-restricted B. mallei. The loss of flagella has been noted in other bacteria, including Bordetella pertussis and Bordetella parapertussis during their host specialization, derived from the strains of Bordetella bronchiseptica, [9] and Yersinia pestis during its conversion from a gut to a systemic pathogen [30]. Additional disrupted genes not present in all strains were found at approximately the same levels as in B. pseudomallei, suggesting that there were no significant increases in mutation rates in B. mallei after geographical segregation. There also was no significant level of erosion of these, so called, pseudogenes by purifying selection at levels high enough to contribute to the actual genome size reduction (data not shown).
The extensiveness of the genome-wide reduction and rearrangements as well as additional nucleotide-level mutations may suggest that there is a potential for altered gene expression patterns in B. mallei. A total of 341 potential regulatory genes survived the general IS-mediated genomic reduction in B. mallei (not taking into account the diverse strain-specific deletions that occurred after speciation). Among these genes, only a small fraction (about 10) in each strain had deleterious (e.g. frameshift, null, or IS-insertion) mutations (for the list of the genes, see Table S5). In addition, none of the predicted operons in B. mallei, which correspond to the putative operons previously found in B. pseudomallei K96243 [31], were disrupted by IS elements (data not shown). We also estimated the potential for changes in promoters. There were 2,473 upstream sequences of genes, many of which may overlap or contain promoters, in the reference genome of B. pseudomallei K96243 that have homologous sequences (with at least 95% identity over at least 95% of their lengths) in all other strains of B. pseudomallei. We found that up to 99% of these sequences also matched the corresponding regions in B. mallei ATTC23344 at the same homology levels (see Table S6 for the list of the 2,473 upstream sequences, associated gene information, and the blast data). Together, all these data from the analyses of the conserved genomic regions suggest that there is only a low potential for the genes in B. mallei to have significantly divergent gene expression patterns from B. pseudomallei.
By contrast, there were 56 genes with putative regulatory functions that were lost along with the commonly deleted genomic fragments of the B. mallei genome. These genes include potential global regulatory genes, such as those encoding a quorum-sensing system (genes BPSS1176 and BPSS1180 in the reference genome of B. pseudomallei K96243), a two-component regulatory system (the pair BPSS1994 and BpSS1995 in B. pseudomallei K96243), and a number of regulators of various families (Table S7). Whether the loss of any of these 56 regulatory genes affects the expression of the remaining genes in the B. mallei genome was yet to be examined.
To experimentally estimate the possible transcriptomic divergence between B. pseudomallei and B. mallei, we infected female BALB/c mice with B. mallei ATCC 23344 or B. pseudomallei K96243, employing the previously established aerosol models of acute glanders and melioidosis [32]. Gene expression was compared in the bacteria that colonized the lungs and the spleens of the mice. Both B. mallei- and B. pseudomallei-challenged animals showed increases in the bacterial loads within these organs over time, with B. pseudomallei having slightly faster growth rates (Fig. 5). In our experience, B. pseudomallei also grew faster than B. mallei in vitro (data not shown). Unlike the mice infected by B. mallei, sampling the B. pseudomallei-challenged animals after 72 hr was not possible due to animal mortality from the more rapid disease progression. When gene expression profiles in the spleens and lungs were compared between B. mallei and B. pseudomallei at middle- (i.e., 24 hr for both bacteria) and late stages (i.e., 48 hr for B. pseudomallei and 72 hr for B. mallei) of infection (a total of four comparison pairs), conserved B. mallei and B. pseudomallei orthologs showed nearly identical patterns with high Pearson correlation coefficient (R) values ranging from 0.94 to 0.97, regardless of the host tissue type (Fig. 5). Therefore, there was no indication of significant modifications of the expression schemes in the genes required by B. mallei to thrive in BALB/c mice compared to those in B. pseudomallei. This is consistent with the findings of our previous gene expression studies in culture and in vivo, which also showed similar gene expression patterns in B. mallei and B. pseudomallei [20], [33], [34], [35]. These data suggest that, during the early stage, genomic reduction proceeds conservatively, not seriously affecting the indigenous gene expression patterns. In contrast to B. mallei, most of the transcription units in the insect symbiont Buchnera were altered, most likely due to complex genomic alterations accumulated over a long period of time [2].
In this study, we unveiled the mechanics of genomic deletions and rearrangements that occur in the early stage of bacterial specialization in the host, by conducting comparative analyses of B. mallei and its parental species, B. pseudomallei. It became clear that stepwise IS intervention was the main driving force mediating a large genomic reduction in B. mallei. Expansion of ISBma1 and ISBma2 in a clone of B. pseudomallei set the stage for the wide spread of IS407A, allowing its proliferation to sites, to which the element itself may rarely target. Actual genomic deletions and rearrangements occurred through recombination reactions mainly among IS407A and also among ISBma2 (Fig. 2). These processes achieved highly efficient deletions of dispensable genomic regions, causing only small disruptions to the portions of the genome that were maintained. This was possible due to the guidance by selective forces in the host and via the intrinsic flexibility of the compactly IS-blended genome. The B. mallei genome currently appears to still be structurally flexible with regard to deletions but is now less flexible with regard to genomic rearrangements and additional transpositions. This may indicate that the genomic evolution in B. mallei has been moving into a second stage, in which large-scale genomic alterations are reduced and nucleotide-level erosion has become more important. On the other hand, a large number of genes disrupted by frameshift mutations in SSRs were found in the B. mallei genome. The loss of function encoded by these genes and of flagella via disruption in fliP by IS407A (Table S1), could be part of the adaptive evolution for survival in the host environment, which will eventually lead to genome size reduction by erosion over time. Widespread relics of IS elements found in diverse symbionts and obligate pathogens [1], [3], [8] clearly suggest that a similar sequential IS intervention, modeled in Figure 6, may illustrate a general mechanism, by which elaborate genome transition occurs during early bacterial evolution after establishing constant association with the host.
All research involving live animals was conducted in compliance with the Animal Welfare Act and other federal statutes and regulations relating to animals and experiments involving animals and adhered to the principles stated in the Guide for the Care and Use of Laboratory Animals, National Research Council, 1996. All mouse experiments conducted in the USAMRIID (US Army Medical Research Institute of Infectious Diseases) were approved by the Association for Assessment and Accreditation of Laboratory Animal Care International.
The type strains for B. mallei (ATCC23344) [20] and B. pseudomallei (K96243) [21] were previously sequenced. Strains FMH, JHU, and GB8 horse 4 were direct derivatives of strain ATCC 23344 after passages in the human or horse, and these strains were also sequenced previously [24]. B. mallei strains NCTC10229, NCTC10247, and SAVP1 were sequenced with full closure and manually annotated as previously described [20]. The remaining three strains (2002721280, ATCC10399, and PRL-20) were sequenced to 8× Sanger sequence coverage by the whole genome shotgun method [36] without closure, and assembled using the Celera Assembler [37], and contigs were oriented by alignment to the reference strain ATCC23344 using PROMER [38]. ORFs were predicted and annotated automatically using GLIMMER [39], [40]. Pseudo-chromosomes were constructed from the ordered scaffolds, using manual examination where necessary. Similarly, B. pseudomallei strains 1106a, 1710b, and 668 were sequenced with full closure and manual annotation, while 1655, 406e, S13, and Pasteur 6068 were sequenced without closure and annotated automatically.
For the analyses of genomic deletions and rearrangements in B. mallei and B. pseudomallei, 5,799 predicted protein sequences from the B. pseudomallei type strain K96243 were compared with the nucleotide sequences of the genomes of B. mallei (ATCC 23344, 2002721280, ATCC 10399, FMH, JHU, GB8 horse 4, PRL-20, NCTC 10229, NCTC 10247, SAVP1) and the other strains of B. pseudomallei (1106a, 1106b, 1655, 1710a, 1710b, 406e, 668, Pasteur, S13) using tblastn (http://blast.wustl.edu). For the mapping of the insertions of ISBma1, ISBma2, and IS407A in the genomes of B. mallei and B. pseudomallei, the entire sequences of the IS elements were searched against the 20 genomes using blastn (http://blast.wustl.edu). For the analysis of association of the IS elements with genomic deletions and rearrangements in B. mallei and of the target sequences in the genomes, strain ATCC 23344 represented all of its immediate derivatives, FMH, JHU, and GB8 horse 4, to avoid redundancy in the data, because the three strains showed identical patterns. To compare the patterns of genome rearrangements in the B. mallei strains, the positions of the BRUs in each strain of B. mallei relative to B. pseudomallei K96243 were visualized using a genome-comparative software tool ACT ([41]; http://www.sanger.ac.uk/Software/ACT), and the displays were compared in parallel among the strains.
We also examined B. mallei and B. pseudomallei for intergenic regions that potentially containing promoters, putative regulatory genes, and disruptions of putative operons to estimate the possibility of causing gene expression divergence. For intergenic region comparisons, up to 100 bp upstream of the start codon, or up to as much as available if the neighboring gene was closer, of the genes that contain at least 50 bp of an untranslated upstream region were retrieved from the genome of B. pseudomallei K96243. Then, these sequences (2,268 and 1,566 from chromosomes 1 and 2, respectively) were searched against the genomes of B. mallei and B. pseudomallei using blastn (http://blast.wustl.edu), and the length-match as well as the identity values of the orthologous regions were calculated. Putative operons reported by Rodrigues et al. from the genome of B. pseudomallei K6243 [31] were used to match the orthologous gene clusters in the genome of B. mallei ATCC 23344, and these gene clusters were examined for any disruptions caused by IS elements. All the genome sequences of B. mallei and B. pseudomallei used in this study are available through the Pathema web site (http://pathema.jcvi.org/cgi-bin/Burkholderia/PathemaHomePage.cgi) at the J. Craig Venter Institute (http://www.jcvi.org/).
A phylogenetic tree was constructed with the strains of B. mallei and B. pseudomallei based on the insertion patterns of and the role played in the genomic deletions and rearrangements by the three major IS elements, ISBma1, ISBma2, and IS407A. All the data used are shown in Tables S1 and S2 and Figure 2. Bootstrapped maximum parsimony trees were calculated using the PAUP package with default parameters, and a consensus tree was produced from the bootstrap replicates. Branches with bootstrap scores of less than 50 were collapsed in the tree.
Among the duplicated target regions encompassing the IS elements ISBma1, ISBma2, and IS407A, those regions that had perfectly matching sequences were first collected. Then, among the sequences from unmatched pairs, those that occurred in more than two strains were assumed to be un-mutated valid sequences and, therefore, were added to the data pool for the analysis. Strain ATCC 23344 represented all its direct derivatives (FMH, JHU, and GB8 horse 4) in this analysis to avoid redundancy in the data. The collected sequences were aligned with Clustal X, and the alignments were graphically visualized using Sequence logos [42].
Exposure of mice to bacterial aerosol was performed as described by Roy et al. [43]. Fresh overnight cultures of B. pseudomallei DD503 [44] and B. mallei ATCC 23344 were prepared in LB or in LBG (LB supplemented with 4% glycerol), respectively, at 37°C with aeration (250 rpm). Thirty female BALB/c mice six to eight weeks old (National Cancer Institute, Frederick, MD, USA) were infected with these bacteria: nine mice each with B. pseudomallei and B. mallei for the gene expression studies, and six mice each for the bacterial load assays. The mice infected with B. mallei received an inhaled dose of 7.2×103 cfu (7.2×LD50), and those mice infected with B. pseudomallei received 1.8×104 cfu (18×LD50), as estimated by colony counting on agar plates. The infected mice were provided with rodent feed and water ad libitum and maintained on a 12-hr light cycle. After 24 and 48 hr (for both B. mallei and B. pseudomallei) or 72 hr (for B. mallei) of infection, five mice from each point in time were euthanized in a CO2 chamber, and their spleens and lungs were removed. Due to animal mortality, a 72 hr point in time was not possible for B. pseudomallei. The organs from two randomly picked mice were saved for bacterial load estimations, and the rest were homogenized in 1 ml of Trizol (Invitrogen Corp., Carlsbad, CA, USA) using a Tissue-Tearor (BioSpec Products, Bartlesville, OK, USA). Total RNA was purified according to the manufacturer's recommendations (Invitrogen Corp., Carlsbad, CA, USA). The bacterial load in the mouse organs was estimated as described by Ulrich and DeShazer [32].
Total RNA, both bacterial and mouse, from the same organ types from three mice was pooled to compensate for potential individual variation. These pooled RNA samples were used for the experiments without further purification of the bacterial RNA because RNA from mice does not cross-hybridize to the B. mallei microarray at a level affecting the legitimate interactions between the B. mallei array and the Burkholderia transcriptome [35]. The B. mallei whole genome array used in this study for both B. mallei and the closely related B. pseudomallei (average gene identity at the nucleotide level of 99%) was described in detail previously [33]. The B. mallei- and B. pseudomallei-infected organ samples were paired for the hybridization reactions based on early and late pathological states. A total of eight hybridization reactions or four different comparisons were performed, each of which was replicated in flip-dye pairs and the final ratios were calculated as log2 (B. pseudomallei gene expression intensity/B. mallei gene expression intensity). Labeling of the probes, slide hybridization, and slide scanning were carried out as previously described [35]. The independent TIFF slide images from each channel were analyzed using TIGR Spotfinder to assess the relative expression levels, and the data were normalized using a local regression technique LOWESS (LOcally WEighted Scatterplot Smoothing) with the MIDAS software (<http://www.jcvi.org/cms/research/software>, The J. Craig Venter Institute, Rockville, MD, USA). The resulting data were averaged from triplicate genes on each microarray and from duplicate flip-dye arrays for each experiment.
|
10.1371/journal.pcbi.1004909 | A MAPK-Driven Feedback Loop Suppresses Rac Activity to Promote RhoA-Driven Cancer Cell Invasion | Cell migration in 3D microenvironments is fundamental to development, homeostasis and the pathobiology of diseases such as cancer. Rab-coupling protein (RCP) dependent co-trafficking of α5β1 and EGFR1 promotes cancer cell invasion into fibronectin (FN) containing extracellular matrix (ECM), by potentiating EGFR1 signalling at the front of invasive cells. This promotes a switch in RhoGTPase signalling to inhibit Rac1 and activate a RhoA-ROCK-Formin homology domain-containing 3 (FHOD3) pathway and generate filopodial actin-spike protrusions which drive invasion. To further understand the signalling network that drives RCP-driven invasive migration, we generated a Boolean logical model based on existing network pathways/models, where each node can be interrogated by computational simulation. The model predicted an unanticipated feedback loop, whereby Raf/MEK/ERK signalling maintains suppression of Rac1 by inhibiting the Rac-activating Sos1-Eps8-Abi1 complex, allowing RhoA activity to predominate in invasive protrusions. MEK inhibition was sufficient to promote lamellipodia formation and oppose filopodial actin-spike formation, and led to activation of Rac and inactivation of RhoA at the leading edge of cells moving in 3D matrix. Furthermore, MEK inhibition abrogated RCP/α5β1/EGFR1-driven invasive migration. However, upon knockdown of Eps8 (to suppress the Sos1-Abi1-Eps8 complex), MEK inhibition had no effect on RhoGTPase activity and did not oppose invasive migration, suggesting that MEK-ERK signalling suppresses the Rac-activating Sos1-Abi1-Eps8 complex to maintain RhoA activity and promote filopodial actin-spike formation and invasive migration. Our study highlights the predictive potential of mathematical modelling approaches, and demonstrates that a simple intervention (MEK-inhibition) could be of therapeutic benefit in preventing invasive migration and metastasis.
| The majority of cancer-related fatalities are caused by the movement of cancer cells away from the primary site to form metastases, making understanding the signalling mechanisms which underpin cell migration and invasion through their local environment of paramount importance. Much has been discovered about key events leading to invasive cell migration. Here, we have taken this prior knowledge to build a powerful predictive model based on simple ON/OFF relations and logic to determine potential intervention targets to reduce harmful invasive migration. Interrogating our model, we have identified a negative feedback loop important to the signalling that determines invasive migration, the breaking of which reverts cells to a slower, less invasive phenotype. We have supported this feedback loop prediction using an array of in vitro experiments performed in cells within 2-D and physiologically relevant 3-D environments. Our findings demonstrate the predictive power of such modelling techniques, and could form the basis for clinical intervention to prevent metastasis in certain cancers.
| An estimated 90% of cancer deaths are caused by metastatic secondary tumours [1], a process instigated as certain cells escape the primary tumour to migrate in, and invade through, the local micro-environment. Cancer cells can adopt a range of different migratory mechanisms to achieve such invasion [2]: some migrate in co-operation with near neighbours in whole sheet like structures or chains following initial ‘guerrilla’ cells [3], while others migrate individually, using distinct but interchangeable motility mechanisms. In most cases, the mechanisms which coordinate cell migration are dictated by Rho GTPases [4], of which Rac1 and RhoA are the most well-defined. Rho GTPases are molecular switches which can be in a GTP-bound 'on' state, or a GDP-bound 'off' state [5] in response to activating guanine nucleotide exchange factors (GEFs) and inhibiting GTPase activating-proteins (GAPs) [6]. Rac1 is considered the dominant GTPase acting at the leading edge of lamellipodia, polymerising actin via the Arp2/3 complex to form a dendritic actin network [7,8], while RhoA dominates at the rear of the cell to activate ROCK driven contractility and rear-retraction [8,9]. More recently, RhoA activity has been observed immediately at the leading edge in cells migrating in 2D, with Rac active in a zone immediately behind this [10]. Rac1 and RhoA are thought to be mutually antagonistic [11,12], and studies suggest that cyclic bursts of RhoA and Rac1 activity in a pseudo-oscillatory manner may drive the leading edge of some cells forward by producing a necessary push-pull mechanism [13,14].
In 3D and in vivo, single ‘mesenchymal’ cells, and leader cells in collective migration, migrate in a Rac-driven manner, and the mechanisms of actin polymerization, protrusion and force generation are thought to be analogous to lamellipodial migration in 2D [2,15,16]. However, lamellipodium-independent 3D migration strategies have also been identified. Single cells can adopt amoeboid modes of migration [2,17] and fibroblasts can move in an adhesion- and contractility-dependent lobopodial mode of migration [18,19]. Furthermore, Rab-coupling protein (RCP) dependent endocytic recycling of the fibronectin (FN) receptor α5β1 integrin promotes formation of filopodial actin-spike protrusions to drive invasive migration in FN rich 3D-ECM and in vivo in response to inhibition of αvβ3 integrin or expression of gain-of-function (GOF) mutant p53 in cancer cells [20,21].
Expression of gain-of-function mutant p53 (e.g. R273H, R175H), or inhibition of αvβ3 integrin (with small-molecule inhibitors, e.g. cRGDfV), promotes the association of RCP with α5β1 and leads to rapid recycling of this integrin to promote invasive migration [20,21,22]. Rather than directly influence the adhesive capacity of the cell, RCP and α5β1 recruit epidermal growth factor receptor 1 (EGFR1) to a recycling complex, controlling their co-trafficking to the tips of invasive pseudopods in motile cells [20]. This potentiates EGFR signalling specifically at the tips of invasive pseudopods, activating PKB/Akt which phosphorylates RacGAP1 to allow it’s recruitment to the leading edge by the cytoskleletal scaffold IQGAP1 [23]. RacGAP1 directly inactivates Rac1, allowing RhoA activation specifically at the front of invading cells [23], and RhoA-ROCK mediated activation of formin homology domain-containing 3 (FHOD3) leads to polymerisation of filopodial actin spike protrusions which promote motility in 3D ECM [24]. Whereas cells under basal conditions exhibit slow but directionally persistent migration in 2D and low invasiveness in FN-rich 3D-ECM, RCP-α5β1 trafficking promotes rapid, random migration in 3D and invasion in FN-rich ECM and in vivo [20,21]. Interestingly, mutant p53 expressing cancers are more metastatic in a number of contexts [21,25,26,27,28], and high RhoA activity has been observed at the leading edge of pancreatic cancer cells harbouring such mutations [29], suggesting that RhoA mediated protrusion could underlie the protrusive and invasive characteristics of a variety of metastatic tumours. Understanding the mechanisms by which the pro-invasive Rac1 to RhoA switch is potentiated at the leading edge of a polarised cell in further detail is of paramount importance to in turn identify potential therapeutic targets which may abrogate invasive cell migration leading to metastasis.
Given that RhoA and Rac1 are contained in a highly connected complex network that regulates their relative activities [30,31], it is likely that systems biology analyses will reveal regulatory interactions and feedback loops which are involved in altering GTPase activity and migratory phenotype. Boolean logic is an attractive modelling tool in the context of a large and essentially poorly characterised setting, since the detailed kinetic parameters (activation/inhibition rates, initial molecular concentrations) crucial to the generation of continuous models based on ordinary differential equations are not required [32,33]. Instead, the activity of all variables in the model is binarised into simple ON or OFF states, and all reactions are assumed to take the same arbitrary length of time. Boolean models can therefore encompass broad network topologies and order events to describe the underlying biology and predict outcomes and phenotypes in diverse cellular contexts, including cell migration [14,34]. Here we report a detailed model based on simple Boolean logic to connect EGF/EGFR1 to Rac1 and RhoA in the context of RCP/α5β1 trafficking, identifying nodes/directed links from stringent literature mining to build upon existing EGF signalling pathway models. This model accurately recreated existing experimental findings, and moreover identified a previously unanticipated negative feedback loop, involving MAP kinase dependent control of the Rac1 activating Sos1-Eps8-Abi1 complex which determines the RCP/α5β1 dependent Rac to RhoA switch. Hence, mathematical modelling has revealed a targetable aspect of RhoGTPase activity control, which could be exploited as an anti-invasive approach in mutant p53 expressing cancers.
RCP and α5β1 recruit EGFR1 into a recycling complex, and promotes the return of this receptor to the plasma membrane where it can re-engage ligand to amplify the EGFR1 signalling. Because the signalling response (PKB/Akt and RhoA activation, Rac inactivation) localises to the front of the cell [23], we considered signalling only within this domain, and modelled the EGFR signalling network from the initial stimulation with EGF (the external input) to the Rac and Rho GTPases (the nominated outputs) by generating a Boolean network that included over 40 nodes and interactions mined from the literature (curated as described in Methods; Fig 1A). At the initiation of signalling within the model, PDK1, mTor, c-Src, Pip2 and Rac1 are ON, as this is the assumed activity in basal conditions prior to EGF signalling. All other nodes were initially set to OFF. Under these conditions the first 50 time increments were simulated to assess the effect on the RhoA and Rac1 outputs to allow for steady state or steady limit cycle to be reached (Fig 1B and 1C).
Simulation results depended a priori on the unknown hierarchical binding affinity of the Rac1 activators/inhibitors. In general, a GEF and a GAP can never bind to and activate/inactivate a Rho GTPase in an identical spatiotemporal state [6]. For example, if the Rac inhibitor preferentially binds to Rac1 compared to the Rac activator, then the Rac inhibitor must be OFF for the Rac activator to activate Rac1. Three Rac1 activators were included in the model: Vav2, RalBP1 and Sos1E (a moniker used to describe the Sos1-Eps8-Abi1 complex which shows Rac-specific GEF activity [35]); and one Rac inhibitor: the phosphorylated form of RacGap1 (which is localised to the leading edge upon RCP-driven integrin/EGFR trafficking and hence positioned to inactivate Rac [23]). All logical hierarchies involving the Rac1 activators and inhibitor and subsequent outputs are summarised in S3 Table. Hierarchies exhibited three distinct categories of RhoA/Rac1 binary output dynamics (Fig 1D–1F): 1) Rac1 remains ON and RhoA remains OFF for all time increments (Fig 1D); 2) cyclic bursts of RhoA and Rac1 activity (Fig 1F); and 3). Rac1 switches OFF, RhoA switches ON and these activities are sustained, Fig 1E). If Vav2, in any combination with the other activators, affected Rac1 more readily than RacGAP1 then Rac1 remained active for all time, and due to the mutual antagonism between the two different GTPases, RhoA remained inactive (Fig 1D), which contradicts the premise that RCP-mediated α5β1-EGFR1 co-trafficking promotes a pro-invasive Rac1 to RhoA switch in conjunction with EGF stimulation. This hierarchy was therefore rejected based on modelling results. Logical relations RalBP1 > pRacGAP1 > Vav2, Sos1E, RalBP1 > pRacGAP1 > Vav2 and Sos1E > pRacGAP1 > Vav2, RalBP1 all gave rise to outputs exhibiting both Rac1 and RhoA activity in a cyclic manner. Crucially however only the relation Sos1E > pRacGAP1 > Vav2, RalBP1 exhibited clear RhoA dominance in the output, whereby RhoA activity is ON for significantly more time increments than Rac1 (1C, F). RalBP1 > pRacGAP1 > Vav2, Sos1E and Sos1E, RalBP1 > pRacGAP1 > Vav2 did not exhibit sufficient RhoA activity to agree with previous findings [23] so was rejected as a plausible output. Finally, the hierarchy in which RacGAP preferentially inactivates Rac1 induced a simple Rac1 to RhoA switch (Fig 1B and 1E) which was a further permissible output to describe the known GTPase dynamics.
The two different plausible sets of output dynamics are in close agreement with previous experimental findings. Whether there is simply a switch output in Rac1 to RhoA activity at the leading edge during RCP-driven invasive cell migration (because Rac inactivation dominates over Rac activators) or whether the migration is driven by RhoA dominant pseudo-oscillatory Rho/Rac dynamics at the leading edge (because Sos1-Eps8-Abi1 complex preferentially affects Rac1) as has been suggested for some migratory schemes [14] is not known. We therefore performed individual in silico node knockouts, wherein each node was set to an OFF value for all time points by removing all edges leading into said node (or turning the nodes ON in the case of the inhibitory proteins SHIP2, PTEN and PP2a), to ascertain the effect on the RhoA and Rac1 dynamics for both the Sos1E > pRacGAP1 > Vav2, RalBP1 hierarchy (the cyclic pseudo-oscillatory output) and the pRacGAP1 > Sos1E, RalBP1, Vav2 hierarchy (the switch output) (Fig 2A).
For both hierarchies, the in silico knockouts accurately mimic previous experimental evidence: knockout of the node IQGAP1 for example results in persistent Rac1 ON and RhoA OFF activities, as seen previously in FRET-FLIM experiments in IQGAP1 knockdown cells [23]. Furthermore, upstream activators of PKB/Akt are an absolute requirement for plausible Rac1/RhoA outcomes (Fig 2A). Further to replicating known priors regarding certain proteins contained within the network, the node knockout workflow made predictions for the perturbation effect of all nodes in the model. For the pRacGAP1 > Sos1E, RalBP1, Vav2 case, the only proteins which are implicated as vital for the pro-invasive Rac to Rho switch downstream of RCP-dependent trafficking and subsequent EGF stimulation were those which act upstream of Akt activation, such as the well-known Akt activators PI3K and PIP3 [36,37] or the proteins upstream of RhoA activation, which in this model were Vav2 and its activator PIP3 (Fig 2A). For the pseudo-oscillatory inducing Sos1E > pRacGAP1 > Vav2, RalBP1 hierarchy the in silico knockouts provided more diverse predictions. As well as the same predictions concerning proteins upstream of Akt and RhoA, a Sos1 negative feedback loop was also predicted to affect GTPase dynamics (Fig 2A). Sos1 initiates a signalling path in which Ras [38], Raf1 [39], MEK1/2 [40], ERK1/2 [41] and p90RSK (in the presence of active PDK1 [42]) become active. Serine/threonine phosphorylation of Sos1 by active ERK1/2 or p90RSK however has been shown to cause dissociation of Sos1 from Grb2-EGFR and down-regulate Sos1 activity [43], which in Boolean terms translates to active ERK1/2 or p90RSK inhibiting Sos1 (i.e. switching it from ON to OFF). Once Sos1 is switched OFF, all proteins activated downstream of Sos1 are switched OFF, including the Sos1-Eps8-Abi1 complex (therefore preventing activation of Rac1) as well as all the proteins involved in the Ras, Raf1, MEK1/2, ERK1/2, p90RSK pathway. Once both ERK1/2 and p90RSK are OFF, Sos1 activation is allowed and the negative feedback loop initiates again. Breaking the feedback loop at different points in silico elicited altered GTPase activity: if MEK1/2 was removed, Sos1 remained ON, hence the Sos1-Eps8-Abi1 complex remained ON to keep Rac1 ON, hence there was no Rac1 to RhoA switch (Fig 2B); however if Eps8 was removed, the GTPase output changed from cyclic pseudo-oscillations to a Rac1 to RhoA switch as the Sos1-Eps8-Abi1 complex no longer formed, and thus the MEK1/2-driven feedback loop no longer had any effect on Rac1 activation (Fig 2C). This provided testable predictions which could be used to determine both the veracity of the model described and the hierarchy of GEF and GAP activity within the Boolean network.
To test model robustness and determine that the RhoA dominant cyclic activity of Rac1 and RhoA was not caused by the synchronous updating scheme used in Boolean simulations, we introduced inherent stochasticity into the model by using random asynchronous updates (whereby in any time increment, the next executable reaction will be chosen at random). Results using this updating scheme were in close agreement with the oscillations observed previously (Fig 1C), whereby the crucial RhoA dominance that has been observed experimentally remained present in all asynchronous simulations (S1A Fig) and all simulations showed some cyclic RhoA and Rac1 activity (S1B Fig). Moreover, all predictions concerning MEK1/2 and Eps8 perturbation remained intact. This verification shows that the synchronous update algorithm used in our Boolean simulations does not affect the core findings of the simulations or cause spurious oscillations.
Modelling results predicted that if the Sos1-Eps8-Abi1 complex preferentially affects Rac1, then removal/inhibition of MEK1/2, ERK1/2 or p90RSK will abrogate the invasive migration driven by the Rac1 to RhoA switch. To test this prediction experimentally we imaged cells migrating into a 2-D scratch-wound in the presence or absence of MEK1/2 inhibitors in two different carcinoma cell lines: A2780s, an ovarian cancer cell line where RCP driven α5β1/EGFR trafficking was promoted by inhibition of αvβ3 with cRGDfV; and H1299s, a non-small cell lung cancer cell line in which α5β1/EGFR trafficking was promoted via mutant p53 expression. For both cell lines, both PD184352 and AZD6244, two mechanistically different MEK1/2 specific inhibitors [44,45], showed clear suppression of p44/42 MAPK phosphorylation (Fig 2D and 2E). Under basal conditions, cells migrated slowly and persistently (Fig 2F–2I) with broad lamellipodia at the leading edge (Figs 2J and S2A, S1 Movie) consistent with high levels of active Rac at the leading edge. However when RCP-driven α5β1 trafficking was promoted, cells exhibited a more rapid and random migratory phenotype (Fig 2F–2I, S1 Movie) with narrow ruffling leading edge protrusions (Figs 2J and S2A, S1 Movie) rather than Rac-driven lamellipodia, consistent with suppression of Rac activity. MEK1/2 inhibition had no effect on speed, persistence or protrusion morphology under basal conditions (Figs 2F–2J and S2A, S1 Movie). However, in the context of RCP-mediated integrin-RTK co-trafficking and signalling, MEK1/2 inhibition reduced speed and increased persistence levels comparable to those seen under basal conditions in both A2780 and H1299 cell models (Fig 2F–2I). The phenotypic reversion was also evident in protrusion morphology, whereby narrow ruffling protrusions were replaced with broad lamellipodial structures (Figs 2J and S2A, S1 Movie), consistent with activation of Rac as predicted by the Boolean model with the Sos1E > pRacGAP1 > Vav2, RalBP1 hierarchy. Moreover, this suggests that a simple Rac to RhoA switch does not accurately describe the dynamics of RhoGTPase activity in invading cancer cells.
Imaging of fixed A2780 cells within a 3D cell-derived matrix (CDM) environment at high magnification enabled the resolution of F-actin structure in migrating cells. In line with previous findings [24], and analogous to 2-D scratch wound data, cells switch from lamellipodial protrusions (Fig 3A and 3B) upon induction of RCP/α5β1 trafficking with cRGDfV to actin-spike protrusions which are made up of numerous long filopodia which lack veils of dendritic actin (Fig 3A and 3B). Strikingly, upon inhibition of MEK1/2, protrusions revert back to lamellipodia with fewer and shorter filopodia similar to the leading edge structures seen in the basal, unstimulated case (Fig 3A, 3B, 3C and 3D). These data demonstrate that MEK inhibition prevents the establishment of actin-spike protrusions that form as a consequence of RCP-mediated trafficking of α5β1 and EGFR1, and suggest that in this context cells are unable to implement the Rac to RhoA switch in a physiologically relevant 3-D environment.
To investigate whether the observed MEK1/2 inhibition phenotypes, slowing migration and promoting stable lamellipodial protrusions, were indeed via an effect on the Rho GTPases Rac1 and RhoA as predicted by the model, we used FRET based biosensor experiments for cells individually migrating in 3D CDM. We developed a method to determine the mean average FRET ratio readout for the leading edge of motile cells by isolating a ‘ring’ 40 pixels wide in the front 25% of each cell (S6 Fig, also see Methods). Using this unbiased quantification method, we could show that Rac activity was low at the leading edge of cells moving on 3D ECM, whereas RhoA activity was high when RCP-driven trafficking of α5β1 and EGFR1 was promoted in A2780 cells upon cRGDfV treatment (Fig 3E, 3G and 3I, S2 Movie) or in H1299 cells upon expression of mutant p53 (S3A, S3C and S3E Fig), in agreement with our previous observations [23]. However, upon MEK1/2 inhibition with PD184352, Rac1 activity at the leading edge of a cell was significantly increased (Figs 3F and 3I and S3B and S3E, S2 Movie) whereas RhoA activity at the cell front (but not rear) was reduced (Figs 3H and 3I and S3D and S3E, S2 Movie). These data directly demonstrate that, in agreement with the model, inhibition of MEK prevents the Rac to RhoA switch observed in cells migrating in an RCP/ α5β1/EGFR1 driven manner, and suggest that MEK-ERK signalling could indeed provide a feedback loop to periodically supress Rac activity and permit RhoA activation.
Our Boolean model predicts that MEK-ERK signalling intervenes in the activation of Rac by influencing the activity of the Sos1 within the Sos1-Eps8-Abi1 complex [35,43]. Because MEK1/2 has been implicated in various cellular processes [46], we used knockdown of Eps8 as a method to remove the Rac-activating Sos1-Eps8-Abi1 complex in order to confirm the model prediction. One round of siRNA was sufficient to knock down Eps8 activity for both A2780 and H1299 cell lines using SMARTpool or single oligo siRNA reagents (Figs 4A and 4B and S5C). Eps8 siRNA had little effect on the lamellipodial migration of cells under basal conditions (Figs 4C–4F and 4H and S4A and S5A). RCP/α5β1/EGFR1 stimulated cell migration (A2780 +cRGDfV; H1299-mutant p53) in the absence of MEK1/2 inhibition was similarly unaffected by Eps8 knockdown, as a more rapid, random migratory phenotype remained. Whilst MEK inhibition was able to revert this rapid, random migratory phenotype in control knockdown cells (Figs 4C–4F and 4G and S4A and S5A, S3 Movie), Eps8 knockdown cells retained their fast migrating non-directional mode of movement in 2D even in the presence of MEK inhibitor (Figs 4C–4F and 4H and S4A and S5A, S3 Movie). These data provide further evidence that that a feedback loop initiated by MEK-ERK signalling acts to modulate migratory behaviour through a negative effect on the Sos1-Eps8-Abi1 complex, in agreement with the model prediction.
To validate that the effect of Eps8 siRNA was indeed due to the downstream effects on Rac1 and RhoA, we again used FRET based biosensors to directly ascertain GTPase activity at the leading edge of cells migrating in 3D CDM. For control siRNA, the FRET ratios again showed that MEK1/2 inhibition promotes high Rac1 activity and low RhoA activity in lamellipodium-like protrusions (Figs 4I, 4K and 4N and S5B, S4 Movie). However Eps8 knockdown (Fig 4M) reduced Rac1 activity but concomitantly increased RhoA activity within the protrusive regions of migrating cells (Figs 4J, 4L and 4N and S5B, S4 Movie). These findings are again consistent with the model prediction that MEK1/2 inhibition affects localised Rac1 and RhoA activation dynamics via the Sos1-Eps8-Abi1, but has no effect when formation of this Rac1 activating complex is prevented by Eps8 siRNA.
In cancer cells, the promotion of RCP-dependent α5β1/EGFR1 co-trafficking and signalling increases the invasive migration within FN-rich ECM hydrogels by mediating a Rac to RhoA switch at the leading edge to promote filopodial actin spike protrusions [20,21,23,24]. We therefore investigated whether the MEK1/2-ERK mediated feedback loop, which is required to maintain RhoA activity at the leading edge (Fig 3G–3I) influenced cell motility within 3D ECM that resembles interstitial matrix encountered by metastatic cancer cells. In the absence of Eps8 knockdown, MEK1/2 inhibition supressed the RCP-driven invasion of two different cell lines in high concentration collagen and FN hydrogels to <50% of uninhibited levels (Fig 5A–5D), consistent with previous observations [47] and providing more evidence to support MEK1/2 as a potential target to abrogate invasive cell migration. Strikingly, Eps8 knockdown A2780 and H1299 cells were unaffected by MEK1/2 inhibition, and showed a similar capacity to migrate and invade a 3D ECM when compared to untreated control knockdown cells (Fig 5A–5D), indicating that the Sos1-Eps8-Abi1 complex is critical MEK inhibition-induced phenotype reversion. These data demonstrate the regulation of the Rac to RhoA GTPase switch is critical for cancer cell invasion, and further support the model prediction that a Sos1-Ras-Raf-MEK-ERK negative feedback loop affects Rac1 and RhoA via the Sos1-Eps8-Abi1 complex to determine the mode of cell migration in 3D microenvironments.
We have used Boolean logic to model EGFR1 signalling at the front of the migrating cell in the context of RCP/ α5β1-mediated EGFR1 recycling, including over 40 nodes and interactions mined from the literature, which led to the prediction that a Sos1-Ras-Raf-MEK-ERK negative feedback loop affects localised Rac1 and RhoA dynamics at the leading edge of migrating cells via the Rac1 activating Sos1-Eps8-Abi1 complex. Using either of two MEK1/2 inhibitors with distinct mechanisms for suppression of target activity [44,45], we demonstrated that perturbing the MAP kinase pathway reverts the RCP/α5β1/EGFR1-driven rapid, random migration in 2D to a slower, more persistent, lamellipodial phenotype previously associated with dominant Rab4-αvβ3 integrin recycling [48]. In 3D-ECM, a similar reversion to lamellipodial protrusion was observed upon MEK inhibition, accompanied by increased Rac (but decreased RhoA) activity at the leading edge and a decrease in the ability to invade in FN-rich collagen hydrogels. Both the switching of GTPase activity and phenotypic reversion was dependent on Eps8, suggesting that the Sos1-Eps8-Abi1 complex is the target of the MEK-ERK driven feedback loop.
Our manually curated model was based on existing extensive EGFR networks found in the literature [32,49]. Proteins and interactions included in the final simulations are generally thought to be highly expressed and conserved across various different cell types, and proteins/interactions which resulted in biologically implausible results were rejected a priori. Based on experimental data, we accepted the Sos1E > pRacGAP1 > Vav2, RalBP1 Rac1 activator/inhibitor hierarchy. This particular hierarchy gave rise to cyclic activity of Rac1 and RhoA in a pseudo-oscillatory manner in the unperturbed simulation (Fig 1F). This was deemed plausible as it exhibited RhoA dominance at the leading edge of the cell downstream of RCP-α5β1 trafficking, as RhoA was ON for longer than Rac1. As a consequence of the negative feedback loop feeding into Sos1 activity however, cyclic activity was a natural consequence of this hierarchy. Using in silico node knockout workflow we identified a proposed mechanism through which the MAP kinase pathway affects GTPase activity and subsequent invasion. MEK1/2 is involved in a Sos1 negative feedback loop whereby cyclic activity of all the proteins in the loop is induced by ERK1/2 and/or p90RSK inhibiting Sos1 activity while also being activated downstream of Sos1 via a Ras-Raf-MEK cascade of activation. Following phosphorylation of Sos1 by ERK1/2 or p90RSK the activity of Sos1 has been shown to decrease, and the activity of other proteins involved in the MAP kinase pathway (including Ras) is also abrogated [43]. Sos1 has also been shown to activate Rac1 when in a complex with Eps8 and Abi1 [35,50,51], and Ras is thought to be involved in activating the Sos1-Eps8-Abi1 complex via PI3K [6,52]. This negative feedback loop may therefore affect the Sos1-Eps8-Abi1 indirectly via Ras as well as by directly downregulating Sos1 activity and availability in the leading edge (as explicitly modelled). In order to confirm the robustness of the model, and ensure that cyclic activity was not introduced as a consequence of the synchronous update algorithm used in our Boolean simulation, we used a random asynchronous updating scheme and confirmed that cyclic activity was observed in all simulations, and importantly RhoA activity remained dominant (S1A Fig). The model predicts that breaking the negative feedback loop activity via MEK1/2 removal leads to persistent (rather than cyclic) activity of the Sos1-Eps8-Abi1 complex which causes subsequent persistent activation of Rac1 and suppression of RhoA activity and thus a reduction in invasive cell migration. This feedback loop prediction was supported by the MEK1/2 effects as outlined above. Moreover, Eps8 knockdown via siRNA desensitises cells to MEK1/2 such that MAP kinase perturbation is no longer able to reduce cell migration speed or invasive capacity. FRET data indicates that these observed migration/invasion effects of Eps8 siRNA are also via leading edge Rac1 and RhoA dynamics, as predicted (Figs 4I–4N and S5B, S4 Movie).
Many studies have proposed possible roles for MEK1/2 in driving cell migration and invasion [53,54] via the Arp2/3 complex [55]. RCP-α5β1/EGFR1 trafficking and signalling promotes invasive cell migration in an Arp2/3-independent manner [24] and is thus a fundamentally different migratory system, yet it is interesting to note that in this context MEK inhibition promotes Rac activation and lamellipodia formation, presumable via the Arp2/3 complex. It is important to note that the effects driven downstream of RCP-mediated trafficking are localised specifically to the leading edge of cells because this is the location of vesicular trafficking [23,24], and hence global effects throughout the cell might not be expected. Previous studies have suggested that in cells harbouring K-RASG13D alleles, which lead to chronic stimulation of downstream signalling, ERK phosphorylates the RhoA GEF GEF-H1 and MEK inhibition in this context increases RhoA activity globally to promote actomyosin contractility and amoeboid migration [56]. In this study, we have employed cells lines that express wild-type K-Ras (A2780 ovarian cancer cells and H1299 non-small cell lung cancer cells [57,58]), and MEK inhibition leads to a localised decrease in RhoA activity, most likely due to a spatially restricted increase in Rac activation at the leading edge (Fig 3E–3I). We have previously shown that RCP-mediated α5β1/EGFR trafficking controls localised signalling [23] and hence it is likely that this apparent discrepancy is due to spatially restricted effects on cell signalling in cells that lack constitutive ERK activation.
Eps8 has been shown to have a role in actin filament capping [59] as well as in Rac1 activation when in complex with Sos1 and Abi1 [50]. Eps8 has been suggested to be an ERK effector, whereby both proteins must cooperate to drive migration via blebs [60]. Another study meanwhile has suggested that Eps8 regulates ERK activity which affects migration of breast cancer cells [61]. We find contrarily that MEK1/2 and Eps8 perturbation have opposite effects, whereby Eps8 siRNA desensitises cells to the anti-invasive effect of MEK1/2 inhibition. We propose that in our network, the main role of Eps8 is via its Rac1 activating capacity as part of the Sos1-Eps8-Abi1 complex as opposed to any effect on actin capping. Furthermore, knockdown of Eps8 has no effect on ERK activation in response to EGF (or MEK inhibitor induced suppression, S7 Fig).
Oscillation is an inherent part of the activation cycle of RhoGTPases at the single molecule level. Others have suggested that cyclic activity of RhoA and Rac1 is important in determining motility [13,14], and our model is also suggestive of this. However, direct evidence which supports pseudo-oscillatory Rac1 and RhoA activity is lacking at present. This is particularly due to current limitations in microscopy methods: i) wavelength constraints and the availability of suitable FRET pair fluorescent proteins dictate that only either Rac1 or RhoA activity, but not both, can be visualised in any single cell; ii) the channels required for FRET ratio calculations decree that GTPase activity can only be sampled at certain discrete timepoints (every 15 seconds in our experiments) making it difficult to reveal oscillations unless exactly in phase. This being said, FRET ratio results suggest that there is some fluctuation in both Rac1 and RhoA activity as cells migrate. Furthermore, there may be some cooperation of Rac1 and RhoA activity at the leading edge of motile cells as neither GTPase is completely ‘switched off’ (with a ratio of 1.0) or ‘switched on’ (with a ratio near 2.0). Therefore the ‘cyclic’ activity as simulated by the Boolean model is plausibly representative of experimental results–as all protein activity is binarised into ON or OFF statuses however, the only way to represent antagonistic proteins’ cooperation would be with a steady oscillation. It has previously been shown that negative circuits can generate cyclic output behaviour [62], and moreover it has been proven that negative circuits in a regulatory graph may be sufficient for such an observation [63]. These mathematical rules are observable in our model findings, whereby the RhoA dominant oscillations of Rac1 and RhoA are caused by the Sos1 negative feedback loop.
Our findings indicate that targeting MEK1/2 using specific drugs (such as Selumetinib—AZD6244 and PD184352 used here) may abrogate the harmful invasive migration and metastasis in certain contexts, particular for patients with tumours that harbour gain-of-function p53 mutation or which exhibit an upregulation in α5β1 expression. Targeting of the MAP kinase pathway has been ongoing in various clinical trials in thyroid, melanoma, ovarian and lung tumours due to the perceived effect that this greatly reduces tumour growth via decreased proliferation and increased apoptosis [64]. We propose MAP kinase targeting may additionally reduce metastasis to further benefit treatment of certain tumours.
Harnessing Boolean logic is a simple and effective method to study systems in an unbiased manner without the need for any parameters to be found or estimated. EGFR signalling activates a myriad of downstream effectors [65]. Moreover, the regulation of RhoA and Rac1 is a highly involved process involving the balance of GEFs, GAPs and their own regulators. A modelling approach based on known priors was consequently ideal to uncover further details regarding this pro-invasive signalling network. In particular, in silico workflows like those employed here could predict plausible therapeutic targets as well as derive further mechanistic insight. Such a modelling approach could be extended and adapted to pose specific questions regarding other growth factor signalling effects such as regarding proliferation or differentiation; alternatively, different GTPase networks or migratory events could also be recreated with Boolean logic.
Studying the signalling network in an integrated and connected manner as opposed to abstracting down to the minutia details of every molecular interaction permitted us to test complete unknowns with novel and unexpected experiments, including the Rac1 activator/inhibitor hierarchy and determining feedback via the Sos1-Eps8-Ab1 complex. With the advent of increasingly effective imaging techniques and ‘omics’ type large dataset derivation, more is understood in vitro about the complex signalling events at play in cancer cells. Mathematical approaches can complement the more traditional lab-based techniques, in particular to collate and rationalise wide-ranging data as well as to make novel and testable predictions to expand our present knowledge.
The Boolean model shown in Fig 1A was manually curated where all proteins and interactions were incorporated as follows. All proteins and interactions found to contribute to integrin-driven cell migration in [23] were first included in the model. Then several (>10) previously published Epidermal Growth Factor Receptor (EGFR) pathway maps were mined to find all proteins involved in interactions in pathways between (i) EGFR and Akt; and (ii) EGFR and Rac1/Cdc42 (note RhoA was not observed in any EGFR pathways), as EGFR, Akt and the Rho GTPases were key proteins implicated in [23]. In particular, the comprehensive pathway map produced by [65], which was subsequently adapted and formulated into a large scale logical model by [32], and the NetPath map by [49], contributed many proteins and interactions into the model. The model was then ‘cleaned’ by independent verification or rejection of proteins/interactions from studying the original papers cited in [32,49,65], rejecting interactions which give implausible output data following simulations, and omitting some proteins whose functions in the EGFR pathway are ambiguous and superfluous to model outputs. Reactions involving RhoA which were not included in [65] (such as activation by the known Rho/Rac GEF Vav2) were included based on literature evidence [31] and the necessity of full incorporation of RhoA.
In the signalling network (Fig 1A), each node corresponds to a species or model variable (e.g. proteins in this case), and each directed edge corresponds to a different directed interaction between the nodes from and to which the edge is joining. Each node in the set of all nodes N is either in an active “ON” state or an inactive “OFF” state, denoted 1 and 0 respectively in a binary domain ∀ n ∊ N; t ∊ Z+: n(t) = 0 or 1. All reactions were simplified into activation or inhibition, and built using the 3 Boolean operators AND, OR, and NOT which are sufficient to represent any logical relationship. All reactions were assumed to take equal time, t = 1, then given the set of initial conditions in which at least one node is initially active, ∃ n ∊ N: n(0) = 1, nodes became active (transition from 0 → 1), inactive (1 → 0), or remained the same (0 →0 or 1→1) after each time increment depending on the state of upstream nodes and the interactions which the nodes were affected by. Where there was some ambiguity as to whether a reaction should involve “AND” or “OR” gates (particularly involving proteins found in [49] and not [32]) cited literature was studied to see if there was an explicit mention of two or more proteins acting together to affect a downstream node (in which case and “AND” gate was used) or not (in which case an “OR” gate was used).
All Boolean simulations were performed in CellNetAnalyzer [66], a Matlab plugin. Reactions were coded according to the network graph in Fig 1A (except for reactions involving activation of Rac1 in which the role of RacGAP1 was important, see S3 Table) and then synchronous updates were performed for the first 50 time increments for all figures in the main text, giving rise to the full ON/OFF binary heat maps for all variables (Fig 1B and 1C). Where explicitly stated (S1A and S1B Fig discussion), random asynchronous updates were instead performed for the first 2000 time increments, whereby in any time increment at most one executable reaction will occur at random according to a uniform distribution. In silico knockouts and subsequent node classification was performed by removing all edges which lead into a single node and observing the effects on the output variables RhoA and Rac1. In silico knockouts were performed individually for all nodes in the model.
A2780 cells were cultured in RPM1 1640, H1299 (expressing control vector or mutant p53-273H—[21]) and TIF cells were cultured in DMEM supplemented with 10% FCS and grown at 37°C and 5% CO2. Transient transfections of Raichu-RhoA and Raichu-Rac1 constructs and siRNA knockdowns of Eps8 were performed using the nucleofector (Solution T; 3 μg plasmid DNA or 1 μM siRNA; program A-23; Amaxa; Lonza), according to the manufacturer’s instructions. Both biosensor expression and Eps8 knockdown were obtained using one round of nucleofection, and Raichu-Rac1/RhoA and Eps8 siRNA were combined in the same nucleofection where required. All experiments were performed ~24 hours after nucleofection.
MEK inhibitors used were PD184352 [44] and AZD6244 [45], used at 1 μM in all experiments alongside DMSO as a control vehicle. Knockdown of Eps8 was performed using siRNA pools from an siARRAY (Dharmacon) and validated using two independent individual siRNAs: Eps8-A: 5’ GCGAGAGTCTATAGCCAAA; and Eps8-B: 5’ GCCAACTTCTAATCGCCATAT in comparison with a control siRNA. Biosensors used for FRET microscopy were kindly provided by Prof. M. Matsuda: Raichu-1011X (Rac1) [67] and Raichu-1237X (RhoA) [68]. Rabbit anti–phospho-p44/42 MAPK (ERK1/2) (Thr202/Tyr204) was purchased from Cell Signalling Technology, Rabbit anti–ERK2 (C-14) and Rabbit anti-Eps8 were purchased from Santa Cruz Biotechnology and Mouse anti–Eps8 was purchased from BD Biosciences. cRGDfV was purchased from BACHEM. Alexa Fluor 488 phalloidin used for actin staining was purchased from Life Technologies.
Cells were treated exactly as in the conditions used for imaging experiments before being lysed in non-denaturing lysis buffer (200 mM NaCl, 75 mM Tris-HCl, pH 7.4, 15 mM NaF, 1.5 mM Na3VO4, 7.5 mM EDTA, 7.5 mM EGTA, 1.5% (v/v) Triton X-100, 0.75% (v/v) NP-40, 50 μg/ml leupeptin, 50 μg/ml aprotinin, and 1 mM 4-(2-aminoethyl)-benzenesulfonyl fluoride). Lysates were clarified by centrifugation at 10,000 g for 10 min at 4°C. Cell lysates were resolved under denaturing conditions by SDS-PAGE (4–12% Bis-Tris gels; Invitrogen) and transferred to nitrocellulose membrane. Membranes were blocked with 1x Blocking Buffer (Sigma) and incubated overnight at 4°C with the appropriate primary antibody in 5% BSA and then at room temperature for 1 h with the appropriate fluorophore-conjugated secondary antibody in 1x Blocking Buffer. Membranes were scanned using an infrared imaging system (Odyssey; LI-COR Biosciences). For quantification of western blots, mean intensity of each relevant band was measured using ImageJ. Loading was normalised to an appropriate loading control and backgrounds were subtracted. All conditions were normalised to the band with the highest intensity in each repeat, and then the mean normalised intensity was calculated across three independent repeats. Quantification of western blots was performed where explicitly stated.
Cells were grown to confluence and after 24 hours were scratched with a pipette tip before being washed once in appropriate serum containing medium and placed into fresh serum containing medium. Cells were treated with MEK inhibitors or DMSO at 1μM and cRGDfV at 2.5 μM as appropriate. Images were taken every 10 minutes for >84 time points (15 hours) for 6 different areas per condition. Cells were imaged using an AS MDW live cell imaging system (Leica) using a 20x/NA 0.50 Plan Fluotar Ph2 objective in brightfield. Micromanager imaging software and point visiting mode were used to allow multiple positions to be imaged within the same timecourse while cells were maintained at 37°C and 5% CO2. Images were collected using a Cascade II EM CCD camera (Photometrics). At least 5 cells per time-lapse position (giving 30 cells tracked per condition) were individually manually tracked using the ImageJ plugin MTrackJ [69] for the nucleus position every 3 frames (i.e. using 30 minute timepoint intervals). Cells were unbiasedly chosen and in general either started or ended at the leading edge of the scratch wound. The Chemotaxis and Migration Tool [70] was used to calculate the average speed and directional persistence for each given condition, where persistence is the ratio (displacement from initial position)/(total distance travelled).
Inverted invasion assays were performed based on the protocol as described previously [71]. Collagen I (final concentration ∼5 μg/ml; BD Biosciences) supplemented with 25 μg/ml fibronectin was allowed to polymerise in inserts (Transwell; Corning) for 1 h at 37°C. Transwells were then inverted for cells to be seeded directly onto the underside at a concentration of 5 x105 cells/ml for 2~4 hours at 37°C. Transwells were then re-inverted, washed twice in serum-free medium and placed in 0.1% serum medium treated with 2.5 μM cRGDfV (for A2780s). Medium supplemented with 10% FCS and 30 ng/ml EGF, treated with 2.5 μM cRGDfV was placed on top of the matrix, thereby providing a chemotactic gradient for invasion. After ~24 hours, the medium below the matrix was treated with 1 μM MEK inhibitor or DMSO as required, while the medium above the matrix was treated with 2 μM MEK inhibitor or DMSO. After a further ~24 hours, all cells were stained with Calcein-AM ~1 hour prior to imaging and visualised by confocal microscopy with serial optical sections being captured at 14.97-μm intervals using an inverted confocal microscope (TCS SP5 AOBS; Leica) using a 20× objective. Invasion was quantified using the area calculator plugin in ImageJ, where the invasive proportion was obtained by measuring the fluorescence intensity of cells invading > 45 μm for A2780s (the 5th Z-stack onwards) and > 30 μm for H1299-273s (the 4th Z-stack onwards), and dividing this by the total fluorescence intensity in all Z-stack images.
All RhoA and Rac1 activity was determined by Forster resonance energy transfer (FRET) microscopy. Cells were transfected with RhoA or Rac1 probes and then seeded onto CDMs after ~18 hours at 5x104 cells/ml and left for a further ~4 hours to spread appropriately for 3-D imaging. All growth medium was replaced by Ham’s F12 supplemented with 10% FCS and 25 mM Hepes buffer prior to imaging to achieve optimal visualising conditions and treated with 2.5 μM cRGDfV, 1 μM PD184352 or DMSO as required 1–2 hours pre-capture. Bio-sensor expressing cells were imaged every 15 seconds for the emission bands CFP (445nm, donor and acceptor), YFP (515nm donor and acceptor) and FRET (445nm CFP donor, 515nm YFP acceptor) with exposure times as follows in order: CFP– 800 ms, YFP– 400 ms, Fret– 800 ms for cells on CDMs. Images were acquired using a CSU-X1 spinning disc confocal (Yokagowa) on a Zeiss Axio-Observer Z1 microscope with a 63x/1.40 Plan-Apochromat objective, Evolve EMCCD camera (Photometrics) and motorised XYZ stage (ASI). The 445nm and 561nm lasers were controlled using an AOTF through the laserstack (Intelligent Imaging Innovations (3i)) allowing both rapid ‘shuttering’ of the laser and attenuation of the laser power. Images were captured using SlideBook 6.0 software (3i).
Using custom software written in Python and NumPy, the following semi-automated method was devised to quantify RhoA/Rac1 activity using ratio imaging (CFP donor—YFP acceptor channel over CFP donor—CFP acceptor channel) in a ring crescent region at the front of the cell: First and for each time points in the image stack, the channels were aligned to a 100th of a pixel by cross-correlation subpixel image registration. As the channel with the best signal to noise ratio, the YFP channel was used to create binary masks: the YFP donor—YFP acceptor Raichu probe stained images were then individually band-pass filtered (A trous wavelet, linear 3x3 filter, keeping scales 2–8) to remove both high frequency noise and stationary background. Images were thresholded using a fixed threshold (pixel grey value above 2000) and all the objects in the resulting binary image were identified by 8-connected component labelling. All but the largest object in the image were discarded, thus defining a time coded “cell mask” object for each image stack. Morphological operations (erosion, subtraction from the mask) were then combined to create a “ring mask” object, 40 pixels wide.
The axis aligned minimum bounding box of the “ring mask” objects was calculated and divided into four equal parts along the longer axis. The two extremal regions define two crescents of interest (front of the cell, back of the cell). Then at each subsequent time point, the newly calculated extremal regions were aligned to those found at the first time point. However manual user intervention was required to identify the actual front of the cell.
In order to calculate mean pixel ratios in this “front cell crescent” region of interest, the CFP donor—YFP acceptor and CFP donor—CFP acceptor images were smoothed using Gaussian Blur with a standard deviation of 1.0 pixel. This typically reduces wide pixel-to-pixel variations in the ratio image, with little effect on mean ratio values over larger areas. Only ratio image pixels with locations within the “front cell crescent” were calculated and averaged. We further avoided divisions by zero by ignoring pixels where the CFP donor—CFP acceptor values equal zero. Images shown are of the CFP-YFP/CFP-CFP ratio in the whole cell with a custom look-up table (LUT) set between values of 0.0 and 2.0.
For quantification, the leading edge FRET ratio was calculated as the average FRET ratio in the aforementioned masked area for each timepoint of every 20 frame, 5 minute movies. For comparison between conditions, the average of n = 15–20 cells across three independent repeats of the average FRET across every timepoint was chosen to give a single comparable Rac1/RhoA readout (S6 Fig).
A2780 cells were seeded on CDMs at 5x104 cells/ml and allowed to spread for >4 hours. Cells were then treated with cRGDfV/DMSO/PD184352 and left for a further >1 hour before being fixed in 4% paraformaldehyde and permeabilised with PBS + 0.2% Triton X. Cells were stained with Alexa Fluor 488 phalloidin over > 16 hours before being mounted in Prolong Gold antifade reagent.
Cells were imaged using a Leica TCS SP8 STED 3X microscope with an HC PL APO 100x/1.40 oil objective and further 3.0x or 4.0x confocal zoom. A pinhole with 0.7 Airy units was used to further improve resolution. Images were collected using a HyD detector with acceptance spectrum 504–595nm, 499nm laser line for emission. For whole cell images, one Z-plane was captured and images were not further deconvolved. For leading edge images, 14 Z-slice images were captured covering a total of Z distance of 2.99 μm. Deconvolution was then performed using Huygens Professional software with default settings. Images shown are maximum projections of deconvolved Z-stacks.
All quantifaction of filopodia was performed on deconvolved maximum projections of leading edge regions. Cells were cropped by a common sized rectangular ROI across different conditions and individually manually thresholded using ImageJ software such that pixels inside the cell (i.e. regions of high actin signal) were assigned a value >1 and regions outside the cell (background) were assigned a value of 0. Thresholded images were analysed using the Matlab application CellGeo [72] with a critical length = 20.0, critical width = 12.0 to automatically calculate number of filopodia per leading edge and average length of filopodia.
|
10.1371/journal.ppat.1004291 | Plasmacytoid Dendritic Cells Suppress HIV-1 Replication but Contribute to HIV-1 Induced Immunopathogenesis in Humanized Mice | The role of plasmacytoid dendritic cells (pDC) in human immunodeficiency virus type 1 (HIV-1) infection and pathogenesis remains unclear. HIV-1 infection in the humanized mouse model leads to persistent HIV-1 infection and immunopathogenesis, including type I interferons (IFN-I) induction, immune-activation and depletion of human leukocytes, including CD4 T cells. We developed a monoclonal antibody that specifically depletes human pDC in all lymphoid organs in humanized mice. When pDC were depleted prior to HIV-1 infection, the induction of IFN-I and interferon-stimulated genes (ISGs) were abolished during acute HIV-1 infection with either a highly pathogenic CCR5/CXCR4-dual tropic HIV-1 or a standard CCR5-tropic HIV-1 isolate. Consistent with the anti-viral role of IFN-I, HIV-1 replication was significantly up-regulated in pDC-depleted mice. Interestingly, the cell death induced by the highly pathogenic HIV-1 isolate was severely reduced in pDC-depleted mice. During chronic HIV-1 infection, depletion of pDC also severely reduced the induction of IFN-I and ISGs, associated with elevated HIV-1 replication. Surprisingly, HIV-1 induced depletion of human immune cells including T cells in lymphoid organs, but not the blood, was reduced in spite of the increased viral replication. The increased cell number in lymphoid organs was associated with a reduced level of HIV-induced cell death in human leukocytes including CD4 T cells. We conclude that pDC play opposing roles in suppressing HIV-1 replication and in promoting HIV-1 induced immunopathogenesis. These findings suggest that pDC-depletion and IFN-I blockade will provide novel strategies for treating those HIV-1 immune non-responsive patients with persistent immune activation despite effective anti-retrovirus treatment.
| Persistent expression of IFN-I is correlated with disease progression in HIV-1 infected humans or SIV-infected monkeys. Thus, persistent pDC activation has been implicated in contributing to AIDS pathogenesis. To define the role of pDC in HIV-1 infection and immunopathogenesis in vivo, we developed a monoclonal antibody that specifically and efficiently depletes human pDC in all lymphoid organs in humanized mice. We discover that pDC are the critical IFN-I producer cells in response to acute HIV-1 infection, because depletion of pDC completely abolished induction of IFN-I or ISG by HIV-1 infection, correlated with elevated level of HIV-1 replication. When pDC were depleted during chronic HIV-1 infection in humanized mice, pDC were still the major IFN-I producing cells in vivo, which contributed to HIV-1 suppression. Despite of higher level of viral replication in pDC-depleted mice, we found that HIV-induced depletion of human T cells and leukocytes was significantly reduced in lymphoid organs, correlated with reduced cell death induction by HIV-1 infection. Our findings demonstrate that pDC play two opposing roles in HIV-1 pathogenesis: they produce IFN-I to suppress HIV-1 replication and induce death of human immune cells to contribute to HIV-induced T cell depletion and immunopathogenesis.
| Chronic immune activation induced by HIV-1 infection is highly correlated with CD4 T cell depletion and immunodeficiency [1], [2], [3]. The level of T cell activation (HLA-DR+CD38+CD8+ T cells) is correlated with disease progression independent of HIV-1 viral load and CD4+ T cell count [4]. It is also proposed that immune activation drives AIDS development in simian immunodeficiency virus (SIV) infected monkeys. In SIV-infected Asian monkeys (Rhesus macaques and pigtail macaques, e.g.) AIDS develops, associated with persistent immune activation and rapid CD4+ T-cell loss. In contrast, SIV infection of African monkeys (African Green monkeys and sooty mangabeys, e.g.) leads to no AIDS progression, correlated with only a transient and self-limiting immune activation despite similar levels of viral replication as pathogenic SIV infections [2], [5], [6]. In mice, repeated treatments with Toll like receptor (TLR)-9 [7] or TLR7 [8] ligands lead to AIDS-like immune dysregulation, correlated with immune activation and lymphoid organ destruction. In SIV-infected African green monkeys, treatment with lipopolysaccharide (LPS) results in CD4+ T-cell loss [9]. Finally, anti-inflammatory treatment with chloroquine [10] or hydroxychloroquine in combination with antivirals [11] inhibits immune activation in HIV-1 infected patients, correlated with elevated CD4+ T cells [11].
The mechanism by which HIV-1 infection leads to immune activation is not fully elucidated [2]. Several mechanisms have been proposed, including loss of gut tissue integrity and microbial products translocation [12] or persistent production of IFN-I [13], [14]. Sustained IFN-I production is correlated with HIV-1 induced immune activation and disease progression both in HIV-1 infected patients [15] and pathogenic SIV infected monkey models [16], [17], [18]. Although IFN-I inhibits HIV-1 replication in vitro [19], the high level IFN-I in HIV-1 patients is not correlated with viral control but is predictive of HIV-1 disease progression [20], [21]. IFN-I is induced during acute phase of SIV infection in both pathogenic and non-pathogenic hosts. However, the IFN-I induction is controlled during nonpathogenic persistent SIV infection, while the pathogenic SIV infection is featured by sustained IFN-I production during chronic infection, correlated with immune activation and AIDS development [17], [22], [23] [16].
Plasmacytoid dendritic cells (pDC) are the major IFN-I producing cells [24]. They preferentially express TLR7 and TLR9 in the endosome, sensing viral RNA and DNA respectively during infection. Upon viral infections and other stimulations, pDCs produce large amount of IFN-I and inflammatory cytokines. However, it is still not clear if pDCs are the major source of IFN-I during acute or chronic HIV-1 infection [25], and the role of pDC in HIV-1 replication or disease progression is not well defined. HIV-1 infection can stimulate pDCs to express TNF-Related Apoptosis-Inducing Ligand (TRAIL) [26], [27], [28]. However, the induction of CD4+ T-cell death by TRAIL expressing pDC remains controversial [29]. On the other hand, pDC are also reported to be decreased and functionally impaired in peripheral blood of HIV-1 infected individuals [30], [31], [32], [33]. The decline of IFN-I producing capability of pDC is correlated with opportunistic infection but not CD4 T-cell counts [31], [34], [35]. These reports highlight that pDC may play important but complex roles in HIV-1 infection and immunopathogenesis.
Humanized mice transplanted with human immune tissues or cells have been developed to study HIV-1 infection [36]. In the recent improved humanized mouse models, HIV-1 infection can be established by inoculating through intraperitoneal [37], [38], intravenous [39], or mucosal routes [38], [40]. HIV-1 infection results in persistent viral replication, CD4 depletion in peripheral blood and lymphoid organs. Importantly, HIV-1 infection results in T cell depletion, correlated with immune activation in lymphoid organs of humanized mice [41].
We and other groups have reported that functional human pDC are developed in lymphoid tissues in humanized mouse models [41], [42], [43]. Human pDC are rapidly activated by HIV-1 infection [41] and the level of pDC activation is reversely correlated with CD4+ T-cell numbers [41], which is consistent with the observation from HIV-1 infected patients [15], [20], [21] and SIV infected monkeys [16], [23]. To define the role of human pDC in HIV-1 replication and immunopathogenesis in vivo, we developed a monoclonal antibody that specifically and efficiently depletes human pDC in all lymphoid organs in humanized mice in vivo. Thus we were able to characterize the role of human pDC in HIV-1 infection and immunopathogenesis during acute and chronic phases of HIV-1 infection.
We have previously reported that human pDC are rapidly activated by HIV-1 infection [41] in humanized mice and the level of pDC activation is correlated with CD4+ T-cell depletion [41]. The pathogenic CCR5/CXCR4 dual-tropic HIV-R3A strain efficiently established infection in humanized mice (Figure 1A), associated with IFN-I induction and ISG expression (Figure 1B&C), increased HLA-DR+CD38+ CD8 T cells (Figure 1D), and CD4 T-cell depletion (Figure 1E). As in HIV-1 infected human patients, a decrease of total number of human leukocytes was induced by HIV-1 infection in humanized mice, as measured by cell numbers of human CD4, CD8 T cells and total CD45+ leukocytes in lymphoid organs (Figure 1F). As shown with chronic infection with JRCSF below, similar IFN-I induction, immune activation and depletion of human immune cells were observed in humanized mice (Figure S1A–E). We have reported that pDC frequency did not change during acute HIV-R3A infection [41]. During persistent JR-CSF infection, pDC percentage was also not significantly altered (data not shown). Thus, the humanized mouse model provides a relevant in vivo model for studying the role of pDC in HIV-1 infection and immunopathogenesis.
In order to delineate the role of human pDC in HIV-1 infection and pathogenesis in vivo, we developed and screened a number of pDC-reactive monoclonal antibodies (mAb) and identified an anti-BDCA2 (CD303) mAb (15B), which could specifically deplete human pDC in lymphoid organs in humanized mice. After 15B injection, pDC were specifically depleted in both peripheral blood (Figure 2A) and lymphoid organs (Figure 2B). Importantly, human T, B, myeloid dendritic cells and monocytes\macrophages were not perturbed by 15B mAb (Figure 2C–E and Figure S2&S3).
To test the role of pDC in early acute HIV-1 infection, we injected 15B and isotype control antibody into humanized mice on -5, -3 and -1 days before infection, and then infected them with HIV-R3A (a highly pathogenic dual-tropic HIV-1 strain, [44], [45]) on day 0. The infected mice were treated with 15B or control antibody two more times on 3 and 6 days post-infection (dpi). We found that pDC remained depleted in blood and lymphoid organs of the infected mice (Figure 3A), when terminated on 8 dpi. Interestingly, the induction of plasma IFN-I was completely blocked by pDC depletion in HIV-1 infected mice (Figure 3B). The suppressed expression of different subtypes of human IFN-I was also confirmed at RNA level by real time PCR (Figure 3C). In addition, the up-regulation of ISGs such as Mx1 and TRIM22 was also blocked (Figure 3D and data not shown). We confirmed similar blocking of IFN-I induction by pDC-depletion prior to infection with the CCR5-tropic JRCSF HIV-1 strain (data not shown). These data demonstrate that pDC are the critical IFN-I producing cells during early HIV-1 infection in humanized mice in vivo.
Consistent with the antiviral activity of IFN-I, HIV-1 replication reached higher levels in pDC-depleted mice in vivo (Figure 4). The average plasma viral load was increased about 10-fold comparing with mice treated with isotype control antibody (Figure 4A, p<0.01). We repeated the pDC depletion experiment with the CCR5 tropic HIV-1 JR-CSF. Similar to HIV-R3A infection, JR-CSF replication was increased in pDC-depleted mice (about 5-fold, Figure 4B). The increase of viral replication was further confirmed in the spleen by immunohistochemistry (Figure 4C) or flow cytometry (Figure 4D&E) of HIV p24 protein positive cells. Therefore, pDC are the critical IFN-I producer cells in response to acute HIV-1 infection, and they are required to significantly inhibit early HIV-1 replication. The expression of HLA-DR and CD38 was increased in mice with pDC depletion (Figure S4), correlated with increased level of HIV-1 replication. These data suggest that the elevated level of HIV-1 replication in the absence of pDC can lead to upregulation of T-cell activation markers.
As shown in Figure 1, HIV-R3A infection leads to rapid immunopathogenesis including depletion of human total leukocytes and CD8 T cells, as well as CD4 T cells. Despite the increased HIV-R3A viral replication in pDC-depleted mice, the absolute numbers of human CD4+ T cells in blood and spleen were comparable to those of control mice (Figure 5A). More surprisingly, CD8+ T cells as well as total human CD45+ leukocytes were partly preserved in blood and spleen in pDC-depleted mice (Figure 5B&C). Consistently, this is correlated with decreased levels of cell death of CD45 or CD8 T cells (Figure 5D&E). The similar cell death induction of CD4 T cells in both groups of mice may be due to the highly fusogenic activity of the HIV-R3A Env [45], [46]. The elevated HIV-R3A replication in pDC-depleted mice, combined with its highly pathogenic direct killing activity, may contribute to the observed CD4 T cell death induction and depletion.
During early phase of JR-CSF infection, the decrease of CD4+ T cells and other human leukocytes was not significant (Figure S1). To define the role of pDC in HIV-1 replication and immunopathogenesis during chronic HIV-1 infection, we performed pDC depletion during chronic HIV-JRCSF infection. Humanized mice were infected with JR-CSF for 11 weeks, and then 15B was applied to deplete pDC for additional 10 weeks. In agreement with the data from acute infection, we observed a significant increase in plasma viremia (Figure 6A). The percentage of HIV-1 infected cells (HIV-1 p24 positive) was also significantly increased (Figure 6B&C). Interestingly, plasma IFN-α2 decreased significantly in the pDC-depleted mice (by 70%, Figure 6D). The mRNA induction of different IFN-I subtypes (Figure 6E) and ISGs (Figure 6F) in human leukocytes in spleens was almost completely suppressed by real-time PCR and cDNA array (Figure S5). Thus, pDC are still a major source of IFN-I, and contribute to suppressing HIV-1 chronic infection in humanized mice.
In spite of the persistently higher viremia during 10 weeks of pDC depletion treatment, human CD4+ T cell numbers increased significantly in the spleen, comparing to the control group (Figure 7A, p<0.05). In addition, human CD8 T cells and CD45+ leukocyte numbers in spleens were also increased (Figure 7B&C, p<0.01). Interestingly, the relative depletion of human CD4, CD8 T cells and CD45 leukocytes was the same in the blood (Figure 7A–C). The increase of human CD45+ cells in the spleen was also confirmed by immunohistochemistry staining of spleen sections (Figure 7D). Accordingly, pDC depletion significantly reduced the percentage of dying cells in T cells and total human CD45+ cells in the spleen and other lymphoid organs (Figure 7E&F and data not shown). Therefore, persistent activation of pDC in lymphoid organs during HIV-1 chronic infection, although still producing IFN-I to suppress HIV-1 replication, contributes significantly to HIV-1 induced depletion of human leukocytes including human CD4 T cells.
HIV-1 infection induces a systemic immune activation [1], [47], which has been proposed to contribute to HIV-1 disease progression [1], [3], [4], [47], [48], [49]. Persistent activation of pDC and IFN-I have been correlated with HIV-1 infection induced immune activation both in HIV-infected patients and in SIV-infected rhesus monkeys [15], [20], [21],[16], [18], [23], [49], [50], [51]. The role of pDC in HIV-1 infection and immunopathogenesis, however, remains unclear [52], [53], [54]. Using the humanized mouse model of HIV-1 infection and pathogenesis in vivo, we developed a novel pDC-specific mAb that specifically and efficiently depletes human pDC in various lymphoid organs in vivo. We report here that, in response to acute HIV-1 infection, pDC are the critical IFN-I producer cells and contribute to suppressing HIV-1 replication during early HIV-1 infection. During chronic HIV-1 infection, pDC are still the major IFN-I producing cells and contribute to controlling HIV-1 replication. Most surprisingly, depletion of pDC during chronic HIV-1 infection rescued human leukocytes including human CD4 T cells in lymphoid organs but not in blood, in spite of higher levels of HIV-1 replication. We conclude that pDC play two opposing roles during HIV-1 infection and pathogenesis: they produce IFN-I to inhibit HIV-1 replication, but enhance HIV-1 pathogenesis by promoting cell death of human leukocytes including human CD4 and CD8 T cells.
Study of human pDC has been hampered by the difficulty of isolating sufficient number of human pDC cells and our inability to culture or expand pDC in vitro. Using humanized mice, we screened a number of human pDC-specific mAb and identified the 15B mAb that specifically and efficiently depletes human pDC in all lymphoid organs in vivo. With this mAb, we were able to define the role of pDC in HIV-1 replication and immunopathogenesis during different phases of HIV-1 infection. Our findings show that pDC are the critical or only IFN-I producer cells in response to acute HIV-1 infection in humanized mice. The pDC-independent production of IFN-α in plasma and ISG expression after pDC depletion during chronic HIV-1 infection may be due to the contribution of other cell types such as mDC and macrophages [55]. A recent report showed that TLR7 and TLR9 blockade had minimal impact on plasma IFN-α and expression of ISG in SIV-infected monkeys and did not alter viral load and T cell activation in vivo [56]. The discrepancy may be due to an incomplete blocking of pDC activation by the TLR7 and TLR9 antagonist in vivo or different contribution of other IFN-I producing cells in SIV-infected monkeys [56]. A recent report shows that pDC are the major IFN-I producing cells during primary SIV infection but the pDC pool is replenished by their precursors lacking IFN-I production capacity post acute phase infection [57]. However, we found that the depletion of pDC abolished or dramatically reduced IFN-I production during acute or chronic HIV-1 infection, suggesting that pDC are the major IFN-I producing cells during both acute and chronic HIV infection in humanized mice. The 15B mAb does not bind the BDCA2 receptor on monkey pDC (data not shown). The contribution of pDC in SIV-infected monkeys needs to be reexamined by pDC depletion when the appropriate depleting antibody is available.
The gut associated lymphoid tissue (GALT) is a major site of human T-cell depletion by HIV-1 infection. It is important to point out that the GALT in humanized mouse models are not structurally developed, with impaired intestinal lymphoid organ structure and human cell engulfment [58]. We thus can not examine the effect of pDC depletion in GALT in the current model. It will be interesting to examine the effect of pDC depletion on SIV pathogenesis in the GALT of SIV-infected monkeys, which recapitulates GALT pathogenesis as in HIV-infected humans.
HIV-1 disease progression is associated with gradual depletion of human leukocytes including other lineages as well as CD4 T cells [59], [60], [61]. Multiple mechanisms have been reported to account for the pathogenic activity of HIV-1 infection. Besides direct killing of HIV-1 infected CD4 cells, bystander cells including hematopoietic progenitor cells and uninfected human mature cells are also depleted, associated with immune hyperactivation during pathogenic HIV-1 infection. Several recent reports suggest that pDC may contribute to HIV-1 induced immune activation and subsequent immunopathogenesis. Repeated administrations of TLR7 ligands in mice induce AIDS-like lymphopenia, with reduced CD4+ T cells, CD8+ T cells and B cells [8]. It is reported that IFN-I triggers proapoptotic and antiproliferative effect on T cells [62].
We found that pDC-depletion not only rescued CD4 T cells but also CD8 T cells and total CD45+ leukocytes in lymphoid organs such as spleen and LN, but not in the blood. Therefore, persistent activation of pDC by HIV-1 infection in lymphoid organs contributes to HIV-1 immunopathogenesis by accelerating the death of all human leukocyte cells. Similarly, recent findings showed that treating SIV-infected rhesus macaques with the TLR7 and TLR9 antagonist DV056 led to a significant increase in levels of proliferating memory CD4 and CD8 T cells in the blood [56]. Interestingly, the anti-malaria drug chloroquine, which inhibits IFN-I production by pDC in vitro [63], appears to rescue human T cells in HIV-1 infected patients, correlated with reduced immune activation [10], [11]. However, two recent reports with similar chloroquine treatment failed to demonstrate the significant beneficial effect in HIV-1 patients [64], [65].
Two recent reports showed that blocking IFN-I signaling during LCMV persistent infection could improve antiviral T cell response and accelerate clearance of chronic LCMV infection via an IL-10-related mechanism [66], [67]. It is not clear if IFN-I also plays a similar critical role in HIV-1 infection and immunopathogenesis. It will be of great interest to test, when the human IFN-IR antagonistic mAb is available, if blocking human IFN-IR can similarly improve anti-HIV immunity and lead to better control of HIV-1 infection in humanized mice or in SIV-infected monkeys. However, pDC depletion in HIV-infected humanized mice showed very distinct outcomes as blocking IFN-I signaling in LCMV-infected mice. First, pDC depletion led to increased HIV-1 replication, and it did not affect IL-10 expression (Li, G. and Su, L., unpublished results). In addition, pDC depletion may remove IFN-I and additional factors expressed by “bad” pDC, including other inflammatory cytokines and cell death ligands such as TRAIL (killer pDC, [27], [28], [68]). Therefore, persistent pDC activation by HIV-1 infection contributes to HIV-1 induced depletion of human immune cells and leads to HIV-1 disease progression. Our findings suggest that depletion of “bad” pDC transiently during HIV-1 chronic infection may provide an effective treatment to preserve human immune cells in HIV-1 infected patients or in those HAART-treated immune non-responder patients [69], [70]. Recently, it was reported that blocking IFN-I with a non-signaling IFNα during acute SIV infection promoted viral replication, which is consistent with our data in the current study. However, they also observed accelerated CD4+ T-cell depletion and AIDS progression [71]. This may be due to the distinct experimental systems between pDC depletion in humanized mice and IFN-I blocking in SIV-infected monkeys. pDC have multiple functions beside IFN-I production, including production of other inflammatory cytokines, direct killing of T cells through TRAIL expression [68] and inhibition of antiviral immune response by Treg induction [18]. Therefore, pDC depletion in SIV-infected monkeys should be performed in future experiments to clarify the role of pDC in SIV infection and pathogenesis in various NHP hosts.
The reports followed NIH research ethics guidelines. For the humanized mouse construction, human fetal liver were obtained from elective or medically indicated termination of pregnancy through a non-profit intermediary working with outpatient clinics (Advanced Bioscience Resources, Alameda, CA). The use of the tissue in research had no influence on the decision regarding termination of the pregnancy. Informed consent of the maternal donor is obtained in all cases, under regulation governing the clinic. We were provided with no information regarding the identity of the patients, nor is this information traceable. The project was reviewed by the University's Office of Human Research Ethics, which has determined that this submission does not constitute human subjects research as defined under federal regulations [45 CFR 46.102 (d or f) and 21 CFR 56.102(c)(e)(l)] and does not require IRB approval. The University of North Carolina at Chapel Hill Institutional Animal Care and Use Committee (IACUC) has reviewed and approved this research. All animal experiments were conducted following NIH guidelines for housing and care of laboratory animals and in accordance with The University of North Carolina at Chapel Hill in accordance with protocols approved by the institution's Institutional Animal Care and Use Committee (IACUC ID: 11-103.0).
Approval for animal work was obtained from the University of North Carolina Institutional Animal Care and Use Committee (IACUC). We constructed Balb/C rag2-gammaC (DKO) mutant DKO-hu HSC or Nod-rag1-gammaC (NRG) NRG-hu HSC mice similarly as previously reported [39]. Briefly, human CD34+ cells were isolated from 16- to 20-week-old fetal liver tissues. Tissues were digested with Liver Digest Medium (Invitrogen, Frederick, MD). The suspension was filtered through a 70 µm cell strainer (BD Falcon, Lincoln Park, NJ) and was centrifuged at 150 g for 5 minutes to isolate mononuclear cells by Ficoll. After selection with the CD34+ magnetic-activated cell sorting (MACS) kit, CD34+ HSCs (0.5 x 106) were injected into the liver of each 2- to 6-days old DKO or NRG mice, which had been previously irradiated at 300 rad. More than 95% of the humanized mice were stably reconstituted with human leukocytes in the blood (10%–90% at 12–14 weeks). Each cohort (Humanized mice reconstituted from the same human donor fetal liver tissue) had similar levels of engraftment. All mice were housed at the University of North Carolina at Chapel Hill.
NL4-R3A, generated by cloning a highly pathogenic dual tropic envelope into NL4-3 backbone [44], [45], [46], was used for acute infection experiment. An R5 tropic strain of HIV-1, JR-CSF, was used for both acute and chronic infection. All viruses were generated by transfection of 293T cells. Humanized mice with stable human leukocyte reconstitution were infected with NL4-R3A (5 ng p24/mouse) or JR-CSF (10 ng p24/mouse), through intravenous injection (i.v.). Humanized mice infected with 293T mock supernatant were used as control groups.
A monoclonal antibody specific to blood dendritic cell antigen-2 (BDCA2), 15B, was used to treat humanized mice through intraperitoneal injection (i.p., 4 mg/kg). For acute HIV-1 infection, humanized mice were injected three times with 15B on -5, -3 and -1 days before infection. For acute R3A infection, mice will be treated on 3 and 6 days post-infection. For acute JR-CSF infection, mice will be treated every three days until termination. For chronic JR-CSF infection, 15B was applied to mice at 11wpi by injecting twice every week for 10 weeks.
For HIV-1 gag p24 staining, cells were stained with surface markers first, and then permeabilized with cytofix/cytoperm buffer (BD Bioscience, cat#554714), followed by intracellular staining. Human leukocytes (mCD45−huCD45+) were analyzed for human CD3, CD4, CD8, CD123, HLA-DR and CD38 by CyAn FACS machine (Dako). FITC-conjugated anti–human HLA-DR (clone:L243, cat#307604), PE-conjugated anti-human CD38 (clone:HIT2, cat#303506), PE/Cy5-conjugated anti-human CD4 (clone:RP4-T4, cat#300510), PE/Cy7-conjugated anti-human CD3 (clone:HIT3a, cat#300316), Pacific blue-conjugated anti-human CD3 (clone:UCHT1, cat#300431), PE/Cy7-conjugated anti-human CD8 (clone:HIT8a, cat#300914), APC-conjugated human CD123 (clone:6H6, cat#306012) and APC/Cy7-conjuaged anti-human CD45 (clone:H130, cat#304014) were purchased from Biolegend; PE-conjugated anti-human caspase-3 (clone:C92-605, cat#51-68655X) was purchased from BD Bioscience. Pacific orange–conjugated anti–mouse CD45 (clone:HI30, cat#MHCD4530), PE/Texas red–conjugated anti–human CD4 (clone:S3.5, cat#MHCD0417) or CD8 (clone:3B5, cat#MHCD0817), and LIVE/DEAD Fixable Aqua Dead Cell Stain Kit (cat#L34957) were purchased from Invitrogen. FITC-conjugated anti-HIV p24 (clone:FH190-1-1, cat#6604665) was purchased from Beckman Coulter. The cells were analyzed on a CyAn ADP (Dako).
Paraffin-embedded spleen sections from humanized mice were stained with the mouse anti–human CD45 (Dako, cat#N1514) or HIV-1 p24 antibody (Dako, cat#M0857), washed in PBS, then incubated with Mouse-&-Rabbit-on-Rodent Double Stain Polymer (BIOCARE MEDICAL, cat#RDS513H) and substrate DAB (BIOCARE MEDICAL, cat#BDB2004 H, L, MM). Images were captured using a QImaging Micropublisher 3.3 CCD digital camera and QCapture software version 3.0 (QImaging, Surrey, BC).
Interferon alpha-1/13 (IFNα1/13), interferon alpha-2 (IFNα2), interferon beta (IFNβ) [72], interferon gamma (IFNγ) [73] and tumor necrosis factor alpha (TNFα) [74] were detected. IFN-I stimulated genes, MxA [75] and TRIM22 [76], were detected to confirm pDCs depletion effect on type I IFN production. Real-time PCR assay was performed (ABI Applied Biosystem). All samples were tested in triplicate using the human GAPDH gene [77] for normalization.
Data were analyzed using GraphPad Prism software version 5.0 (GraphPad software, San Diego, CA, USA). The methods used for analysis of microarray data were described above. The data from different infection groups of mice were compared using a 2-tailed Mann-Whitney U test. For gene expression, mean-ΔCT was calculated as the average (± SD) of all ΔCT values within each group of samples and 2-way ANOVA method was used. Correlations were estimated with a Spearman test. All results were considered significant when p<0.05.
|
10.1371/journal.pgen.1007616 | Artificial selection on storage protein 1 possibly contributes to increase of hatchability during silkworm domestication | Like other domesticates, the efficient utilization of nitrogen resources is also important for the only fully domesticated insect, the silkworm. Deciphering the way in which artificial selection acts on the silkworm genome to improve the utilization of nitrogen resources and to advance human-favored domestication traits, will provide clues from a unique insect model for understanding the general rules of Darwin's evolutionary theory on domestication. Storage proteins (SPs), which belong to a hemocyanin superfamily, basically serve as a source of amino acids and nitrogen during metamorphosis and reproduction in insects. In this study, through blast searching on the silkworm genome and further screening of the artificial selection signature on silkworm SPs, we discovered a candidate domestication gene, i.e., the methionine-rich storage protein 1 (SP1), which is clearly divergent from other storage proteins and exhibits increased expression in the ova of domestic silkworms. Knockout of SP1 via the CRISPR/Cas9 technique resulted in a dramatic decrease in egg hatchability, without obvious impact on egg production, which was similar to the effect in the wild silkworm compared with the domestic type. Larval development and metamorphosis were not affected by SP1 knockout. Comprehensive ova comparative transcriptomes indicated significant higher expression of genes encoding vitellogenin, chorions, and structural components in the extracellular matrix (ECM)-interaction pathway, enzymes in folate biosynthesis, and notably hormone synthesis in the domestic silkworm, compared to both the SP1 mutant and the wild silkworm. Moreover, compared with the wild silkworms, the domestic one also showed generally up-regulated expression of genes enriched in the structural constituent of ribosome and amide, as well as peptide biosynthesis. This study exemplified a novel case in which artificial selection could act directly on nitrogen resource proteins, further affecting egg nutrients and eggshell formation possibly through a hormone signaling mediated regulatory network and the activation of ribosomes, resulting in improved biosynthesis and increased hatchability during domestication. These findings shed new light on both the understanding of artificial selection and silkworm breeding from the perspective of nitrogen and amino acid resources.
| Like other domesticates, nitrogen resources are also important for the only fully domesticated insect, the silkworm. Deciphering the way in which artificial selection acts on the silkworm genome to improve the utilization of nitrogen resources, thereby advancing human-favored domestication traits, will provide clues from a unique insect model for understanding the general rules of Darwin's theory on artificial selection. However, the mechanisms of domestication in the silkworm remain largely unknown. In this study, we focused on one important nitrogen resource, the storage protein (SP). We discovered that the methionine-rich storage protein 1 (SP1), which is divergent from other SPs, is the only target of artificial selection. Based on functional evidence, together with key findings from the comprehensive comparative transcriptome, we propose that artificial selection favored higher expression of SP1 in the domestic silkworm, which would influence the genes or pathways vital for egg development and eggshell formation. Artificial selection also consistently favored activated ribosome activities and improved amide and peptide biosynthesis in the ova, like what they may act in the silk gland to increase silk-cocoon yield. We highlighted a novel case in which artificial selection could directly act on a nitrogen resource protein associated with a human-desired domestication trait.
| The silkworm, Bombyx mori, is the only fully domesticated insect species, originating from its wild ancestor, B. mandarina, approximately 5000 years ago. During the domestication process, the domestic silkworm evolved rapidly under human-preferred selection. Deciphering the way in which artificial selection acts on the silkworm genome to produce human-favored domestication traits will provide clues from a unique insect model for understanding Darwin's theory of artificial selection [1]. Recently, through genome-wide screening of selection signatures in a large batch of domestic and wild silkworms, we identified candidate domestication genes that enriched nitrogen and amino acid metabolism pathways, specifically in glutamate and aspartate metabolism. Knockout of two involved genes resulted in abnormal metamorphosis and decreased cocoon yield [2]. These findings suggest that, like domestic plants and animals, domestic silkworms also tend to have efficient utilization of nitrogen resources to adapt to human-preferences [2–4]. In addition to the glutamate and aspartate metabolism, which is an ammonia re-assimilated system [5], we further wonder whether other kinds of nitrogen supplies are also affected by artificial selection. If this is the case, how have they contributed to silkworm phenotypic changes during domestication?
Insect storage proteins (SPs) are another important resource of amino acids and nitrogen. Specifically, SPs are repositories of stored amino acids that belong to a special conserved arthropod hemocyanin superfamily [6]. Most insects have at least two main types of storage proteins, i.e., arylphorin and methionine-rich storage proteins; some species have other atypical SPs [7]. SPs have been cloned or predicted in many insect species, including Lepidoptera moths and butterflies [8–11]. Insect SPs are believed to serve as a source of amino acids and nitrogen for pupae and adults during metamorphosis and reproduction [12], however there is little solid functional evidence of their biological significance [10, 13]. In plants, storage proteins are mainly reserved in seeds, where, along with other nutrients such as oil and starch, they supply energy for seed germination and growth [14, 15]. Particularly in crops, seed SPs act to provide energy for humans and animals and thus are of great interest and a target for breeding and improvement [14–16].
In the domestic silkworm, previous studies preliminarily characterized the gene and protein expression patterns of four SPs [8, 17–19]. SP1 is female-biased expressed when entering the last instar, and only accumulates in the female pupa [8, 20, 21]. It has been suggested that SP1 contributes to adult female characters and is related to the synthesis of vitellogenin (Vg), the precursor of yolk protein [17]. SP2 couples with SP3 to form a heterohexamer and has inhibitory effects on cell apoptosis [18, 19]. SSP2 was a heat resistant protein and suggested has a cell-protective function [19]. The determination as to whether or not SPs are also important in silkworm domestication, as is the case in domesticated plants, awaits a thorough exploration of their biological and evolutionary significance.
Development of genomics and genome-editing techniques provide tools for efficiently deciphering the evolutionary and functional significance of particular genes [22, 23]. In this study, we conducted a genome-wide identification of the silkworm SPs. Taking advantage of the genomic data resource of a batch of representative domestic and wild silkworms [2], we performed selection signature screening of the silkworm SPs followed by functional verification via the CRISPR/Cas9 knockout system and comprehensive comparative ova transcriptomes of wild-type and mutant silkworms as well as domestic and wild silkworms. Our findings suggest that artificial selection on SP1 contributes to increased egg hatchability during silkworm domestication, possibly by promotion of vitellogenin, influence of hormone synthesis and egg development and eggshell formation. These results provide a novel case with functional evidence for the determination of a regulatory framework on a silkworm domestication gene, revealing that artificial selection acting on the nitrogen and amino acid supply is also required for improved silkworm reproduction.
In total, we identified 8 SPs in the silkworm genome by means of a blast search. Among which SP2 were not annotated in the gene list. Of these, SP1 exhibited the highest methionine content (10.98%) (Table 1). Phylogenetic analysis showed that SP1 was located in one distinct clade, whereas other SPs were in another, indicating an obvious divergence between SP1 and the remaining SPs (Fig 1A). SP1 is located on chromosome 23 while the other SPs are clustered on chromosome 3, suggesting possible tandem duplication events during evolution. Interestingly, by screening artificial selection signatures on the genomic region bearing SP1 and the other SPs respectively, we detected a strong selection signature in the SP1 region of the domestic silkworm (see Material and methods), since there was notably reduced nucleotide diversity in the domestic silkworm group (Fig 1B and 1C). Furthermore, we detected strong differentiation in allelic frequency upstream of SP1 (Fig 1D). Correspondingly, SP1 was differentially expressed in the ova of domestic and wild silkworms, with higher expression in the domestic one (Fig 1E, S1 Fig). We also detected 11 SNPs that caused amino acid changes in the coding sequence of the gene (S2 Fig); the biological significance of these SNPs requires further evaluation. These results suggest that artificial selection acting on SP1 during silkworm domestication may affect the function of this gene in domestic silkworms. At the very least, we can infer that selection may favor higher expression in the domestic silkworm.
To explore the possible phenotypic influence of artificial selection of SP1 acting on the domestic silkworm, we first investigated the biological role of this gene in the silkworm using CRISPR/Cas9 knockout system. For the single guide RNA (sgRNA) design, we selected highly specific targets in the first exon, close to the translation starting site; namely, S1 and S2 (Fig 2A). We chose another site S3 close to the end of the first exon, more than 60 bp downstream from S1 and S2 (Fig 2A and Table 2) to obtain a potentially large fragment deletion by injecting the pool of three gRNAs. After mutation screening of the injected eggs (G0 generation), the gRNAs targeting the above three sites successfully guided DNA editing and generated a variety of mutation types, including 4–9 bp deletions or small insertions followed by a large deletion (Fig 2B).
Through screening of the exuviae of the fifth instar larvae in the G0 cocoons, we successfully identified 26 mosaic mutant G0 moths. We then generated pairwise crosses of those G0 mutants with similar mutant genotypes from the G1 populations. After mutation screening of the G1 eggs, we selected two populations with large deletions for further feeding and mutation screening (see Material and methods). Finally, in the G2 generation we obtained two types of homozygous mutants, i.e., MU1 and MU2 (Fig 2C). In MU1, there was an 8 bp insertion followed by a 63 bp deletion in the SP1 coding sequences. In MU2, there was a 4 bp insertion followed by a 65 bp deletion. The mutations occurred at +29 and +26 bp of the first SP1 exon in MU1 and MU2, respectively (Fig 2C), resulting in reading frame shift mutations and severe premature termination close to the translation starting site, with stop signals at +10 aa and +37 aa of the SP1 protein (Fig 2D).
We selected and maintained the MU1 population for assay on phenotypes related to reproduction and metamorphosis, such as the number of eggs, hatching rate, pupa weight, and cocoon weight. Compared with the wild-type, which exhibited hatching rates of approximately 90%, the hatching rates of the SP1 mutants were dramatically lower, with a mean value of about 40% (Fig 2E), although neither the number of eggs produced nor the whole pupa weight or cocoon shell weight were noticeably affected (Fig 2E). Given that the data were obtained from large replicates (83 replicates for the hatchability assay and 240 replicates for the pupa and cocoon weights), the results are robust. Loss-of-function mutation resulted in significantly decreased expression of SP1 and Vg in the ova, based on the RNA-seq data (Fig 2F). These results suggest that in the silkworm SP1 may positively affect the expression of ova Vg and contribute to silkworm egg development.
Given that knockout of SP1 resulted in a reduced hatching rate (Fig 2E) and that it is female-specific expressed in the pupa and adult stages, we suspect that it plays an important role in ova development, thus contributing to an efficient hatching process. During domestication, artificial selection preferred higher expression of SP1, thus may improve the silkworm hatching rate. As expected, we found that the hatching rate of the domestic silkworm was significantly higher than that of the wild one (Fig 3A). No obvious differences in egg production was detected between wild and domestic silkworms (Fig 3A). The lower hatching rate of the wild silkworm has also been reported in other studies [24, 25]. We further tested expression of Vg in the ova and discovered that consistently, it was significantly higher expressed in the domestic silkworm than in the wild silkworm (Fig 3B, S1 Fig). Promotion of SP1 expression in the domestic silkworm thus results in the corresponding up-regulation of Vg, which further contributes to increased hatchability during silkworm domestication.
In order to further explore the regulation network and possible molecular mechanisms of female-specific SP1 on egg hatchability, we generated comprehensive ova comparative transcriptome analyses between the wild-type and the mutant, as well as the domestic and wild silkworm (Bombyx mandarina), with 4.87~9.15 Gb RNA-seq data for each sample (S1 Table). We chose ova instead of fertilized eggs as target because silkworm SP1 is female-biased expressed when entering the last instar and only accumulated in the female [8, 20, 21]. Comparative transcriptomics in this target tissue would directly focus mechanism of SP1 on female reproductivity and avoid potential influence from the male. In total, there were 561 genes identified as differentially expressed genes (DEGs) in the SP1 knockout mutants (MU1) compared to the wild-type silkworm, with significantly more down-regulated genes (341) than up-regulated (220) (p = 0.0003, Chi-squared test with Yates' continuity correction) (Fig 4A and S2 Table). As expected, we found many more DEGs (2882) between the wild and domestic silkworms, since wild silkworms are much more genetically and phenotypically different from the domestic one, compared with the silkworm mutant from the wild-type. It is interesting that in the 2882 DEGs, there were also significantly more lower expressed genes (1761) than higher expressed (1121) (p = 2.2e-16, Chi-squared test with Yates' continuity correction) (Fig 3A and S3 Table) in the wild silkworm. These results suggest that transcriptome repression in ova might be an output of SP1 depletion in the SP1 mutant (Fig 2F and Fig 4A) and a low expressional level of SP1 in the wild silkworm (Fig 1E; Fig 4A and S1 Fig).
We identified 302 common genes in the two sets of DEGs. KEGG enrichment analysis indicated that these common DEGs were significantly enriched in pathways related to cell proliferation, such as ECM-receptor interaction and folate biosynthesis, as well as the hormone synthesis pathway, which is important in adult ovary development and female production [26, 27] (Fig 4B). Gene ontology (GO) enrichment analysis indicated that the common genes were enriched in reproduction related biological processes, such as chorion-containing eggshell formation (Fig 4B). These genes were also enriched in the molecular functions of the structural constituents of chorion (Fig 4B). In fact, they are all annotated as chorionic proteins, including 3 chorion class CB protein M5H4-like genes (BGIBMGA009720, BGIBMGA009719, BGIBMGA009715) as well as a chorion class B protein PC10 gene (BGIBMGA009721). All these chorion like genes showed significantly higher expression in the domestic silkworm, compared with both the mutant and the wild silkworm (Fig 4B, S2 Table, S3 Table). This pattern was also supported by Real-Time PCR validation in the domestic and wild silkworm (S1 Fig). Genes in ECM-receptor interaction pathway include collagens and integrins (S3 Fig) and those in folate biosynthesis include folylpolyglutamate synthase, which involves in 7,8-Dihydrofolate (DHF) and 5,6,7,8-Tetrahydrofolate (THF), substrates for subsequent one carbon pool mediated by folate (S4 Fig). The enriched hormone synthesis pathway includes genes functioning in both juvenile and molting hormones (S5 Fig). Extend to all the enriched genes, it is notable that most of these enriched genes were relatively highly expressed in the domestic silkworm, compared with SP1 mutant and the wild silkworm (Fig 4B).
We further generated enrichment analyses on DEGs on the two sets of DEGs independently and observed consistent pattern (Tables 3 and 4). Functional enrichment analysis of DGEs between wild-type silkworm and SP1 mutant silkworm revealed a significantly enriched the KEGG pathway “ECM-receptor interaction” as well as other three pathways with marginal significances, insect hormone biosynthesis, Glycine, serine and threonine metabolism and Folate biosynthesis (Table 3). The enriched GO terms included eggshell formation process and structural constituent of chorion. Most of the genes in these two GO terms were down-regulated in the mutant (Table 3). Consistently, These GO items were also in the top rank with the lowest p values when analyzing the DEGs between wild and domestic silkworm (Table 4). Nearly all of the genes involved in these KEGG and GO terms showed significant lower expression level in the wild silkworm, i.e., up-regulated in the domestic silkworm (Table 4). These results further supported that in the wild silkworm, low expression level of SP1 may be associated with suppressed expression of genes in the eggshell formation process.
DEGs between domestic and wild silkworms were significantly enriched in function of structural constituent of ribosome. The related genes are mostly ribosome proteins (S6 Fig). We also noted that the related biological processes, such as amide and peptide biosynthesis, was also in the top rank with the lowest p value (Table 4). The related genes were also up-regulated in the domestic silkworm. During domestication, there might be other factors that contribute to improved hatchability, such as, improved amide and peptide biosynthesis and activated ribosome activities in the ovaries.
Nitrogen resources are very important for silkworm domestication. The domestic silkworm tends to efficiently utilize nitrogen resources to yield protein outputs to adapt to human-preference, such as the economically important product, the cocoon. In this study, we discovered that artificial selection could directly act on a nitrogen resource gene, i.e, storage protein 1 (SP1), to improve silkworm hatchability. SPs are also target loci of breeding in crops [28]. However, with edible crops, human can directly benefit from the nutrients of these improved SPs [16], whereas in the silkworm, the SPs benefit is in the form of increased silkworm reproductive capacity.
Among all the SPs identified, SP1 is quite divergent and somewhat unique from the others, both in terms of genomic location and phylogenetic position. A similar pattern was also observed in other Lepidoptera species, such as the tobacco hornworm, Manduca sexta [29], suggesting that SP1 may have evolved independently, while the other types of SPs might have experienced duplication during Lepidoptera evolution. Methionine-rich SP1 seems to be of special interest, since methionine is reported to be an important amino acid in the trade-off between growth and reproduction [30]. In Drosophila, dietary methionine restriction extends lifespan [30], while in grasshoppers, a reduced reproduction-induced increase in expression methionine-rich protein occurred during life extension [31]. Similarly, in the beet armyworm, silencing of SP1 by RNA interference (RNAi) decreases larval survival, which indicates the role of the methionine-rich SP in growth and metamorphosis[13]. We therefore added a new evidence that different to grasshopper [31] and the beet armyworm[13], but similar to Drosophila[30], silkworm methionine-rich SP1 functions in the reproduction process but does not obviously affect growth.
Given that in cocoon-producing silk moths, other nitrogen utilization system such as the glutamate /glutamine cycle, have been reported to be vital in metamorphosis silk-cocoon production [2, 5, 32], we suspect that the strategy of nitrogen resource allocation via storage proteins may have diverged or modified during Lepidoptera insect evolution. In the silkworm, the function of SP1 is limited to influencing the egg hatching rate. Artificial selection acted only on SP1 rather than other SPs, suggesting the importance of SP1 for human-preferred domestication traits, i.e., increased hatchability.
Ova comparative transcriptome analyses further illustrated a framework of regulatory network of SP1 on hatchability. Firstly, there are many genes near the bottom of the regulatory network, including vitellogenin(Vg), chorion proteins, structural component proteins in the extracellular matrix (ECM)-interaction pathway such as collagen and integrins, and synthetase in folate biosynthesis are all generally repressed in both the SP1 mutant and the wild silkworm. Thus, artificial selection acts on SP1 for increased hatchability, possibly associated with the influence of those genes, pathway or biological processes, and finally contributes to an improved performance of ovary. Vg is the main nutrient for silkworm egg formation and embryonic development[33]. It appears and accumulates at the stage when SP1 rapidly declines and disappears in the fat body, shortly before the emergence of the adult silkworm [17, 34]. SP1 may supply amino acids for the synthesis of Vg, as previously reported in Plutella xylostella [35]. We therefore suspect that deficiency of SPs might directly trigger an as yet unknown regulatory pathway for the expression or synthesis of Vg. Chorion proteins are the major component of the silkworm eggshell and perform the essential function of protecting the embryo from external agents during development, while simultaneously allowing gas exchange for respiration. Eggshell (chorion) is constructed by the ovarian follicle cells. The follicle cell epithelium surrounds the developing oocyte and, in the absence of cell division, synthesizes a multilayer ECM [36]. Eggshell ECM was usually linked by integrins, a family of transmembrane receptor proteins to the cytoskeleton of the oocyte. Via a series of signal transductions, ECM-integrins function in oocyte movement, differentiation, and proliferation [36]. Integrins were reported to function in formation of actin arrays in the egg cortex [37] and they were also involved in tracheole morphogenesis which affects respiration [38]. Repression of these genes are directly associated with deficient development and function of the ovary. Loss of function of SP in the mutant or low expression level in the wild silkworm of SP might influence development and function of the ovary, further reducing the expression of Vg and chorion genes.
Secondly, folate is known to be important for human fetal development [39]. In insects, folate also plays an important roles in egg development, possibly promoting the biosynthesis of nucleic acids in the ovaries, and evoking mitoses in cells of the collicular epithelium [40–42].
Last and interestingly, we found that the hormone synthesis pathway was also repressed in response to SP1 deficiency. Recent advances in hormone signaling indicate that in the adult insect, juvenile (JH) and molting hormones may cooperate to promote Vg expression and oocyte development [27, 43, 44]. Therefore, hormone signaling pathway might function in the regulatory network of SP in these downstream genes, although there are still black boxes in the regulation connections of these genes, which require further in-depth experimental exploration.
Notably, increased hatchability during domestication may not be solely attributed to the increased expression of SP1 and the associated downstream genes, given that artificial selection acts on hundreds of gene loci in the silkworm genome [2, 45] and that the ova comparative transcriptome between wild and domestic silkworms identified many more genes than that between SP1 mutant and wild-type silkworm. We observed significantly enriched pathway and structural constituent of ribosome, the protein translation machinery and the biological processes involved in nitrogen metabolism and, are generally up-regulated in the domestic silkworm compared with the wild one (Table 4). These results again supported the importance of nitrogen and amino acids in silkworm domestication, not only for silkworm protein output [2], but also for productivity.
Similar to other domesticates, hatchability of silkworm eggs directly determines the quantity of offspring, and thus it is an important productivity trait for human to favorably select during domestication. Based on the above results and the discussion, we propose that artificial selection, favors higher expression of SP1 in the domestic silkworm, which would subsequently up-regulate the genes or pathways vital for egg development and eggshell formation. On the other hand, artificial selection consistently favors activated ribosome activities and improved nitrogen metabolism in the ova, as it might act in the silk gland for increased silk-cocoon yield [2]. In result, the domestic silkworm demonstrates improved egg hatchability compared with it wild ancestor.
A multivoltine silkworm strain, Nistari, was used in all experiments. Larvae were reared on fresh mulberry leaves under standard conditions at 25°C. The wild silkworms were collected in Zhejiang Province, China and maintained as laboratory population in our lab.
The genomic single nuclear polymorphic data (SNP) file (the VCF) for the domestic and wild silkworm obtained from DEYAD platform (https://doi.org/10.5061/dryad.fn82qp6) [2].
Reference genome and the annotation file used for RNA-seq data mapping were obtained from the Ensemble database (http://metazoa.ensembl.org/Bombyx_mori/Info/Index).
The reference sequences of B. mori storage proteins (SP1, SP2, SSP2 and SP3) were retrieved from the NCBI GenBank. These sequences were used as query, searching for homologs in the B. mori genome by tblastn with e-value <10−7. Other insect homologs of the silkworm SPs were searched in GenBank (https://blast.ncbi.nlm.nih.gov/) by BLASTP with an e-value <10−7. We selected sequences from several representative Lepidoptera species and Drosophila melanogaster as candidate proteins for further analyses. The sequences of the SP1 homologs were aligned using MEGA 6.0 software [46]. A gene tree was constructed using MrBayes-3.1.2 with GTR + gamma substitution model [47]. The gene-ration number was set as 1000000 and the first 25% was set as burn-in. Other parameters were set as default.
Based on the available whole genomic single nuclear polymorphic data (SNP) of domesticated and wild silkworm populations [2], (https://doi.org/10.5061/dryad.fn82qp6), we screened the selection signatures of the silkworm SPs, according to Xiang et al’s pipeline [2]. Specifically, data from 19 samples of the early domesticated group (i.e., trimoulting local strains [CHN_L_M3]) of the domestic silkworm Bombyx mori and 18 samples of wild silkworm B. mandarina were used. The SNP data of the two chromosomes that SP1 (Chromosome 23) and the cluster of the other SPs (Chromosome 3) were located were used to screen for the domestication signature. Chr. 23 is 20,083,478 bp in length and 2,046,397 SNPs were identified. Chr. 3 is 14,662,804 bp in length and 1,448,852 SNPs were identified, based on the published data. Allelic frequency and SNP annotation were calculated using in-house Perl scripts.
For the detection of selection signature during silkworm domestication, we set a very stringent threshold to screen out regions significantly deviated from the overall distribution. We only used windows within the top 1% of selective signatures (the corresponding p value of a Z test < 0.001) and applied Fst (fixation index) between the two groups to represent the selective signatures, taking the highest 1% value as the cutoff. The selection in the domestic silkworm group (i.e., the early domesticated group) was further confirmed by limiting π at a relatively low level (the lowest 5%).
The 20 bp sgRNA targets immediately upstream of PAM were designed by the online platform CRISPRdirect (http://criSpr.dbcls.jp/) [48]. The sgRNA DNA template was synthesized by PCR, with Q5 High-Fidelity DNA Polymerase (NEB, USA). The PCR conditions were 98°C for 2 min, 35 cycles of 94°C for 10 s, 60°C for 30 s, and 72°C for 30 min, followed by a final extension period of 72°C for 7 min. The sgRNA were synthesized based on the DNA template in vitro using a MAXIscript T7 kit (Ambion, Austin, TX, USA) according to the manufacturer’s instructions. The Cas9 construct was a kind gift provided by the Shanghai Institute of Plant Physiology and Ecology (Shanghai, China). The Cas9 vector was pre-linearized with the NotI-HF restriction enzyme (NEB, USA). The Cas9 mRNA was synthesized in vitro with a mMESSAGE mMACHINE T7 kit (Ambion, Austin, TX, USA) according to the manufacturer’s instructions. All related primers are shown in Table 2.
Fertilized eggs were collected within 1 h after oviposition and microinjection was within 4 h. The Cas9-coding mRNA (500 ng/μL) and total gRNAs (500 ng/μL) were mixed and injected into the preblastoderm Nistari embryos (about 8 nl/egg) using a micro-injector (FemtoJet, Germany), according to standard protocols (Tamura, 2007). The injected eggs were then incubated at 25°C for 9–10 d until hatching.
To calculate the effect of Cas9/sgRNA-mediated gene mutation in the injected generation (G0), we collected ~10% of the eggs (64 out of 600) 5 d after injection to extract genomic DNA for PCR, with primers Sp1-F and Sp1-R (Table 2). The amplified fragments were cloned into a pMD19-T simple vector (Takara, Japan) and sequenced to determine mutation type.
When the injected G0 silkworms pupated, we collected silkworm exuviae from fifth instar larvae in each cocoon. Genomic DNA was extracted using a TIANamp Blood DNA Kit (Tiangen Biotech, Beijing) according to the manufacturer’s instructions. Individual mutation screening was generated with PCR at 94°C for 2 min, 35 cycles of 94°C for 30 s, 57°C for 30 s, and 72°C for 45 s, followed by a final extension period of 72°C for 5 min. The PCR products were cloned into the pMD19-T simple vector (Takara, Japan) and sequenced.
Mosaic mutant moths were obtained from the above mutation screening of exuviae DNA from fifth instar larvae. Moths with the same mutation site were pairwise crossed with each other to acquire G1 offspring. About 7 d after the G1 eggs were laid, we collected ~30 eggs from each offspring population from one parental pair and pooled them to extract genomic DNA for mutation screening by PCR. The amplified fragments were cloned into a pMD19-T simple vector (Takara, Japan) and sequenced to determine the exact mutation type. Two G1 offspring populations with large deletions in BmSp1 were selected for further breeding. At the pupa stage, 20 randomly selected individuals within each population were subjected to mutation screening of exuviae DNA. Homozygous mutant moths with the same identified mutant genotype were crossed to acquire G2 offspring. Mutation effects on proteins were evaluated using MEGA 6/0 software[46] through codon alignment of the wild-type and the mutant.
On the fourth day of pupation (P4), we weighed and recorded the whole cocoon weight, pupa weight, and cocoon shell weight. In total, data from 240 SP1-MU1 mutants and 110 wild-type silkworms were recorded respectively.
Offspring of the homozygous mutants and wild-type silkworms were incubated at 25°C for 9–10 d until hatching. The number of eggs produced and hatched from each female moth were recorded respectively. The egg hatching rates were then determined. Eighty-three replicates were set for the SP1 mutant and 17 for the wild-type populations, respectively. These assays were also generated for 20 wild silkworms.
Student’s t-test was used to analyze the significance of the differences. For comparisons of datasets with unbalanced size, Student’s t-test with FDR (false discovery rate) correction was used. Specifically, for the cocoon- and pupal- related traits, we divided the samples of SP1 mutant to two groups, consisting of 120 samples for each, and generated t-test with the wild type respectively. The average p value followed by FDR correction was used to verify the significance.
As for the analysis of the number of eggs and hatching rates, we divided the samples of SP1 mutant to 4 groups, consisting of 20 or 21 samples for each, and generated t-test with the comparable data from the wild-type silkworms, respectively. Average p value followed by FDR correction was used to verify the significance.
Ova from newly emerged virgin moth of the domestic wild type silkworm, SP1 mutant and the wild silkworm were dissected and collected for RNA extraction with three replicates set for each. Total RNA were isolated using TRIzol (Invitrogen). For each sample, RNA were sent to Novogene Bioinformatics Institute (Beijing, China) for cDNA library construction and RNA-seq by Illumina Hiseq 2500 (Illumina, San Diego, CA, USA) with 125 bp paired-end reads according to the manufacturer’s instructions.
Raw data were filtered with the following criteria: (1) reads with ≥ 10% unidentified nucleotides (N); (2) reads with > 10 nt aligned to the adapter, allowing ≤ 10% mismatches; and (3) reads with > 50% bases having phred quality < 5. The clean data were mapped to the Bombyx mori reference genome using Tophat with 2 nt fault tolerance and analyzed using Cufflinks [49]. The relative expression value of each gene was calculated using the widely used approach, i.e., fragments per kilobase of exon per million pair-end reads mapped (FPKM) [49], using Cuffdiff In order to identify differentially expressed genes (DEGs), Cuffdiff was further used to perform pairwise comparisons between wild-typed and SP1 mutant samples, as well as the wild and domestic silkworm, respectively, with corrected P-value of 0.05 <5 and Log2-fold change>1.
KEGG and GO enrichment analyses of DEGs were performed with an online platform (http://www.omicshare.com/tools/), using all the expressed genes (FPKM >1) in the ova of virgin moth of Bombyx mori as background.
We used real-time PCR to evaluate the results of RNA-seq data. The Ova of the domestic wild type silkworm and the wild type were dissected from newly emerged virgin moths. Total RNA was digested with DNase I (Takara) to remove the remaining DNA. For Complimentary DNA synthesis, 1ug of total RNA was used in the ReverAid First Strand cDNA Synthesis kit (Takara). Primers for real-time PCR were as follows: 5′ -GGCTTCACTGTCACCAGCACTT-3′ (BGIBMGA009715_f) and 5′ -ACCACAGCCGTAAGACACCAGA-3′ (BGIBMGA009715_r) for BGIBMGA009715; 5′ -GGGCTTATGATGCCGTAGGA-3′ (BGIBMGA009719_f) and 5′ -CGGTGGGAGTTATTGGTGATGT-3′ (BGIBMGA009719_r) for BGIBMGA009719; 5′ -ACCAGCATATCACCAATAGCACC-3′ (BGIBMGA009720_f) and 5′ -ATCGCCGCAGCCATACAGAA-3′ (BGIBMGA009720_r) for BGIBMGA009720; 5′ -GGCTTCATCTATCATCGCTCCAC-3′ (BGIBMGA009721_f) and 5′ -GCCACACCCATACGCCACTTCT-3′ (BGIBMGA009721_r) for BGIBMGA009721; 5′ -GGCAATTATAGCCGCCGTGTCC-3′ (Vg_f) and 5′ -GGCCAGGACTCTTTACCCGGAT-3′ (Vg_r) for Vg; 5′ -GACTCGTCGTGTAATGGAAAGC -3′ (SP1_f) and 5′ -ATGTGGGCAAGAGCATACCG -3′ (SP1_r) for SP1 and 5′ -CAGGCGGTTCAAGGGTCAATAC-3′ (RP49_f) and 5′-TACGGAATCCATTTGGGAGCAT-3′ (RP49_r) for the internal control, the ribosomal protein 49 (Bmrp49, AB48205.1). Real-time PCR was performed in three duplicates with SYBR Green PCR Mix (Bio-Rad) and subjected to the Roche LightCycler 480 Real-Time PCR System. The messenger RNA quantity of each gene was calculated with the 2-ΔΔCT method and normalized to the abundance of RP49.
|
10.1371/journal.pbio.2001492 | Dramatic and concerted conformational changes enable rhodocetin to block α2β1 integrin selectively | The collagen binding integrin α2β1 plays a crucial role in hemostasis, fibrosis, and cancer progression amongst others. It is specifically inhibited by rhodocetin (RC), a C-type lectin-related protein (CLRP) found in Malayan pit viper (Calloselasma rhodostoma) venom. The structure of RC alone reveals a heterotetramer arranged as an αβ and γδ subunit in a cruciform shape. RC specifically binds to the collagen binding A-domain of the integrin α2 subunit, thereby blocking collagen-induced platelet aggregation. However, until now, the molecular basis for this interaction has remained unclear. Here, we present the molecular structure of the RCγδ-α2A complex solved to 3.0 Å resolution. Our findings show that RC undergoes a dramatic structural reorganization upon binding to α2β1 integrin. Besides the release of the nonbinding RCαβ tandem, the RCγ subunit interacts with loop 2 of the α2A domain as result of a dramatic conformational change. The RCδ subunit contacts the integrin α2A domain in the “closed” conformation through its helix C. Combined with epitope-mapped antibodies, conformationally locked α2A domain mutants, point mutations within the α2A loop 2, and chemical modifications of the purified toxin protein, this molecular structure of RCγδ-α2A complex explains the inhibitory mechanism and specificity of RC for α2β1 integrin.
| In animals, collagen-mediated platelet aggregation is an essential component of the blood’s clotting response following vascular injury. A small group of snake venom toxins belonging to the C-type lectin protein family exert their harmful effects by directly targeting this pathway. Rhodocetin (RC) is a heterotetrameric protein found in the venom of the Malayan pit viper (C. rhodostoma). RC specifically binds α2β1 integrin, the key protein required for collagen-mediated platelet aggregation. In this study, we describe the interaction between RC and α2β1 integrin at atomic resolution. This study reveals that RC undergoes a massive structural reorganization upon α2β1 integrin binding, such that RC’s αβ subunit is released from its γδ subunit and a γδ-α2β1 integrin complex is formed. The inhibitory nature of this complex can be readily explained as RC binding along the top surface of the α2β1 integrin and directly above the collagen binding site. As a result, access of collagen to its binding site is blocked, thereby preventing collagen-mediated platelet aggregation.
| Most cellular processes depend on the formation of interactions between cells and the extracellular matrix (ECM). Key facilitators of these interactions are the integrins. They consist of 2 subunits, α and β, each of which has multiple isoforms [1,2]. The different subunit composition between integrins determines their ligand-binding specificity and functionality. Integrins are cell adhesion molecules, which are involved in a broad range of cell functions, such as proliferation, differentiation, adhesion, and migration. Defect or dysfunction of integrins, in particular of α2β1 integrin, a prominent collagen binding receptor of many cell types [3] and the only collagen binding integrin on platelets [4], may affect vascular development and angiogenesis [5], epithelial cell differentiation [6], wound repair and fibrosis [7], inflammation [8,9], and cancer and cancer therapy [10], as well as collagen-induced platelet activation, hemostasis, and thrombosis [4,11]. Therefore, α2β1 integrin has become a prominent target in drug research [12–14].
The collagen binding site is located within the α2A domain of α2β1 integrin, which is homologous to the A-domain of von Willebrand factor (vWF). The α2A domain contains a metal ion that is required for collagen binding as it is part of the binding site for the collagen triple helix [15]. In order to bind to collagen, the α2A domain undergoes a series of concerted conformational changes. In short, helix C unwinds, the N-termini of helices 6 and 7 simultaneously turn away from each other, and, finally, helix 7 moves downward against helix 1 to give the collagen binding “open” conformation, which contrasts with the previous “closed” conformation [15,16]. This likely general mechanism of molecular movement of integrin A-domains was subsequently confirmed by introducing a disulfide bridge into the A-domain of the integrin αL subunit such that this interconversion was blocked with the protein locked in either the “open” or “closed” state [17].
Integrin function can be blocked by two major classes of snake venom proteins, the disintegrins [18,19] and the C-type lectin-related proteins (CLRPs) [20,21]. In contrast to the disintegrins, which can target multiple integrins, CLRPs specifically inhibit α2β1 integrin activity [21]. The high selectivity and affinity of these snake venom proteins for α2β1 integrin make them ideal lead compounds for drug development [22–24]. Current members of the CLRP family include the proteins rhodocetin (RC), EMS16, vixapatin, sochicetin-B, lebecetin, flavocetin, and rhinocetin [25–31]. As more CLRP structures become available, it is clear that, although the supramolecular structure can vary from the basic heterodimer of EMS16 [27] to the ring-like (αβ)4 structures of flavocetin and convulxin [32,33], the underlying basic unit is a heterodimer consisting of 2 subunits, usually named α and β, which dimerize via their characteristic index finger loops [20,34]. Interestingly, in the case of the RC heterotetramer (αβγδ) structure [26], the αβ and γδ subunits form 2 heterodimeric pairs that are oriented orthogonally towards each other in a cruciform shape. Despite these differences, the subunits of CLRP family members are highly homologous with each other. Evolutionarily, the CLRP fold has developed from a carbohydrate recognizing domain (CRD) into a structure that specifically targets clotting factors IX and X, α2β1 integrin, and other platelet adhesion receptors [20,34–36]. Among the latter, the vWF receptor and the 2 collagen binding receptors, glycoprotein GPIV and α2β1 integrin, are targets for snake venom CLRPs, thereby inhibiting or activating platelet activation and aggregation [37,38]. Consequently, these snake venom proteins severely interfere with hemostasis [36,39]. However, the nature of the molecular mechanism by which CLRPs inhibit α2β1 integrin and by which CLRPs implement specificity towards α2β1 integrin has remained undetermined.
RC is a CLRP of the Malayan pit viper C. rhodostoma [26], and together with EMS16 from Echis multisquamatus, they are the only known CLRP family members proven to target the α2A domain for which atomic resolution structures are available [27,40]. Unlike the α2β1 integrin–collagen interaction, which is metal ion-dependent, the binding of RC to α2β1 integrin does not require a metal ion, which implies a different mechanism of action. In a previous study, we demonstrated that the RCαβγδ heterotetramer binds to α2β1 integrin before releasing the αβ subunit (RCαβ) from the complex [40]. In the current work, we present the molecular structure of this RCγδ-α2A domain complex and unravel the molecular mechanism of this interaction. The RC binding site overlaps with that of collagen, including the key metal ion site, thereby sterically blocking collagen binding. Moreover, a comparison with the previously determined RC structure [26] reveals that, in addition to the release of the RCαβ subunit, the RCγδ subunit undergoes a major conformational change upon integrin binding, which causes it to snap into a bent conformation like a mouse trap. In this final state, RCγδ holds the α2A domain in the “closed” conformation, allosterically unable to bind to collagen. The result is a highly efficient inhibition of α2β1 integrin-mediated attachment and signaling in cells and platelets.
To isolate RC in complex with the integrin α2A domain, recombinant α2A domain was immobilized to Ni Sepharose resin via its His6-tag. Thereafter, an RC-rich protein fraction of C. rhodostoma venom was applied to this column, resulting in the formation of the complex of α2A with tetrameric RC (RCαβγδ) that still bound to the column. Treatment with 5 mM EGTA resulted in the dissociation of the α2A domain bound RC tetramer and the release of RCαβ from the complex, which was eluted from the column. In contrast, RCγδ remained firmly attached to the column bound α2A (Fig 1). This RCγδ-α2A complex was then eluted with a linear gradient of imidazole (Fig 1A). Its His6-tag was cleaved by trypsinolysis, and the excess α2A was removed by size-exclusion chromatography. The close physical contact of both partners within the RCγδ-α2A complex was proven by cross-linkage with 0.5 mM bis(sulfosuccinimidyl)suberate (BS3) (Fig 1B).
The crystal structure of the RCγδ-α2A complex was determined at 3.0 Å resolution by molecular replacement using the previously determined RCαβγδ structure (pdb:3GPR) as a search template (Fig 2). The RCγδ-α2A structure clearly showed that the RCγδ subunit bound to the top of the α2A domain directly above the metal ion-binding site, thereby sterically blocking access of collagen (Fig 2A). Both chains of RCγδ are typical CLRP folds, characterized by a globular core domain interlinked mutually by extended index finger loops. The A-domain of α2β1 integrin assumed the “closed” conformation with its central β-sheet flanked by the α-helices 3, 1, and 7 and 4, 5, and 6 on either side. The crystal structures contain 6 RCγδ-α2A complexes per asymmetric unit (S1 Fig).
We determined the total interaction surface between RCγδ and α2A in the complex to be 965 Å2. There were 2 interface areas on the surface of RCγδ in contact with α2A (Fig 2B–2D). First, the larger interaction site (715 Å2) consisted of 2 adjacent patches of 3 residues each on the RCδ subunit, K59-Y60-K101 (Fig 2C), and R92-Y94-K114 (Fig 2D), which were largely hydrophilic. Second, a smaller hydrophobic site (280 Å2) on the RCγ subunit consisted of the triad L66-R109-W110 that interacted with helix 3, helix 4, and loop 2 of α2A (Fig 2B).
Two complementary contact surfaces on the α2A domain extended down from helix C and the metal ion-binding site (top face) to the loop 2 sequence S214QYGGD219 (lateral face) to form an almost contiguous interface that interacted with the RCγδ subunit. The top face of α2A was approached by the RCδ subunit with its larger 2 patches containing interface (Fig 2C and 2D). The first patch comprised residues K59, Y60, and K101 of RCδ interacting with residues D292 and T293 together with the adjacent helix C of α2A. The side chains of K59 and Y60 were countered by complementary carboxylate and hydroxyl groups of D292 and T293 of α2A, while the amino group of K101 pointed towards the backbone carbonyl groups at the C-terminus of helix C. The second patch had the side chains of R92, Y94, and K114 of the RCδ subunit pointing into the collagen binding crevice of α2A. The long side chain of K114 of this protuberance sat at the entrance to the divalent cation binding site (Fig 2D) and was positioned 7.7 Å above the magnesium ion, whereas the positively charged guanidino group and the phenolic hydroxyl group of R92 and Y94 contacted the main chain carbonyl of D219 in loop 2 of α2A.
The second contact surface is the loop 2 sequence S214QYGGD219 at the lateral face of α2A, which interacted with the amino acid side chains of L66, R109, and W110 of the RCγ subunit (Fig 2B). For example, the aromatic indole ring of W110 contributed to a hydrophobic surface and interacted with the backbone chain of the glycine residues G217 and G218 together with the adjacent aspartate residue D219 within loop 2 of the α2A domain (Fig 2B). In addition, L66 of RCγ contacted N154 of loop 1 of the α2A domain. The final RCγ residue of the triad R109 made contact with the S214 side chain of α2A. Taken together, the hydrophobic patch of the RCγ subunit predominantly interacted with the loop 2 sequence S214QYGGD219 of α2A. This loop 2 sequence immediately preceded residue T221, which was part of the metal ion binding site of α2A. A key residue with regard to the interface between the RCγδ subunit and the α2A domain in the RCγδ-α2A complex was the loop 2 D219 of α2A, as it was part of both RC contact sites. In addition, it connected the loop 2 sequence with the collagen binding crevice and helix C of α2A. The presence of helix C in the RCγδ-α2A complex structure indicated that RC had trapped the α2A domain in the “closed” conformation, which is not capable of binding collagen [15].
To test whether RC exclusively binds the closed conformation of α2A, we generated 2 conformationally distinct mutants in which the A-domain was held by a disulfide bridge between K168C-E318C and K168C-A325C in the open and closed conformations, respectively (S2 Fig) [17,41]. Before introducing cysteine residues at these positions, it was necessary to replace the naturally occurring original cysteine residues at position 150 and 270 with alanines. No change in binding affinity to RC was observed for this α2A-C150A,C270A double mutant. In this cysteine-free α2A domain, K168 of α-helix 1 was replaced by a cysteine residue, with a second cysteine residue introduced into α-helix 7 at either position E318 or A325. As a consequence of the newly formed disulfide bridge, the movement of helices 1 and 7 with respect to each other that occurs when α2A shifts between the “open” and “closed” conformation was blocked. Thus, the α2A domain was held in the “open” (K168C-E318C) and “closed” (K168C-A325C) conformation, respectively. The α2A mutant with the “open” conformation hardly bound to RC (Fig 3A), while RC binding to the “closed” conformation of α2A (Kd-value: 0.21 ± 0.03 nM) was similar to that obtained with wild-type α2A (Kd-value: 0.29 ± 0.02 nM). Our structural findings revealed that the sidechain moiety of Lys101 is oriented towards the negatively charged dipole of helix C, stabilizing the closed conformation of the α2A domain (Fig 3B).
Among several monoclonal antibodies raised against the RCγδ subunit [40], IIIG5 belonged to the subgroup that only recognized its epitope within RCγδ after its complexation with α2A and the subsequent release of the RCαβ subunit (Fig 4A). This became evident when the antibody was immobilized and its ability to capture RCαβγδ, RCγδ- α2A, or RCγδ out from solution was probed. IIIG5 gave a binding signal with the RCγδ- α2A complex and RCγδ but not with the RC tetramer alone. Of the 2 RC species capable of binding the IIIG5 antibody, the RCγδ subunit gave the highest binding signal (Fig 4A). The most probable explanation for these results was that the IIIG5 epitope was fully accessible in RCγδ, and so, we observed what approximates the maximal binding. At the other extreme, we had no binding of RCαβγδ, as the epitope was entirely masked in the tetramer. Between these 2 extremes was the RCγδ-α2A complex, in which the epitope is sufficiently exposed for IIIG5 to bind but not to the same extent as for RCγδ due to the nature of the RCγδ-α2A interaction.
The sequence epitope of IIIG5 was isolated from a tryptic digestion of RCαβγδ by affinity chromatography on an IIIG5 column and subsequently by reversed-phase high-performance liquid chromatography (HPLC). Mass spectrometry (MS) identified the γ chain sequence 94–106 as the IIIG5 epitope (S3 Fig), which was mainly located within the index finger loop of RCγ (Fig 4B). This result can be clearly explained by comparing the native RCαβγδ structure with the newly determined RCγδ-α2A complex structure. The IIIG5 epitope was covered by the RCαβ subunit in the RCαβγδ structure and only became accessible upon formation of the RCγδ-α2A complex. Moreover, the index finger loop of the RCγ underwent a major conformational change upon formation of the RCγδ-α2A complex, leading to increased accessibility of the IIIG5 epitope.
The dramatic conformational changes that took place within the RCγδ subunit were readily apparent upon comparing the molecular structures of the RCγδ-α2A complex with the native RCαβγδ tetramer (Fig 5). The binding face of RCαβγδ changes from a flat surface into a concave binding surface to embrace the α2A domain (Fig 5A and 5B). This was implemented via (i) a rigid body movement of both core segments of chains γ and δ, (ii) a dramatic re-orientation of the index finger loop of the γ subunit, which harbors the IIIG5 epitope, and, consequently, (iii) local re-orientations of key binding residues in both RC subunits (Fig 5C and 5D).
The rigid body arrangement can best be described as a flipping of helices 1 and 2 between the RCγ and RCδ subunits whilst maintaining the same relative orientation of the 2 helices within their respective core domains. An additional consequence of this rigid body movement is a conformational shift of the connecting finger loop to track the motion of the opposing core domain. As a result, the 2 core domains flipped over with respect to each other and bent towards the α2A domain to form a concave binding surface such that the RCγδ residues involved in α2A binding were brought into the correct orientation for binding the α2A domain.
The apical ends of the index finger loops were in close contact with the CLRP core element of the opposite subunit, forming the 2 interfaces: loop γ–core δ and loop δ–core γ. Whereas the former hardly changed (Fig 5E and 5F), the latter showed a dramatic shift within the RCγδ-α2A complex as compared to the RCαβγδ tetramer (Fig 5C and 5D). In the loop δ–core γ interface of the RCαβγδ tetramer (Fig 5C), a tryptophan core composed of 3 residues (W76δ, W71δ, and W116γ) together with a salt bridge between R92δ and D74γ stabilized the index finger loop of the RCδ subunit and oriented it towards the RCγ subunit core sequence connecting helices 1 and 2. However, in the RCγδ-α2A complex, the salt bridge between R92δ and D74γ found in the RCαβγδ tetramer (Fig 5C) was broken. R92δ now formed a hydrogen bond to the main chain of D219 in the α2A loop 2, and a new salt bridge was observed between R75γ and E77δ and D81δ (Fig 5D). In addition, the RCδ subunit index finger loop became embedded within the antiparallel sheet S3–S4–S5 of the RCγ core such that the indole moiety of W76δ now made van der Waals contacts to Q105γ and Y118γ (see inset Fig 5C and 5D).
As a result of these enormous conformational changes, especially at the loop δ–core γ interface, the rigid cores of the 2 RCγδ subunits swung towards each other by about 40°–50° around a hinge located in the center of the index finger swap domain between the cores. This global movement had 2 major consequences. First, as the RCδ subunit snapped into its new position, the 3 key residues of RCγ (L66, R109, and W110) underwent a local conformational change that transformed them into an orientation that is competent for α2A binding (Fig 6A). Second, as a consequence of the index finger loop tracking the movement of the RCγ subunit, the contact site between the RCα and RCγ subunits changed its 3D structure due to the formation of the new salt bridge between R75γ and E77δ and D81δ (Fig 5D). Consequently, the previous interface between the RCγ subunit (K77EQQC81) and the RCα subunit (N74KQQR78) became sterically blocked [26]. The movement of the RCγ subunit would also produce steric clashes with the RCβ subunit, and it is likely the combination of these 2 events that resulted in the dissociation of the RCαβ subunit from its RCγδ counterpart. In contrast, the contact site within the RCδ subunit would allow integrin binding irrespective of the conformational change of RC, as their local positions and orientations remained almost unchanged (Fig 6B). In fact, the distance between Y60δ and Y94δ within the RCδ contact sites only changed slightly, from 21.7 Å to 20.4 Å (Fig 6B), while their distances towards W110γ of the RCγ contact site were reduced from 47.5 Å to 31 Å and from 28.4 Å to 18.6 Å, respectively when comparing the structure of RCαβγδ and RCγδ-α2A complex. This illustrated how significant a reorganization of the RCγδ is required to facilitate the formation of the ultimate inhibitory RCγδ-α2A complex.
Unlike helix C, the docking site S214QYGGD219 did not change its conformation between the “open” and “closed” conformation of the α2A domain. To analyze its role, we challenged RC binding to α2A with the monoclonal antibody JA202. Its epitope had previously been mapped to the sequence QTS214QY [42] and thus overlapped with the RCγ subunit docking site. Among different antibodies against distinct epitopes within α2A, JA202 was the only monoclonal antibody which sterically inhibited RC binding to the α2A domain in a dose-dependent manner (Fig 7A).
A comparison of integrin α2 chains from different species showed a high interspecies homology of the loop 2 sequence, S214QYGGD219LT221 (S4 Fig). In contrast, this sequence was absent in A-domains of other integrin α subunits, suggesting that it served as a selective docking site for RC on α2β1 integrin (S5 Fig). Therefore, we replaced the α2A sequence S214QYGGD219L with the corresponding sequence VGRGGRQ of the α1A-domain and tested binding of RC to this α2A-L2α1 mutant. Although this α2A mutant was still able to bind RC, the binding affinity was reduced, as indicated by an increase of the Kd-value from 0.76 ± 0.12 nM to 2.70 ± 0.39 nM (Fig 7B). In parallel to the α2A-L2α1 mutant, we exchanged residues in the loop 2 that interacted with RC (Fig 7C), specifically S214, Y216, and D219, as well as the G217 and G218 that are conserved in both integrin α1 and α2 loop 2 sequences, to see which residues were functionally important for the RCγδ-α2A binding. The S214G and D219A mutants, which are located at the outer edges of loop 2, gave Kd values of 0.77 ± 0.32 nM and 5.2 ± 1.36 nM, respectively, while the Y216G mutant in the center of the loop gave a Kd value of 1.98 ± 0.64 nM (Fig 7D and 7E). In contrast, mutating either of the conserved glycine residues of loop 2 by generating G217K and G218L resulted in a complete loss of RC binding (Fig 7D). This result is in agreement with our structure findings (Fig 7C), which showed that anything larger than a glycine at either position 217 or 218 would sterically clash with the indole side chain of W110γ. In addition, we chemically modified the solvent-exposed W110γ of RC with 2-nitrophenyl sulfenylchloride (NPS-Cl), which introduced a bulky 2-nitro-phenylsulfenyl (NPS) group onto the indole side chain. The modified W110γ is no longer able to stack above the 2 glycines G217 and G218, causing a loss of RC binding to the α2A domain (Fig 7F). Taken together, these results demonstrated that the interaction of W110 of RCγ and the loop 2 of α2A is highly specific and essential for the formation of the high-affinity and inhibitory RCγδ-α2A complex.
Our study reveals not only the interaction sites within RC and its molecular target, the integrin α2A domain, but also the conformational changes that take place within the RCγδ subunit upon α2A binding and the relevance of the 2 contact sites within α2A for RCγδ binding. Moreover, these data suggest a molecular mechanism for the avid and selective interaction of this CLRP and its target.
CLRP dimers recognize other target molecules, such as factor IX/X, and the A-domain of vWF by forming a bay region with their joint index finger loop swap domain and 2 flanking core domains. This concave face shapes the binding sites for clotting factors IX and X [43,44] and the vWF-factor A-domain [45]. Due to their importance in hemostasis, clotting factors and vWF are valid targets for CLRPs from snake venoms. Bitiscetin and botrocetin interact with the vWF–A1 domain without or together with the glycoprotein Ib (GPIb) receptor [27,45,46]. These studies showed that these snake venom toxins can approach the A-domain from different orientations [35,45,46]. In yet another orientation, EMS16 approached the α2A domain of α2β1 integrin, which is homologous to the vWF–A1 domain, along its top face directly above the metal binding site and collagen binding crevice, thus preventing collagen from binding [27]. EMS16 and RC are the 2 α2β1 integrin-binding CLRPs whose crystal structures in both the unliganded and the CLRP in complex with the A-domains have been resolved so far [26,47]. Although RC approached the α2A domain in a similar orientation to EMS16, our data revealed that RC, in contrast to any known CLRP structure [27,45,46], undergoes a dramatic conformational change to form a concave binding surface. In contrast, the heterodimeric EMS16 did not alter its molecular structure upon α2A binding [27,47], as the concave binding surface required for α2A binding was already preformed. This difference in mode of α2A binding between EMS16 and RC is determined by the distinct quaternary structures of the dimeric EMS16 versus the tetrameric RC and/or by the different purification protocols. When we employed the same purification procedure for RC as for EMS16 and other CLRPs [28–30,48] using reversed phase chromatography performed in 0.1% trifluoroacetic acid (TFA) solution, the RC tetramer dissociated into its subunits α, β, and γδ [49]. The RCγδ subunit alone was still able to bind α2A and to block α2β1 integrin-mediated platelet aggregation specifically [50], albeit with a different kinetics [40]. Only when applying a milder purification protocol could we obtain a stable RC tetramer and the RCγδ-α2A complex, whose different conformational structures are presented here.
Our crystal structure of the RCγδ-α2A complex reveals a geometry of interaction similar to the α2A-bound EMS16, suggesting that the α2β1 integrin-blocking CLRPs may have a more uniform binding mechanism than the vWF binding CLRPs (Fig 8). Both CLRPs share the same 2 contact sites within the α2A domain: the conformationally stable loop 2 sequence (Fig 8C) and the helix C of the “closed” conformation (Fig 8D). Helix C is recognized by the structurally robust contact area of the RCδ subunit or the homologous EMS16 subunit β (or B). Apart from slight variations of the K59δ side chain and the loop 2 Y216 side chain (Fig 8D) adopting an alternate conformation to form a hydrophobic interaction with L66γ, the structures of both complexes are almost identical in this region. In our studies, the role of the loop 2 sequence S214QYGGD219 was reinforced by the JA202 antibody, whose epitope overlaps with this loop 2 sequence and inhibits RC binding completely, presumably due to steric hindrance by the bulky antibody. More subtly, recombinant exchange of the respective loop 2 sequence with the homologous sequence of integrin α1 showed that the loop 2 sequence changes the affinity of the venom component towards the integrin α2 subunit. Similar reductions in the affinity of RC for α2A were also observed with the loop 2 mutants Y216G and D219A. However, a loss of binding was obtained with the G217K and G218L mutants. These 2 glycine residues form part of a shallow dimple on the α2A surface that is covered by W110 of the RCγ subunit. In the molecular structure of the RCγδ-α2A complex, there is not any space to accommodate anything larger than a glycine at either of these 2 positions, which explains the loss of function of these 2 mutants. The loop 2 sequence of the integrin α2A domain is evolutionary conserved between different animal species, especially the GG motif at positions 217 and 218, but varies remarkably between other integrin α subunits. This suggests that RC’s specificity is mediated by the integrin α2-specific loop 2 sequence, as RC affects α2β1 integrin-mediated platelet blockage in various potential preys but does not affect biological functions mediated by other integrins. Our conclusion—that this cluster of RCγ W110 and G217/G218 of the α2A loop 2 sequence is a key to the RCγδ-α2A interaction—is further supported by the fact that the RC binding is completely lost if the bulky chemical adduct of 2-nitrophenylsulfenyl is introduced to the indole side chain. It is noteworthy that the loop 2 sequence is also relevant for collagen binding, as it forms a hydrophobic contact for the phenylalanine side chain of the middle strand of the trimeric integrin recognition motif of collagen [15], albeit not as close a contact as with the RCγ W110 side chain.
Based on our findings, we suggest the following mode of action (Fig 9). RCαβγδ interacts with helix C of the α2A domain through the RCδ subunit, where the interacting residues are already in binding-competent orientation. This stabilizes the “closed” conformation of α2A. As a consequence of the movement of RCγ, the RCαβγδ tetramer changes conformation such that RCαβ dissociates from the heterotetrameric assembly. Coupled to this dissociation is the reorganization of L66, R109, and W110 of RCγ to interact with loop 2 sequence S214QYGGD219. Having established both interaction sites, RCγδ firmly binds to α2A and holds it in the “closed” conformation, thereby blocking collagen binding and antagonistically turning off α2β1 integrin signaling. After its release upon formation of the high-affinity RCγδ-α2β1 complex, the RCαβ subunit plays another important role in blocking GPIb and, consequently, vWF-induced platelet aggregation [49]. Moreover, our biochemical data showed that the RCαβ subunit is significantly more soluble than the RCγδ subunit [40]. Therefore, it likely acts as a solubility enhancer to ensure that the RCγδ subunit is delivered to α2β1 integrin. Once RCγδ has bound to its target and the RCαβ subunit has been released, RC effectively shuts down the 2 platelet receptors, α2β1 integrin and GPIb, thereby effectively blocking both collagen-induced and vWF-induced platelet activation and aggregation.
In summary, a comparison of the RCγδ-α2A structure with the EMS16-α2A integrin complex [27] shows that the residues involved in the binding of RC and EMS16 to α2β1 integrin are highly conserved. The formation of the inhibitory RC-α2A complex requires both the interaction of RCδ with the helix C of α2A and RCγ with the α2A loop 2 sequence. Furthermore, the presence of helix C in our structure confirms that we have trapped α2A in the “closed” conformation, which is not able to bind collagen and explains why RC is able to block collagen-mediated platelet aggregation. Finally, the requirement of 2 separate sites within the α2A domain for both function and specificity may be instrumental for the design of novel α2β1 integrin inhibitors.
RC and its γδ subunit were isolated as previously described [40,51]. The monoclonal antibodies (mAbs) against RC, among them IIIG5 from mice and IC3 from rats, were generated and isolated as previously described [40]. The murine mAbs against the human α2A domain, JA202 and JA218, were a generous gift from D. Tuckwell (formerly of the University of Manchester, United Kingdom) [40,42]. PCR primers were obtained from Eurofins (Eurofins Genomics, Germany) and are written in 5′-3′ direction. Restriction enzymes and molecular biology reagents were from Thermo Fisher Scientific (Germany) unless otherwise stated. Cloning products and expression vectors were validated by DNA sequencing (Eurofins Genomics).
RC, dissolved at 110 μM in 30% acetic acid solution, was treated with 9.2 mM 2-nitrophenyl sulfenylchloride (NPS-Cl, TCI Chemicals, Germany) or left untreated for 1 h at 20 °C in the dark according to [52], subsequently dialyzed against 0.1% TFA (RP-solution) and separated on a Supercosil C18 column (Supelco, Germany) by reversed-phase chromatography as described [26]. The RCγδ-containing fractions were pooled, lyophilized, and dissolved in RP-solution containing 30% acetonitrile. Purity was assessed by SDS-PAGE. Spectroscopic evaluation at 365 nm according to [52] confirmed the covalent modification of RC tryptophan residues with 2-nitro-phenylsulfenyl (NPS)-groups.
The His6-tagged α2A domain was generated as previously described [26,53]. It was loaded onto a HiTrap Ni Sepharose column (GE Healthcare; 5 ml) previously equilibrated with PBS/MgCl2-buffer, pH 7.4 (20 mM sodium phosphate, pH 7.4, 150 mM NaCl, 1 mM MgCl2). After washing with the same buffer, the RCαβγδ-containing fractions from the RC isolation with MonoS column [51] were applied to the α2A domain loaded Ni Sepharose column after having been treated with 0.5 μM phenylmethylsulfonyl fluoride (PMSF) and 1 μg/ml aprotinin to prevent proteolytic digestion by potentially contaminating snake proteases. After RCαβγδ had bound to the Ni Sepharose-immobilized α2A domain, the HiTrap Ni Sepharose column was washed with PBS/MgCl2-buffer, pH 7.4. Then, the column was washed with PBS/EGTA-buffer, pH 7.4 (5 mM EGTA in 20 mM sodium phosphate, pH 7.4, 150 mM NaCl) and the RCαβ subunit eluted. After another washing step with PBS/MgCl2-buffer, pH 7.4, the RCγδ-α2A complex was eluted with a linear gradient of 0–200 mM imidazole in PBS/MgCl2-buffer, pH 7.4 from the HiTrap Ni Sepharose column. Protein concentration in the imidazole eluate was determined using the Bradford reagent (BioRad). For crystallization, the complex-containing fractions were pooled and digested with TPCK-treated trypsin (Sigma-Aldrich) at an enzyme:substrate ratio of 1:100 at 37 °C for 1 h. The digest was stopped with 1 mM PMSF, concentrated and separated by gel filtration to remove excess α2A domain, trypsin and contaminating peptides from the RCγδ-α2A complex. The TSK G2000SWXL chromatography was performed in 10 mM HEPES, pH 7.4, 100 mM NaCl buffer. The RCγδ-α2A complex was concentrated by ultrafiltration and its protein concentration determined with the Bicinchoninic Acid Protein Assay (BCA, Thermo Fisher Scientific). To analytically prove the physical contact of both partners, the complex was cross-linked with 0.5 mM bi-sulfosuccinimidyl-suberate (BS3, Thermo Fisher Scientific). Its IEP was determined to be pH 6.5–6.8 and pH 6.7 by isoelectric focusing in precast ZOOM pH 3–10 gels (Thermo Fisher Scientific) and by analytical chromatofocusing on a MonoP column (GE HealthCare) with a pH gradient of 7.4 to 4.0, respectively.
Crystals of 10 mg of RCγδ-α2A were grown by hanging-drop vapor diffusion at 293 K by mixing 2 μL of protein solution with 2 μL reservoir solution containing 2.65 M ammonium sulfate and 100 mM Tris pH 8.0. Crystals appeared after 6 weeks and were soaked in mother liquor containing 20% glycerol for 5–10 min before being flash frozen in liquid nitrogen. Diffraction data was collected at the Canadian Light Source CMCF-08ID-1 beamline (λ = 0.97949Å) at 100 K using a Rayonix MX225 CCD detector. The dataset was indexed, integrated, and scaled with MOSFLM [54] and the CCP4-package [55]. The spacegroup is P41 with 6 molecules in the asymmetric unit (see also Table 1). The phases were determined by rigid body refinement using the previously solved RC structure (PDB code 3GPR) in Refmac [56,57]. The model was built and refined without NCS restraints using Coot [58] and refined with the Phenix software package [59]. The crystallographic data and refinement statistics are summarized in Table 1. The final coordinates and structure factor amplitudes were deposited in the PDB (RCSB-code: 5THP).
The human α2A domain and its mutants were produced in a bacterial expression system. The expression vectors encoding the disulfide-locked conformation mutants of α2A were generated using a previously described pET15b-His6-α2A construct (residues 142 through 337 of human integrin α2). To replace the endogenous cysteine residues at 150 and 270, this plasmid was used as template for a 2-step PCR with the 3 primer pair sets (i) HTfwd(CTCTCCATGGGCTCTTCTCATCATCATCATCATCATTC) and R1(C11A) (CATCAGCCACAACCACAAC), (ii) F2(C11A) (TTGTGGCTGATGAATCAAATAG) and R2(C131A) (TTGGCTTGATCAATCACAGC), and (iii) F3(C131A) (ATTGATCAAGCCAACCATGAC) and α2Arev (CGGACATATGCTAACCTTCAATGCTGAAAAATTTG) in the first set of reactions. The 3 amplicons were purified and again PCR-amplified with the outer primer pair HTfwd and α2Arev to a 670 bp amplicon, which, after A-tailing with Taq DNA polymerase, was intermediately ligated into pCR2.1 TOPO, excised with NdeI and NcoI, and the restriction fragment was subcloned into the linearized, NdeI, NcoI-cleaved pET-15b expression vector. The final expression plasmid pET-15b-His6-α2A(C150,270A) was transformed into Escherichia coli BL21 (DE3).
To generate the disulfide-locked conformation mutants of α2A, which share the same K168C mutation but differ in E318C (“open” conformation: K168C, E318C) or A325C (“closed” conformation: K168C, A325C), 3 rounds of PCR amplification were performed. In the first, site-directed mutagenesis K168C was introduced by amplifying the entire plasmid with the back-to-back primer pair K168C fw (AAGGCCTGGATATAGGCCCC) and K168C rev (GTACAAAGCATTCCAAAAAATTCTTTACTGC). Based on this mutation, the final 2 mutants (K168C, E318C; K168C, A325C) were similarly generated using the primer pairs E318C fw (GTCTGATTGCGCAGCTCTACTAGAAAAG)/E318C rev (ACATTGAAAAAGTATCTTTCTGTTGGAATAC) and A325C fw (ATTAGGAGAACAAATTTTCAGCATTGAAG)/A325C rev (GTCCCGCACTTTTCTAGTAGAGCTG). For each site-directed mutagenesis, only 1 primer contained the specific mutation. The PCR products were amplified by the Phusion Hot Start II polymerase and covered the whole template vector (6307 bp) with the mutation. After the original, methylated vector had been digested with DpnI, the amplicons were purified using the DNA Clean & Concentrator Kit (Zymo Research), followed by 5′-phosphorylation with T4 polynucleotide kinase and religated using T4 DNA ligase. For protein expression, E. coli strain BL21 (DE3) were transformed with the validated plasmid constructs encoding the α2A domain in its “open” (pET-15b-His6-α2A-C150/270A-K168C/E318C) and “closed” (pET-15b-His6-α2A-C150/270A-K168C/A325C) conformations.
The α2A-L2α1 mutant, in which the sequence S214QYGGDL is replaced by the corresponding loop 2 sequence V214QRGGRDQ of the integrin α1 A-domain, was generated by 2-step PCR. The pET15b-construct encoding the His-tagged α2A domain [26] was used as a template. The primer pairs α2A fw (GGATATCTGCAGAATTCGCCCTTC) and R1_a1insert into a2 (CTTTACTAACATCGTTGTAGGGTCTGTCACGTCGCGCCACCAGCGGTC), F1_a1insert into a2 (GTGCAGCGCGGTGGTCGCCAGACAAACACATTCGGAGCAATTC), and α2A rev (AGGCCATATGCTAACCTTCAATGCTGAAAATTTG) amplified the N- and C-terminal halves of the cDNA. The 2 amplicons were mixed and amplified with the outer primer pair. The resulting 680 bp amplicon was trimmed with NcoI and NdeI, ligated into a correspondingly cut pET-15b vector, verified by sequencing, and transformed into E. coli BL21(DE3).
Point mutations within the loop 2 sequence were also generated by a 2-step PCR using the wild-type α2A-encoding cDNA as template. First, cDNA fragments encoding the N- and C-terminal halves of α2A were amplified by using the 2 pairs of forward outer and reverse inner primers and of forward inner and reverse outer primers, respectively, as summarized in Table 2.
The amplicons were purified and taken as template for a second PCR with the outer primer pair to obtain the wild-type and mutant α2A domains encoding cDNAs, which were digested with NdeI and BamHI and ligated into the likewise-cut pET-15b vector. After verification by sequencing, the expression vectors were transformed into E. coli BL21(DE3). All α2A domain mutants were purified using HiTrap Ni Sepharose column (GE HealthCare) as per the wild type.
The wells of a half-area microtiter plate (Costar) were coated with 10 μg/ml His-tagged α2A domain in TBS/Mg buffer (50 mM Tris/HCl, pH 7.4, 150 mM NaCl, 3 mM MgCl2) at 4 °C overnight. After washing twice with TBS/Mg buffer, the wells were blocked with 1% BSA in TBS, pH 7.4, 2 mM MgCl2 for 1 h at room temperature. The immobilized α2A domain was titrated with a serial dilution of RCαβγδ or RCγδ without and with NPS-modified tryptophans in the blocking buffer for 1.5 h. For the mAb inhibition experiment, RC at a constant concentration of 2 nM was added to the wells in either the absence or presence of mAb JA202 against RC. After washing twice with HEPES-buffered saline (HBS) (50 mM HEPES/NaOH, pH7.4, 150 mM NaCl, 2 mM MgCl2), bound RC was fixed with 2.5% glutaraldehyde in the same solution for 10 min at room temperature. After 3 additional washes with TBS/Mg buffer, bound RC was quantified by ELISA using a primary rabbit antiserum against RC and a secondary alkaline phosphatase conjugated anti-rabbit–IgG antibody, each diluted 1:2,000 in 1% BSA/TBS/Mg. Conversion of para-nitrophenyl phosphate (pNpp) to para-nitrophenolate was stopped with 1.5 M NaOH and measured at 405 nm. The titration curves were evaluated as described below. The inhibition curves were approximated by GraphPad Prism software using the inhibition vs. log [inhibitor]-approximation. To compare independent inhibition and binding experiments, the dynamic ranges were normalized to the mAb-free control and to the saturation value of the wild-type α2A domain, respectively.
Alternatively, the α2A domains, either wild-type or mutants, were captured using the mAb JA218 at a ligand-binding–irrelevant epitope, thereby avoiding any conformational changes due to adsorption to the plastic. To this end, 2.5 μg/ml JA218 was immobilized to a microtiter well at 4 °C overnight. After the wells were washed twice with TBS/Mg buffer, wells were blocked with 1% BSA in the same buffer for 1 h, and then, the α2A domain was added at 10 μg/ml for 1 h. After washing the wells, RC was titrated and detected as described above.
The mAb IIIG5 was coated to the wells of a microtiter plate at 3 μg/ml in TBS/Mg buffer overnight. After 2 washing steps, wells were blocked with 1% BSA in TBS/Mg buffer for 1 h and then titrated with either RCαβγδ, RCγδ, or RCγδ-α2A complex for 1.5 h at room temperature. Bound RC was fixed and quantified as described above. A mathematical approximation of the titration curve, including determination of Kd-values, is described below.
IIIG5 was immobilized to cyanogen bromide-activated sepharose according to the manufacturer’s instruction (GE Healthcare). RCαβγδ-containing fractions from the Mono S purification of C. rhodostoma venom [26] were reduced with 4 mM tris(hydroxymethyl)phosphine (THP, Calbiochem) for 20 min at 60 °C, and free thiol groups were alkylated with 16 mM iodoacetic acid. The protein was precipitated with trichloroacetic acid, washed with acetone twice, resuspended in 87.5 mM sodium bicarbonate/0.5 M urea and digested with TPCK-trypsin for 23 h at 37 °C. After addition of 1 mM PMSF, the digest was diluted with TBS/HCl buffer, pH 7.4 and loaded onto the IIIG5 column. The RC peptide harboring the IIIG5 epitope was eluted in a pH gradient from pH 7.5 to 3.0 in 20 mM citrate buffer and further purified by reversed phase on a Supercosil C18 column in a 0%–28% acetonitrile gradient in 0.1% TFA/water. Lyophilized HPLC fractions were dissolved in 40% methanol containing 0.5% formic acid and analyzed by nano-electrospray ionization (nanoESI) MS and MS/MS. Peptide structures were deduced from the corresponding fragment ion spectra. NanoESI MS experiments were carried out by using a SYNAPT G2-S mass spectrometer (Waters, Manchester, UK) equipped with a Z-spray source in the positive ion sensitivity mode. Typical source parameters were as follows: source temperature, 80 °C; capillary voltage, 0.8 kV; sampling cone voltage, 20 V; and source offset voltage, 50 V. For low-energy collision-induced dissociation (CID) experiments, the peptide precursor ions were selected in the quadrupole analyzer, subjected to ion mobility separation (IMS; wave velocity 850 m/s, wave height 40 V, nitrogen gas flow rate 90 ml/min, and helium gas flow rate 180 ml/min), and fragmented in the transfer cell using a collision gas (Ar) flow rate of 2.0 ml/min and collision energies up to 100 eV (Elab).
In titration curves, a signal S, usually the extinction at 405 nm caused by the alkaline phosphatase-catalyzed conversion of pNpp, is measured in response to the total concentration c0 of added titrant. Based on a Michaelis–Menten-like binding mechanism, we deduced the following equation to approximate titration curves, if the signal S and the total concentration c0 of added ligand (RC) is known:
S(c0)=(SM−Sm)∙((c0+cR+K)−(c0+cR+K)2−4∙c0∙cR2∙cR)+Sm+B∙c0
with SM and Sm, maximum and minimum signals, respectively; cR, the concentration of ligand binding site (equals the receptor concentration for monovalent receptors); and K, the dissociations constant Kd. The term B·c0 takes into account a linear change in the signal due to nonspecific binding of the ligand. The 5 parameters SM, Sm, cR, K, and B are calculated by nonlinear regression from titration curves.
The data from titration and inhibition curves were statistically evaluated using GraphPad Prism software. Values were usually compared with the values obtained for the wild-type α2A or nonmodified RC with Student t test, where the significance level was set at 1% unless otherwise stated.
|
10.1371/journal.ppat.1004295 | Cytolethal Distending Toxins Require Components of the ER-Associated Degradation Pathway for Host Cell Entry | Intracellular acting protein exotoxins produced by bacteria and plants are important molecular determinants that drive numerous human diseases. A subset of these toxins, the cytolethal distending toxins (CDTs), are encoded by several Gram-negative pathogens and have been proposed to enhance virulence by allowing evasion of the immune system. CDTs are trafficked in a retrograde manner from the cell surface through the Golgi apparatus and into the endoplasmic reticulum (ER) before ultimately reaching the host cell nucleus. However, the mechanism by which CDTs exit the ER is not known. Here we show that three central components of the host ER associated degradation (ERAD) machinery, Derlin-2 (Derl2), the E3 ubiquitin-protein ligase Hrd1, and the AAA ATPase p97, are required for intoxication by some CDTs. Complementation of Derl2-deficient cells with Derl2:Derl1 chimeras identified two previously uncharacterized functional domains in Derl2, the N-terminal 88 amino acids and the second ER-luminal loop, as required for intoxication by the CDT encoded by Haemophilus ducreyi (Hd-CDT). In contrast, two motifs required for Derlin-dependent retrotranslocation of ERAD substrates, a conserved WR motif and an SHP box that mediates interaction with the AAA ATPase p97, were found to be dispensable for Hd-CDT intoxication. Interestingly, this previously undescribed mechanism is shared with the plant toxin ricin. These data reveal a requirement for multiple components of the ERAD pathway for CDT intoxication and provide insight into a Derl2-dependent pathway exploited by retrograde trafficking toxins.
| Cytolethal distending toxins (CDTs) are produced by several bacterial pathogens and increase the ability of these bacteria to cause disease. After being taken up by host cells, CDTs are trafficked to the endoplasmic reticulum (ER) where they must translocate across the ER membrane to gain access to their intracellular target; however, this translocation process is poorly understood for CDTs. Here we provide evidence that CDTs require components of the ER-associated degradation (ERAD) pathway, a normal cellular process utilized to translocate terminally misfolded ER lumenal and membrane proteins across the ER membrane for degradation in the cytosol. Deletion of a key member of this pathway, Derl2, makes cells resistant to multiple CDTs. Interestingly, two domains within Derl2 which are required for ERAD of misfolded proteins are dispensable for intoxication by CDT. Further, we report two previously uncharacterized domains within Derl2 that are each required for intoxication. Consistent with a role of Derl2, abrogation of two other members of the ERAD pathway, Hrd1 and p97, results in retention of CDT in the ER and resistance to intoxication. Taken together, these data provide novel insight into how CDTs exit the ER and therefore gain access to their cellular targets.
| Cytolethal distending toxins (CDTs) are produced by a variety of Gram-negative pathogens including the oral pathogen Aggregatibacter actinomycetemcomitans, the sexually transmitted pathogen Haemophilus ducreyi, and the gastrointestinal pathogens, Escherichia coli and Campylobacter jejuni. These toxins belong to a larger, emerging group of intracellular-acting “cyclomodulins” whose expression is associated with increased persistence, invasiveness and severity of disease [1]–[7]. Rather than inducing overt cytotoxicity and tissue damage, cyclomodulins drive more subtle alterations in the host through changes in cell cycle progression. CDTs cause DNA damage in susceptible host cells, resulting in the induction of DNA repair signaling mechanisms including phosphorylation of the histone H2AX, cell cycle arrest at the G2/M interface and disruption of cytokinesis [8]. Inhibiting the cell cycle interferes with many functions of rapidly dividing eukaryotic cells, including lymphocytes and epithelial cells, which play a role in immunity and provide a physical barrier to microbial pathogens [5],[9],[10]. In cultured cells, the DNA damage response ultimately leads to apoptotic cell death, while in vivo, persistent DNA damage may give rise to infection-associated oncogenesis [11]. Although the cellular response to CDTs is well characterized [8], [12], the mechanism by which CDTs bind to host cells and ultimately gain access to their nuclear target is less clear.
CDTs generally function as complexes of three protein subunits, encoded by three contiguous genes (cdtA, cdtB, cdtC) in a single operon [13]. Consistent with the AB model of intracellular acting toxins [14], CdtB functions as the enzymatic A-subunit and possesses DNase I-like activity responsible for inducing DNA damage within the nuclei of intoxicated cells [15], [16]. CdtA and CdtC are thought to function together as the cell-binding B-moiety of AB toxins to deliver CdtB into cells [17]–[20].
To exert their cyclomodulatory effects, CDTs must be taken up from the cell surface and transported intracellularly in a manner that ultimately results in localization to the nucleus. Recent data suggest that the endosomal trafficking pathways utilized by CDTs from unrelated pathogens are different, but that all CDTs are trafficked in a retrograde manner through the Golgi apparatus and into the ER [21], [22]. CDTs and other retrograde trafficking toxins lack the ability to translocate themselves across the ER membrane and must therefore rely on host cellular processes to access their intracellular targets. Toxins such as cholera toxin, Shiga toxin, and ricin use a host-encoded protein quality control process known as ERAD [23]–[31]. ERAD is a normal physiological process by which misfolded proteins in the ER lumen and membrane are translocated to the cytoplasm for degradation by the proteasome. The core machinery driving ERAD in mammalian cells consists of the Hrd1/Sel1L ubiquitin ligase complex, the Derlin family of proteins and may also involve Sec61 [32]. Translocation of misfolded proteins across the ER membrane is energetically unfavorable and is facilitated by the AAA-ATPase p97 [33]–[35]. While toxins use various components of the ERAD pathway to exit the ER lumen, they avoid proteasomal degradation, thereby hijacking the host quality control mechanism to gain access to the cytosol.
In contrast to other retrograde trafficking toxins, several reports have suggested that ERAD does not play a role in the translocation of CDT across the ER membrane. Mutant cell lines deficient in the retrotranslocation of several retrograde trafficking toxins, such as cholera toxin, Pseudomonas aeruginosa exotoxin A, E. coli heat labile-toxin IIb, plasmid encoded toxin, and ricin were sensitive to CDT [22], [36]. Overexpression of Derlin-GFP fusions, which can act as dominant negative proteins to inhibit ERAD, did not block CDT intoxication [22]. Thermal stability of CdtB suggested that this catalytic subunit does not unfold prior to translocation and thus may not be an ERAD substrate [37]. Finally, CdtB was not found in the cytoplasm of intoxicated cells prior to nuclear localization, but rather was localized with ER membrane projections into the nucleus (i.e. nucleoplasmic reticulum), leading to the model that CDTs translocate directly from the ER lumen into the nucleoplasm [37]. Contrary to these data, others have described requirements for nuclear localization signals within the CdtB subunits, implicating a requirement for retrotranslocation to the cytosol prior to trafficking to the nucleus [38]–[40]. Identifying host factors required for translocation of CDT across the ER membrane would provide insight into mechanism of toxin entry; however, these data have been elusive [22], [41], [42].
Here we describe the results of two genetic screens aimed at identifying host genes required for intoxication by CDT from four human pathogens. These results implicate key components of the ERAD pathway in retrotranslocation of CDT and thereby provide insight into the mechanism by which host cells are intoxicated by this family of bacterial toxins.
In order to identify genes that confer sensitivity to CDT, we performed two separate forward somatic cell genetic screens. First, we utilized the frameshift mutagen ICR-191 to induce mutations in ten separate pools of CHO-pgs A745 cells (A745). Each pool of 1×106 cells was selected with 20 nM A. actinomycetemcomitans CDT (Aa-CDT), a toxin concentration high enough to cause death in parental cells. Five of the ten pools yielded Aa-CDT resistant clones; the most resistant clone isolated (CHO-CDTRA2) was resistant to the highest dose of Aa-CDT tested (Fig. 1a). Interestingly, CHO-CDTRA2 cells were also resistant to the highest dose of H. ducreyi CDT (Hd-CDT) tested (Fig. 1b) and more modestly resistant to CDTs from E. coli (Ec-CDT; Fig. 1c) and C. jejuni (Cj-CDT; Fig. 1d). To identify the gene responsible for CDT resistance in CHO-CDTRA2 cells, we utilized a high throughput cDNA expression-based complementation approach. A custom cDNA library consisting of approximately 3.7×103 arrayed clones was prepared from the mammalian gene collection [43]. Plasmid DNA was isolated from the library, normalized for concentration, plated individually into 384-well plates and reverse transfected into CHO-CDTRA2 cells. After 72 hours, the transfected cells were intoxicated with 20 nM Aa-CDT and immunostained using fluorescent anti-pH2AX antibodies to identify activation of CDT-mediated DNA damage response. Cells were stained with Hoechst 33342 to enumerate nuclei, imaged by automated fluorescence microscopy and scored using automated image analysis software. We identified Mus musculus Derlin-2 (Genbank ID: BC005682), a gene involved in the ERAD pathway, as able to complement the sensitivity of CHO-CDTRA2 cells to Aa-CDT. CHO-CDTRA2 cells were transduced with a retroviral vector encoding Derl2 to verify this finding and test whether Derl2 was able to complement resistance to the remaining three CDTs. CHO-CDTRA2 cells expressing Derl2 regained sensitivity to all four CDTs tested to near parental levels (Fig. 1a–d).
In a parallel effort to identify genes required for CDT intoxication, a retroviral mutagenesis approach was employed [44]. Approximately 1×107 A745 cells expressing the tetracycline repressor protein fused to the Krüppel associated box from human Kox1 (A745TKR) were transduced with murine leukemia virus (MLV) encoding the tetracycline repressor element at a multiplicity of infection of 0.1 and selected with 5 nM Hd-CDT, a toxin concentration high enough to cause death in parental cells. Two independent pools produced Hd-CDT-resistant clones. Subsequent characterization of one clone from each pool, CHO-CDTRC1 and CHO-CDTRF1, revealed that they were resistant to cell killing by the highest concentrations of the four CDTs tested (Fig. 2a–2d) as well as cell cycle arrest induced by lower CDT concentrations (Fig. S1). The site of mutational proviral integration was determined using a combination of sequence capture, inverse PCR and sequencing [44]. Proviral integration sites in the mutants were distinct; the mutagenic integration in CHO-CDTRC1 cells occurred between the first and second Derl2 exons and occurred in the opposite orientation in CHO-CDTRF1 cells between the fourth and fifth Derl2 exons (Fig. 2e). Overexpression of Derl2 in these mutants complemented sensitivity to all CDTs tested (Fig. 2a–2d, S2). In contrast, overexpression of the functionally related Derl1, which shares 51% homology and 35% amino acid identity with Derl2, failed to complement sensitivity to Hd-CDT in CHO-CDTRC1 cells (Fig. 2f). Both CHO-CDTRC1 and CHO-CDTRF1 mutant cells displayed decreased Derl2 expression by immunoprecipitation followed by western blot (Fig. 2g). Targeted deletion of Derl2 was performed in HeLa cells using the Cas9 clustered regularly interspaced short palindromic repeats (CRISPR) system [45]. HeLa cells lacking Derl2 were resistant to Hd-CDT (Fig. 2h, 2i). Additionally, siRNA mediated knockdown of Derl2 in HeLa cells rendered them resistant to Hd-CDT (data not shown). Although the demonstration of a direct physical interaction between Derl2 and CDT would support the hypothesis that Derl2 is part of a retrotranslocation apparatus, attempts to co-immunoprecipitate CDT with Derl2 were unsuccessful, likely due to very small quantities of CDT reaching the ER during intoxication.
Although Derlins have been most intensely studied as important factors in the translocation of ERAD substrates, these proteins have also been implicated in the trafficking of the plant toxin ricin from endosomes to the Golgi apparatus [46]. To identify which step of the CDT retrograde trafficking pathway was blocked in Derl2-deficient cells, the intracellular trafficking of Hd-CDT in parental A745TKR and mutant CHO-CDTRC1 and CHO-CDTRF1 cells was assessed by immunofluorescence microscopy as a function of time. After 10 minutes of intoxication, Hd-CdtB was clearly internalized into all the cell types tested (Fig. 2j–2l, S3). However, after 60 minutes, significantly more CdtB had localized to the nucleus of the parental A745TKR cells than in the Derl2-deficient CHO-CDTRC1 and CHO-CDTRF1 cells. In the CHO-CDTRC1 and CHO-CDTRF1 cells, Hd-CdtB was clearly localized to the ER, even after 60 minutes, but nearly absent within the ER of the parental A745TKR cells. Together, these data support a model that Derl2 is required for retrograde translocation of Hd-CdtB from the ER lumen.
Derl2 is part of the Hrd1-containing “retrotranslocon”, a protein complex that mediates retrotranslocation of ERAD substrates [47]. Indeed, Hrd1 was co-immunoprecipitated with Derl2 from wildtype but not Derl2-deficient cells (Fig. 3a). Similarly, Derl2 could be co-immunoprecipitated from wildtype cells, but not from cells in which Hrd1 was targeted by CRISPR (Fig. 3b–3c). Intoxication of Hrd1-deficient cells revealed that this gene, like Derl2, is required for cell killing by multiple CDTs (Fig. 3d–3g, S4). Interestingly, cells lacking Hrd1 displayed full sensitivity to intoxication by Cj-CDT (Fig. 3g). Similar to Derl2 deficient cells, deletion of Hrd1 resulted in retention of Hd-CDT in the ER 240 minutes post-intoxication (Fig. 3h–3j). These data suggest that the Derl2 and Hrd1-containing retrotranslocon is required for intoxication by multiple CDTs, implicating a role for the ERAD pathway in cellular entry for a subset of this family of toxins.
Derlins have been implicated in retrotranslocation of misfolded proteins out of the ER [35], [48]. In order to evaluate whether Derl2 might function by a similar mechanism to retrotranslocate CDTs, we investigated the importance of several Derlin functional motifs required for the retrotranslocation of previously characterized ERAD substrates. A carboxyl terminal SHP box (FxGxGQRn, where n is a non-polar residue) was recently demonstrated to be required for the interaction of Derlins with the AAA ATPase p97 [49], which provides energy to extract ERAD substrates from the lumen into the cytosol [33], [34], [50]. To assess the importance of p97-Derl2 interactions for the escape of CdtB from the cytosol, we tested whether Derl2 with a deletion of the C-terminus (Derl2ΔC) that removes the SHP box could complement Derl2 deficiency in CHO-CDTRC1 cells. Additionally, we tested a dominant negative form of Derl2 with a C-terminal GFP tag (Derl2-GFP)[22], [48]. Similar to what had been shown previously, Derl2ΔC was unable to bind p97 (Fig. 4a) [47], [49]. Further, Derl2-GFP was also unable to bind p97 (Fig. 4a). Surprisingly, intoxication studies revealed that despite failing to interact with p97, Derl2-GFP did not act as a dominant negative inhibitor, and that both Derl2-GFP and Derl2ΔC complemented sensitivity to Hd-CDT (Fig. 4b–d). These results suggest that Hd-CDT has evolved to use a Derl2-dependent retrotranslocation pathway that is independent of interaction between Derl2 and p97.
Although the interaction between Derl2 and p97 is not required for Hd-CDT retrotranslocation, this does not preclude a requirement for p97 in intoxication. To investigate this, dominant negative (R586A) and control (R700A) versions of p97 were overexpressed in 293 cells. Activity of the dominant negative p97 was confirmed by an increase in fluorescence signal from the ERAD substrate TCRαGFP [51] (Fig. 4e). Expression of dominant negative p97 caused a reduction in cell cycle arrest in G2 mediated by Hd-CDT, compared to control p97 (Fig. 4e). Consistent with a role for p97 in egress of CdtB from the ER lumen, expression of the dominant negative p97 resulted in retention of Hd-CDT in the ER after 240 minutes of intoxication (Fig. 4g–4i).
We next evaluated the importance of a second functional domain required for Derl2-mediated retrotranslocation of ERAD substrates. Derlins were recently classified as members of the rhomboid protease family of proteins, although they lack key residues required for proteolytic activity [49]. Rhomboid proteases are unique in that they contain an aqueous membrane-embedded cavity that allows for hydrolytic catalysis within the lipid bilayer [52]. Similar to other rhomboid proteases, Derl2 contains a “WR motif” (Q/ExWRxxS/T) in the sequence between the first and second transmembrane domains and a GxxxG motif in the sixth transmembrane domain. The WR motif protrudes laterally into the bilayer and plays a role in rearrangement of the local lipid environment [52], [53] while GxxxG motifs enable intra- and inter-molecular dimerization of transmembrane domains [52], [53]. Mutation of either of these domains in Derl1 renders it unable to retrotranslocate a constitutively misfolded protein to the cytosol for proteosomal degradation [49]. To test for a role for these motifs in CDT egress from the ER, Derl2 variants with single point mutations in the residues that comprise the WR and GxxxG motifs were expressed in Derl2 deficient CHO-CDTRC1 cells. Expression of Derl2 variants Q53A, W55A and T59A complemented the resistance to Hd-CDT in CHO-CDTRC1 cells to the same levels as that of wildtype Derl2 (Fig. 4f). One point mutant in the WR domain (R56A) and mutants in either residue of the GxxxG domain (G175V, G179V) failed to complement CHO-CDTRC1 cells; however, these mutants were poorly expressed as determined by immunoprecipitation and western blot, and therefore no conclusion can be made regarding a role for these residues (data not shown). These data suggest that although the WR motif is required for retrotranslocation of misfolded proteins by Derl1 [49], it is not required for retrotranslocation of Hd-CDT.
In order to provide insight into the mechanism by which Derl2 supports intoxication, we set out to identify Derl2 domains that are required for intoxication by Hd-CDT. Taking advantage of the knowledge that Derl1 is sufficiently divergent from Derl2 such that it cannot complement Derl2 deficiency (Fig. 2f), we constructed chimeric proteins comprised of fusions between homologous segments of Derl1 and Derl2 to map Derl2 segments that support intoxication by Hd-CDT. Replacing the C-terminal cytoplasmic tail of Derl2 with that from Derl1 (Derl21–187:Derl1189–251) gave a chimera that retained function and complemented sensitivity to Hd-CDT in CHO-CDTRC1 cells, consistent with a dispensable role for this domain (Fig. 5a). Likewise, CHO-CDTRC1 cells expressing a fusion protein in which the third ER luminal loop of Derl2 was replaced with that from Derl1 (Derl21–112:Derl1114–121:Derl1120–239) were sensitive to Hd-CDT, indicating that this domain is not required for intoxication (Fig. 5b).
In contrast, two distinct domains were identified in Derl2 that were each independently required for intoxication by Hd-CDT. Three fusion proteins comprised of Derl1 from the N-terminus through the second, fourth and fifth transmembrane domains respectively fused to the remaining portions of Derl2 (Derl11–88:Derl288–239; Derl11–138:Derl2138–239; Derl11–162:Derl2162–239) were unable to complement sensitivity to Hd-CDT in CHO-CDTRC1 cells, implicating a Derl2-specific sequence within the first 88 N-terminal residues as required for CDT intoxication (Fig. 5c). Second, a fusion protein consisting of Derl2 with the second ER luminal loop of Derl1 (Derl21–161:Derl1163–171: Derl2171–239) was unable to complement sensitivity, demonstrating that one or more of the six amino acids in the second luminal loop unique to Derl2 were also required for intoxication by Hd-CDT (Fig. 5b). We attempted to express several other Derl1:Derl2 chimeric proteins; however, these were expressed at levels lower than their wildtype counterparts and therefore these results were deemed inconclusive (data not shown). Taken together, these data identify two distinct domains of Derl2 required for Hd-CDT intoxication.
Similar to CDT, several other protein toxins such as ricin, Shiga toxin and cholera toxin rely on retrograde trafficking from the cell surface through the ER in order to gain access to the cytoplasm [25], [54]. Recently, RNAi-mediated repression of members of the Derlin family was shown to cause a slight resistance to ricin [26], [46] that was attributed to reduced trafficking from endosomes to the Golgi apparatus [46]. Similarly, the Derl2 deficient mutant cell line CHO-CDTRC1 displayed four-fold resistance to ricin, which was complemented by transduction with Derl2 (Fig. 6a). CRISPR mediated deletion of Hrd1 in 293 cells caused resistance to ricin, albeit to a lesser degree than resistance to Hd-CDT (Fig. 6b, 2h). This low-level resistance to ricin suggests that Derl2 and Hrd1 contribute to, but are not absolute requirements for ricin intoxication. In contrast, a high level of resistance to multiple CDTs resulted from Derl2 or Hrd1 deficiency (Fig. 2). Interestingly, the novel Derl2 SHP box- and WR motif-independence characterized for CDT was shared with ricin. Derl2ΔC and Derl2 WR mutants were able to restore sensitivity to Derl2 deficient CHO-CDTRC1 cells (Fig. 6c, 6d), suggesting that Derl2 may have multiple functions that are independent of the conserved WR motif and SHP box-mediated interactions with p97.
In order to gain access to their intracellular targets, retrograde trafficking toxins such as CDT bind the plasma membrane, are endocytosed and then trafficked though endosomes, the Golgi apparatus and ultimately the ER. At this point they must cross the formidable barrier posed by the host cellular membrane. The current model is that retrograde trafficking toxins commandeer the host ERAD pathway to cross the ER membrane, thereby gaining access to the cytosol. Various components of the ERAD machinery have been identified for cytoplasmic delivery of ricin, Shiga, and cholera toxins as well as for Pseudomonas aeruginosa exotoxin A [23]–[29]. These ERAD components include members of the HRD ubiquitin ligase complex, Hrd1 and Sel1L [27], [28], Derlins 1–3 [26], [30], [31], ER proteins involved in substrate recognition and unfolding of ERAD substrates [23]–[25], and the Sec61 translocon [26], [29]. Interestingly, different toxins appear to require distinct ERAD components, suggesting that multiple pathways exist by which toxins are translocated out of the ER lumen [26]. In contrast to these toxins, the pathway(s) by which CDTs exit the ER and ultimately gain access to the host nucleus was previously unknown. An ERAD-independent pathway was suggested based on failure of Derl1-GFP and Derl2-GFP fusion proteins to block intoxication by Hd-CDT, as well as susceptibility of mutant cells to CDT that were resistant to multiple other retrograde trafficking toxins [22], [36]. Here we provide evidence that three core components of the ERAD machinery, Derl2, Hrd1 and p97, are in fact required for intoxication by multiple CDTs and that abrogation of these key members of the ERAD pathway leads to Hd-CDT accumulation in the ER, consistent with a role in retrotranslocation.
The inability of Derl1 to complement Derl2 deficiency further enabled identification of novel domains within Derl2 required for intoxication by CDT. Derl2 is a six-pass transmembrane protein with three predicted loops in the ER lumen [49]. Replacing the third luminal loop from Derl2 with Derl1 sequences supported intoxication, indicating that this loop is not required, though we cannot exclude a more minor role. However, replacing the second luminal loop, which consists of just eight amino acids, two of which are conserved with Derl1, resulted in loss of function. This finding supports a key role for specific amino acids within this small domain in sensitivity to Hd-CDT. The first luminal loop may also be important, though chimeras consisting of this loop from Derl1 swapped with Derl2 and vice versa were not expressed and thus this could not be tested directly. However, replacing the first 88 N-terminal residues, inclusive of the first two transmembrane domains and the first luminal loop, with those from Derl1 did express well but failed to support Hd-CDT intoxication. This N-terminal region also contains the WR motif conserved among rhomboid proteases and required in Derl1 for retrotranslocation of misfolded proteins. However, the WR motif is conserved between Derl1 and Derl2 and point mutations within this WR motif in Derl2 still supported intoxication. These findings suggest that another functional domain exists within this region that is required for intoxication by Hd-CDT. Further studies are needed to determine whether additional requirements for intoxication map to the first luminal loop, the two transmembrane domains, or perhaps the N-terminal tail that extends into the cytosol.
In addition to identifying Derl2, Hrd1, and p97 as host factors usurped by CDTs to exit the ER, the studies presented here provide insight into the mechanism by which Derlin-GFP fusions act as dominant negative proteins. These constructs have been used to study the role of derlin family members in retrograde translocation of misfolded proteins, cytomegalovirus mediated degradation of class I MHC, infection by murine polyomavirus, and intoxication by ricin and cholera toxin [25], [30], [48], [49], [55]; however, the mechanism by which these constructs inhibit ERAD function was unknown. Interestingly, overexpression of Derl1-GFP or Derl2-GFP was previously shown to have no effect on the intoxication of HeLa cells by ricin or Hd-CDT, leading the authors to conclude that derlins are not required for these toxins [22], [25]. Similarly, we found that overexpression of Derl2-GFP (Fig. 4b) or Derl1-GFP (not shown) had no effect on CDT intoxication of parental A745TKR cells. Rather, overexpression of Derl2-GFP actually complemented sensitivity to Hd-CDT in Derl2-deficient CHO-CDTRC1 cells. Expression of Derl2ΔC complemented resistance to both ricin and CDT. The data presented here suggest that Derlin-GFP constructs act in a dominant negative manner by blocking interactions mediated by the C-terminus such as SHP box-mediated interactions with p97, and therefore may only exert dominant negative effects on ERAD and trafficking processes that require these interactions. Interestingly, although the interaction of p97 with Derl2 is not required for CDT interaction, p97 activity is indeed required for intoxication as expression of dominant negative p97 causes reduced sensitivity to Hd-CDT. p97 may supply energy for the retrotranslocation process that is common to both misfolded proteins and CDT through interactions with other proteins such as Hrd1 [50], or may be required for other entry or trafficking steps [56]. Determining the precise roles for this multifunctional protein requires more detailed studies and it remains possible that p97 contributes to more than one step in the intoxication pathway.
Previous somatic cell genetic screens identified twelve host genes required for intoxication by CDTs and ricin, but failed to identify Derl2, Hrd1 or p97 [41], [42]. The reason for this difference is unclear, though any single genetic model system is unlikely to provide a complete picture of such a complex biological process. Indeed, the host genes identified thus far only begin to explain the host processes required for cellular binding and entry by CDTs [21], [41], [42]. Only ten of the fifteen host factors identified thus far are required for intoxication by more than one CDT and of these, only two, sphingomyelin synthase 1 (SGSM1) [42] and Derl2 (Fig. 1, 2) have been shown to be required for all four CDTs tested here. These results suggest that various members of the CDT family have evolved distinct strategies to gain access to the host nucleus [21], [57]. Cj-CDT is the most evolutionarily divergent CDT studied here and displays unique requirements for host factors compared with Ec-, Aa-, and Hd-CDTs [42], [57]. Consistent with these prior findings, Cj-CDT had the least dependence on Derl2 and no requirement for Hrd1 (Fig. 2d). Future studies will likely identify many more host requirements for this family of toxins and provide further insight into their cellular entry pathways. Comparison of multiple members of the CDT family will elucidate a core set of host factors required for entry of all CDTs, but will also provide insight into unique solutions evolved by distinct CDTs to gain access to the host nucleus.
Chinese hamster ovary cells (CHO) and derivatives were maintained in F-12 media (Gibco) supplemented with 10% fetal bovine serum (Sigma Aldrich), 100 U/mL penicillin, 100 µg/mL streptomycin, 5 mM L-glutamine (Invitrogen) and 1 µg/mL doxycycline (Sigma Aldrich). HeLa and 293 cells (American Type Culture Collection) were maintained in Dulbecco's Modified Eagle Medium (DMEM; Cellgro) containing 25 mM HEPES, 4.5 g/L sodium pyruvate, 4.5 g/L glucose, 10% fetal bovine serum (Sigma Aldrich), 100 U/mL penicillin, 100 µg/mL streptomycin, and 5 mM L-glutamine (Invitrogen). In some cases, 293 culture medium was supplemented with 1% non-essential amino acids (Gibco). All cells were cultured at 37°C in a humid atmosphere containing 5% CO2.
To isolate chemically mutagenized CDT-resistant clones, ten pools of CHO-pgs A745 cells (A745, provided by Jeff Esko, UCSD) were treated with ICR191 (Sigma Aldrich) at a concentration high enough to kill 90% of the cells [58]. The resulting cells were counted, seeded at 1×106 cells per 10 cm plate and selected with 20 nM Aa-CDT. Resulting resistant cells were subjected to limiting dilutions to obtain single cell clones, expanded and reselected with Aa-CDT.
Selection of retrovirally mutagenized CDT-resistant clones was performed similar to a previously reported protocol [44]. Briefly, an Hd-CDT-sensitive clonal A745 cell line expressing tetR-KRAB (A745TKR) was established. Ten pools of 1×106 A745TKR parental cells were mutagenized by transduction with murine leukemia virus encoding the transcription response element TetO7 in the long terminal repeat (pCMMP.GFP-NEO-TRE) at a multiplicity of infection of 0.1. These pools were transcriptionally repressed at proviral integration sites for 96 hours in the absence of doxycycline then selected with 5 nM Hd-CDT for 24 hours. After selection, two of the ten pools yielded colonies; these colonies were picked, expanded and reselected with Hd-CDT. None of the CDT-resistant clones displayed doxycycline dependant sensitivity to CDT, so they were further maintained in the presence of doxycycline.
Mammalian cells were trypsinized, counted and seeded at approximately 1×103 cells per well in 384-well plates. The following day, medium was removed and toxin containing medium was added for 48 hours, followed by addition of ATPlite 1-step reagent (Perkin Elmer). Recombinant CDTs were cloned, expressed, and purified as described previously [57] and ricin was purchased commercially (List Biological Laboratories). Each biological replicate intoxication was performed in triplicate. Analysis of intoxication was performed either by quantitation of pH2AX immunofluorescence (as described previously [57]) or by using ATPlite reagent (Perkin Elmer) according to manufacturer recommendations. Intoxication data obtained by ATPlite reagent was normalized by dividing the luminescence relative light unit (RLU) signal of each replicate by the average of the unintoxicated control cells. All intoxication results presented are representative of at least three biological replicates.
In order to identify the location of the provirus in the CDT-resistant clones, genomic DNA was purified from each clone according to manufacturer recommendations (Qiagen), followed by digestion of 2 µg of genomic DNA with BamHI restriction enzyme (New England Biolabs). Digested genomic DNA was purified by column chromatography (Qiagen) and resuspended in 100 mM Tris-HCl, 150 mM NaCl, 50 mM EDTA, pH 7.5, containing 10 pmol biotinylated oligonucleotide complimentary to the 3′ pCMMP long terminal repeat (Sigma Aldrich; [Biotin]GTACCCGTGTTCTCAATAAACCCTC). The samples were heated to 95°C for 5 minutes then plunged on ice, followed by end over end rotation at 55°C for 14 hours.
Streptavidin coated magnetic beads were washed three times with 10 mM Tris-HCl, 2 M NaCl, 1 mM EDTA, pH 7.5 and added to the samples. Samples were vortexed for 0.5 hours at room temperature then the beads were immobilized on a magnet and supernatant removed, followed by three washes with 5 mM Tris-HCl, 1 M NaCl, 0.5 mM EDTA, pH 7.5 and resuspension in 100 µL water. The tubes were heated to 95°C in the presence of the magnet and the supernatant was removed and self-circularized with T4 DNA ligase according to manufacturer recommendations (Fermentas). PCR was performed using the following primers (GAGGGTTTATTGAGAACACGGGTAC and GTGATTGACTACCCGTCAGCGGGGTC) followed by nested PCR with the following primers (CGAGACCACGAGTCGGATGCAACTGC and GTTCCTTGGGAGGGTCTCCTCTG). Amplicons were run in a 1% agarose gel, bands were cut out, column purified (Qiagen) and sequenced (Genewiz).
In order to confirm that the MLV proviral integration occurred at the Derl2 locus, PCR amplification was performed on the genomic DNA from the retrovirally induced CDT resistant clones and the parental A745TKR cells. The primers used for amplification annealed to the fifth exon in the Derl2 open reading frame (CCATGAGCACCCAGGGCAGG) and either forward proviral elements (TGATCGCGCTTCTCGTTGGG) or reverse proviral elements (AGCGCATCGCCTTCTATCGC).
Murine Derl1 and Derl2 cDNA were subcloned by PCR amplifying using the following primers (restriction sites and kozak consensus sequences shown underlined and capitalized, respectively): Derl1 forward aaaagatctTCCACCATGtcggacatcggggactggttcagg; Derl1 reverse aaactcgagctggtctccaagtcggaagc; Derl2 forward aaaagatctTCCACCATGgcgtaccagagcctccggctgg; Derl2 reverse aaactcgagcccaccaaggcgctggccctcacc. The amplicons and the empty retroviral vector pMSCVpuro (Clontech) were digested with BglII and XhoI (New England Biolabs), gel purified (Qiagen) and ligated with T4 DNA ligase (Fermentas). The Gibson assembly reaction was utilized to construct the chimeric Derl1:Derl2 and Derl2:Derl1 [59]. Briefly, primers (Table S1) were designed to span the ends of the segments to be cloned by using the NEBuilder (TM) tool (New England Biolabs). PCR amplification and gel purification were performed to isolate segments to be cloned. Segments were assembled and cloned into pMSCVhygro (Clontech) by using Gibson assembly mastermix according to manufacturer's protocol (New England Biolabs). In order to generate retroviral vectors, plasmid DNA was purified and transfected into human 293 cells along with MLV gag/pol and vesicular stomatitis virus G-spike protein expression plasmids, as previously described [44]. 48 and 72 hours later, resulting retroviral particles were harvested, filter sterilized and used to transduce target cells in the presence of 8 µg/mL polybrene (Sigma Aldrich).
Approximately 1×107 cells were lysed in 1% digitonin, 25 mM Tris-HCl, 150 mM NaCl, 5 mM EDTA, 1 U/mL DNAse (Promega), and protease inhibitors (Roche), pH 7.0. The lysates were centrifuged at 14,000× G and supernatants were mixed with either 1 µg/mL rabbit α-Derl2 antibody or 5 µg/mL mouse anti-Hrd1/SYVN1 monoclonal antibody (Sigma Aldrich) and incubated overnight at 4°C with agitation. Protein-A sepharose beads (Santa Cruz Biotechnology) were washed, blocked with 5% bovine serum albumin (EMD Millipore) and incubated with the lysates for 1 hour at room temperature with agitation. Following incubation, the beads were washed three times, mixed with SDS reducing buffer and subjected to SDS-PAGE followed by transfer to PVDF membranes. Membranes were probed with either rabbit anti-Derl2 antibody (Sigma Aldrich) or rabbit anti-Hrd1 polyclonal antibody (Novus Biologicals) at a 1∶2000 dilution followed by HRP conjugated α-rabbit antibody (Invitrogen) to allow detection.
To test interactions between Derl2 and p97, 293 cells were seeded at 1×106 per 10 cm plate and allowed to adhere overnight. The following day, cells were transfected with 10 µg of plasmid DNA by calcium phosphate method. Seventy-two hours post-transfection, the cells were lysed in 1% digitonin lysis buffer (as described above). S-protein agarose beads were blocked in 5% bovine serum albumin for 1 hour and incubated with the lysates overnight at 4°C. The beads were washed with 0.1% digitonin, 25 mM Tris-HCl, 150 mM NaCl, 5 mM EDTA, pH 7.0 and protease inhibitors and then mixed with 1X SDS reducing buffer. Samples were subjected to SDS-PAGE, transferred to PVDF membranes then probed with rabbit anti-S-tag antibody (Cell Signal Technologies) and mouse anti-p97 antibody (Santa Cruz Biotechnology).
One hundred thousand Hela or 293 cells were transfected with 1 µg Cas9 expression plasmid (AddGene) [45] and 1 µg DNA derived from RT-PCR amplification of gRNA (Integrated DNA Technologies; Derl2 target sequence: AAGAAGTTCATGCGGACAT; Hrd1 target sequence: TGATGGGCAAGGTGTTCTT) using lipofectamine 2000 (Invitrogen) according to the manufacturer's protocol in a 12-well plate. Twenty four hours following transfection, cell culture medium was aspirated and replaced with complete DMEM containing 300 µg/mL of G418 to select for cells successfully transfected with the human codon optimized pcDNA3.3 TOPO vector carrying the Cas9 gene sequence and neomycin resistance cassette. After 72–96 hours under G418 selection the remaining viable cells were expanded to 10 cm tissue culture plates in complete DMEM without G418 and allowed to reach ∼80% confluence, after which toxin resistant cells were selected by intoxication with 5 nM Hd-CDT holotoxin. Cells surviving Hd-CDT intoxication were further expanded and the loss of either Derl2 or Hrd1 was confirmed by IP-western blot.
8 well-chambered slides (Nunc) were seeded with cells and allowed to adhere overnight. The following day, they were chilled on ice for 30 minutes then incubated on ice with 100–200 nM Hd-CDT for 30 minutes. The monolayers were washed with ice-cold PBS pH 7.4 (Lonza), and then incubated at 37°C with complete medium. After 60 minutes at 37°C, the cells were washed with ice-cold PBS pH 7.4, and fixed with ice-cold 2% formaldehyde (Sigma). After fixing for 30 minutes at room temperature, the cells were permeabilized by incubating in PBS 7.4 containing 0.1% Triton X-100 for 15 min, and blocked with 3% BSA (Sigma) for 30 minutes. To probe for Hd-CdtB, cells were incubated with rabbit polyclonal anti-Hd-CdtB antibodies (generated by The Immunological Resource Center, University of Illinois, Urbana, IL) at 4°C overnight, followed by incubation with goat anti-rabbit antibody labeled with either Alexa Fluor 488 or Alexa Fluor 568 (Invitrogen) at room temperature for 2 hours. Where indicated, the ER is labeled with either Alexa Fluor 594 conjugated Concanavalin A (Invitrogen) or mouse monoclonal anti-calreticulin antibody (Abcam) at 4°C overnight, followed by incubation with goat anti-mouse Alexa Fluor 647-labeled antibody (Invitrogen). Where indicated, nuclear counterstaining was performed by either incubating with DAPI for 30 minutes at room temperature or transfecting with 1 µg of plasmid encoding Histone-GFP (pH2B-GFP; Addgene, Cambridge, MA). The slides were mounted with ProLong Gold antifade reagent (Invitrogen) and images were collected using DIC/fluorescence microscopy and deconvoluted by using SoftWoRX constrained iterative deconvolution tool (ratio mode), and analyzed using Imaris 5.7 (Bitplane AG). For each cell, images were collected from an average of 30 z-planes, each at a thickness of 0.2 µm. Nuclear localization analysis was conducted by using the DeltaVision SoftWoRx 3.5.1 software suite. For nuclear localization, the percentage of Hd-CdtB localization into nucleus in parental and Derl2 deficient cells were calculated from approximately 30 cells from each group over at least two independent experiments. To test the colocalization of Hd-CdtB with the endoplasmic reticulum, results were expressed as the localization index, which was derived from calculating the Pearson's coefficient of correlation values, which represent the colocalization of Hd-CdtB and the ER in each z plane of the cell. In these studies, a localization index value of 1.0 indicates 100% localization of Hd-CdtB to the ER, whereas a localization index of 0.0 indicates the absence of Hd-CdtB localization to the ER. The localization index was calculated from the analysis of a total of 30 images collected over at least two independent experiments.
One hundred thousand 293 cells expressing T-cell receptor alpha fused to green fluorescent protein were seeded the day prior to transfection with 1 µg of plasmid encoding either dominant negative p97 (R586A) or control p97 (R700A) co-expressed with CD4 as a surface marker of positive expression (plasmids generously provided by Ron Kopito, Stanford University). Seventy-two hours after transfection, the cells were intoxicated with a concentration of Hd-CDT sufficient to cause cell cycle arrest in 48 hours. Intoxicated cells were rinsed with PBS, detached from the wells with PBS+1 mM EDTA, rinsed with PBS again and incubated with phycoerythrin conjugated rabbit anti-CD4 antibody (Invitrogen) in PBS+3% bovine serum albumin on ice for 30 minutes. Following staining, the cells were washed with PBS, fixed with 1% formaldehyde, washed with PBS again and stained with Hoechst 33342 for 10 minutes. Cells were then washed with PBS, resuspended in PBS and analyzed for phycoerythrin, Hoechst and GFP fluorescence by flow cytometry (LSR II; Becton Dickinson). Cell cycle analysis was performed on CD4 expressing cells.
The half maximal lethal dose (LD50) of ricin intoxication was calculated by log transforming ricin concentrations and calculating sigmoidal variable slope dose response curves using the least squares (ordinary) fitting method. Paired t-tests were performed on average LD50 values calculated from three independent experiments performed in triplicate to determine two tailed p-values. Data analysis was performed using Prism version 5.0d (GraphPad software).
|
10.1371/journal.pntd.0003457 | Effectiveness of Routine BCG Vaccination on Buruli Ulcer Disease: A Case-Control Study in the Democratic Republic of Congo, Ghana and Togo | The only available vaccine that could be potentially beneficial against mycobacterial diseases contains live attenuated bovine tuberculosis bacillus (Mycobacterium bovis) also called Bacillus Calmette-Guérin (BCG). Even though the BCG vaccine is still widely used, results on its effectiveness in preventing mycobacterial diseases are partially contradictory, especially regarding Buruli Ulcer Disease (BUD). The aim of this case-control study is to evaluate the possible protective effect of BCG vaccination on BUD.
The present study was performed in three different countries and sites where BUD is endemic: in the Democratic Republic of the Congo, Ghana, and Togo from 2010 through 2013. The large study population was comprised of 401 cases with laboratory confirmed BUD and 826 controls, mostly family members or neighbors.
After stratification by the three countries, two sexes and four age groups, no significant correlation was found between the presence of BCG scar and BUD status of individuals. Multivariate analysis has shown that the independent variables country (p = 0.31), sex (p = 0.24), age (p = 0.96), and presence of a BCG scar (p = 0.07) did not significantly influence the development of BUD category I or category II/III. Furthermore, the status of BCG vaccination was also not significantly related to duration of BUD or time to healing of lesions.
In our study, we did not observe significant evidence of a protective effect of routine BCG vaccination on the risk of developing either BUD or severe forms of BUD. Since accurate data on BCG strains used in these three countries were not available, no final conclusion can be drawn on the effectiveness of BCG strain in protecting against BUD. As has been suggested for tuberculosis and leprosy, well-designed prospective studies on different existing BCG vaccine strains are needed also for BUD.
| After tuberculosis and leprosy, Buruli Ulcer Disease (BUD) is the third most common human mycobacterial disease. The only available vaccine that could be potentially beneficial against these diseases is BCG. Even though BCG vaccine is widely used, the results on its effectiveness are partially contradictory, probably since different BCG strains are used. The aim of this study was to evaluate the possible protective effect of BCG vaccines on BUD. The present study was performed in three different countries and sites where BUD is endemic: in the Democratic Republic of the Congo, Ghana, and Togo from 2010 through 2013. The large study population was comprised of 401 cases with laboratory confirmed BUD and 826 controls, mostly family members or neighbors. Considering the three countries, sex, and age, the analysis confirmed that the BCG vaccination did not significantly decrease the risk for developing BUD or for developing severe forms of BUD. Furthermore, the status of BCG vaccination was also not significantly related to duration of BUD or to time to healing of lesions. In our study, we could not find any evidence of a protective effect of routine BCG vaccination on BUD.
| Buruli Ulcer Disease (BUD), caused by Mycobacterium ulcerans, is an infectious disease affecting skin, subcutanous adipose tissue, and in rare cases, bones. It is one of the 17 neglected tropical diseases as defined by the World Health Organization (WHO). BUD has been reported in 33 countries, with a major endemic focus in West and Central Africa. The exact mode of transmission of M. ulcerans is still unknown. However, recent studies suggest that the pathogen is acquired from the environment with different modes of transmission in different geographic areas and epidemiological settings, as shown in a systematic review [1]. Consequently, except for early case detection, confirmation, and treatment, primary measures to prevent BUD are currently lacking. Furthermore, no effective vaccine against BUD is available so far [2].
After tuberculosis and leprosy, BUD is the third most common mycobacterial disease among immunocompetent human hosts. The only available vaccine against these diseases contains live attenuated bovine tuberculosis bacillus (M. bovis), also called Bacillus Calmette-Guérin (BCG), named after its inventors [3]. Calmette and Guérin began their research for an antituberculosis vaccine at the Pasteur Institute in Lille, France, in 1900. The first use in humans dates from 1921, when Turpin and Weill-Hallé vaccinated infants at the Charité Hospital in Paris by oral and later also by subcutaneous and intracutaneous routes [4], [5]. From 1924 to 1928, 114,000 infants were vaccinated without serious complications, however with limited effectiveness on preventing tuberculosis [6].
From the late 1940s onward, many studies appeared providing evidence for the effectiveness of BCG for tuberculosis, with widely varying results ranging from 0% to 80% effectiveness for vaccinated adults [5], [7]. Due to these disparate results, two principal hypotheses were discussed. The first one stated that exposure to various environmental mycobacteria could itself provide some protection against tuberculosis and affect the immune system in various ways, implying that BCG could not improve greatly upon that background [5], [8]. The second hypothesis attributed the differences to variation between strains of BCG [5], [9]. It was recognized that strains produced by diverse manufacturers differed in microbiological properties, as shown in a review [10]. Hence it was not unreasonable to suggest that these might be reflected in differences in immunogenicity [5], [11]. However, in children, the effectiveness of BCG was estimated to be 50%, or even up to 80% effective in preventing tuberculous meningitis and miliary tuberculosis as shown in a meta-analysis [12] and two other publications [13]–[14].
Worldwide, over 90% of children are immunized with BCG, making it the most commonly administered vaccine, with more than 12 million doses being used each year [15]. Although BCG has been administerd to more people than any other vaccine, its history has been clouded by variable efficacy and reports of strain variability [16]. BCG has never been cloned, and there are now several different BCG seed strains in use, produced by more than 40 manufacturers [17]. Nineteen major vaccine strains are described in the literature, whereas the original vaccine from 1921 was lost: BCG-Moreau (“Brazilian strain”: 1924), BCG-Russia (BCG-Moscow or “Russian strain”: 1924; genetically identical to BCG-Bulgaria or BCG-Sophia: 1950s), BCG-Japan (“Tokyo strain 172”: 1925), BCG-Romania (1925), BCG-Sweden (“Goethenburg strain”: 1926), BCG-Birkhaug (1927), BCG-Danish (BCG-Denmark or BCG-Copenhagen or “Danish strain 1331”: 1931), BCG-Tice (BCG-Chicago or “Tice strain”: 1934), BCG-Frappier (BCG-Montreal: 1937), BCG-Phipps (BCG-New York, BCG-Park, BCG-Philadelphia: 1938), BCG-Prague (“Czechoslovakian” strain: 1947), BCG-China (BCG-Beijing: 1947 or 1948), BCG-Shanghai (1948), BCG-Lanzhou (1948), BCG-Connaught (BCG-Toronto or “Theracys strain”: 1948), BCG-Polish (1950s), BCG-Glaxo (“BCG-London F10” or “Glaxo strain 1077”: 1954), BCG-Pasteur (“Pasteur strain 1173P2”: obtained in 1961), BCG-Mexico (1970), BCG-Mérieux (1989).
The following eight strains are the most common BCG strains in present use: Moreau, Russia, Japan, Danish, Tice, Connaught, Glaxo, and Pasteur. These five BCG strains represent more than 90% of the global BCG production: Russia, Japan, Danish, Glaxo, and Pasteur [16], [18]. According to Ritz et al., for some BCG strains (Russia, Japan, Danish, Prague, Glaxo, and Pasteur) results from at least nine studies were published from each strain, whereas for others, very little or no study results were found in the literature [15]. Studies and observations have shown that BCG-Pasteur and BCG-Danish are “strong” vaccines with higher immunogenicity and with greater complication rates than BCG-Japan or “weak” vaccines as BCG-Russia or BCG-Glaxo [18], [19].
Each of these BCG vaccines is produced in a different manner, and they are recognized to differ in various qualities, such as the proportion of viable cells per dose [5], [10]. However, the majority of the world's population is supplied with BCG vaccines procured by UNICEF (The United Nations Children's Fund) on behalf of the GAVI Alliance (formerly “Global Alliance for Vaccines and Immunization”). UNICEF uses only four BCG vaccine suppliers, who produce only three different BCG vaccine strains: BCG-Russia, BCG-Japan, and BCG-Danish [5].
BCG is also recognized to cause cross-protection against leprosy, as shown in a review [20] and in a meta-analysis [21]. That meta-analysis found that experimental studies demonstrated an overall protective effect of 26% (95% CI 14–37%) and that observational studies overestimated the protective effect [21]. Over the years, several vaccine trials using BCG have been performed to establish its limited protective effect against leprosy, often in combination with M. leprae or related mycobacterium vaccines. BCG was as good as, or superior to the other mycobacterium vaccines [22], [23].
Additionally, cross-protection of BCG against BUD was also shown in several studies, but their results are partially contradictory. An earlier clinical trial in Uganda showed an immune protection by BCG vaccination lasting six months [24]. The findings are consistent with another clinical trial in Uganda concluding that BCG vaccination provides only short-term protection against BUD [25]. In two studies in Benin, BCG was shown to be protective against more severe BUD, notably osteomyelitis [26], [27]. A study performed in Cameroon concluded that BCG appeared to protect children against more severe forms of BUD with multiple lesions [28]. However, none of these studies described the BCG strain used for vaccination.
In a mouse model experiment, the potential mechanisms for cross-protection were studied. A study identified and characterized the M. ulcerans homologue of the important protective mycobacterial antigen 85 (Ag85A) from BCG. This antigen was sufficiently conserved to allow cross-reactive protection, as demonstrated by the ability of M. ulcerans-infected mice to exhibit strong cellular immune responses to both BCG and its purified Ag85 complex [29]. It was also shown, that the BCG vaccine offered short-term protection against experimental footpad infections of mice with M. ulcerans, and that duration of this protection could not be prolonged by a booster vaccination [30]. Another experiment using a mouse model observed that BCG vaccination significantly delayed the onset of M. ulcerans growth and footpad swelling through the induction of an earlier and sustained IFN-γ triggered T cell response in the draining lymph node. BCG vaccination also resulted in cell-mediated immunity in M. ulcerans-infected footpads [31].
Two epidemiological studies, performed in Benin, could not find any evidence of a protective effect of routine BCG vaccination against BUD. In the second study, in persons aged >5 years, a BCG scar even resulted in a risk factor of 2.5 for BUD compared with those without a BCG scar [14], [32]. The first two epidemiological studies on the effectiveness of BCG vaccines on BUD performed in Ghana did not show any significant difference between cases and controls regarding their BCG vaccination status [33], [34]. None of these studies described the BCG strain used for vaccination.
Although many studies on the BCG vaccine were performed, the results regarding the vaccine's effectiveness against mycobacterial diseases including BUD differ immensely. Based on this unclear situation, the present case-control study was conducted with a large study population in the Democratic Republic of the Congo (DR Congo), Ghana, and Togo. In these three countries, only three different BCG strains were used since BCG was introduced from 1978 through 1984: BCG-Russia, BCG-Japan, and BCG-Danish. In the context of the EC-funded research project “BuruliVac” (FP7/2010–2013; grant agreement N° 241500), the aim of the present study is to evaluate possible protective effectiveness of routine BCG vaccination containing live attenuated bovine tuberculosis bacillus M. bovis on BUD in the DR Congo, Ghana, and Togo.
BuruliVac was founded in 2009 as consortium of 16 European and African partners. As there is currently no existing vaccine lead candidate available, BuruliVac aimed to identify and develop new vaccine candidates of three different types: (1) Mycolactone-directed vaccines, (2) attenduated live vaccines, and (3) subunit protein vaccines. Furthermore, BuruliVac evaluated the resulting vaccine candidates using bioinformatics, applied genomics and proteomics, and subjected them to consecutive test systems. BuruliVac was funded by the European Commission under the 7th Framework Programme of the European Union [35].
The present study was performed in the DR Congo, Ghana, and Togo. These three countries follow the WHO recommendations for routine immunization, which are part of their national immunization programs. This includes the advice to administer the one-time BCG vaccine intracutaneously, as soon as possible, either at birth or directly after, but not later than twelve months after birth, because at that age the vaccination is usually of limited benefit, although it is not harmful or contraindicated. Booster shots are not recommended [36]. The WHO estimates the BCG coverage rates in these three African countries as follows: 78% in the DR Congo, 98% in Ghana, and 97% in Togo [37].
This study consists of data collected at the following three sites, which are members of BuruliVac. The Institut Médical Evangélique (IME) de Kimpese in the DR Congo has implemented the “Project Ulcère de Buruli”. Since 1999, the General Reference Hospital (GRH) of the IME, located in the Songololo Territory, 220 km southwest of Kinshasa, regularly admits BUD cases. In 2004, the GRH launched a specialized BUD program offering in-patient treatment free-of-charge and supplementary aid. The principal aims of this project are the improvement of patient care for BUD patients admitted to the IME and the promotion of early community-based detection of suspected BUD cases. Patients and controls were recruited from Kimpese and Nsona-Mpangu health zones, both located in the Songololo Territory, Province of Bas-Congo [38], [39].
The Department of Medicine and the Kumasi Centre for Collaborative Research (KCCR) of the School of Medical Sciences at the Kwame Nkrumah University of Science and Technology (KNUST) are based in Kumasi, Ghana. They are involved with BUD in the areas of training, diagnostic confirmation, provision of specialist care for BUD patients in disease endemic districts, recruitment of patients and controls from the Ahafo Ano North, Asante Akim North, Atwima-Nwabiagya, and the Upper Denkvira districts, which are all within 70 km of the Ashanti regional capital Kumasi [40], [41].
The Centre Hospitalier Régional Maritime (CHR Maritime) in Tsévié, Togo, collaborates since 2007 with the German Leprosy and Tuberculosis Relief Organization, Togo office (DAHWT). This collaboration is supported by the Togolese National Buruli Ulcer Control Program (“Programme National de Lutte contre L'Ulcère de Buruli – Lèpre et Pian” [PNLUB-LP]), in the area of training, active case finding, laboratory confirmation, and treatment of BUD. In 2007, the CHR Maritime was appointed National Reference Centre for BUD in Togo [42], [43].
In the DR Congo, BCG vaccination was routinely introduced in 1984. The following BCG strains were used for vaccinations: 1984–2003: BCG-Russia (equivalent to “BCG-Bulgaria”; produced by Bulbio [BB-NCIPD], Sofia, Bulgaria, and by Serum Institute of India); 2004: BCG-Japan (produced by Japan BCG Laboratory); 2005–2009: BCG-Japan (produced by Japan BCG Laboratory) and BCG-Russia (produced by Serum Institute of India); 2010–2011: BCG-Japan (produced by Japan BCG Laboratory); 2012: BCG-Japan (produced by Japan BCG Laboratory) and BCG-Russia (produced by Serum Institute of India); January to July 2013: BCG-Russia (equivalent to “BCG-Bulgaria”; produced by Bulbio [BB-NCIPD], Sofia, Bulgaria); August and September 2013: BCG-Russia (produced by Serum Institute of India); October 2013 to date: BCG-Bulgaria which is BCG-Russia (produced by Bulbio [BB-NCIPD], Sofia, Bulgaria).
In Ghana, BCG vaccination was routinely introduced in 1978. The following BCG strains were used for vaccinations: 2007: BCG-Danish (produced by Danish Statens Serum Institute); 2008–2009: BCG-Bulgaria = BCG-Russia (produced by Bulbio [BB-NCIPD]; 2010 to date: BCG-Japan (produced by Japan BCG Laboratory, Tokyo, Japan). Exact data on BCG strains used in Ghana from 1978 through 2006 are not available.
In Togo, BCG vaccination was routinely introduced in 1980. The following BCG strains were used for vaccinations: 2004: BCG-Japan (produced by Japan BCG Laboratory, Tokyo, Japan); 2004–2009: BCG-Russia (produced by Serum Institute of India); 2010 to date: BCG-Russia (equivalent to “BCG-Bulgaria”; produced by Bulbio [BB-NCIPD], Sofia, Bulgaria, and by Serum Institute of India). Exact data on BCG strains used in Togo from 1980 through 2003 are not available.
In these three study sites, the recruitment of both BUD cases (among patients presenting with “clinically suspected” BUD lesions) and healthy controls was conducted. The present retrospective case-control study defined cases (CA) as patients affected by BUD, whose diagnosis was confirmed in laboratory by microscopy, IS 2404 polymerase chain reaction (PCR), or culture. Any CA had at least one positive test result. Patients who were “clinically suspected” (CS) for BUD, but without laboratory confirmation (i.e. none of the tests results was positive) were not considered in the study population. The controls (CO) were defined as healthy persons without any history of BUD in the past, who were in close relationship with the CA (see in next chapter).
In the time period from February 2010 through April 2013, data from 1,335 individuals were collected. Out of them, 406 (30.41%) were CA, 103 (7.72%) were CS, and 826 (61.87%) were CO. From these data, 622 participants (128 CA: 20.58%; no CS; 494 CO: 79.42%) were from the DR Congo, 504 participants (196 CA: 38.89%; 65 CS: 12.90%; 243 CO: 48.21%) were from Ghana, and 209 participants (82 CA: 39.23%; 38 CS: 18.18%; 89 CO: 42.58%) were from Togo. Four CA from Ghana and one CA from Togo had unknown BCG status and were excluded out of the study.
Consequently, the study population was comprised of 1,227 participants (401 CA: 32.68%; 826 CO: 67.32%), including 622 from the DR Congo (128 CA: 20.58%; 494 CO: 79.42%), 435 from Ghana (192 CA: 44.14%; 243 CO: 55.86%), and 170 from Togo (81 CA: 47.65%; 89 CO: 52.35%). The 826 CO were in the following relationship with the CA: 225 (27.24%) were family members, 518 (62.71%) neighbors, 32 (3.87%) friends or classmates, and 51 (6.17%) were others or those with unspecified relationship.
Data collection was conducted by means of the WHO “BU01” form, and standardized project-specific “BuruliVac” laboratory data entry forms (Form S1). All socio-demographic, clinical, and laboratory data were entered in a web-based database specifically designed for the “BuruliVac” project [43]. Following WHO guidance, the categories of BUD were defined as follows: Category I were single lesions <5 cm in diameter; Category II were single lesions between 5 and 15 cm in diameter; Category III were single lesions >15 cm in diameter, multiple lesions, lesions at critical sites or osteomyelitis [44].
The BCG vaccination status was assessed from all CA and CO of the study population by examining both sides of the arms or shoulders, and if they presented a scar typical for vaccination with BCG or not, but not by documents such as vaccination certificates or hospital registers. Former studies that evaluated the presence or absence of BCG scars to determine vaccination status reported that scars develop in most vaccinated persons, with scarring rates of >80% [14], [45]–[47].
In the DR Congo, fine needle aspirates were only collected from non-ulcerative lesions. Routinely, a direct smear was conducted at peripheral health centers from the first fine needle aspiration (FNA) and then the sample was stored in transport media (7H9 and PANTA liquid) and forwarded to IME for microscopy and culture. The second FNA (if possible) or a suspension was forwarded to the Institut National de Recherche Biomédicale (INRB) in Kinshasa via IME, where microscopy and IS 2404 real-time PCR was performed. Similar procedures were applied for swabs and tissue biopsies, however stored in semi-liquid transport medium (Dubos and PANTA semi-liquid).
In Ghana and Togo, diagnostic samples were collected according to standardized procedures [43]. Briefly, swabs were collected by circling the entire undermined edges of ulcerative lesions. Fine needle aspirates were collected from the center of non-ulcerative lesions or from undermined edges of advanced ulcerative lesions with scarred edges. Punch biopsy samples were only collected from advanced ulcers with scarred edges if fine needle aspirates were tested negative by PCR according to recent WHO recommendation [48].
Standardized specimen collection bags including swabs, biopsy punches, syringes and needles, slides, containers with transport media (700 µl [swab and punch biopsy samples], 300 µl [FNA samples] CLS [cell lysis solution, Qiagen, Hilden, Germany] for PCR samples; 4 ml PANTA transport medium for mycobacterial cultures [Ghana only]) and data entry forms were provided to the study sites in Ghana and Togo [49]–[57].
Samples for PCR analysis in CLS and for mycobacterial culture in PANTA transport medium were transported at ambient temperature in an upright position in custom-made specimen collection bags from the field to the laboratories from the two study sites in Ghana and one study site in Togo, within a maximum of 48 hours and stored at 4–8°C until further processing. Slides for microscopy were transported in slide boxes at ambient temperature to the laboratory.
Direct smears for microscopy were prepared from swab and fine needle aspirates at the laboratory (Ghana: KCCR; Togo: CHR Maritime), and were subjected to Ziehl-Neelsen staining. Slides were analyzed according to the WHO recommended grading system [56], [58] including quality assurance measures (re-reading of slides at INH and DITM). For PCR analysis, DNA was prepared using the Gentra Puregene DNA extraction kit (Qiagen) with minor modifications of the manufacturer's protocol [59], [60].
In the study site in the DR Congo, the Maxwell 16 DNA extraction procedure was carried out with the Maxwell 16 Tissue DNA Purification Kit and the Maxwell 16 Instrument, according to manufacturer's instructions: 200 µl of specimen was added to 200 µl of lysis buffer (10 mM Tris-HCl pH 7.5, 10 mM NaCl, 10 mM EDTA, 50 ml 10% SDS solution) and 10 µl proteinase K (20 mg/ml) and incubated overnight at 60°C in a shaker incubator. IS2404 qPCR was performed on an Applied Biosystems 7500 Fast Real-Time PCR System using the method previously described by Fyfe et al. [61].
In the study sites in Ghana and Togo, the dry-reagent-based (DRB) IS 2404 PCR (INH, KCCR) was applied, accompanied by external quality assurance through IS 2404 qPCR at DITM. Briefly, for DRB-PCR the oligonucleotides MU5 and MU6 were lyophilized in reaction tubes. Illustra PuReTaq Ready-To-Go PCR beads (GE Healthcare, Munich, Germany) were added and dissolved in water before adding template DNA [50], [51], [60]. IS2404 qPCR was performed as recently described using a BioRad CFX96 real-time PCR detection system [61], [62]. All PCR assays included negative extraction controls, as well as positive, negative (no template) and inhibition controls.
In Kimpese, the DR Congo, the ethical clearance was obtained through the “Comite d'Éthique” of the “Ecole de Santé Publique” of the University of Kinshasa (Ref. No. ESP/CE/057/2010). In Kumasi, Ghana, the ethical clearance was obtained through the Committee on Human Research Publication and Ethics of the College of Health Sciences of the Kwame Nkrumah University of Science and Technology (Ref. No. CHRPE/91/10). In Tsévié, Togo, the ethical clearance was obtained through the national Togolese ethics committee (“Comité de Bioéthique pour la Recherche en Santé”) at the University of Lomé (14/2010/CBRS) and the study was approved by the “Ministère de la Santé de la République Togolaise” Lomé, Togo (Ref. No. 0009/2011/MS/DGS/DPLET). All samples analyzed in this study were collected for diagnostic purposes within the EC funded research project “BuruliVac”. Written informed consent was obtained from all study participants, or their guardians if aged <18 years, according to the recommendations of the respective ethical committees. In case of illiterates, informed consents were countersigned by means of thumb prints.
All data assessed at these three study sites were entered into the web-based database of BuruliVac and descriptively analyzed with Excel 2007 (Microsoft, Redmond, WA). The hypothesis of the present study was to evaluate associations between the presence of BCG scars (independent variable), which are caused by BCG vaccinations, and risk for BUD (dependent variable). Bivariate approximative tests (χ2-tests) and exact test (Fisher's tests) were conducted using EpiInfo, version 3.3.2. (Centers for Disease Control and Prevention, Atlanta, GA) and multiple logistic regression by Stata software, version 9.0. (Stata Corporation, College Station, TX) and. Significant differences were defined as p-values below 0.05.
Among the study population of 1,227 individuals (401 CA and 826 CO) males comprised 45.56% (559), which was not significantly (p = 0.57) different between CA (44.39%: 178) and CO (46.13%: 381). Stratification by the three countries found no significant differences in the proportion of males among CA and CO. Among the 401 CA, the range of age was 1 to 78 years (y) and the median of age was 13 y (25% percentile: 8 y, 75% percentile: 27 y). Among the 826 CO, the range of age was 1 to 90 y and the median of age was 16 y (25% percentile: 9 y, 75% percentile: 30 y). Age distribution in CA and CO was significantly (p = 0.01) different, as the CA were younger than the CO: Age group (AG) 0–9 y (30.42% in CA vs. 26.63% in CO), AG 10–19 y (34.66% vs. 28.81%), AG 20–39 y (21.95% vs. 29.78%), and AG 40–90 y (12.95% vs. 14.77%). Stratified by the three countries, significant differences (p<0.01 each) of the proportions of these four AG among CA and CO were found in Ghana and Togo, but not in the DR Congo (p = 0.97) (Table 1).
Among the 401 CA, 383 (95.50%) were detected with a single lesion, 15 (3.74%) with two lesions each, two (0.50%) with three lesions each, and one (0.25%) with four lesions. Out of them, 167 (41.65%) CA had non-ulcerative and 234 (58.35%) ulcerative lesions. The proportion of detected non-ulcerative lesions was as follows: nodules (74: 18.45%), plaques, (58: 14.46%), edema only (27: 6.73%), papules (7: 1.75%), and osteomyelitis (1: 0.25%).
Among the 401 CA, microscopy was performed for 399 (99.50%), PCR for 384 (95.76%), and culture for 159 (39.65%). The sensitivity of the three tests was as follows: PCR 97.14% (373/384), microscopy 69.42% (277/399), and culture 35.22% (56/159). Of 384 (95.76%) CA with known lesion sites, 2.86% (11/384) were on the face, 41.41% (159) on the upper limbs, 11.46% (44) on the trunk, and 44.27% (170) on lower limbs. The right lower limb (26.30%: 101) was significantly (p<0.01) more frequently affected than the left lower limb (17.97%: 69), whereas no significant differences where found between presence of lesions on the right and left side of the body for the face, upper limbs, or trunk.
Among 401 CA, 175 (43.64%) had no BCG scar (CAscar), whilst 226 (56.36%) had BCG scar (CAno_scar). Among 826 CO, 277 (33.54%) had no BCG scar (COscar), whilst 549 (66.46%) had BCG scar (COno_scar). The proportion of those with a BCG scar was significantly (p<0.01) higher among the CO than among CA. When stratified by the three countries, a significant difference of the proportion of individuals with a BCG scar among CA and CO was only found in Ghana (p = 0.03), and not in the DR Congo (p = 0.22) or in Togo (p = 0.67) (Table 1).
Stratified by four age groups, a significantly higher proportion of those with a BCG scar among CO was only found in AG 10–19 y and AG 40–90 y (p<0.01 each). Stratified by the three countries and four age groups, a significantly higher proportion of those with BCG scar among CO was only found in Ghana in 10–19 y (p = 0.03) (Table 1). Multivariate analysis confirmed that the independent variables country (p<0.01), age (p<0.01), and status of BCG vaccination (p = 0.02) did significantly influence the dependent variable, if an individual develops BUD (CA) or not (CO).
Stratified by sex, a significantly higher proportion of those with a BCG scar among CO was only found among females (p<0.01), but not males (p = 0.09). When stratified by sex and by country, no significant difference of that proportion was found. After stratification by three countries, two sexes, and four age groups, no significant correlation was found between the presence of BCG scar and BUD status of individual (CA or CO).
Among the 175 CAscar representing 85.14% (149/175) and 226 CAno_scar, representing 77.43% (175/226) the BUD category was recorded. The proportions of CA with category I, II and III among CAscar were respectively 48.99% (73), 41.61% (62), and 9.40% (14), whereas these proportions were respectively 60.57% (106/175), 27.43% (48), and 12.00% (21) among CAno_scar. Consequently, among the CAscar, the proportion of those with categories II and III was 51.01% (76), which was significantly (p = 0.04) higher than among CAno_scar (39.43%: 69). Stratified by the three countries, no significant correlation was found between presence of BCG scar and categories (I or II/III). Among the CAscar, the proportion of those detected with multiple lesions was with 4.57% (8) not significantly (p = 0.94) higher than those of 4.42% (10) among the 226 CAno_scar (Table 2).
Among individuals with known BUD category (149 CAscar and 175 CAno_scar), the proportion of males was 44.75% (145), which was not significantly (p = 0.55) different between CAscar (42.95%: 64) and CAno_scar (46.29%: 81). Stratified by sex, no significant correlation was found between presence of BCG scar and categories (I or II/III). Among the CAscar, the range of age was 1 to 78 y and the median of age was 18 y. Among the CAno_scar, the range of age was 2 to 70 y and the median of age was 12 y. Age distribution in CAscar and CAno_scar was significantly (p<0.01) different, as the CAscar were younger than the CAno_scar: AG 0–9 y (20.81% in CAscar vs. 37.71% in CAno_scar), AG 10–19 y (34.23% vs. 34.29%), AG 20–39 y (24.83% vs. 19.43%), and AG 40–90 y (20.13% vs. 8.57%) (Table 2).
After stratification by the three countries, two sexes and four age groups, no significant correlation was found between presence of BCG scar and categories (I or II/III). Multivariate analysis confirmed, that the independent variables country (p = 0.31), sex (p = 0.24), age (p = 0.96), and presence of BCG scar (p = 0.07) did not significantly influence the dependent variable, if an individual develops BUD category I or category II/III.
Among the 175 CAscar, the proportions of individuals with duration of 0–30 days (d), 31–60 d, 61–90 d, 91–180 d, and >180 d were 46.29% (81), 21.71% (38), 12.57% (22), 13.14% (23), and 6.29% (11), whereas these proportions were 46.02% (104), 22.12% (50), 13.72% (31), 11.06% (25), and 7.08% (16) among the 226 CAno_scar. The difference was not significant (p = 0.97), neither after stratification by the BUD categories.
Among the 401 CA, 305 (76.06%) were treated adequately by only antibiotics, 87 (21.70%) by antibiotics and surgery, seven (1.75%) by surgery only, and from two (0.50%) CA, no data on treatment were available. Among the 175 CAscar representing 82.29% (144/175) the time to healing (i.e. the time difference between onset of treatment up to the point of time of macroscopic healing of BUD lesion) was known, by contrast with those of 80.97% (183/226) of 226 CAno_scar. Among the 144 CAscar, the proportions of time to healing of 7–90 d, 91–180 d, and >180 d were 27.08% (39), 45.83% (66), and 27.08% (39), whereas these proportions were 32.79% (60), 33.33% (61), and 33.88% (62) among the 226 CAno_scar. The difference was not significant (p = 0.07), and neither after stratification by the BUD categories of lesions.
This is one of the largest observational studies on the effectiveness of Bacillus Calmette-Guérin (BCG) vaccines on Buruli Ulcer Disease (BUD). The aim of the present retrospective case-control study was to evaluate possible protection of routine BCG vaccination with live attenuated bovine tuberculosis bacillus Mycobacterium bovis against BUD in the DR Congo, Ghana, and Togo. Since the first human vaccination with BCG in 1921, many studies of BCG vaccines have been performed to estimate their effectiveness, but their results differed immensely. These discrepancies are explained by three main factors: the BCG strain used for vaccination, the population vaccinated, and the mycobacterial disease or its manifestation.
The past and continued use of both strong and weak vaccine strains makes interpretation and comparison of clinical trials extremely difficult, thus no conclusions can be made that one BCG strain is clearly superior to another in the protection of humans against tuberculosis or other mycobacterial diseases [17], [63]. More than 20 different BCG seed strains are in use for vaccination, which are produced by more than 40 manufacturers. African countries like the DR Congo, Ghana and Togo, were mainly supplied with BCG vaccine procured by UNICEF as BCG-Russia, BCG-Japan, and BCG-Danish. As explained above, the BCG vaccines used in these three countries changed very often, so it was not possible to figure out retrospectively with which BCG strain a certain study participant was vaccinated if that person has shown a typical BCG scar. As no documentation in hospital files or on vaccination cards was performed, no data on exact time of vaccination could have been assessed. Consequently, the present study could not consider the BCG strain used for vaccination even though it is known that strong strains as BCG-Danish, less strong strains as BCG-Japan and weak strains as BCG-Russia were in use in these three countries. This classification refers only to tuberculosis and it is completely unknown if this might be also conferrable on BUD [17], [19].
This study assessed the effectiveness of BCG vaccination on BUD only. Tuberculosis, leprosy or any other disease which might influence the data, were not considered. The study population included 401 laboratory confirmed BUD cases and 826 adequate controls. To minimize confounding, the association between presence of BCG scar and BUD status (case or control) were calculated after stratification by the three countries, two sexes, and four age groups, and by multiple analysis.
Several studies have shown that the effectiveness of BCG is dependent on the population in which the vaccination is used. Age plays a role, as effectiveness among children is much higher in preventing tuberculous meningitis and miliary tuberculosis [12]–[14]. On the other hand, BCG vaccines seem to be more effective against leprosy among adults [20], [21]. To avoid influence of age, all analyses were performed after stratification by four age groups. The age distribution of cases in the present study was comparable with those in others [43], [53].
It is completely unknown if there is any age-depending vaccine effectiveness against BUD like found against tuberculosis and leprosy. After stratification into three countries and four age groups, the present study found only a significant higher proportion of those with BCG scar among CO in Ghana in AG 10–19 y (p = 0.03), but confounded by sex. After stratification by three countries, two sexes and four age groups, no significant correlation was found between the presence of BCG scar and BUD status of individual (CA or CO).
Furthermore, that vaccine effectiveness was calculated to be different in populations with high or low exposure to environmental mycobacteria. High exposure to mycobacteria affects the immune system in various ways and thus, BCG might not improve greatly upon that background [5], [8]. In the three study sites of the present study, it was assumed that there was equal, or at least comparable, exposure to mycobacteria among the populations. To avoid influence of country specific populations in general, all analyses were performed after stratification by the three countries.
In the present study, multivariate analysis has shown that country, sex, age, and presence of BCG scar did not significantly influence whether an individual develops BUD category I or category II/III. Furthermore, the status of BCG vaccination was also not significantly related to duration of BUD before initial presentation of patients nor to time of healing. These results underline those of four studies performed in Benin [14], [32] and in Ghana [33], [34], which did not reveal any significant difference between cases and controls regarding their BCG vaccination status. These results contradict those of two other studies performed in Benin which generated the hypothesis that BCG vaccination might protect children against more severe forms of BUD, notably osteomyelitis [26], [27], and another study performed in Cameroon which concluded that BCG appeared to protect children against more severe forms of BUD with multiple lesions [28]. None of the studies considered the BCG strain used for vaccination, and they could not answer the question if certain BCG strains might protect better than others against BUD.
The present study has the same limitation. Exact data on BCG vaccination among the study participants could not be assessed by documents, such as vaccination certificates or hospital registers. Thus, the status of BCG vaccination of every case and control was assumed by detection of a typical scar on one shoulder or anterior side of the forearm, based on the fact that scars develop in most vaccinated persons as described before [14], [45]–[47]. Probably a certain proportion of individuals were defined as “vaccinated”, even though the scar was caused by something other than a BCG vaccination (“false positive”). On the other hand, also a certain proportion might have been defined as “not vaccinated”, if no scar was found on the shoulder or anterior side of the forearm, because BCG vaccination did not lead to a “typical scar” (“false negative”). The number of such “false positive” and “false negative” cases and controls is not known and could not be estimated in the present study. Furthermore, no other data on the BCG vaccination (e.g. method of application, booster vaccination, and side effects) could be assessed. This inaccuracy cannot be estimated either, but might be equally distributed among cases and controls. To minimize this bias, we have chosen a case-control-design.
From the time since the first studies were conducted on the effectiveness of the BCG vaccine, the results are varying and will continue to vary as long as retrospective studies with little precise data are performed. As a consequence of this, we recommend to conduct prospective studies, with an exact documentation as to which vaccine was administered. Given the fact that some BCG strains might have a short-time protection against BUD in certain populations as shown in some studies [24], [25], this effect would have little impact on the overall incidence of BUD. A safe and effective specific vaccine with long-term protection against BUD which could be used in several populations of the most BUD endemic countries would be an adequate preventive tool to reduce the risk for this disease.
Given the fact that some BCG strains might provide protection to avoid more severe forms of BUD, notably osteomyelitis [26], [27] and multiple lesions [28], this effect would also not decrease the incidence of BUD, because only a small proportion of BUD cases are diagnosed with osteomyelitis (in the present study <1%) and only a small proportion of BUD cases are diagnosed with multiple lesions (in the present study <5%).
Even though only a limited number of studies on BCG effectiveness for the prevention of BUD have been conducted, the probability of finding an effective BCG strain against BUD is low, and thus efforts to research specific vaccines against BUD should be accelerated like approached by the BuruliVac consortium.
In our study, we did not observe significant evidence of a protective effect of routine BCG vaccination with Mycobacterium bovis on the risk of developing either BUD or severe forms of BUD. Since accurate data on BCG strains are used in these three countries were not available, no final conclusion can be drawn on the effectiveness of BCG strain in protecting against BUD. As has been suggested for tuberculosis and leprosy, well-designed prospective studies on different existing BCG vaccine strains are needed also for BUD and further research on safe and specific vaccines against BUD should be supported.
|
10.1371/journal.pntd.0006723 | Comparison of individual and pooled diagnostic examination strategies during the national mapping of soil-transmitted helminths and Schistosoma mansoni in Ethiopia | Laboratory-based studies have highlighted that pooling stool and urine samples can reduce costs and diagnostic burden without a negative impact on the ability to estimate the intensity of soil-transmitted helminth (STH, Ascaris lumbricoides, Trichuris trichiura and hookworms) and schistosome infections (Schistosoma mansoni and S. haematobium). In this study, we compare individual and pooled stool examination strategies in a programmatic setting.
Stool samples were collected from 2,650 children in 53 primary schools in Amhara Regional State, Ethiopia, during the national mapping of STHs and schistosome infections. Eggs of STHs and S. mansoni were quantified in both individual and pooled samples (pools were made from 10 individual samples) using a single Kato-Katz smear.
A pooled diagnostic examination strategy provided comparable estimates of infection intensity with higher fecal egg count (expressed in eggs per gram of stool (EPG)) than those based on individual strategy (Ascaris: 45.1 EPG vs. 93.9, p = 0.03; Trichuris: 1.8 EPG vs. 2.1 EPG, p = 0.95; hookworms: 17.5 EPG vs. 28.5 EPG, p = 0.18; S. mansoni: 1.6 EPG vs. 3.4 EPG, p = 0.02), but had lower sensitivity (Ascaris: 90.0% vs. 55.0%; Trichuris: 91.7% vs. 16.7%; hookworms: 92.6% vs. 61.8%; S. mansoni: 100% vs. 51.7%, p < 0.001). A pooled approach resulted in a ~70% reduction in time required for sample testing, but reduced total operational costs by only ~11%.
A pooled approach holds promise for the rapid assessment of intensity of helminth infections in a programmatic setting, but it is not major cost-saving strategy. Further investigation is required to determine when and how pooling can be utilized. Such work should also include validation of statistical methods to estimate prevalence based on pooling samples. Finally, the comparison of operational costs across different scenarios of national program management will help determine whether pooling is indeed worthwhile considering.
| Infections with intestinal (roundworms, whipworm and hookworms) and blood-dwelling (schistosomes) worms pose a significant public health burden in developing tropical countries. To optimize control programs against these worms, large-scale surveys are required to determine the worm distribution to initiate control and to monitor the success of the programs. These large-scale surveys come at an important cost for governments, which in resource-limited countries present major challenges. During a nationwide survey in Ethiopia, we assessed whether the examination of pooled rather than individual samples could be a cost-saving strategy to assess prevalence and intensity of worm infections (which is an indicator of worm-related morbidity and success of the control program). We showed that a pooled examination strategy was useful in estimating the intensity of worm infections, but that it underestimated prevalence. Examination of pooled samples significantly reduced laboratory time, but it only resulted in limited financial gain.
| Neglected tropical diseases (NTDs) are a group of parasitic, bacterial and viral infections that pose an important burden on public health, particularly in tropical and sub-tropical countries [1]. In 2015, this group of 17 diseases resulted in approximately 26 million disability-adjusted life years (DALYs) and are considered an issue of global importance hindering progress towards the Sustainable Development Goals [2]. Soil-transmitted helminthiasis (STH) and schistosomiasis (SCH) are two of seven NTDs that are amenable to control through regular mass drug administration (MDA) [3]. Millions of people are infected worldwide, with each disease attributable for more than 10% of the overall NTD burden (schistosomiasis: 11%; soil-transmitted helminthiasis: 14%) [4]. Both are targeted primarily through school-based treatment programs, during which anthelmintic drugs (albendazole or mebendazole for STH and praziquantel for SCH) are administered to all school age children [5]. Fueled by the London Declaration on NTDs (January 2012; [6]), the global treatment coverage of school-aged children has increased since 2011 for SCH (2011: <20% vs. 2016: 53.7%) and STH (2011: ~30% vs. 2016: 69.5% [7–9], with the ultimate goal of treating at least 75% of school-aged children in all endemic countries by 2020 [10].
The pharmaceutical industry donates anthelmintic drugs at-scale (albendazole: GlaxoSmithKline, mebendazole: Johnson & Johnson, praziquantel: Merck KGaA) [11]. However, MDA programs require substantial political and financial investments from endemic countries [12]. Additionally, there are costs for periodically assessing the epidemiology of the diseases. Prior to treatment, nationwide epidemiological surveys are used to target treatment appropriately. Periodic follow-up surveys are required to measure progress and determine whether scaling-down of MDA is justified [13].
Ethiopia published its first National Master Plan for NTDs in 2012 (2012–2015), outlining plans to scale-up MDA efforts for eight priority NTDs. For STH and SCH, Ethiopia mobilized financial resources through a range of partners to support its nationwide baseline surveys. Given the substantial cost of such surveys, there is a need to identify cost-effective mapping approaches to further drive country ownership. The examination of a pooled stool sample strategy (ten individual samples) rather than using individual samples is a potential cost-saving strategy. Evidence from veterinary medicine shows that pooling can reduce diagnostic burden and costs, without having a negative impact on estimating the intensity of helminth infections [14]. Pooling has been evaluated for the assessment of STH and SCH in humans [15,16], highlighting that a pooled approach holds promise for rapid assessment of infection, but lacks diagnostic sensitivity. However, previous studies have focused on a small-scale (number of samples: 116–840), confined geographical area (Ethiopia: Jimma Town [15,16] and Amibara District [17]; Côte d’ Ivoire: Azaguié health district [18]), moderate and high transmission areas (STH prevalence ~50% [15,16]; SCH prevalence ~25% [15,16] and ~50% [15,16]), and provided limited information on operational costs. In the current study, we compared an individual and pooled examination strategy for the detection and quantification of soil-transmitted helminth and Schistosoma mansoni infections in 2,650 children in 53 primary schools across 35 woredas of Amhara Regional State in Ethiopia. In addition, we compared the time for sample testing for a subset of the samples, and the total operational costs for both strategies.
This study was embedded in the national mapping of STH (caused by Ascaris lumbricoides, Trichuris trichiura and the hookworms, Necator americanus and Ancylostoma duodenale) and SCH (caused by S. mansoni and S. haematobium) in Ethiopia. The study protocol was reviewed and approved by the Scientific and Ethical Review Office of the Ethiopian Public Health Institute (ref. no.: SERO-128-4-2005). The Regional Health and Education Bureau were informed about the survey. During the survey, school directors, teachers, and students were informed of its purpose, including its benefits, potential risks and operational procedures. Participation in the study was entirely voluntary. Written informed consent was obtained from the directors of all participating schools, and verbal consent was obtained from all subjects. Subjects who provided a stool sample were given a single dose of mebendazole 500 mg and subjects excreting eggs of S. mansoni or S. haematobium were provided a single dose of praziquantel (40 mg/kg of body weight).
This study was conducted in Amhara Regional State in the North of Ethiopia (9° - 14° N and 36° - 40° E). Amhara consists of 10 zones and 157 woredas and is divided into three major ecological zones: the highlands (>2,300 m above sea level [asl]), midlands (1,500 to 2,300 m asl) and the lowlands (<1,500 m asl). The annual mean temperature is between 15°C and 21°C. The mean annual rainfall is 1,165 mm, with the highest rainfall from June to September. The region’s population is 17.2 million of which 2.4 million (14%) are 10 to 14 years [19].
This study was part of a school-based cross-sectional study to map the distribution of STH and SCH in Amhara Regional State and was conducted from February to March 2015. Ten schools were randomly selected in each district. From this list, five schools were purposively selected based on (i) reports of schistosomiasis, (ii) the presence of water bodies close to the schools and (iii) practices of irrigation and fishing in the community. Finally, 50 grade-five pupils (9 to 14 years; 25 girls and 25 boys) were randomly selected. In schools with fewer than 25 boys or girls in the appropriate grades, children (9 to 14 years) from lower grades (grade four) or higher grade (grade six) were included.
Each subject was asked to provide a stool sample of approximately 3 g in order to examine samples individually and to subsequently make pools of 10 individual samples. All laboratory procedures were performed in the nearest District Health Facility. At the collection site, the stool samples were immediately stored in a cool box (at 4°C) to avoid development of hookworm eggs. On average, samples were kept in cool box for 2hrs prior to processing. At the District Health Facility, samples were processed individually using a single Kato-Katz thick smear, as described elsewhere [20]. Subsequently, stool samples were combined into pools of 10 samples. The procedure of pooling was based on the methodology described by Kure et al., 2015 [15] and is illustrated in S1 Fig. In summary, 50 individual samples per school were placed in 5 rows of 10 samples. From each individual, 1g of stool was transferred into a new pre-labeled stool cup and thoroughly mixed with a wooden spatula until the color of the mixture became uniform. Finally, the pools were processed applying a single Kato-Katz thick smear.
The sensitivity and specificity (based on faecal egg counts (FEC) expressed in eggs per gram of stool (EPG)) of the pooled examination strategy was determined. The sensitivity was calculated using the combined results of both strategies as the diagnostic ‘gold’ standard, against which the sensitivity of the different individual approaches were compared. Therefore, the specificity of both strategies was set at 100%, as indicated by the morphology of the eggs. The sensitivity was determined at the level of the pools and at the level of schools by comparing against the combined strategies. Differences in sensitivity between examination strategies were assessed by a permutation test taking into account the dependency of results within samples (10,000 iterations) [21]. The variation in sensitivity of a pooled examination over different levels of egg excretion was explored for each of the four helminths. The classification of the levels of egg excretion were based on the 33th and 66th quantile (q33 and q66) of the mean of the corresponding individual FECs, resulting in 3 levels of egg excretion (level 1: mean FECs ≤ q33; level 2: q33 < mean FECs ≤ q66; level 3: mean FECs > q66). Differences in sensitivity between levels was assessed by a permutation test taking into account the dependency of results within samples (5,000 iterations). Tukey’s method was applied for multiple comparisons [21]. The agreement between FEC obtained by examining a pooled sample with the mean FEC of the corresponding individual FECs was evaluated by a permutation test (5,000 iterations) based on Pearson correlation coefficient and differences in FEC. For the assessment of correlation, FECs of the pooled examination strategy and the mean FECs were log transformed. For the assessment of the difference, no transformation was applied.
A comparison of time taken to prepare and examine a subset of the samples (n = 2,450) was conducted. The steps were (i) the preparation of Kato-Katz thick smears, (ii) the pooling of stool, and (iii) the examination of the Kato-Katz thick smear. Given that timing of the preparation for each individual Kato-Katz thick smear would slow down the workflow, we recorded the total time to make batches of 10 Kato-Katz thick smears. The preparation of pools and the examination of a Kato-Katz thick smear were recorded on an individual basis. The mean time and corresponding standard deviation was calculated for preparing and reading of the individual and pools across the five survey teams.
The total operational cost to map soil-transmitted helminth and S. mansoni infections were estimated for both strategies. The operational costs were assumed to depend on (i) human resources utilised, (ii) the number of schools that could be screened in one day, and (iii) the time during which no activities linked to the survey could be performed. In the present study, five field teams were involved, each consisting of three laboratory technicians and one nurse supported by one vehicle. Our experience from previous surveys suggested that number of schools that can be screened by one team per day ranges from one to three depending on the schools’ accessibility. The teams obtained the permission of the Health and Education Office to operate within each district. Teams were on the road during the entire survey, but were not able to work over weekends as both schools and Health and Education Offices were closed.
We estimated the operational cost for one team to be on the road for twelve weeks across different scenarios of school accessibility. We first calculated the cost for one day of work at each school (e.g. driving to schools, sample collection and processing samples), administration (e.g. obtaining the permission to conduct the survey), travel, and days-off (a day without survey related activities) across the three levels of school accessibility. These costs included the expenses for materials, salary, transport, and fees to facilitate the work at the schools and data entry. The costs for the materials included equipment, supplies and reagents, based on an itemized cost assessment considering the cost per unit, the usage over a one year period, the life expectancy (in years), and the number of samples that can be processed per day. For simplicity, the number of samples that can be processed per day was fixed at 100. This assumption implies that the cost per sample will not increase or decrease when fewer or more samples per day are screened. It was estimated that the cost of materials for 50 individuals or pooled samples equaled US$ 4. For more details on the itemized cost assessment see S1 Table. The daily salary equaled US$ 13.7 for each team member. When samples were individually processed, a team consisted of three laboratory technicians and one nurse. When a pooled examination strategy was applied, we assumed two technicians were required rather than three. Data entry clerks were paid US $1.1 per 100 data records entered. The daily cost of car rental (including driver) was US$ 66.0 and the cost for fuel (20 L) was estimated at US$ 15.8. Two schoolteachers were paid US$ 7.2 each to support the survey team in informing the students about the survey, facilitating the selection of students, and sample collection.
We determined the number of days required for each of the four activities (work at school, administration, travel, and days off) within a period of 12 weeks (84 days) for three scenarios of school accessibility. S2 Fig illustrates the activities over a 12-week period when school accessibility was low. In this scenario, there are 45 days of work at school, 21 days off, and 9 days for either travel or administration. S3 and S4 Figs illustrate the activities over a 12-week period when school accessibility is moderate and high, respectively. In these scenarios, the total number of five schools per district remained unchanged, and hence there are days that fewer schools per day are visited. For example, in the scenario of moderate school accessibility, two schools per day will be surveyed in the first two days, followed by one day where only one school is visited. In that case, the daily cost for a poorly accessible school was used. Finally, the number of days and the cost per days were then multiplied for each activity separately, to obtain the total operational cost.
We performed a one-way sensitivity analysis in which the cost for materials, salary, fee for school teachers, data entry, rent of car and fuel transport was increased and decreased by 10%. This univariate sensitivity analysis was applied for both individual and pooled examination strategies and the three scenarios of school accessibility, separately.
Eggs of STH or S. mansoni were found in 354 out of the 2,650 (13.4%) subjects. The predominant helminth species was hookworm (6.2%, n = 113), followed by A. lumbricoides (4.3%, n = 164) and S. mansoni (2.6%, n = 70). The least prevalent was T. trichiura (0.8%, n = 22). The overall mean FEC was 45.1 EPG for A. lumbricoides, 17.5 EPG for hookworm, 1.6 EPG for S. mansoni and 1.8 EPG for T. trichiura. At the level of the schools, at least one parasite was found in 41 out of 53 (77.4%) schools, and as illustrated in S2 Table, both prevalence and intensity of infections ranged widely across the schools (T. trichiura: 2.0–26.0%, 0.5–55.2 EPG; S. mansoni: 2.0–36.0%, 0.5–35.5 EPG; A. lumbricoides: 2.0–50.0%, 0.5–1,168.8 EPG; and hookworms: 2.0–78.0%, 0.5–535.2 EPG).
Tables 1 and 2 summarize the sensitivity at the level of the pools and the schools, respectively. The sensitivity was set at 100% when at least one parasite egg was detected in a sample or school for both strategies and the sensitivity of both strategies were compared against this target. Generally, the examination of pooled samples resulted in a significantly lower sensitivity for each of the four helminths. At the level of the pools, the sensitivity for a pooled examination strategy ranged from 16.7% for T. trichiura to 61.8% for hookworms, whereas for an individual examination strategy the sensitivity was at least 90% for all helminths. When determined at the school level, the sensitivity remained roughly unchanged, for a pooled examination strategy it ranged from 22.2% for T. trichiura to 75.0% for S. mansoni and was significantly lower than the sensitivity of individual examination strategy (sensitivity >89%).
As illustrated in Fig 1, the probability of detecting helminth eggs increased with higher levels of egg excretion for A. lumbricoides (level 1: 40.7% vs. level 2: 42.3% vs. level 3: 81.5%), hookworms (level 1: 52.0% vs. level 2: 50.0% vs. level 3: 82.6%) and S. mansoni (level 1: 30.8% vs. level 2: 50.0% vs. level 3: 80.0%). The pair-wise comparison revealed only a significant difference between levels 1 and 3 (p = 0.006), and levels 2 and 3 (p = 0.012) for A. lumbricoides. Since there were only 12 cases of T. trichiura, variation in sensitivity across the levels of egg excretion was not assessed.
Overall, there was a significant positive correlation between the mean FECs of individual samples and the FECs of the pooled samples for A. lumbricoides (0.68, p <0.001), hookworm (0.65, p <0.001), and S. mansoni (0.75, p <0.001) (Fig 2). Given the low number of cases at the pooled level (n = 12), the correlation in FECs between examination strategies was not determined for T. trichiura.
The mean FEC based on a pooled examination strategy were generally higher than those based on an individual strategy (Table 3). A significant difference between FECs derived from individual and pooled examination strategies was observed for A. lumbricoides, (FECindividual = 45.1 EPG vs. FECpooled = 93.9 EPG, p = 0.03) and S. mansoni (FECindividual = 1.6 EPG vs. FECpooled = 3.4 EPG, p = 0.02).
The total time to prepare and read 2,450 stool samples individually equaled 198h 16min, compared to 53h 50min when samples were processed in pools; a reduction of 72.8%. Considerable inter-team variation was observed in time required to prepare and screen samples (Table 4). The mean time for the preparation of ten individual Kato-Katz thick smears and pools of ten individual stool samples ranged from 10.3 min to 31.7 min, and from 2.4 min to 8.8 min, respectively. The reading of the Kato-Katz thick smears ranged from 2.2 min to 4.4 min for the examination of individuals, and 2.7 min to 8.2 min for the examination of pools. Across the teams the reduction in time when using a pooled strategy ranged from 50.1% to 82.0%.
The cost per day of the four activities when samples were individually examined and when schools are poorly accessible are summarized in Table 5. The estimated daily costs were US$ 155.6 for work at school, US$ 136.6 for a day of administration and a day of travel, and US$ 128.7 for a day-off. The differences in daily cost per activity can be explained by differences in usage of material (only required for work at school), payment of fees to school teachers (only applicable for work at school), and the amount of fuel (less fuel required on days off). The cost breakdown for one day of work at school across the three levels of school accessibility when samples are individually examined is summarized in Table 6. The daily cost increases from US 155.6 for poorly accessible schools to US$ 190.3 for moderately accessible schools to US$ 225.1 for highly accessible schools. Table 5 reports the cost breakdown for the four activities at a poorly accessible school using a pooled strategy. Table 6 reports the costs of work comparing the different levels of accessibility under a pooled strategy. These two tables show that costs are lower compared to an individual examination strategy, and these differences are due to lower usage of materials and less data entry (one tenth of an individual examination strategy). Moreover, since the workload to process samples can be covered by only 2 laboratory technicians, salary costs were reduced. The number of days of work at school, administration, travel and days-off over an 84-day period are summarized in S2 Fig (poor accessibility of schools), 3 (moderate accessibility of schools) and 4 (high accessibility of schools). When schools are poorly accessible, a team will spend 45 days (53.6%) working in schools and 9 days for administration (10.3%) and travel (10.3%), and will have 21 days-off (25.0%). When the schools are moderately and highly accessible the number of days of working in schools equaled 36 (42.9%) and 32 (38%), respectively. The number of days dedicated to administration and travel equaled 12 (14.3%; moderate school accessibility) and 16 days (19%, high school accessibility), the number of days-off equaled 24 (28.6%; moderate school accessibility) and 20 (23.8%; high school accessibility).
The total operational costs for one team to be on the road for 12 weeks across the different levels of school accessibility are summarized in Table 7. When samples are processed individually, the estimated operational costs were US$ 12,161.3 when schools are poorly accessible, US$ 12,801.0 when schools are moderately accessible and US$ 13,590.8 when schools are highly accessible. Applying a pooled examination strategy reduced the costs by approximately 11%, regardless of the accessibility of schools.
The one-way sensitivity analysis, presented in the Figs 3 and 4, demonstrates the main cost drivers. Overall, car hire had the largest impact on total costs, followed by salaries (Fig 3). Varying these parameters resulted in a relative change in operational costs of approximately 4% and 3%, respectively. The impact of the other parameters did not exceed 2% (fuel: 0.7%– 1.6%; material: 0.0–0.2%; data entry: 0.0–0.0%).
The impact of car rental costs was consistent across examination strategies and levels of school accessibility. The impact of salary variance was similar for the three levels of accessibility, but was slightly less pronounced for an individual examination strategy (~3.0%) compared to a pooled examination strategy (~3.6%). Despite these differences in total operational costs, both parameters had little impact on the cost-savings effect of a pooled examination strategy (less than 1% reduction; (Fig 4)).
SCH and STH programs rely heavily on large-scale surveys to initiate and monitor their success [22]. Initial treatment frequency is determined by the prevalence of infection at baseline, prior to treatment, based on WHO guidelines [1]. Ongoing monitoring and evaluation primarily uses intensity of infection as an indicator for the program’s progress. Under either the individual or pooled approach, large-scale surveys constitute an important cost for governments and funders, which in resource-limited countries present major challenges. Various studies have provided insights on minimizing the operational costs for individual diagnosis of helminths, while ensuring accurate results and subsequently, correct programmatic decisions [23]. To date, the diagnostic strategy of pooling stool to reduce costs has not been fully explored for STH and SCH in humans [15–18], as studies were based on a small number of samples collected in confined geographical areas where transmission was moderate to high. Therefore, our group tested the applicability of a pooling strategy under field conditions during the national mapping of STH and SCH in Ethiopia. We compared individual and pooled examination strategies for the detection and quantification of STH and intestinal schistosomiasis (caused by S. mansoni) at a scale that is unprecedented in the literature and in area were transmission was low. Finally, we compared the time for sample testing, and the total operational costs for both strategies.
Overall, our findings on the diagnostic performance are in line with previous small-scale laboratory studies [15–18], confirming that a pooled strategy provides comparable estimates of population infection intensity, but that it often fails to detect infections, particularly those that are light. At this stage, it remains premature to make any formal recommendations on a pooled approach in a programmatic setting. We evaluated a pooling approach during an early phase of a STH and SCH program (mapping of disease) in a low transmission area applying only one diagnostic method (a single Kato-Katz thick smear) and one pool size (10 individual samples). It has been shown that the sensitivity of a pooled examination strategy is a function of the number of individual samples pooled (sensitivity inversely correlated with the number of individual samples) and the intrinsic sensitivity of the diagnostic technique [17,18]. As a consequence of this, pooling 10 individual samples and testing with a single Kato-Katz thick smear, a technique with poor sensitivity [24], may not be ideal to assess the intensity and prevalence of infections in all possible scenarios of STH and SCH epidemiology and phases of the program. Complementary studies evaluating pooling of samples in varying scenarios of endemicity, program phase, and diagnostic effort (number of samples pooled and analytic sensitivity) are welcomed to inform program managers on when and how to best pool samples. Given that it would be impossible to field test each of these scenarios, one could complement field studies with in silico approaches. Such an approach are best illustrated by the recent study by Lo et al. (e.g., reference 18). In this study, field data were used to inform a micro-simulation study. This in silico study was designed to verify whether pooling held promise for drawing programmatic conclusions across scenarios of endemicity other than those observed in the field. For application of a pooled approach in assessing the prevalence of infections, it is also necessary to develop and validate statistical approaches that allow the estimation of the true underlying prevalence based on the results of a pooled examination strategy. A variety of methods have been described for this, and they differ based on how the inference is drawn (frequentist vs. Bayesian approach), assumptions on the diagnostic performance (perfect vs. imperfect diagnostic techniques), number of samples pooled (fixed number vs. variable number) and input data (binary inputs vs. counts) With a few exceptions, these methodologies were initially developed for diseases other than STH or SCH [25–28].
Our results on the time required for testing demonstrated that a pooling strategy reduced the time to prepare and read slides under field settings by 72.8%. A previous study which pooled five individual samples reported a similar reduction in time (~70%; Kure et al., 2015), and this highlights that the reduction in laboratory time is likely not a linear function of the number of samples pooled. Between field teams there was a large variation in reduction in laboratory time between the examination strategies, ranging from 50.1% to 82.0%. This variation can by explained by a series of factors, such as the experience of technicians, school set-up, and varying issues related to the working environment. In general, a large proportion of the time for sample testing is dedicated to reading slides: reading a slide takes more than half of the total time to test an individual stool sample (Table 4). Efforts to develop and validate easy-to-use and point-of-care technology that allows electronic imaging of slides, and subsequently automated egg counting should be further encouraged [29].
Our results indicate that, despite a reduction in sample testing time of ~73%, the pooling strategy has relatively little impact on total survey costs (total operational costs were reduced by ~11%). The cost of any survey likely depends on diagnostic technique/s used and survey design [24,30–32]. The one-way sensitivity analysis on the different sources of costs revealed little to no variation in the relative cost-savings when a pooled examination strategy was used. Rather, our results indicated that the total operational costs were mainly impacted by logistical factors such as obtaining permission from the district offices and being constrained to the days children are at school. These factors incur additional costs for vehicle rental and survey team salaries, which affected the total operational cost for both strategies. In this regard, our observations indicated that under the different scenarios of school accessibility the teams spend between 36% and 44% of the total days off work when one and three schools are sampled per day. As recently highlighted by Turner and colleagues [33] the cost of a Kato-Katz thick smear varies considerably due to factors such as the method of collection (processing samples on site vs. examining samples next day off-site), the number of sites sampled per day (increases cost), the number of samples collected per site (decreases cost), variation in personnel, and adjustment of microscope costs (microscope used for other activities vs. microscope exclusively used for the STH/SCH survey). Given the number of schools per woreda (n = 5), subjects per school (n = 50), the operational steps in the field (S1 and S2 Figs), their corresponding costs (Table 7) and a survey period of 12 weeks, the estimated cost per single Kato-Katz thick smear on an individual stool sample varies from US$ 3.4 when 3 schools are surveyed per day (total number of school children = 4,000) to US$ 5.4 when one school is surveyed per day (total number of children = 2,250). Under the same scenario, the cost for a single Kato-Katz thick smear when a pooled examination strategy is applied increases approximately tenfold (US$ 48.1 when one school is surveyed per day: US$ 30.2 when 3 schools are surveyed per day), indicating that the way samples are examined (individual vs. pooled) should also be considered when costs of the Kato-Katz thick smear are estimated. These differences are explained by the low number of Kato-Katz thick smears (1x a single Kato-Katz thick smear is processed from one pooled sample vs. 10x a single Kato-Katz thick smears from 10 individual samples) and the relatively low reduction in total operational cost when samples are pooled. Finally, the total operational costs were estimated for this specific survey in an Ethiopian setting, and care should be taken when extrapolating to any other national programs. Consequently, it is necessary to compare operational costs of both strategies across a variety of scenarios of national program management to determine whether and when pooling is worthwhile considering.
In conclusion, we identified that a pooled strategy provided comparable results for infection intensity, but that it lacks sensitivity and therefore may perform poorly at estimating infection prevalence. A pooled examination strategy resulted in a reduction of 73% of time spent for sample testing, but this only resulted in a reduction of 11% in total operational costs. Based on these findings we conclude that a pooled examination strategy holds some promise for the rapid assessment of intensity of STHs and schistosome infections in a programmatic setting, but that does not result in a major cost-saving opportunity. To make any formal recommendations on a pooled approach, further investigation is required to determine when and how pooling can be utilized. For prevalence-based assessments, such work should also include validation of statistical methods to estimate prevalence based on a pooled examination strategy. Finally, operational costs should be compared different scenarios of national program management.
|
10.1371/journal.pntd.0002178 | Viperin Is Induced following Dengue Virus Type-2 (DENV-2) Infection and Has Anti-viral Actions Requiring the C-terminal End of Viperin | The host protein viperin is an interferon stimulated gene (ISG) that is up-regulated during a number of viral infections. In this study we have shown that dengue virus type-2 (DENV-2) infection significantly induced viperin, co-incident with production of viral RNA and via a mechanism requiring retinoic acid-inducible gene I (RIG-I). Viperin did not inhibit DENV-2 entry but DENV-2 RNA and infectious virus release was inhibited in viperin expressing cells. Conversely, DENV-2 replicated to higher tires earlier in viperin shRNA expressing cells. The anti-DENV effect of viperin was mediated by residues within the C-terminal 17 amino acids of viperin and did not require the N-terminal residues, including the helix domain, leucine zipper and S-adenosylmethionine (SAM) motifs known to be involved in viperin intracellular membrane association. Viperin showed co-localisation with lipid droplet markers, and was co-localised and interacted with DENV-2 capsid (CA), NS3 and viral RNA. The ability of viperin to interact with DENV-2 NS3 was associated with its anti-viral activity, while co-localisation of viperin with lipid droplets was not. Thus, DENV-2 infection induces viperin which has anti-viral properties residing in the C-terminal region of the protein that act to restrict early DENV-2 RNA production/accumulation, potentially via interaction of viperin with DENV-2 NS3 and replication complexes. These anti-DENV-2 actions of viperin show both contrasts and similarities with other described anti-viral mechanisms of viperin action and highlight the diverse nature of this unique anti-viral host protein.
| Viperin is a virally induced host protein that has been previously shown to have antiviral activity against a variety of viruses. Here we have demonstrated that viperin is also anti-viral against the medically significant arbovirus, dengue virus. Viperin was able to inhibit dengue virus at the level of viral replication, and cell lines unable to produce normal levels of viperin grew the virus to higher titres. These anti-dengue effects of viperin were mediated by amino acid residues in its C-terminus, and did not require structural domains of the N-terminal region as has been previously shown by us and others for the related virus, hepatitis C virus. Viperin was also demonstrated to co-localise and interact with the dengue capsid protein on the surface of lipid droplets, as well as with the NS3 protein and viral RNA. Viperin's association with NS3 was further demonstrated to be involved in its anti-dengue activities. The anti-viral activities of viperin presented in this manuscript show both similarities and contrasts with other described anti-viral mechanisms for the protein and highlight the diverse nature of this unique anti-viral host protein.
| The interferon (IFN) response is triggered in cells infected by RNA viruses, including members of the Flaviviridae family via a number of RNA recognition pathways that ultimately act to limit viral replication [1], [2]. Production of type I interferons (IFNs; IFN-α and IFN-β) by virus infected cells results in up-regulation of anti-viral IFN-stimulated genes (ISGs) and cytokines [3]. Infection of cells with the flavivirus, dengue virus (DENV) is recognised by the toll-like receptor-3 (TLR3), retinoic acid inducible gene–I (RIG-I) and melanoma differentiation associated gene-5 (MDA5) pathways to induce the IFN response [4], [5]. Microarray studies have shown up-regulation of ISGs, including viperin during DENV infection in cell lines and patient peripheral blood mononuclear cells (PBMC) [6], as well as during DENV infection of macaques in both macrophages and B-cells [7]. We and others have demonstrated that viperin is induced by infection with a number of diverse viruses as well as able to limit viral infection in most instances, including the first reported up-regulation of viperin in human cytomegalovirus (HCMV) infected cells [8], [9], [10]. Subsequently, viperin has been shown to have anti-viral actions in other viral infections such as hepatitis C virus (HCV), influenza virus, human immunodeficiency virus (HIV), sindbis virus (SINV), the flaviviruses Japanese encephalitis virus (JEV) and West Nile virus (WNV) [9], [10], [11], [12], [13], [14], [15] and more recently, Bunyamwera virus [16] and Chikungunya virus [17]. The roles and actions of viperin in these different viral infections appear diverse and multifaceted with anti-viral activity in some cases dependent on alterations to lipid rafts (influenza, [14]), membrane localisation (HCV, [9]), the radical S-adenosylmethionine (SAM) enzymatic activity of viperin (HIV, [10]), negated by viral proteins (JEV, [11]) and even an enhancing role under some conditions for HCMV [18]. Viperin also has anti-viral activity against DENV infection [6], [13] however the interaction of DENV and viperin has not been thoroughly investigated.
In this study we have further defined the induction of viperin and its mechanisms of anti-viral actions in DENV-2-infected cells, using an infectious DENV-2 in vitro replication model and including primary monocyte-derived macrophages (MDM) which represent a target cell type for DENV in vivo. Results show that DENV-2 infection induces viperin mRNA and protein, that expression of viperin is anti-viral, requiring the C-terminal but not N-terminal regions of viperin protein, and restricts DENV-2 infection by reducing viral RNA production. Viperin co-localised and interacted with DENV-2 CA, viral RNA and NS3 proteins. The interaction of viperin and NS3 but not membrane association, however, is necessary for viperins anti-viral actions. These results show both similarities and differences to our recent data suggesting that the anti-viral actions of viperin relate to its interaction with HCV NS5A and VAP-A in HCV replication complexes [9] and supports the growing evidence for both conserved and unique mechanisms of action of viperin against viral infections, even within the closely related Flaviviridae family of viruses.
Vero African green monkey kidney cells, A549, a human lung carcinoma cell line, Huh-7 and Huh-7.5 human hepatoma cells and primary monocyte-derived macrophages (MDM) were used for DENV-2 infection studies and maintained as previously described. Primary MDM were generated by adherence from PBMC that were isolated from voluntary blood donation at the Australian Red Cross Blood Service. Blood was provided anonymously and used with approval from the Southern Adelaide Clinical Human Research Ethics Committee. Infections utilised DENV-2, Mon601, a derivative of the New Guinea C strain [19] that was produced from in vitro transcribed RNA, transfected into BHK-21, baby hamster kidney cells, amplified in C6/36 insect cells and titred in Vero cells. Viperin shRNA and control cells were generated in Huh-7 cells, as previously described [9].
Cells were infected at a multiplicity of infection (MOI) of 0.1 or 1 for cell lines and an MOI of 3 for MDM for 90 min at 37°C, as described previously [20], [21], [22]. At the indicated time points post infection (pi) cell culture supernatants were collected, clarified by centrifugation and stored at −80°C prior to performing a plaque assay in Vero cells as previously described [21].
MDM were generated and DENV-2 infected, as above and at 48 h pi cells lysed and lysates subjected to SDS-PAGE. Proteins were transferred to nitrocellulose membranes and probed for viperin (in house rabbit anti-viperin antibody, 1/1000 [12]) with detection of complexes with goat-anti-rabbit-HRP conjugate and chemiluminesence. Protein loading was normalised by re-probing filters for β-actin (anti-rabbit β-actin, 1/500, BioVision). Images were captured with a LAS-4000 imaging system (Fuji Corp) and quantitated using Carestream Molecular Imaging Software 5.02 (Carestream Health Inc).
Wild type (WT) viperin and viperin mutant constructs were as described previously [9]. The DENV-2 NS3-GFP and pEPI–GFP CA constructs were a kind gift from Professor David Jans (Monash University, Australia). Viperin mCherry fusion proteins were created utilising pLenti6-mCherry; WT and viperin mutants were cloned in frame (XhoI/SacII) into the construct using previously described primers [9]. Cell lines were transfected using FuGene6 (Roche, IN) as per manufacturer's instructions. The viperin coding region was cloned into the lentiviral vector pLenti6/V5-D-TOPO (Invitrogen, CA) and the control lentiviral plasmid pLenti6/V5-D-TOPO-tdTomato was obtained from Dr Yuka Harata-Lee (University of Adelaide, Adelaide). Infectious lentivirus was generated as previously described [23]. Primary MDM were transduced with tdTomato control or viperin expressing lentivirus for 90 min at 37°C and were DENV-2 challenged at 24 h post transduction.
Total cellular RNA was isolated from cells using Trizol (Invitrogen), DNase treated and quantitated by spectrophotometry. For DENV-2 strand specific RT-PCR, 100 ng of denatured RNA was reverse transcribed at 37°C for 1 h with 20 pmol of DENV-2 specific primer (DENV5.1 or DENV3.2 [21]) attached to a 19mer long sequence (Tag) [24] and 10 U MMuLV RT (Promega, WI). The tagged DENV-2 cDNA was then subjected to real time PCR using SYBER Green PCR mix (Applied Biosystems, CA) and 20 pmole of each primer, Tag, DENV3.2 for negative (−ve strand) and DENV5.1 for positive (+ve strand), as previously described [21]. Real time PCR for viperin and the control gene RPLPO was performed as previously described on the ABI 7000 prism [12]. Primer sequences for IFIT1 were 5′ AACTTAATGCAGGAAGAACATGACAA and 5′ CTGCCAGTCTGCCCATGTG.
Cells were cultured on gelatin coated glass coverslips, and fixed in either 1% (v/v) formaldehyde for MDM and HeLa, or acetone∶methanol (1∶1) for Huh-7 cells and stored at −20°C. Slides were washed in PBS, and the formaldehyde fixed cells permeabilised with 0.05% (v/v) IGEPAL before blocking in 4% (v/v) goat serum, 2% (v/v) human serum, 0.4% (w/v) bovine serum albumin (BSA) in Hanks buffered salts solution (Gibco BRL, NY). Cells were immunolabelled using mouse anti-DENV-2 (serotypes 1–4, Santa Cruz Biotechnology Inc, 1/100 dilution), a rabbit anti-viperin, a mouse anti-FLAG (Sigma, MO) or a mouse anti-DENV CA antibodies (a kind gift from Prof David Jans, Monash University, Australia). Immunoreactivity was detected with goat anti-mouse IgG-Alexa 488, a goat anti-rabbit IgG-Alexa 647 or a goat anti-mouse IgG-Alexa 555 secondary antibodies (Molecular probes, CA). Nuclei were labelled with Hoechst 33342 (Molecular Probes, CA). BODIPY 493/503 (Invitrogen) was prepared as a stock solution of 1 mg/ml in ethanol. Fluorescence was visualised by confocal laser scanning microscopy (Biorad Radiance 2100 or Leica SP5 Spectral Confocal Microscope). For some experiments, quantification of intensity of immunofluorescence labelling was performed using ImageJ software (National Institutes of Health). The mean grayscale value was obtained for each channel for all cells where the image plane passed through the nucleus and excluding any cells at edge of the image and clusters of overlapping cells. Thresholds for detection of DENV-2 immunoreactivity were set at a grayscale value of three standard errors of the mean above the mean grayscale value measured in mock infected cells.
293T cells were transfected with pLenti6/V5-D-TOPO-viperin for 3 h then allowed to recover for 2 h prior to infection with DENV-2 at an MOI = 3. At 24 h pi cells were lysed (10 mM Tris, pH 7.5, 100 mM NaCl, 0.5% (v/v) Triton X-100+complete mini protease inhibitors [Roche]), lysates clarified and incubated for 1 hr with rabbit anti-FLAG antibody. Complexes were recovered with protein A-sepharose, washed six times (10 mM Tris, pH 7.5, 100 mM NaCl, 0.05% [v/v] Triton X-100+protease inhibitors) and resuspended in water. Precipitates were analysed for proteins by western blot and total DENV-2 RNA by RT-PCR with DENV5.1 and DENV3.2 primers, as described above with the exception that the reverse transcription step was non-primer directed.
Acceptor photobleaching was carried out as previously described in [9] with the use of GFP and mCherry tagged protein constructs. Pre and post-bleaching images were aligned using ImageJ and the difference in fluorescence (DIF) analysed in 5–10 regions of each cell where lipid droplets and/or cytoplasmic stained structures were positive for both proteins. At least 10 different cells in each of at least two independent experiments were analysed to ensure reproducibility. Negative slides were prepared by imaging cells with only the donor molecule present and treated in parallel photobleaching experiments.
Student t-tests were utilised to analyse the distributions of 2 normally distributed data sets and experiments were performed a minimum of three times, in triplicate or duplicate. Statistical analysis was performed using SPSS 10.
As mentioned previously, a number of viruses are able to induce viperin expression, and to extend these observations we infected cell lines and primary MDM with DENV-2. Cells were lysed at the indicated time points pi, RNA extracted and analysed by RT-PCR for viperin and DENV-2 negative strand (−ve) RNA, which is a marker of productive DENV-2 replication. Viperin mRNA was significantly induced in DENV-2 infected human cell lines with approximately 25 fold induction in A549 lung carcinoma cells (Figure 1A) and a lesser 4 fold increase in Huh-7 hepatoma cells co-incident with high level DENV-2 −ve strand RNA production (Figure 1B). In contrast viperin mRNA was not increased following DENV-2 infection of Huh-7.5 cells, a cell line which is defective in dsRNA signalling via a mutation in the pathogen-recognition receptor RIG-I [25] (Figure 1C). Up-regulation of viperin by DENV-2 infection was also demonstrated in primary MDM, with a much greater, approximately 1000 fold induction of viperin mRNA at 24 h pi (Figure 1D).
We next assessed up-regulation of viperin protein in MDM since these showed the most significant change in viperin mRNA. Cells were DENV-2 infected, lysed and analysed for viperin by western blot with IFN-α treated cells used as a positive control. Results show increased levels of viperin protein in DENV-2 infected primary MDM, at levels greater than that induced by IFN alone (Figure 2A). We further characterised viperin protein in DENV-2-MDM by confocal microscopy. As can be seen in Figure 2B, at 24 h pi viperin was elevated in DENV-2 infected compared with mock infected MDM. Interestingly MDM positive for DENV antigen displayed reduced amounts of viperin protein, whereas DENV-antigen negative bystander cells (indicated via arrows, Figure 2B, upper panel) were shown to express significantly increased levels of viperin. The intensity of viperin staining in these different populations was quantitated. Results support the visual up-regulation of viperin in DENV-2 compared to mock infected cells, but most predominantly in the antigen negative, bystander cells of the DENV-2 infected population (Figure 2C).
Ectopic expression of viperin has been previously shown to inhibit DENV infection using virus and reporter virus-like particles in vitro [6], [13], however neither the anti-viral mechanism(s) or the interaction with viperin has been fully explored. Here we first transfected HeLa cells with plasmid to express WT viperin and at 24 h post transfection infected with DENV-2. Cells were fixed 24 h pi and immunolabelled for dsRNA and viperin. Results are suggestive of anti-DENV activity of viperin, with individual cells expressing viperin harbouring either no or very little DENV-2 RNA (Figure 3A). We next quantitated this potential anti-DENV-2 activity of viperin using a panel of viperin mutants that have previously been used to investigate viperins anti-HCV activity [9]. Although little is known about the structure/function relationship of viperins anti-viral activity, recent work by us and others has demonstrated that both the localisation of viperin to the ER membrane through its N-terminal amphipathic helix, as well as its C-terminal residues are essential for its ability to limit the replication of HCV [9], [26]. Transient expression of WT viperin in Huh-7 cells significantly inhibited DENV-2 −ve strand RNA levels by up to 61% (Figure 3B). A significant reduction of DENV-2 −ve strand RNA was also observed for cells transfected with viperin mutants in the SAM (SAM 1-3) domain, leucine zipper (LZ) and N-terminal deletions from 17 and up to 50 amino acid residues, suggesting that these regions play no role in the anti-viral activity of viperin (Figure 3B). In contrast, C-terminal deletions, as small as 17 amino acids, completely abolished the anti-DENV-2 activity of viperin (Figure 3B). These C-terminal deletion mutants have previously been shown in our laboratory to retain the same expression level and localisation as WT viperin [9]. Mutation of the single C-terminal residue of viperin (CTM) partly abrogated the anti-DENV-2 activity of WT viperin, although compared with the no viperin control the CTM still produced a significant reduction in DENV-2 −ve strand RNA levels (Figure 3B). These results highlight the importance of the C-terminal end of viperin for anti-viral activity and the C-terminal, 3′Δ17 mutant is used as a control in subsequent experiments.
Huh-7 or A549 cells were transfected to transiently express WT or the C-terminal 3′Δ17 mutant viperin lacking anti-viral activity, infected with DENV-2 and were analysed at 24 h pi (Figure 3C and D respectively). Results confirmed a significant reduction in both infectious virus release as determined by plaque assay of media from infected cells and production of −ve strand RNA induced by WT but not 3′Δ17 viperin expression in these two different cell types.
An important cell type for DENV infection in vivo are cells of the monocyte-macrophage lineage. Additionally, these cells are major contributors to the IFN response. As such we have analysed the anti-viral actions of viperin in primary MDM. Given the difficulty in transfecting MDM, we expressed viperin via lentivirus-mediated transduction. MDM were transduced with a td-Tomato-red fluorescent protein control or viperin encoding lentivirus expression vector, infected with DENV-2 and infection analysed. Results show a significant reduction in infectious virus release from lentivirus-viperin transduced MDM compared with lentivirus td-Tomato transduced control MDM, with a significant 30 and 4 fold decrease seen at 24 and 48 h pi respectively (the 8 h time point is considered a measure of input virus and is not significantly different between control and viperin transduced cells) (Figure 4A). At 48 h pi cells were fixed and immunostained for viperin and DENV-2 antigens, followed by confocal microscopy. Enumeration of >600 cells from 10 different fields and two different infections showed a significant reduction in the number of DENV-2 antigen +ve cells in viperin compared with tdTomato transduced cells (images not shown, 5.9%±0.8 vs 9.5%±0.6, p<0.05, Students unpaired t-test). Additionally, we observed dramatically higher levels of viperin protein in the DENV-2 infected cells compared with mock-infected viperin-lentivirus transduced cells (Figure 4B). Further, this up-regulation of viperin was again only observed in the DENV-2 antigen −ve bystander cells of this population (Figure 4B). This likely represents up-regulation of endogenous viperin, as demonstrated previously in DENV-2 infected MDM (Figure 2B).
We next assessed the requirement for induction of viperin to restrict DENV-2 replication using a well characterised viperin shRNA Huh-7 cell line. Cells were DENV-2-infected at a lower MOI (0.1) to avoid DENV-induction of viperin mRNA, as in Figure 1, potentially overwhelming the capacity of the viperin shRNA. DENV-2 infection of viperin shRNA cells resulted in a significant, approximately 2 fold enhancement of infectious DENV-2 release at 24 h pi compared with control shRNA expressing cells (Figure 5A). By 48 h pi, infectious virus release was comparable between viperin shRNA and control shRNA expressing cells, possibly due to enhanced cytopathic effects in viperin shRNA cells associated with the earlier and higher level of DENV-2 replication, although this was not specifically quantitated. DENV-2-infection did not induce viperin mRNA in viperin shRNA expressing cells at 24 h pi (Figure 5B), although expression was detected at 48 h pi in some instances. In all cases, the unrelated ISG, IFIT1 mRNA, utilised as a control, was induced to comparable levels in both DENV-2-infected viperin shRNA and control shRNA expressing cells, demonstrating effective induction of other anti-viral responses in the absence of viperin (Figure 5C).
The lower levels of DENV-2 −ve strand RNA and viral release observed in HeLa, Huh-7 cells and primary MDM following infection of viperin expressing cells could be consistent with restriction of DENV-2 entry. To investigate this possibility Huh-7 and A549 cells were transfected to express viperin and following DENV-2 infection cells were immediately lysed and +ve strand DENV-2 RNA, indicative of intracellular genomic input RNA, quantitated by RT-PCR. Results showed no difference in intracellular levels of +ve strand DENV-2 RNA between cells transfected to express viperin and the inactive viperin C-terminal mutant, 3′Δ17 (Figure 6A). Additionally, DENV-2-infections were performed as above and cells lysed at 6 h pi and −ve strand DENV-2 RNA quantitated. Results demonstrated a significant reduction in the intracellular level of DENV-2 −ve strand RNA at this early time point in both A549 and Huh-7 cells transfected with viperin (Figure 6B), demonstrating a post-entry restriction in early DENV-2 RNA replication.
Our prior studies with HCV and viperin have demonstrated a requirement for the anti-viral actions of viperin mediated through lipid droplet and replication complex localisation and association with NS5A [9]. We next assessed the ability of viperin to associate with DENV-2 replication complexes by immunoprecipitation (IP). 293-T cells were transfected to express FLAG-tagged viperin, infected with DENV-2 then at 24 h pi cells lysed and IPed with anti-FLAG antibody. Precipitates were analysed for the presence of total DENV-2 RNA by RT-PCR. Results demonstrate successful co-precipitation of DENV-2 RNA with FLAG-antibody (Figure 6C). Concurrent analysis of precipitates by western blot, however failed to detect co-precipitated DENV-2 NS3 protein (data not shown).
The above data suggests the association of viperin with complexes containing DENV RNA (ie. replication complexes). Such complexes reportedly are also associated with cellular membrane structures and DENV-2 NS3 protein [27], [28]. The maintenance, however of the anti-DENV-2 activity of viperin containing N-terminal deletion mutants suggests that the ability of viperin to associate with membranes is not required for its restriction of DENV-2 infection (Figure 3A). We thus assessed the cellular localisation of viperin in DENV-2 infected cells. Viperin primarily localises to lipid droplets in Huh-7 cells as we have shown previously [9], and as can be seen in Figure 7A; this distribution remains unaltered in DENV-2 infected Huh-7 cells. The DENV capsid (CA) protein has also been demonstrated to localise to lipid droplets [29], [30] and consistent with these previous reports we observed partial co-localisation between DENV-2 CA and viperin at the interface of lipid droplet-like structures (white arrows, Figure 7B). Interestingly, viperin and the CA protein appear to coat the surface of the droplet at distinct loci, with small overlapping areas of co-localisation. Huh-7 cells were also co-transfected with a DENV-2 NS3-GFP expression plasmid and viperin. A clear co-localisation between DENV-2 NS3 and viperin is observed at the surface of lipid droplet-like structures as well as in distinct cytoplasmic loci (Figure 7C). The N-terminal viperin deletion mutant (Vip5′Δ33) which loses its ability to localise to lipid droplets [9], [31], but remains anti-viral against DENV-2 (Figure 3A), remained co-localised with DENV-2 NS3, although the pattern of localisation was solely cytoplasmic (Figure 7C). This observation demonstrates that viperin's anti-viral activities may be exerted through a possible interaction with NS3 but not necessarily at the lipid droplet interface.
As described above, although viperin co-precipitated DENV-2 RNA, IP experiments could not demonstrate co-precipitation of NS3 with viperin from either DENV-2 infected or NS3/viperin co-transfected cells (data not shown). These co-precipitation experiments are likely confounded by our observation of low levels of viperin protein in DENV-infected cells (Figure 2B) and low levels of DENV-2 antigens in viperin transfected cells (Figure 3A). Further, viperin is a lipid associated protein, which are notoriously difficult to extract and retain physiological protein-protein interactions.
We therefore investigated the physical interaction of viperin with DENV-2 NS3 and CA by fluorescence energy resonance transfer (FRET). Huh-7 cells were transfected with expression plasmids for DENV-2 CA-GFP or DENV-2 NS3-GFP in conjunction with either mCherry N-terminally tagged viperin-WT, viperin 5′Δ33 or viperin 3′Δ17 and FRET acceptor photobleaching performed. Results show positive FRET for viperin-WT and the DENV-2 CA protein at the surface of the lipid droplet (Figure 8A) demonstrating an interaction of WT-viperin and CA proteins at this site. FRET analysis also demonstrated an interaction of DENV-2 NS3 and WT-viperin in distinct cytoplasmic foci (Figure 8B), similar to that seen in the confocal co-localisation studies of these two proteins (Figure 7C). No positive FRET was detected between DENV-2 NS3 and viperin surrounding lipid droplet like structures, despite our prior observation of co-localisation at these sites (Figure 7C). The ampipathic helix mutant (5′Δ33) of viperin, which retains its anti-DENV-2 activity, but has lost its membrane localisation ability, demonstrated a positive interaction by FRET analysis with DENV-2 NS3, once again at distinct cytoplasmic foci within the cells (Figure 8C). In contrast, the C-terminal viperin mutant, 3′Δ17, which has no anti-DENV-2 activity but maintains WT viperin localisation [9], showed no positive FRET with DENV-2 NS3 suggesting this protein is unable to interact with DENV-2 NS3. These results indicate that the C terminus of viperin mediates its anti-DENV-2 activity through an interaction with DENV-2 NS3 but does not require lipid droplet or membrane association.
Viperin is emerging as an important virus-induced ISG that can be up-regulated by both IFN-dependent and independent pathways and has a diverse array of anti-viral actions. Viperin can be induced in an IFN independent manner via IFN regulatory factor-1 (IRF-1), following infection with the RNA virus, vesicular stomatis virus (VSV) [32]. In contrast, SINV induction of viperin requires IFN but JEV induction of viperin occurs in an IFN-independent manner that requires IRF-3 and AP-1 [11]. In this study we show that viperin is induced early in DENV-2 infection and similar to our observation in HCV infected cells, does not occur in Huh-7.5 cells that are deficient in RIG-I [9]. IRF-3 and AP-1 are transcription factors downstream of RIG-I activation suggesting that the RIG-I pathway has an important role in induction of viperin, at least for the Flaviviridae members JEV, HCV and DENV. Furthermore, our studies in DENV-2-infected viperin shRNA cells suggest that the viperin already present or induced intracellularly in the DENV-2 infected cell acts to restrict or control DENV-2 infection in this initial target cell. Additionally our results from DENV-2-infected MDM show strong induction of viperin protein in DENV antigen negative bystander cells. This indicates that induction of viperin in these bystander cells, probably secondary to the release of IFN from the DENV-2 infected cell, is likely to also be important for restricting DENV-2 spread. Our observation of a far greater level of induction of viperin following DENV-2 infection compared with IFN stimulation of primary MDM suggests that induction of viperin by DENV-2, either in the initial DENV-2 infected cell or the uninfected bystander cell, occurs via factors other than IFN that are yet to be defined.
Viperin protein contains N-terminal ampipathic helical domains that direct viperin cellular localisation to the endoplasmic reticulum (ER) and lipid droplets [9], [31], [33]. The C-terminal portion of viperin is relatively unstructured and highly conserved amongst species; however its current function remains unknown. Previously we have demonstrated that the anti-viral actions of viperin are dependent on a number of functional domains of the viperin protein (i) the N-terminus for intracellular ER and lipid droplet localisation of viperin, (ii) the extreme C-terminus in the context of HCV replication [9], and (iii) the radical SAM domain in the context of HIV egress [10]. Using DENV-1 virus-like particles (VLP) and a luciferase reporter replicon system, a previous study has shown that viperin is induced by DENV-1 infection, inhibits DENV-1 RNA production and requires the N-terminal SAM1 domain of viperin [13]. This same study showed a similar induction of viperin, inhibition of RNA production and requirement for the viperin SAM1 domain and in part, residues within the first 50 amino acids of viperin during WNV infection. In contrast, our study observed anti-DENV-2 activity of viperin SAM1-4 mutants at levels comparable to WT viperin. The differences in the requirement for the viperin SAM1 domain seen in our current study compared with previous results with DENV-1 and WNV may be due to (i) the level of expression of viperin through use of a tet-induction system compared with the transient viperin transfection system in the current study; (ii) the analysis of different markers of infection with viperin SAM1 reducing infectious DENV-1 release [13] but in our study not DENV-2 −ve strand RNA; and/or (iii) the use of DENV-1 compared to DENV-2 herein. Regardless, studies clearly suggest that the anti-viral actions of viperin can be mediated by residues outside of the SAM1 domain, including the N-terminal 50 amino acid residues for WNV [13] and in our previous work with HCV, also the C-terminal regions of viperin [9]. Consistent with this requirement of the C-terminal region of viperin for anti-HCV activity, in our current study we have similarly defined anti-DENV activity to reside in the C-terminal 17 amino acids of viperin.
The specific regions of viperin necessary for anti-viral activity between various viruses differs, as does the biological effect of viperin during different virus infections. Viperin is reported to inhibit release of influenza virus by disruption of lipid rafts [14], to inhibit HIV egress [10] and to diminish viral protein production in HCMV infection [8]. In this study we show that viperin inhibits early post-entry DENV-2 RNA replication consistent with prior reported effects of viperin in inhibiting RNA replication in other Flaviviridae family members, HCV [9] and WNV [13]. In contrast, viperin is induced but is not anti-viral against the related flavivirus, JEV due to mechanisms of JEV that proteolyse and degrade viperin in infected cells [11]. However, contrary to earlier reports of anti-viral activity of viperin against HCMV, a recent study has shown that the vMIA protein of HCMV induces re-localisation of viperin from the ER to mitochondria, resulting in an increase in HCMV infection [18]. These contrasting effects of viperin suggest that its effects on viral infection are multifaceted, virus specific and involve multiple mechanisms of action including alterations in the subcellular localisation of viperin.
Viperin has been shown to localise to lipid droplets and the ER. Similarly we have shown here that viperin co-localises with the lipid droplet marker, BODIPY, in DENV-2 infected cells and thus the cellular localisation of viperin is unchanged during DENV-2 infection. The DENV CA localises to lipid droplets and preventing this CA-lipid droplet association reduces DENV RNA replication and infectious virus particle production [30]. While we confirm that viperin is able to co-localise and interact with the DENV-2 CA at lipid droplet-like structures (Figure 7B, 8B), our observation that viperin N-terminal mutants, which lose the ability to localise to lipid droplets, still retain substantial anti-DENV-2 activity suggests that viperin has significant anti-DENV-2 activities independent of its lipid droplet/CA association. We also demonstrate that viperin co-localises and interacts with the DENV-2-NS3 protein and co-precipitates with DENV-2 RNA, both of which are components of DENV replication complexes [27], [28]. The interaction of viperin and DENV-2 NS3 was independent of the N-terminal ampipathic helix, but reliant on the C-terminus of viperin. Further, the ability of viperin to co-localise and interact with DENV-2 NS3 correlated with anti-viral activity. We propose that viperin has anti-viral activity mediated by a C-terminus interaction with DENV-2 NS3 that reduces early RNA production by interfering with DENV-2 replication complexes. It remains to be determined whether this occurs solely through a direct interaction with the DENV-2 NS3 protein, or also through an intermediate pro-viral host cell factor, such as is the case for HCV, whereby viperin interacts with both NS5A and the pro-viral factor VAP-A [9], [26]. Currently, the only known pro-viral host factor for DENV that interacts with NS3 in the context of functional replication complexes is fatty acid synthetase (FASN) [34]. We feel FASN is an unlikely candidate as a target of viperin's actions since viperin and FASN exist in alternate cellular compartments (Figure S1).
In conclusion, this study has revealed further critical functions of viperin during DENV-2 replication and highlighted similarities and differences in the mechanisms of induction and in the anti-viral actions of viperin between DENV-2 and other medically important Flaviviridae. This data highlights the incredibly diverse anti-viral nature of viperin and the complexity of the viperin-virus interaction.
|
10.1371/journal.pgen.1002557 | DNA Damage in Nijmegen Breakage Syndrome Cells Leads to PARP Hyperactivation and Increased Oxidative Stress | Nijmegen Breakage Syndrome (NBS), an autosomal recessive genetic instability syndrome, is caused by hypomorphic mutation of the NBN gene, which codes for the protein nibrin. Nibrin is an integral member of the MRE11/RAD50/NBN (MRN) complex essential for processing DNA double-strand breaks. Cardinal features of NBS are immunodeficiency and an extremely high incidence of hematological malignancies. Recent studies in conditional null mutant mice have indicated disturbances in redox homeostasis due to impaired DSB processing. Clearly this could contribute to DNA damage, chromosomal instability, and cancer occurrence. Here we show, in the complete absence of nibrin in null mutant mouse cells, high levels of reactive oxygen species several hours after exposure to a mutagen. We show further that NBS patient cells, which unlike mouse null mutant cells have a truncated nibrin protein, also have high levels of reactive oxygen after DNA damage and that this increased oxidative stress is caused by depletion of NAD+ due to hyperactivation of the strand-break sensor, Poly(ADP-ribose) polymerase. Both hyperactivation of Poly(ADP-ribose) polymerase and increased ROS levels were reversed by use of a specific Poly(ADP-ribose) polymerase inhibitor. The extremely high incidence of malignancy among NBS patients is the result of the combination of a primary DSB repair deficiency with secondary oxidative DNA damage.
| Damage to DNA is extremely dangerous because it can lead to mutations in genes that initiate or accelerate the development of a tumor. Evolution has led to highly complex networks of DNA repair enzymes, which for the majority of individuals are extremely effective in keeping our DNA intact. The devastating consequences of DNA damage are manifested in those individuals in which one or other of the repair pathways is non-functional. Several genetic disorders can be attributed to such DNA repair deficiencies and have the common feature of increased tumor incidence as the major life-threatening symptom. Cancer incidence varies amongst these disorders and is probably highest for the disease Nijmegen Breakage Syndrome, where more than 50% of patients develop a hematological malignancy in childhood. We have sought to understand this extremely high incidence by exploiting cells from a mouse model and cells derived from patients. We find that deficiency in the repair of DNA double-strand breaks leads to disturbances in cellular metabolism, leading ultimately to a loss of antioxidative capacity. The ensuing accumulation of highly reactive oxygen species generates further DNA lesions, thus potentiating the initial damage and increasing the likelihood of malignancy.
| Genetic cancer susceptibility disorders, such as Xeroderma pigmentosum and Fanconi anemia, generally have deficiencies in DNA repair and cell cycle regulation leading to tumour initiation. The specific mutagen sensitivities underlying these disorders define a set of enzymes and pathways involved in the DNA damage response. Nevertheless, these pathways clearly overlap and components in one pathway can be critically involved in another. For example, the nucleotide excision repair pathway mutated in Xeroderma pigmentosum is also required for the repair of interstrand crosslinks, to which Fanconi anemia patient cells are particularly sensitive [1].
Nijmegen Breakage Syndrome (NBS), Nijmegen Breakage Syndrome like disorder (NBSLD), Ataxia telangiectasia (AT) and Ataxia telangiectasia like disorder (ATLD) are clinically and biologically overlapping entities. Whilst the underlying proteins are intimately associated, cancer predisposition is a major life threatening feature of NBS and AT only. The proteins mutated in these four disorders are all involved in the sensing and repair of DNA double-strand breaks (DSB). However, if, as seems likely, the mutation rate in patient cells is increased, this may not be solely due to the primary DNA lesion but, rather, to the cumulative effects of auxiliary cellular disturbances. Thus it has been repeatedly shown that AT patient cells and knockout mice have increased oxidative stress [2]–[4] which could contribute to clinical progression of the disease.
Oxidative stress has not previously been associated with NBS, however, our previous proteomic study of null mutant mice suggested disturbances in the redox homeostasis in the livers of irradiated mice [5]. We speculated that this could be due to hyperactivation of members of the Poly(ADP-ribose) polymerase (PARP) family, such as PARP-1, PARP-2 and PARP-3 which rapidly detect DNA strand breaks and regulate/modulate proteins required for an effective cellular response. In cells unable to repair DSBs, the permanent activation or even hyperactivation of PARP enzymes was expected to disturb cellular function and contribute to an increased mutation rate. Interestingly, poly(ADP-ribosyl)ation was reported to be unaffected in both AT patient cells and knock out mouse cells [6]. In view of the close relationship between NBS and AT we sought to examine the situation in NBS. For these investigations we have started with our null mutant mouse cells since, unlike patient cells, they provide a system with complete absence of the affected protein, nibrin.
We find greatly increased levels of reactive oxygen species in both null mutant mouse cells and NBS patient cells after a DNA damaging exposure. Unlike AT cells, we find a parallel increase in the activity of PARP enzymes as measured by examining poly(ADP-ribosyl)ation of proteins. As we have previously hypothesized, depletion of the cellular NAD+ pool accompanies excessive poly(ADP-ribosyl)ation in NBS cells and this severely compromises the anti-oxidant capacity of the cells. Thus the extremely high incidence of hematological malignancies in NBS may be the result of the combination of a primary DSB repair deficiency and accompanying oxidative damage.
The murine fibroblasts used in these experiments have a neomycin insertion in one Nbn allele (Nbnins-6), a null mutation, and loxP sites flanking exon six in the other Nbn allele (Nbnlox-6). Treatment of these cells with cre recombinase leads to cells with biallelic Nbnins-6/del6 null mutations [7]. Henceforth we refer to wild type alleles and alleles with exon 6 flanked by loxP sites as Nbn+ and the null mutant Nbnins-6 and Nbndel-6 alleles as Nbn−. As shown in Figure 1A, 12 hours after introduction of DSBs there is a particularly high level of ROS in fibroblasts completely lacking nibrin due to null mutation of the Nbn gene (Figure 1B). The cells were treated here with 10 µg/ml bleomycin, which is equivalent to irradiation with 2 Gy irradiation causing approximately 60 DSBs per cell, with a ratio of DSBs to single-strand breaks of 1∶9 [8], [9]. As the non-fluorescent compound, CM-H2DCFDA, is converted to fluorescein specifically by hydrogen peroxide, hydroxyl radicals, peroxynitrite anion and peroxyl radicals, the observed over two-fold increase in fluorescence intensity in comparison to heterozygous cells is therefore due to the accumulation of these species [10]–[12]. These radicals are short lived with half-lives of just seconds or less [13], [14]. Therefore, their high concentration 12 hours after treatment with bleomycin suggests that they are being permanently produced in the Nbn−/− cell, presumably as a consequence of its unrepaired DSBs.
The null mutant murine cells examined here are particularly useful since they allow examination of cellular responses in the complete absence of nibrin, a situation not naturally available for human cells. Having seen the importance of full length nibrin for maintenance of cellular redox homeostasis by timely repair of DSBs, we turned to NBS patient cells, in which a truncated and partially functional nibrin fragment, p70-nibrin, is present [15], [16]. As shown in Figure 1C, fibroblasts from NBS patients also show an increased level of ROS after DNA damage. The increase in ROS-induced fluorescence, 1.5(+/−0.27) times that of controls, is less than in the complete absence of nibrin, 2.33(+/−0.9) times, which might indicate partial repair of DSBs or simply reflect differences in murine and human cells in ROS induction.
The results of repeated measurements of ROS levels in Nbn null mutant and NBS patient cells are shown in Figure 2. As indicated in the figure, the differences in ROS levels in comparison to wild type cells after DNA damage are statistically significant in the non-parametric two-tailed Mann-Whitney test.
We have argued previously that the permanent production of ROS in the absence of nibrin is caused by rapid depletion of NAD+ due to hyperactivation of Poly(ADP-ribose) polymerases and consequent loss of cellular antioxidant capacity [5]. In order to test this hypothesis, we treated cells heterozygous and homozygous for Nbn null mutations with bleomycin to induce DSBs and examined the extent and kinetics of poly(ADP-ribosyl)ation by western blot. As shown in Figure 3A, there is rapid and sustained poly(ADP-ribosyl)ation of proteins in the absence of nibrin under conditions in which PARP enzyme activity in heterozygous cells cannot be detected.
In Figure 3B PARP activity is shown for control fibroblasts and fibroblasts from NBS patients. Even in these cells with a partially active nibrin fragment, there is rapid and extensive activation of PARP as evidenced by poly(ADP-ribosyl)ation of proteins. By densitometric analysis of three independent blots we found a 3-fold increase in PAR-modified proteins in control fibroblasts 10 minutes after a DNA damaging treatment and a 17-fold increase in NBS patient fibroblasts (p = 0.05). Levels of protein poly(ADP-ribosyl)ation have returned to near normal 12 hours after treatment (data not shown).
We reasoned that if the increased ROS levels in NBS patient cells are a consequence of increased PARP activity, rather than its cause, inhibition of the enzyme should reduce ROS levels, whilst scavenging of ROS should not affect PARP activity. As shown in Figure 4, scavenging ROS using the antioxidant vitamin E derivative TROLOX reduced ROS levels in damaged NBS patient cells to the same levels as in untreated cells (Figure 4A and Figure 2). However, although cells treated with TROLOX and bleomycin were essentially ROS free, PARP activity remained high (Figure 4B). Inhibition of PARP-1, PARP-2 and PARP-3 with the specific inhibitor KU-0058948 [17], on the other hand (Figure 4B), did reduce ROS levels to normal (Figure 2).
The link between ROS levels and PARP enzyme activity is the latter's requirement for NAD+, an important component of the cells antioxidant capacity. Indeed, numerous reports have shown that PARP inhibition prevents the reduction of NAD+ concentrations in cells subject to genotoxins, with a resulting decrease in cellular necrosis [18], [19]. As shown in Figure 5, we measured NAD+ levels in NBS fibroblasts after a bleomycin treatment in comparison to normal fibroblasts. There is very rapid depletion of NAD+ in the patient cells in comparison to the control cells, confirming the hyperactivation of PARP in these cells. Interestingly, the baseline levels of NAD+ in these NBS patient cells were considerably higher than in controls (1,850(+/−86) vs. 875(+/−15) pmol/106 cells) suggesting that even in the absence of exogenous damaging agents, NAD+ requirements are higher in these repair deficient cells. In line with this observation we note that PARP activity in undamaged NBS cells is apparently higher than in controls (Figure 3B). Even 12 hours after treatment with bleomycin, baseline levels of NAD+ have still not been reached in NBS cells, in agreement with the timing of ROS measurements shown in Figure 1 and Figure 2.
Nibrin is a component of the trimeric MRN complex together with Mre11 and RAD50. This complex is involved in the processing of all DNA double-strand breaks in the cell, whatever their origin: mutagen exposure, physiological processes or simply chromosome ends [20]. The complex is implicated in DSB repair by both non-homologous end joining and homologous recombination [21], [22]. As a sensor of DSBs the MRN complex is involved in the activation of ATM and subsequent downstream targets to induce cell cycle checkpoints [23]. Telomeres, the ends of chromosomes, are maintained by a mechanism in which the MRN complex has also been implicated [24].
Cancer incidence in Nijmegen Breakage Syndrome is extremely high with 40% of patients developing a tumor, mostly lymphoma, before the age of 20 [25], [26]. This contrasts with the related disorder AT in which lifetime cancer risk is 20–30% [27], [28]. For accurate prognosis and improved patient care it is important to establish which factors contribute to this cancer predisposition. In this respect, the role of nibrin in both DNA repair and cell cycle regulation may be significant. It has been shown that mutation of Nbn in mouse models leads to a defect in apoptosis [29], [30], [31] and reduced clearance of damaged cells could clearly contribute to the high cancer incidence [32]. In addition, factors leading to an increased mutation rate, beyond that of the primary double strand break, could be present in NBS. Previous work has suggested that oxidative stress could be one such factor [5].
Unphysiologically high levels of ROS are a hallmark of oxidative stress and can be directly due to mutagenic agents, such as ionizing radiation, or, rather, reflect overburden of the cellular antioxidation mechanisms. These mechanisms can be either direct scavenging of radicals or regeneration of oxidized biomolecules [33]. NADPH and NADH have been reported to be involved in both kinds of antioxidant activity [34]. Thus, reduction in the availability of these essential cellular antioxidants leads inevitably to increased cellular ROS levels and oxidative stress, even in the absence of DNA damage. The radicals detected in this report, peroxynitrite anion, hydroxyl and peroxyl radicals or their metabolites, all react aggressively with DNA to yield oxidized bases and single strand breaks. An increased mutation rate would be the consequence. In lymphocytes, in which DSBs are a prerequisite for immunoglobulin gene rearrangements, their non-repair due to the absence of nibrin could thus lead to redox disturbances and an even higher occurrence of mutations.
NAD+ is the precursor for NAD(P)H and its cellular level is therefore critical for cellular redox homeostasis. NAD+ is also a substrate for the PARP superfamily of enzymes with a common catalytic activity and involved in the DNA damage response [35], [36]. For example, PARP-1 is a nuclear DNA damage sensor and binds to persisting single- and double-strand breaks [37]. PARP enzymes covalently attach ADP-ribose to glutamate, aspartate, and lysine residues of acceptor proteins. Branched ADP-ribose polymers are formed at nuclear acceptor proteins that facilitate DNA repair through modifying and activating structural proteins and enzymes such as histone H2AX, topoisomerase I and II, DNA polymerase α and β, DNA ligase I and II, nuclear factor (NF)-κB, and p53. Hyperactivation of PARP has been frequently described in various systems leading to depletion of the NAD+ pool [38]–[40]. It has also been suggested as a contributing factor in AT [41]. The unrepaired DSBs in Nbn null mutant cells clearly lead to such PARP hyperactivation, as shown here. As previously reported, the Nbn−/− cells attempt to combat the increased ROS levels due to NAD+ depletion by upregulating genes involved in the detoxification of radicals, such as MnSOD. In contrast, enzymes also requiring the NAD+ substrate were downregulated, for example, glyoxylate reductase 6.7-fold [5].
In Nbn−/− cells and also in NBS patient cells, the loss of full nibrin function leads to a delay in the activation of ATM [42]. It has recently been shown that ATM, in addition to its direct role in the DNA damage response, also promotes the pentose phosphate pathway leading to increased NADPH levels and thus improving anti-oxidant defence [43]. In the absence of nibrin, promotion of the pentose phosphate pathway will not occur, indeed in our proteomics analysis of irradiated Nbn−/− mouse livers, transaldolase, a key enzyme of the pentose phosphate pathway, was actually reduced 6-fold [5].
Human cells with null mutation of the NBN gene are non-viable and NBS patients all have hypomorphic mutations and express a truncated nibrin protein [15], [44]. In the case of the major founder mutation, c.657_661del5 (p.K219fsX19), the truncated protein, p70-nibrin, is translated from an upstream start codon brought into frame by the deletion [15]. These proteins clearly have enough partial activity to ensure survival [7], [45], but are severely compromised in the DNA damage response. This is manifest as the increased chromosome breakage, characteristic translocations, radiosensitivity, immunodeficiency and cancer predisposition characteristic of NBS [25]. These partially active proteins all have the carboxy terminal MRE11 and ATM interacting domains but lack the FHA and first BRCT domains of the amino-terminus, which are required for interaction with proteins such as gamma-H2AX, MDC1 and p53BP1 [46], [47], [48].
Here we describe increased ROS levels after DNA damage in NBS patient cells. The truncated p70-nibrin is clearly unable to fully prevent the hyperactivation of PARP, NAD+ depletion and ROS generation. Patient cells showed approximately 1.5 times the level of ROS after DNA damage in comparison to control cells whilst in null mutant mouse cells the levels were more than two fold increased. We have previously described individual variations in the level of p70-nibrin expression [16] which are due to differences in its proteasomal degradation [49]. Low levels of p70-nibrin correlate with cancer incidence and it can be speculated that a contributing factor may be higher oxidative stress.
In conclusion we present evidence for a further detrimental consequence of NBN mutation. In addition to a DSB repair deficiency and failure in cell cycle checkpoints, lack of fully functional nibrin results in increased ROS levels and oxidative stress. This unique combination would lead to an extremely high mutation rate in cells with an underlying apoptosis deficiency. Oncogene activation and tumour initiation are the consequence.
Spontaneous transformed murine fibroblasts were grown from ear explants of Nbnlox-6/ins-6 mice [7]. Cells were cultured in Dulbecco's modified Eagle's medium (DMEM; Gibco, Life technologies) supplemented with 5% glucose (glc) and 10% fetal calf serum (FCS) strictly in the absence of antibiotics. Cell culture conditions were 37°C and 5% CO2. Environmental oxygen was reduced to 10%. Cells were split 1∶10 at least twice a week.
The immortalized human NBS cell lines GM7166VA7 and NBS-1LBI homozygous for NBN657del5/657del5 and a control cell line, LN9, transformed with simian virus 40 (SV40) were cultured using the same conditions described above.
The cre recombinase fusion protein, HTNC [50], was isolated as previously described [7]. Exponentially growing cells were incubated in 2 µM HTNC twice for 6 hours in a 48 hour period. Knock-down efficiency was verified by western blot analysis using a polyclonal rabbit antibody to detect murine nibrin (Pineda Antikörper-Service, Berlin). Murine polyclonal antibody against β-actin served as a loading control (Abcam).
DNA damage was induced by incubating cells for two hours in the radiomimetic drug bleomycin at 10 µg/ml. Cells were then washed with medium and returned to culture for the specified times.
Cells were scraped into ice-cold PBS, pelleted and snap frozen in liquid nitrogen. Cell pellets were stored at −80°C. For analysis, pellets were solubilised in LDS-sample buffer (Invitrogen) and sonified for 60 seconds using a model 450 sonifier (Branson, Emerson Industrial Automation). Proteins were separated on Tris-Acetate gels (3–8%) and transferred to PVDF membranes (Hybond-P, GE Healthcare).
To detect poly(ADP-ribosyl)ated proteins, two different polyclonal antibodies (Abcam) each producing the characteristic smear of PAR modified proteins were used. As a loading control β-actin was detected using a murine polyclonal antibody. Blots were repeated three times using independent lysates. For densitometry, films were scanned using the ScanMaker scanner (Mikrotek) and lanes quantified using ImageQuant software (Molecular Dynamics).
The PARP inhibitor KU-0058948 was kindly provided by KuDOS Pharmaceuticals Ltd. (AstraZeneca PLC). The compound was dissolved in 100% dimethylsulfoxid (DMSO) and stored at −20°C. Cells were treated with 1 µM inhibitor in medium containing 0.5% DMSO for 10 hours before induction of DNA damage and then for a further 12 hours. Control cells were incubated in parallel in medium containing 0.5% DMSO.
In some experiments, ROS were scavenged by treating cells with the antioxidant vitamin E derivative, 6-hydroxy-2,5,7,8-tetramethylchroman-2-carboxylic acid (TROLOX, Hoffman-La Roche). Cells were incubated in 500 µM TROLOX for 12 hours after the bleomycin damaging treatment.
The amount of intracellular ROS was monitored before and after the induction of DNA damage in fibroblasts at 50% confluence. Cells were washed and harvested into PBS and 106 cells were stained in 500 µl PBS with 10 mM 5-(and-6)-chloromethyl-29,79-dichlorodihydrofluorescein diacetate (CM-H2DCFDA; Invitrogen) for 20 min at 37°C in the dark. Samples were subsequently washed using ice-cold PBS and centrifuged for 10 min at 1000 rpm (∼180×g) before being resuspended in FACS dissociation solution (FACSmax, Genlantis) and kept on ice until analysis. Flowcytometry was performed using the FACS-Calibur (Becton Dickinson Bioscience) counting a minimum of 104 cells per sample. The opensource flowcytometry software WinMDI V2.9 was used for data analysis. Gates were placed on dot blots of forward vs. side scatter to exclude apoptotic cells and debris from the fluorescence histograms shown in the figure. In all measurements at least 85% of cells were within this gate. All experiments were repeated independently at least three times. Relative ROS-levels are expressed as [gated mean of bleomycin treated cells]/[gated mean of untreated cells] and were evaluated for statistical significance using the non parametric two-tailed Mann-Whitney U test.
NBS patient fibroblasts and control fibroblasts were treated in triplicate with bleomycin as indicated above. At the timepoints indicated in Figure 5 cells were precipitated with 0.5 M perchloric acid on ice. After 15 min samples were centrifuged at 1500×g for 10 min and the supernatant (500 µl) was combined with 350 µl of 1 M KOH, 0.33 M K2HP04, 0.33 M KH2P04 followed by incubation on ice for 15 min.
Cells were centrifuged at 1500×g for 10 min and the supernatant was frozen at −20°C before NAD+ determination by using an enzymatic cycling assay [51].
|
10.1371/journal.pcbi.1004964 | Phylogenetic Analysis Reveals That ERVs "Die Young" but HERV-H Is Unusually Conserved | About 8% of the human genome is made up of endogenous retroviruses (ERVs). Though most human endogenous retroviruses (HERVs) are thought to be irrelevant to our biology notable exceptions include members of the HERV-H family that are necessary for the correct functioning of stem cells. ERVs are commonly found in two forms, the full-length proviral form, and the more numerous solo-LTR form, thought to result from homologous recombination events. Here we introduce a phylogenetic framework to study ERV insertion and solo-LTR formation. We then apply the framework to site patterns sampled from a set of long alignments covering six primate genomes. Studying six categories of ERVs we quantitatively recapitulate patterns of insertional activity that are usually described in qualitative terms in the literature. A slowdown in most ERV groups is observed but we suggest that HERV-K activity may have increased in humans since they diverged from chimpanzees. We find that the rate of solo-LTR formation decreases rapidly as a function of ERV age and that an age dependent model of solo-LTR formation describes the history of ERVs more accurately than the commonly used exponential decay model. We also demonstrate that HERV-H loci are markedly less likely to form solo-LTRs than ERVs from other families. We conclude that the slower dynamics of HERV-H suggest a host role for the internal regions of these exapted elements and posit that in future it will be possible to use the relationship between full-length proviruses and solo-LTRs to help identify large scale co-options in distant vertebrate genomes.
| Animal genomes contain ancient pathogens known as endogenous retroviruses (ERVs). Though the widespread abundance of ERVs is due to their ability to self replicate, some ERVs are known to have become important to host processes including placentation, and in the case of HERV-H, the functioning of human stem cells. In our study we place the insertion and deletion activity of primate ERV families in direct quantitative comparison. In particular, we show that ERV deletion is an age dependent process, so that as an ERV ages it becomes less likely to be deleted at any given instant. We also find that ERVs from the HERV-H family are unusually slowly deleted, an interesting result that suggests that the exaptation of HERV-H may have involved internal regions of the virus and not just its terminal promoters. Assuming the behaviour of primate ERVs is not unusual, our study suggests that future bioinformatics screening for ERVs with slow deletion dynamics could help identify large-scale exaptations in distant species. Furthermore, as we demonstrate that ERVs are deleted rapidly, we think that such screening could be performed using ratios of conserved to deleted elements and could therefore be applied to single genomes.
| By definition, endogenous retroviruses (ERVs) are the result of the Mendelian (vertical germ line) transmission of retroviruses from parent to progeny. Over many generations it is possible for an ERV to fix in a host population so that in humans, for example, as much as 8% of the genome is thought to be retrovirally derived [1]. Though the majority of fixed ERV loci are thought not to be under selection for function there are remarkable exceptions such as the syncytins[2] that are necessary for placentation, and members of the HERV-H family that are essential to stem cell identity in humans [3, 4].
Successful retroviral insertions (proviruses) are known to initially possess a common structure consisting of viral genes flanked by a pair of identical sequences known as long terminal repeats (LTRs). ERVs that retain this characteristic viral structure are commonly described as full-length ERVs. In addition to full-length ERVs, endogenized viruses are also found in a second, dramatically different form, referred to as a solo-LTR. A solo-LTR is a solitary LTR that is missing its associated partner LTR and adjacent proviral genes. Solo-LTRs are thought to be generated when paired LTRs undergo non-allelic homologous recombination which results in a deletion and an associated acentric fragment [5], a piece of chromosomal material lacking a centromere that is unlikely to persist across many cell divisions. Clearly, like other genomic DNA, both forms of ERVs are also subject to ordinary mutational processes so that over time they may become degraded or fragmented due to point mutations or indel events.
Studies have identified ERV activity dating back over millions of years and in many species e.g. [6–8]. When describing the replicational activity of viruses on evolutionary timescales many studies start by searching for full-length ERVs in a host genome. The resulting ERVs are then dated on an individual basis by comparing the divergence of paired LTRs that are assumed to have been identical at integration time. The aforementioned strategy is certainly reasonable but does have some limitations, including the fact that as the majority of ERVs are present in solo-LTR form [9] they do not contribute to analyses of replicational activity at all—whether or not this is problematic depends on the relationship between full-length proviruses and solo-LTRs.
In this study we analyze ERV insertion and solo-LTR formation in primates. Like other authors, we bring together insertion rates from several viral groups that are present in primates. Unlike other authors we do this by systematically sampling both full-length ERVs and solo-LTRs—in the same way for different ERV families and different host species—and relating these ERVs explicitly via a host genome alignment. By combining our sampling process with a host phylogeny we can then place insertion rates in quantitative comparison.
Our study is particularly concerned with the ratio of full-length ERVs to solo-LTRs. A study of the polymorphic HERV-K (HML2) loci in humans found that the majority of loci are represented by pre-integration sites or solo-LTRs [10]. This suggests that the solo-LTR formation process is very rapid. If this is true for most ERV families for most of the time then it is reasonable to treat a count of full-length ERVs as a constant fraction of a count of all ERVs, no matter the age of the infection. However, if the result does not generalize then such an assumption is not reasonable as the proportion of ERVs that are present in full-length versus solo-LTR form will vary given the age of an infection.
Within this paper we extend prior work by examining a wider variety of ERV families, by considering ERVs over a host phylogeny, and by introducing a likelihood framework that allows model comparison. Using this approach we confirm that there is an age dependent deletion process and are also able to demonstrate that HERV-H has unusual dynamics. Our results lead us to suggest that previously used models of deletion should be abandoned where possible. We also argue that a phylogenetic approach to characterizing ERV activity could help identify novel ERV exaptations in species distinct to those that we study here.
We obtained site patterns describing ERV integrations and deletions that had occurred in the primate lineage (Fig 1) since the split between macaque and marmoset roughly 40 million years ago (Ma). The site patterns were obtained by relating post-processed RepeatMasker annotations to a six-way genome scale alignment of human, chimpanzee, gorilla, orangutan, macaque and marmoset sequence. These annotations were then converted to site patterns using a heuristic method. Our intention was to quantitatively describe the insertion rate of ERVs across branches of the primate phylogeny and also to investigate the process of ERV deletion that converts full-length ERVs into solo-LTRs.
Applying the methods sketched above, we identified 1,197 distinct insertion events that had occurred on the branch hcgom or later. These distinct insertions could be split naturally into groups based on the type of ERV they involved. We investigated the properties of insertions from the four largest families in our sample: ERV9 (245 insertions); HERV-K11 (197 insertions); HERV-H (116 insertions); and HERV-K (59 insertions). Based on a BLAST search against a library of 51 viral sequences, many ERVs from smaller families were assigned to group-I (131 insertions) or group-II (112 insertions). These patterns of insertion and deletion are presented in machine readable format in S1 Tabular data. Below we present the following results: (i) the insertion rate parameters obtained; (ii) the deletion rate parameters obtained under two different models of deletion; (iii) a comparison of the two competing deletion models; and (iv) the results of a simulation that tests the adequacy of the most appropriate deletion model.
We applied our phylogenetic model (see Methods) to obtain the maximum likelihood relative insertion rates per million years (Myr) for the nine branches in our tree (Fig 2 and Table 1). These estimates of insertion activity are independent of the deletion model used. Our results are broadly compatible with descriptions contained in commonly cited studies on ERV dynamics e.g. [9, 12]. For example, the HERV-K insertion rates for branches h (1.36 relative insertions per Myr) and c (0.61 relative insertions per Myr) capture the commonly reported fact that HERV-K has been recently active in both human and chimp and also that the activity in human specific ancestors appears to have been at least 50% greater than the activity in chimp specific ancestors. In general, more detailed comparison is difficult as individual studies vary considerably in methodology and reporting style, a state of affairs that partially motivated us to perform the analysis reported in this paper.
As well as insertion, we are also interested in the process by which ERVs are deleted. The simplest conceivable model of a deletion process is one which has a constant hazard over time, i.e. a process under which the probability of deletion in an infinitesimally small period of time is constant. The unique process with this property is an exponential decay process. Under the exponential model, the probability of deletion of an ERV is independent of its age. Such a model is appropriate if the probability of deletion of an ERV is small and fairly constant across generations, and if the probability of deletion of an ERV has nothing to do with the process by which an ERV ages. For each of the six groups of ERV we obtained a maximum likelihood estimate of exponential rate parameter ψe (Table 2).
Under the exponential model, full-length elements from the HERV-K family would be deleted most quickly, with full-length loci having an average pre-deletion lifetime of approximately 6.25 Myr. Under the same model, the most long-lived group would be HERV-H, for which the average full-length lifetime would be roughly 20 Myr. At an age of 400,000 years (the expected fixation time of a neutral ERV given an effective population size of 10,000 and a generation time of 10 years), the exponential model predicts that 94–98% of ERVs would retain their full-length form. By an age of 25 Myr, a time period comparable to the scope of our phylogeny, the exponential model predicts that only 2% of HERV-K insertions would remain in full-length form while a much larger 29% of HERV-H insertions would.
Beyond the simplest possible deletion scenario, we are also interested in the hypothesis that the formation of solo-LTRs is governed by a process that depends on the age of an insertion i.e. a process with a variable hazard. For example, it may be that as ERVs age, substitutions and gene conversion introduce differences between paired LTRs that substantially reduce their chance of producing solo-LTRs. A process with a hazard rate that changes with time is often modeled using a Weibull distribution. Under this process the rate of deletion is proportional to a power of time so that the probability of the removal of a full-length ERV can decrease with age, given a shape parameter ω < 1, or increase with age, given a shape parameter ω > 1. A shape parameter of ω = 1 implies an exponential decay process so that the exponential model is a nested submodel of the Weibull model. For each of the six groups of ERVs we obtained a maximum likelihood estimate of scale parameter ψw and shape parameter ω (Fig 3, Table 2).
Under a Weibull model, we find that at an early age it is again ERVs from the HERV-H family that are most likely to remain in full-length form. We find that at an age of 400,000 years 58% of HERV-H would remain in full-length form whereas only 19–21% of the other five groups would. These predictions differ by 40–76 percentage points from those of an exponential model. At the longer time scale, maximum likelihood parameter estimation suggests that at an age of 25 Myr we expect 31% of HERV-H to remain in full-length form while we expect less of ERV-9 (5%), HERV-K11 (8%), HERV-K (6%), group-I (6%) or group-II (8%) to do so. Summarizing a Weibull deletion process requires considering the role of the shape parameter. The maximum likelihood estimate of shape parameter ω is less than 1 for all six groups of ERV, and therefore suggests the rate of ERV deletion does decrease monotonically with time. Bootstrap replicates suggest this is unambiguously true for all families apart from HERV-K, for which 12% of bootstrap replicates identify a shape parameter >1. This implies that there is a degree of uncertainty over whether HERV-K has qualitatively different dynamics than the other ERV groups, with these exceptional bootstrap estimates suggesting peak deletion rates occur at ages of up to just over 4 Myr.
We have described the results of fitting two competing models, formalized as the null hypothesis He that the deletion process is an exponential decay, and as the alternative hypothesis Hw that the deletion process is age dependent. To decide which of the models is more appropriate we performed a likelihood ratio test. As He is nested within Hw, which has one additional parameter, we compare 2Δℓ = 2(ℓw−ℓe) with a χ2 cutoff of 10.83. We find that 2Δℓ implies that the Weibull age specific model is clearly more appropriate for all six groups of ERV (p < 0.001).
Our likelihood values show that a Weibull model is a much better description of the ERV deletion process than an exponential model. However, likelihood ratio tests do not provide an assessment of the adequacy of the Weibull model itself. For this reason we conducted a simulation to see whether the Weibull model could explain the empirical site patterns we observed for each of the six ERV groups. Our simulation proceeded as follows. For each of the six ERV groups, we generated 10,000 insertions on branch hcgom, the deepest branch in our phylogeny. We then simulated the history of these insertions according to the Weibull model operating under group specific maximum likelihood estimates of ψw and ω. This generated the group specific frequency distribution of site patterns at the tips of the tree. We performed a goodness of fit test comparing the distribution of empirical site patterns with the frequencies obtained via simulation. We found no statistical evidence that the distribution of observed site patterns differed from those expected under the Weibull model for any of the six groups (Table 3). This suggests a Weibull model is an adequate one.
In this paper we sample ERV site patterns from the primates and present a phylogenetic model (see Methods) which we show captures the deletion process of ERVs in a way that is superior to existing descriptions. Applying this model to data on six large groupings of ERVs we find that HERV-H is the most slowly deleted group of ERVs across the primate phylogeny and that, with the potential exception of HERV-K, ERVs appear to “die young.” Below we discuss the biological implications of our findings, the limits of our approach, and why we think our approach can help identify exapted ERV families in other lineages of vertebrates.
Previous studies of ERV activity (i.e. insertions) have often proceeded by enumerating the full-length ERVs, perhaps of a specific type, from one host species. Meta-studies will then collate the results of primary studies and attempt a synthesis of their contents e.g [9, 12]. Meta studies face the difficult problem of relating various sampling (search) methodologies. They also face the impossible problem of relating counts of full-length ERVs between species when the overlap between counts is unknown. Here we think our approach is helpful. Consider Fig 2, where we provide estimates that allow one to answer quantitative questions about the insertional activity of ERVs from different families and species. Assuming our sampling of viruses has been effective, we can be confident that both group-I and group-II ERVs became less active after the split of the macaque lineage from the ape lineage, but also that ERV9 and HERV-K11 were more active in the ape lineage after this split than before. As we measure insertions using solo-LTRs as well as full-length ERVs, we also suggest that, contrary to a previous conclusion [13], an apparent speedup in HERV-K(HML2) is not an artefact of considering only full-length ERVs. Therefore we think our results complement existing research. We also think that similar approaches will become more useful as newly sequenced genomes give better resolution within phylogenies and allow for improved genome scale alignments.
The topic of solo-LTR formation has been less widely studied in the past, perhaps because it is assumed to be unimportant to host phenotype, or perhaps because an adequate treatment requires phylogenetic data. What is true is that when a deletion process is explicitly mentioned it is often assumed to be exponential or constant over time e.g. [14] or [15]. Our results show that estimates produced assuming a constant deletion model differ dramatically from those produced using an age dependent model. We also show that the assumption of a pan-lineage constant deletion process is clearly inappropriate for five of the six groups of ERVs we examined in primates. We have no a priori reason to consider other deletion scenarios (e.g. branch specific deletion) and conclude that, pending further research, an exponential model is inappropriate more generally.
The question of whether ERV deletion rates vary with age was previously addressed by Belshaw et al. [16] who reported that the deletion rate for recent integrations was 200 fold higher than for integrations that occurred over 6 Ma. This effect was ascribed to mutational divergence between LTRs reducing homologous recombination although background genomic recombination might also play a role [17]. The result of Belshaw et al. [16] is implicitly conditional on loci being retained in full-length form in at least one of human or chimpanzee and, to our knowledge, while commonly cited, has not been followed up elsewhere. Our results generalize the qualitative conclusions of Belshaw et al. [16] to the primate phylogeny and to a variety of ERV families. This is important because the previous results rely exclusively on elements from the HERV-K category which are unusual for two reasons. First, the HERV-K family of ERVs have recently been insertionally active. Second, some recent HERV-K insertions have been shown to have biochemical effects, including virion formation, during early stages of human development [18]. In addition, our results surpass previous ones as we provide a description (a parameterized Weibull model) that gives an indication of the expected survival function of an ERV at various points in its lifetime.
Given that divergence between LTRs is hypothesised to be responsible for an age dependent deletion process [16], it is interesting to consider mutations into the LTRs of unfixed ERVs. Experiments by Datta et al. [19] show that a single nucleotide difference between two 350 base pair (bp) substrates can lower recombination 3-fold in yeast. Opperman et al. [20] find a 4-fold reduction in recombination when using 618 bp substrates in plants. Both studies find that additional mutations have relatively little effect. Bearing these data in mind, simulation using a Wright-Fisher population of 10,000 individuals shows that a (conservative) mutation rate of 10−8 per site per generation will introduce a difference into a neutrally segregating pair of 1,000 bp long LTRs by generation 215 on average. The expected frequency of a full-length ERV at the time a difference is introduced is less than 1%, and an ERV can be expected to have a complementary mutant by the time it has reached a frequency of 4% at most. Further, if we assume deletion does not occur, an ERV with mutation free LTRs will fix on only 33% of occasions and an ERV’s LTRs will contain 1 (24% of occasions), 2 (15% of occasions) or 3 or more mutations (28% of occasions) upon fixation otherwise. Thus, simulation suggests that if, as experimental evidence suggests, divergence plays a substantial role in reducing recombinational deletion then deletion should usually occur quickly. If this were not the case then most ERVs that fixed would already contain mutations that hindered deletion, an outcome that would seem to contradict the empirical fact that most ERVs are found in solo-LTR form. In other words, simulation appears to theoretically support our findings.
Not all ERVs are deleted equally quickly. Among the groups of ERVs we examined we found that HERV-H were unusually slowly deleted. This is interesting for several reasons. HERV-H is notable as a family because there is very strong evidence that some HERV-H loci are essential for the maintenance of stem cell identity in humans [3, 4]: half of the full-length HERV-H loci in the human genome are bound by pluripotency associated transcription factors NANOG, OCT5 and LBP9, causing them to produce chimeric stem cell specific transcripts and long non-coding RNAs. It has also been demonstrated that highly transcribed HERV-H loci have diverged faster since the chimp-human split at the nucleotide level than other ERVs and other repetitive DNA [21]. Here our results show that HERV-H loci are more likely to be preserved in a full-length state than ERVs from any of the other five groups we examined. As well as being consistent with the aforementioned mutational divergence hypothesis, this finding may also suggest that full-length HERV-H are useful to the host in a way that solo-LTRs are not. Both solo-LTRs or LTRs that are part of full-length proviruses can act as promoters; we think an implication of our findings is that HERV-H exaptation may involve more than the provision of an LTR promoter and that additional study of the internal regions of HERV-H loci is necessary.
Of course, although we favour the idea that full-length HERV-H is selected for—in the sense that selection may prefer full-length ERVs to solo-LTRs in some instances—there are certainly other hypotheses that could be put forward. Indeed, while we have previously argued that the rapid divergence of what appear to be functional ERVs is evidence of directional selection [21], one could also argue that there might be some other reason that HERV-H is fast evolving. Additionally, it is also possible that full-length HERV-H might be particularly benign, so that for some reason it is easier for full-length HERV-H to fix than for full-length ERVs of other kinds. This idea has some biological justification because it is known that HERV-H envelope genes—involved in cell entry and immunosuppression in exogenous retroviruses—are generally highly degraded [22, 23]. This is not necessarily the case for other full-length ERVs and may have helped full-length HERV-H spread [24]. Alternatively, the unknown number of HERV-H elements that have been exapted may have caused primates to be more tolerant of HERV-H expression in general, making it easier for non-exapted full-length elements to fix. These discussions highlight complex issues, though if one were able to definitively identify those particular HERV-H that had been exapted then some progress could be made. For example, using independent data to classify ERVs as exapted or otherwise, one could determine whether full-length HERV-H are deleted at the same rate as other ERV groups in the case that they are not exapted. In such a scenario, the ERV deletion parameters we show in Fig 3 would probably turn out to be an underestimate of deletion probabilities for non-exapted elements and an overestimate for the others.
It is important to discuss the limits of our model. While phylogenetic assignment of insertions to branches can potentially be more precise than assignments based on LTR dating, such assignment is limited by the resolution of the phylogeny used. This is clear from Fig 2. The branches leading to human and chimpanzee generally reflect a decrease in ERV activity while the branch leading to macaque can only reflect the average activity over a 30 Myr period. Although the average is similar to the average over internal branches, it is probable that the majority of ERV insertions actually occurred closer to the origin of the macaque lineage than to the present day. It is therefore likely that data from the macaque branch will bias deletion rates in the direction of overestimation. Additionally, our approach assumes that ERVs arrive in full-length form. This assumption was necessary, but it is reasonable to point out that some site patterns—those with xs at every tip—describe ERV integrations that might never have been passed vertically from generation to generation in a full-length form. Empirical studies of active ERV families such as KoRV [25] can potentially tell us what proportion of loci can be expected to endogenize in solo-LTR form. Finally, our model is limited by the availability of full-length insertion data. Branches for which all insertions have resulted in solo-LTRs provide no upper bounds on the rate of deletion. For this reason we performed parameter inference on the interval 10−6 to 102, a range broadly compatible with the resolution of our phylogeny. Larger datasets or improved sampling of site patterns can resolve this problem.
Limitations notwithstanding, we hope to show that ERV activity is well described using phylogenetic techniques that avoid LTR dating, and that previous arguments that younger ERVs are more quickly deleted than older ones can be correctly formalized. We also suggest that our finding that HERV-H is deleted slowly across the primate phylogeny supports a long term biological role for some full-length members of the family. Finally, given HERV-H loci seem to have been subject to slow deletion, fast divergence, and are sometimes actively transcribed, we suggest that the in silico identification of ERV families with similar dynamics in other species might be expected to highlight other large scale co-option events. Indeed, a consequence of generalizing the rapid and age dependent nature of solo-LTR formation is the knowledge that such exaptations may be identifiable by locating families with unusually high full-length to solo-LTR ratios within a single genome. From this perspective HERV-H is an outlier in our dataset (S1 Tabular data: Summary of ratios) and others [9]. Such an approach would be reasonable when comparing families with sufficiently similar past activity and when the insertions under consideration are old enough that one would expect rapid deletion to have converted the majority of non-exapted elements into solo-LTRs.
We sample site patterns describing the state of endogenous retroviruses (ERVs) in six primate species and analyze the process of insertion and deletion. Using a phylogenetic framework, we place the activity of six categories of ERV in direct quantitative comparison for the first time. We find that ERVs “die young” but that HERV-H loci are markedly more long-lived than ERV9, HERV-K11, HERV-K, or other class-I and class-II loci. To us, the lower probability of solo-LTR formation for HERV-H loci suggests a long-term host role for the internal regions of exapted elements.
We show that an age dependent Weibull model is sufficient to describe the ERV deletion process and that the use of an exponential decay process is less accurate. As the observation that solo-LTR formation occurs rapidly can be phylogenetically formalized, and generalizes to the majority of ERVs in primate genomes, we propose that the ratio of full-length to solo-LTRs may be used to help identify exapted ERV families in other vertebrate lineages.
In overview, our method was to collect a sample of ERV site patterns from a variety of primates, and to use those patterns, in combination with a host phylogeny, to find the maximum likelihood parameter values for two variations of insertion and deletion processes. This allowed us to decide which model process fit our data best. Below we first describe the process of sampling site patterns and then introduce our phylogenetic techniques.
We obtained the six-way Enredo-Pecan-Ortheus (EPO) whole genome multiple alignment of primate species that forms part of Ensembl Release 71 [26]. The six species included in the alignment are: human (Homo sapiens), chimp (Pan troglodytes), gorilla (Gorilla gorilla), orangutan (Pongo abelii), macaque (Macaca mulatta) and marmoset (Callithrix jacchus).
The EPO dataset contains 7,224 individual alignments containing exactly one sequence for each of the six species. These specific alignments comprise approximately 32% of all alignments in the dataset, which also covers duplicate regions and regions that are present or alignable in only a subset of the six primates. The 7,224 alignments were usually of the order of 105 to 106 columns in length and had a median LTR content of 8%. To identify LTR retroelements we ran RepeatMasker 3.3 [27] on each ungapped sequence from each individual alignment using the -species mammal -no_is -pa 4 -q -nolow -norna options.
The repeats identified by RepeatMasker are often fragmented so that an LTR element originating from a single insertion event is referenced using several distinct annotations. For this reason we applied REannotate 26.11.2007 [28] using options -c -n -f to our RepeatMasker results. This resulted in the identification of complete ERVs, truncated ERVs, and solo-LTRs, entities corresponding to distinct insertional events as defined by [28]. The application of REannotate also conveniently mapped synonymous RepeatMasker identifiers to an appropriate canonical identifier e.g. identifiers HERVH, LTR7, LTR7Y, RTVL-H, RTVL-H2 and RGH were all mapped to identifier HERV-H.
The result of the above repeat masking and annotation processes was data giving the location, repeat type, and structural status of LTR elements in ungapped coordinates. These ungapped coordinates could be mapped back to the appropriate locations in the original EPO multiple alignment files.
To check that REannotate had assigned the correct identifiers to ERVs we performed a BLAST [29] alignment of a representative sequence underlying each repeat locus against a library of 51 viral sequences drawn from [30] and [31]. For any given alignment locus, a full-length representative sequence was preferred to a solo-LTR where available. Representative sequences were drawn from human or the closest primate to human in preference to those that were more distant. For each repeat assigned to families ERV9, HERV-K11, HERV-H and HERV-K, we also submitted the sequences to Dfam [31], and as BLAST queries to NCBI to examine their structure in detail. This resulted in the removal of 8 SVAs that would have otherwise been erroneously included in our study.
Consider a six-way alignment of length m. We form a corresponding 6 by m classification matrix A = {ai,j} in order to combine the information output from the REannotate program with information contained in the alignment. Each entry ai,j is conceptually of one of the following kinds: an unannotated nucleotide coded as d; an unannotated gap coded as g; the ith solo-LTR having identifier id coded as s- i - id; or the ith partial or full-length ERV having identifier id and coded as c- i - id. For example, positions annotated as belonging to the fourth full-length HERV-H in an alignment would be given the classification c-HERV-H-4.
Consider the primate phylogeny T reproduced in Fig 1 where the branch length immediately below node i is denoted Ti and takes the value given by [11]. We wish to relate subsets of our sampled site patterns U (described above), for example, the subset of patterns relating to HERV-H, to an insertion process on the tree T as well as to one of two potential deletion processes, also on T, between which we wish to discriminate. The insertion process is assumed to be Poisson and the deletion process is assumed to be either Weibull or exponential. The exponential deletion model is a nested submodel of the Weibull deletion model. We first discuss insertion and then discuss deletion.
We have now completely specified the state transition probabilities for individual branches under a strict exponential model and under a Weibull model. To compute the likelihood of a site pattern given a tree and a deletion model we use a dynamic programming algorithm that is similar to Felsenstein’s pruning algorithm [33]. In the case that we use a Weibull deletion model, the algorithm must be modified to keep track of the insertion branch when considering the probability of transitions on post-insertion branches.
The above description showed how to calculate the probability of insertions as well as the probability of all post-insertional state transitions that may occur. We have also described how to sample site patterns from primate genomes. Therefore we are ready to describe how to calculate the probability of a set of site patterns U given a tree T, an insertion model Mi and a deletion model Md. For each of n site patterns U(j) we can identify the insertion branch for that pattern, and hence the subtree T(j) of T that includes only the insertion branch and its descendants i.e. any post-insertion branches. To calculate the likelihood of the site patterns we compute:
Pr ( U | M i , M d ) = ∏ i = 1 9 e - ϕ i ϕ i N i N i ! ∏ j = 1 n Pr ( U ( j ) | M d , T ( j ) ) ,
where the first product gives the likelihood of the insertions and the second product uses the aforementioned dynamic programming method to sum over all possible post insertion state transitions.
The insertion model Mi has 9 parameters, the 9 insertion rates in Φ. The deletion model has one parameter if it is strictly exponential (the rate parameter ψe) or two parameters (ψw and the additional shape parameter ω) in the case that it is Weibull. By repeatedly computing the likelihood of our site patterns we can numerically maximize the logarithm of Pr(U|Mi,Md) using code written in the MATLAB language. In practice we performed simulated annealing using simulannealbnd (limited to 5 minutes per replicate during bootstrapping [34] i.e. when sampling the site patterns with replacement) followed by gradient descent using fmincon. As the insertion process is independent of the deletion process we were able to carefully verify our maximum likelihood results using grid search and gradient descent from random starting points.
|
10.1371/journal.pgen.1001295 | Phosphoinositide Regulation of Integrin Trafficking Required for Muscle Attachment and Maintenance | Muscles must maintain cell compartmentalization when remodeled during development and use. How spatially restricted adhesions are regulated with muscle remodeling is largely unexplored. We show that the myotubularin (mtm) phosphoinositide phosphatase is required for integrin-mediated myofiber attachments in Drosophila melanogaster, and that mtm-depleted myofibers exhibit hallmarks of human XLMTM myopathy. Depletion of mtm leads to increased integrin turnover at the sarcolemma and an accumulation of integrin with PI(3)P on endosomal-related membrane inclusions, indicating a role for Mtm phosphatase activity in endocytic trafficking. The depletion of Class II, but not Class III, PI3-kinase rescued mtm-dependent defects, identifying an important pathway that regulates integrin recycling. Importantly, similar integrin localization defects found in human XLMTM myofibers signify conserved MTM1 function in muscle membrane trafficking. Our results indicate that regulation of distinct phosphoinositide pools plays a central role in maintaining cell compartmentalization and attachments during muscle remodeling, and they suggest involvement of Class II PI3-kinase in MTM-related disease.
| Muscles require strong extracellular attachments to preserve cellular integrity during force-generating contractions. Integrin transmembrane receptors mediate muscle attachments at highly localized sites, but how this pattern of attachments is continuously maintained with muscle use is not understood. Human X-linked myotubular myopathy (XLMTM), a frequently fatal muscle disease, is associated with mutations in the MTM1 lipid regulator. Myotubularin (MTM) lipid phosphatases are implicated in endocytosis, a process of cellular uptake that can traffic transmembrane receptors for redelivery to the plasma membrane or to protein destruction. Here, we address MTM roles in muscle, using the genetically tractable fruit fly for detailed investigation of muscle cellular organization and functions. We show that fly muscle cells depleted for mtm function exhibit hallmarks of human XLMTM. We found that mtm regulates integrin localization through endocytosis and, in this role, is needed to maintain muscle attachments. Co-depletion of Class II PI3-kinase with mtm restores normal integrin localization at muscle attachment sites and fly survival, identifying a potential therapy target in MTM-related disease. Importantly, we show that integrin localization is also disrupted in human XLMTM. Our work shows conservation of MTM function in integrin trafficking and reveals insights into regulation of muscle cell maintenance and human disease.
| Myofibers are large, highly differentiated contractile cells that rely on strong extracellular attachments to preserve their integrity during force-generating muscle contractions. Myofiber attachments are mediated by integrin adhesion complexes (IACs) composed of α- and β- transmembrane heterodimers that associate with cytoskeletal bridging factors, similar to those found in non-muscle cells [1]. IACs are crucial at myotendinous junctions (MTJs), attaching the ends of myofibers to tendons. In addition, IACs concentrated at costameres associated with repeating sarcomeric Z-lines attach peripheral myofibrils to the extracellular matrix. IACs are known to be essential for invertebrate and vertebrate muscle cell attachments and organization [2], [3], but it is unclear how the critical pattern of spatially restricted adhesions is continuously maintained.
In non-muscle cells, integrin turnover through endocytic recycling has clear roles in localization of dynamic adhesion complexes that mediate cell migration and membrane remodeling in cytokinesis. Trafficking pathways that engage specific endocytic adaptors, protein kinases and Rab GTPases for internalization and recycling of specific integrins are emerging, as primarily understood in isolated cells [4]. In contrast, it is not clear how important regulated integrin turnover is in differentiated muscle, or how this turnover is regulated. In isolated myofibers, uptake of markers for endocytic recycling occurred in the vicinity of adhesion sites and trafficked to perinuclear compartments, distinct from a degradative pathway [5], suggesting common trafficking themes shared with non-muscle cells. Experiments using fluorescence recovery after photobleaching (FRAP) in intact flies recently provided the first observation of endocytosis-dependent, growth-regulated mobility of IAC proteins at MTJs [6], underscoring the significance of regulated endosomal integrin trafficking in muscles, as well.
Dynamic membrane compartment identity and functions are in part conveyed through phosphoinositides. Phosphoinositides exist as seven phosphorylated phosphatidylinositol forms interconverted by dedicated lipid kinases and phosphatases [7]. Different phosphoinositide forms can recruit specific binding proteins to distinct membranes in order to elicit spatiotemporal responses that include localized signaling, cytoskeletal reorganization, membrane deformation and trafficking. However, the complex cellular relationships in vivo are less defined.
The control of phosphoinositide balance by the Myotubularin (MTM) phosphoinositide phosphatases is both elaborate and crucial in metazoans. MTMs are encoded as a large family of genes (15 humans, 7 flies, 1 yeast) associated with human disease [8]. Mutations in human MTM1 lead to human X-linked myotubular (centronuclear) myopathy (XLMTM, OMIM #310400) characterized by centrally misplaced nuclei in hypotrophic myofibers [9], [10]. Disruption of related MTMR2 leads to Charcot-Marie-Tooth neuropathy (CMT4B1, OMIM #601382) with abnormal morphology and plasma membrane outfolds in myelinating Schwann cells [11]. Both MTM1 and MTMR2 partially localize to endosomal compartments and are attributed with PI(3)P and PI(3,5)P2 turnover [12]–[18]. However, it is not clear how disruption of MTMs and potential regulation of endosomal phosphoinositides might lead to the morphological defects found in MTM-related disease. The kinase(s) that coregulate the relevant phosphoinositide pool(s) for specific MTM functions in muscle have not been explored. Candidates include both Class II and Class III PI3-kinases (PI3KC2 and Vps34, respectively) that generate PI(3)P [19]. Vps34 has well established conserved roles at endosomal membranes. In contrast, PI3KC2 has less-understood roles mostly related with functions at the plasma membrane [18], [20]–[23].
Here we show that mtm (GenBank NM_078765), encoding the sole D. melanogaster homolog of human MTM1/MTMR2, acts with Class II Pi3K68D (GenBank NM_079304) to maintain attachments upon myofiber remodeling. We found that mtm controls β-integrin turnover and trafficking from perinuclear compartments to maintain spatially restricted adhesions at MTJs and costameres, reflecting a broad mtm requirement for integrin-mediated adhesion also needed in the wing. The defects discovered in flies were substantiated by observing similar integrin mislocalization in human XLMTM myopathy, suggesting a conserved MTM1 function in membrane trafficking and roles for integrin adhesions in maintenance of myofiber organization. Altogether, our results identify specific phosphoinositide regulation important for endocytic recycling and dynamic control of cell compartmentalization.
Given the role for myotubularins in human myopathy, and our discovery of an mtm requirement in muscle essential for fly viability [18], we investigated the contribution of mtm-dependent phosphoinositide regulation to muscle cell function and compartmentalization. Loss of mtm function using either null alleles or muscle-directed RNAi had no visible effects on muscle in larvae, which remained mobile and exhibited normal body wall muscle formation, attachments and growth (Figure 1A and 1B–1B′, Figure S1A–S1C′). However, targeted RNAi depletion revealed muscle requirements for mtm at later developmental stages. Muscle-specific mtm knockdown, as indicated by protein depletion, showed either animal lethality (24B-GAL4) or developmental delay (DMef2-GAL4) around the stage of adult eclosion that was rescued by co-expression of either wildtype mtm or human MTMR2 (Figure 1A and 1C, Figure S1B).
Metamorphosis occurs inside a rigid pupal case that adult flies escape at eclosion with the help of muscle contractions, including supporting contractions from a subset of abdominal persistent larval muscles (PLMs) [24], [25]. The PLMs called dorsal temporary internal oblique muscles (IOMs) are large, individual, multinuceated myofibers that span abdominal segments (Figure 1D; Figure S1D). Consistent with defects in eclosion, there was a decrease in the number of both dorsal IOMs and ventral PLMs in mtm-depleted abdomens (Figure 1D′–1E; Figure S1D′–S1E). The remaining mtm-depleted myofibers were frequently detached and seen as rounded-up balls or as elongated fibers with one completely detached end (Figure 1D′, 1F; Figure S1D′, S1F), never observed in controls.
To explore the developmental requirement for an mtm muscle function, we first characterized myofibers using timelapse microscopy in intact animals. With mtm depletion, GFP-labeled IOMs were properly maintained during early pupal stages, when other larval muscles undergo developmentally regulated cell death (Figure 1A,1G–1G′, 2 days after pupal formation, APF). The mtm-depleted IOMs subsequently underwent normal myofiber thinning (3d APF) and rethickening (4d APF), indicative of developmental turnover and rebuilding of the contractile myofibrils [25]. While no detachment was observed in control animals (n = 19), IOM detachment occurred during remodeling in late pupal stages with mtm knockdown (n = 15) (Figure 1G′, 4d APF). Thus, mtm is not essential for IOM formation or survival, but is important for muscle attachments and maintenance upon remodeling.
To address whether mtm plays a role in other muscles, we examined different developmental stages and myofiber types. Although adult somatic muscles appeared to form normally with muscle-specific mtm knockdown (Figure 1D′ and not shown), 100.0±0.0% of the viable adult flies were flightless (versus 22.2±5.4% control; n = 10, ≥124 flies). Visceral muscles that normally migrate to ensheath the testis [26] were present but also disrupted with mtm muscle-depletion (Figure S1G–S1G′, 24B-GAL4). Taken together, mtm function appears dispensable for myogenesis, but is broadly required in both somatic and visceral muscles for myofiber remodeling, maintenance and function.
Pathological hallmarks of XLMTM are small, rounded myofibers with nuclei displacement and disorganization of the perinuclear compartment [8]. In wildtype IOMs, myofibrils are normally tightly packed around centrally aligned nuclei following myofiber remodeling [25] (Figure 1H). In contrast, in mtm-depleted IOMs, central myofibrils were misaligned or absent around a normal number of centrally-displaced nuclei (2.1-fold increased nuclei distance from midline; Figure 1H′–1J). The nuclei were otherwise normal in size and morphology (Figure S2A–S2A′) and pharate adult IOMs were impermeable to propidium iodide staining (Figure S2B–S2C′), while ultrastructural analysis confirmed normal mitochondrial integrity (Figure S2D–S2D′), all indicating viability of mtm-depleted IOM cells. The peripheral myofibrils appeared normal (Figure S2E–S2H), suggesting that mtm is unlikely to function directly in sarcomere assembly.
We also found that transverse (T)-tubules were disrupted in mtm-depleted myofibers, consistent with defects recently described in vertebrate XLMTM [14], [27]. T-tubules are an extensive membrane network, continuous with the sarcolemma, which mediates excitation-contraction coupling throughout the myofiber interior. Although critical for force-generating contractions, there is little understanding of T-tubule biogenesis and structural regulation. We found that both the Amphiphysin (Amph) BAR-domain protein and Dlg1 membrane-associated guanylate kinase scaffold protein localize to T-tubules in wildtype abdominal myofibers (Figure S3A, S3B, S3C), as in flight muscles [28]. In mtm-depleted IOMs, although longitudinal elements of T-tubule membranes were present, lack of Amph and Dlg indicated that transversal membranes were specifically disorganized or absent (Figure S3A′, S3B′; 9.5% control versus 96.3% mtm RNAi with transversal membranes in <half of IOM; n≥21). These conclusions were confirmed by transmission electron microscopy (Figure S3D–S3D′). Altogether, the conserved mutant phenotypes and timing of onset suggests that mtm-depleted muscles in flies model hallmarks of XLMTM.
Given the muscle detachment and myofibril misalignment observed in mtm mutant myofibers, we considered a possible defect in IACs at MTJs and costameres (Figure 2A–2C). We found that βPS-integrin, the single D. melanogaster β-integrin subunit encoded by mys (GenBank NM_080054), was dramatically mislocalized in mtm-depleted muscles (Figure 2B′). In contrast to wildtype muscle, βPS-integrin was absent at the ends of detached myofibers (Figure 2B″) and from costameres (Figure 2C′), consistent with detachment due to disruption of integrin adhesions. Although an intracellular pool of βPS-integrin protein was detected as small punctae within wildtype myofibers (Figure 2D), upon mtm knockdown, βPS-integrin became enriched along abnormal vacuolar inclusions within the myofiber center (Figure 2D′). Ultrastructural analyses revealed large, lucent membrane-bound compartments within the central regions of the mtm-depleted but not control myofibers (Figure 2E–2E″). Other proteins of the integrin adhesion complex, αPS2-integrin and Talin, were both detected at MTJs but not at the inclusions, suggesting βPS-integrin as a primary target of mtm function (Figure S4A–S4B′).
To address the possible relationship between the appearance of membrane inclusions and muscle detachment, we examined βPS-integrin localization at earlier developmental stages. In mtm null or RNAi depleted larval muscles, normal integrin localization was detected at myofiber attachments, without any βPS-integrin-containing central inclusions (Figure S4C, S4C′). This indicates that appearance of inclusions coincides with detachment, and that mtm function is not needed for initial IAC formation. Although detected at the larval myofiber surface, βPS-integrin was not organized into uniform striations in either wildtype or mutant myofibers, as compared to the costameres observed in wildtype pharate adult IOMs. To address whether mtm function affects integrin trafficking prior to myofiber remodeling and detachment, we performed FRAP analysis of βPS-integrin:YFP along MTJs in intact larvae [6]. The mobile fraction of βPS-integrin:YFP was significantly increased in mtm mutant larval muscles (Figure 2F), indicating that mtm is required to stabilize sarcolemmal βPS-integrin localization, preceding myofiber remodeling.
To explore a basis for the sensitivity to mtm loss of function in pupal stages, we investigated integrin localization during IOM remodeling in metamorphosis. In wildtype myofibers at 2–3 days after pupal formation (APF), we discovered that there was a normal loss of integrin from the cell surface, along with detectable presence of integrin-marked inclusions (Figure 2G–2H). By 4 days APF, integrin was again predominantly absent in the myofiber center, with reappearance at costameres. This result reveals a normal redistribution of integrin that occurs with IOM remodeling, and suggests a distinct requirement for myofiber integrin regulation in pupal stages. To test a temporal requirement for mtm function specifically in pupal stages, we performed temperature shift experiments to induce conditional mtm knockdown. Due to the temperature sensitivity of the GAL4 transcription factor, flies with muscle-targeted mtm hairpin expression maintained normal myofiber attachments when raised continuously at 18°C with low GAL4 activity (Figure S4D). However, when flies were shifted during metamorphosis to 29°C for 1–2 days with increased GAL4 activity, the pharate adults then exhibited myofiber detachment (Figure S4D′–S4E, 0% versus 71% mtm pharates, respectively). Similarly, flies also carrying the temperature sensitive GAL80ts, an inhibitor of GAL4, raised continuously at 18°C did not exhibit integrin-containing inclusions (Figure S4F, 0%). In contrast, flies shifted to 29°C for 3 days (with shorter metamorphosis at higher temperatures), exhibited myofibers with integrin-containing inclusions (Figure S4F′–S4G, 58%). These results indicate a requirement for mtm function in pupal stages that is important for integrin localization at the cell surface following myofiber remodeling, and further supports a primary role for mtm in integrin trafficking.
The requirement for mtm function in myofiber remodeling during development raised the question whether there is a similar mtm requirement during cellular remodeling that may occur with ongoing adult muscle use, repair or ageing. We investigated integrin localization in adult abdominal myofibers, which are derived from a different developmental program from the persistent larval muscles. The long, thin adult ventral abdominal muscles, called lateral transversal muscles, normally exhibit a striated pattern of intense integrin localization at repeating costameres (Figure 2I). In contrast, integrin deviated from this pattern with a diffuse distribution in portions of mtm-depleted myofibers in both six and ten day old adult flies (Figure 2I′). This result points to an important role for mtm in the maintenance of integrin adhesions with ongoing muscle use in adult flies.
To test whether mtm function has a specific role for integrin localization in broader developmental contexts, we assayed function of integrin-mediated adhesions in the epithelial bilayer of the developing fly wing [1]. The low frequency of wing blisters resulting from a hypomorphic allele of αPS2-integrin, if3 [29], was dominantly enhanced when in combination with heterozygous mtm null alleles with reduced function (Figure 2J), similar to interactions seen with components known to be required for integrin adhesions [30]. The interaction in two different tissues suggests a specific and fundamental role for mtm in maintenance of integrin-mediated attachments.
The disruption in mtm mutants of both IACs and T-tubules, and their normal proximity along the sarcolemma, raised the question whether a structural or functional relationship between the two compartments normally exists or is relevant to the abnormal membrane inclusions (Figure 3A). Moreover, in mtm-depleted myofibers, we noted that Dlg, similar to βPS-integrin, appeared along abnormal central inclusions (Figure 3B, 3B′). To characterize the membrane identity of the inclusions, we first tested whether βPS-integrin and Dlg or Amph co-localized, either in normal or mtm mutant muscles. In wildtype myofibers, Dlg and Amph were not detected at IACs. However, internal βPS-integrin frequently co-localized with Dlg (Figure 3C) and occasionally with Amph (Figure S5A, S5B, S5C) along apparent longitudinal elements of T-tubules. Upon mtm depletion, βPS-integrin extensively co-localized with Dlg and Amph on longitudinal T-tubules (Figure S5A, S5B′) and along the central inclusions (Figure 3C′, Figure S5B″), suggesting accumulation of a possible common precursor membrane or trafficking compartment in mtm-depleted muscles.
We next considered whether there is functional co-dependence between integrin adhesions and T-tubules. In amph26 null mutants that lack T-tubules in IOMs (Figure S5D), as in adult flight muscles [28], we observed normal muscle attachments, normal βPS-integrin localization to costameres, and no βPS-integrin- or Dlg-inclusions (Figure 3D–3D′). This suggests that T-tubules are not required for βPS-integrin trafficking in the formation or maintenance of IACs, and that the inclusion defects in mtm mutants do not reflect a general consequence of failed T-tubule formation. Furthermore, we found βPS-integrin mislocalized to internal membrane inclusions upon mtm depletion in amph26 null mutants (Figure 3E), signifying that the abnormal inclusions are independent of transverse tubule membrane and possible misregulation of amph function. Conversely, transverse tubules were present normally in abdominal muscles with hypomorphic mys conditions that were pharate lethal (Figure 3F), suggesting that T-tubule organization does not require normal levels of βPS-integrin protein or IACs. The dramatic defects in both IACs and T-tubule organization upon mtm depletion therefore appear to reflect a requirement for two independent mtm functions.
The disrupted βPS-integrin localization together with the enlarged membrane inclusions suggested defective membrane trafficking in mtm mutant myofibers. Characterization of the central inclusions could point to a specific compartment or trafficking step that normally requires Mtm phosphatase activity in muscle remodeling. The inclusions did not noticeably contain markers of endoplasmic reticulum, the trans-Golgi network or autophagosomes (Figure S5E–S5E′ KDEL; Figure S5F–S5F′ PH-FAPP1; Figure S5G–S5G′ Atg8). In contrast, the majority of inclusions were decorated by the endosome-lysosomal marker, GFP:LAMP (Figure 4A–4A″). The inclusions were frequently colocalized with an indicator of early endosomes, GFP:Rab5 (Figure 4B–4B″), and infrequently by the Rab5 effector, Rbsn5 (Figure S5H–S5H′), but not an indicator of late endosome identity, GFP:Rab7 (Figure S5I–S5I′). Together, these results suggest a relationship between the inclusions and early endocytic traffic, and that mtm depletion disrupts endocytic traffic.
PI(3)P is normally enriched at endosomal membranes. We have recently shown that the normal Mtm PI(3)P phosphatase activity promotes membrane efflux, effecting both endosomal homeostasis and cortical remodeling in macrophages [18]. We therefore explored PI(3)P distribution in muscle with respect to integrin adhesion and T-tubule compartments. In wildtype animals, muscle expression of the PI(3)P biosensor, GFP:2xFYVE, was detected along the sarcolemma and localized to punctae distributed throughout abdominal myofibers, with the greatest concentration in the perinuclear area without obvious overlap with βPS-integrin (Figure 4C) or Dlg (Figure 4D). Upon mtm-depletion, enlarged and more erratically positioned PI(3)P-containing compartments were detected (Figure 4C′, 4D′). In addition, GFP:2xFYVE co-localized with βPS-integrin (Figure 4C″) and with Dlg (Figure 4D″) along the abnormal inclusions in mtm-depleted myofibers, suggesting a possible role for Mtm phosphatase activity in PI(3)P turnover involved in integrin trafficking.
To test if PI(3)P regulation is involved in mtm muscle functions, we investigated the contribution of Class II and III Pi3-kinases (Pi3K68D and Vps34, respectively), known to synthesize PI(3)P, to abdominal muscle maintenance. Muscle-targeted knockdown of Pi3K68D or expression of dominant negative kinase-dead Vps34-KD did not individually disrupt eclosion or animal viability. However, Pi3K68D depletion in combination with mtm RNAi was able to rescue the lethality and delayed development; in contrast, Vps34-KD expression enhanced lethality in combination with mtm depletion (Figure S6A, S6B). Neither Pi3K68D nor Vps34 knockdown rescued the loss of T-tubules with mtm-depletion (Figure S6C), and accordingly adult flies remained flightless (Figure S6D). These results indicate separable Pi3K68D-independent and dependent mtm muscle functions required for normal T-tubules and viability, respectively.
A similar functional relationship was seen between Pi3K68D and mtm for roles related to integrin adhesions, as with viability. Importantly, Pi3K68D, but not Vps34, depletion rescued muscle detachment (Figure 5A, 5B) and loss of βPS-integrin localization at costameres (Figure 5C, 5C′, 5D) that occurs with loss of mtm function. Consistent with rescue of the IACs, co-depletion of mtm and Pi3K68D, and not Vps34, also eliminated the βPS-integrin- and Dlg-containing membrane inclusions (Figure 5E, 5E′, 5F, Figure S6E), indicating a functional relationship between the abnormal central inclusions and IACs at the sarcolemma. The testis visceral muscle function was also restored to normal with Pi3K68D and mtm co-depletion, implicating turnover of integrin-mediated adhesions in the gonadal muscle. Altogether, these results signify that Pi3K68D function mediates mtm RNAi mutant defects in maintenance of IACs, and suggest that Pi3K68D may synthesize a PI(3)P subpool co-regulated by Mtm important for integrin trafficking and localization.
Interestingly, muscle-targeted disruption of Vps34 shared with mtm depletion a similar staged semi-lethality (Figure S6A), IOM detachment (Figure 5A′, 5B) and loss of βPS-integrin at costameres (Figure 5C″, 5D), however, without inducing abnormal inclusions (Figure 5E″, 5F, Figure S6E). This points to possible shared or sequential roles for Vps34 and Mtm in phosphoinositide-mediated steps in IAC maintenance, distinct from trafficking points that involve antagonistic Pi3K68D and Mtm co-regulation. Vps34 is broadly attributed with roles in PI(3)P-regulated endocytosis and autophagy. We found that inhibition of autophagy upon depletion of the central regulator, Atg1, phenocopied the Vps34 integrin defects (Figure S6F, S6G, S6H), indicating an important role for autophagy in IOM remodeling. We tested whether the phenotypes associated with mtm phosphatase depletion are a result of increased PI(3)P-mediated autophagy. The myofiber detachment and integrin localization defects, including integrin-containing inclusions, persisted with co-depletion of Atg1 and mtm (Figure S6F′, S6G′, S6H′), indicating that autophagy is not responsible for the integrin-related defects upon mtm depletion.
Given shared defects observed in mtm- and MTM1-disrupted myofibers in flies and human XLMTM, respectively (Figure 1H′, Figure S3A′–S3D′), we asked whether an mtm function required for integrin adhesions is also shared with MTM1 in human muscle. β1-integrin, the major β-integrin isoform found in vertebrate muscle, was detected along the myofiber sarcolemma in cross-sections of skeletal muscle from control subjects, as expected (Figure 6A). In contrast, β1-integrin localized throughout the perinuclear compartment of centronucleated myofibers in muscle from neonates with XLMTM (Figure 6B). The Dystroglycan adhesion complex (DAC) is a second complex localized to MTJs and costameres with key roles in muscle attachments, and mutations of DAC components are frequently associated with muscular dystrophy. Unlike integrin, the dystroglycan transmembrane protein exhibited only the expected peripheral staining along the sarcolemma in both control and XLMTM myofibers, without any abnormal centronuclear localization or inclusion (Figure 6C–6D). These results show that MTM1 is specifically required for normal β1-integrin localization in human myofibers, and suggests that disruption of integrin trafficking and adhesion complex function is important in XLMTM.
We found that mtm regulates integrin adhesions in muscle and in the developing wing, and that integrin localization was disrupted in human XLMTM, pointing to a central role for Mtm/MTM1 in a trafficking pathway important for localization of β-integrin at the plasma membrane. It is well-established that integrin turnover contributes to cell motility, whereby targeted integrin recycling and reassembly of localized adhesions mediate polarized matrix attachments and signaling responses [4]. Our results reveal that regulated integrin turnover is also important for integrin adhesions in non-motile myofibers, after the establishment of attachments. Importantly, mtm disruption uncovered a demand for βPS-integrin trafficking in the maintenance of adhesions both at MTJs as well as at costameres, a less-understood adhesion site with putative roles in muscle integrity, mechanotransduction, and myofibril assembly [31], [32]. Although integrin was destabilized at larval MTJs in mtm mutants, the most severe consequences occurred later with specific loss of pupal or adult mtm function during developmental myofiber remodeling or adult muscle use, respectively. This is consistent with costamere sensitivity to integrin depletion in adult muscle [33] and the possibility that mtm similarly regulates integrin turnover with myofiber remodeling that occurs both in development and with demands in adult muscle growth, repair and aging.
In fly macrophages, Class II Pi3K68D and mtm co-depletion could revert both an imbalance in PI(3)P and defects in cortical remodeling that impaired macrophage shape and in vivo immune cell distribution [18]. Here, we found Pi3K68D disruption is also a specific and potent suppressor of integrin adhesion defects in mtm-depleted muscle. Despite distinct macrophage and myofiber morphology and function, a shared requirement for a PI3KC2/Mtm pathway highlights common functions during cellular remodeling. Loss of Mtm phosphatase activity could be considered a gain of function condition, analogous to ectopic kinase activity, leading to inappropriate phosphoinositide accumulation. In line with this, either mtm depletion (this study) or Pi3K68D overexpression [34] disrupted integrin adhesion in the fly wing, presumably through imbalanced responses to an accumulation of the same phosphoinositide pool. PI3KC2 and Mtm family members in vertebrates have been associated with antagonistic functions related to regulation of traffic to the plasma membrane. PI3KC2 isoforms are required to promote while overexpression of MTM1 impairs GLUT4 trafficking [20], [35] and integrin-mediated cell motility [21], [36]. Together, the observations point to a broad and conserved relationship for PI3KC2/Mtm co-regulation at the plasma membrane.
How might PI3KC2 and Mtm co-regulate integrin trafficking? One possibility is that the cycle of phosphoinositides co-regulated by PI3KC2/Mtm tunes the balance between endocytic-exocytic flux. The strong genetic interaction between mtm and Pi3K68D, in conjunction with Pi3KC2 ability to create PI(3)P in vivo [18], [20]–[23], supports the possibility that Pi3K68D could generate a PI(3)P substrate pool acted on by Mtm phosphatase. Alternatively, Pi3K68D could act more distantly on an interrelated phosphoinositide pool. We envision that Pi3K68D mediates early endocytic trafficking, tethering or sorting of integrin-containing vesicles. The integrin detected on large inclusions in mtm-depleted and XLMTM muscles in flies and humans, respectively, and evidence that mtm promotes membrane tubulation from PI(3)P compartments [18], [37], point to an Mtm/MTM1 role in membrane efflux for delivery of integrin to the plasma membrane. Mtm phosphatase could act to promote recycling or to negatively regulate retention, for example, through a PI(3)P-mediated fusion of integrin-containing vesicles with endosomes-lysosomes. An accumulation of β1-integrin on enlarged, perinuclear compartments has been observed with certain genetic manipulations in non-muscle cells. These results raise the possibility that normal Mtm phosphatase activity functions antagonistically to Rab21 GTPase or in concert with PKCε kinase, Rab11 and/or Arf6 GTPase, respectively, to control redelivery of β-integrin to the plasma membrane [38]–[40]. We found that class III PI3K, Vps34, also contributes to integrin localization upon myofiber remodeling, but with no effect on integrin-containing inclusions. A requirement for class III Pi3K could be at a shared step with the early endosomal Rab5 GTPase shown to be involved in integrin turnover at larval MTJs [6]. Thus, regulation of distinct PI(3)P pools is important for differential regulation of integrin endosomal trafficking, whereby Pi3KC2 and Mtm are dedicated to specific paired antagonistic functions.
We discovered that mtm is required in muscle for both integrin-mediated adhesions and T-tubule organization. The T-tubule requirement for mtm was similar to but not as severe as that for amph, the sole homolog of human AMPH2 that is also associated with centronuclear myopathy [41]. However, unlike mtm, null alleles of amph did not share a defect of myofiber detachment. Despite localization of βPS-integrin at T-tubules, and the dual requirements for mtm, we found that normal integrin adhesions and abnormal βPS-integrin localization on inclusions are independent of T-tubule organization. This suggests that mtm may serve a common function for integrin turnover and T-tubule formation at a shared precursor compartment, for example, at recycling endosomes, or alternatively, act independently at two distinct sites. β-integrin, Dlg and Amph are known to functionally interact at postsynaptic junctions [42]–[44], and MTMR2 has been shown to interact with Dlg1/SAP-97 and Dlg4/PSD-95 to promote postsynaptic function [45], [46]. Thus, the shared accumulation of βPS-integrin, Dlg and Amph on central membrane inclusions in mtm-depleted myofibers, and their elimination with Pi3K68D co-depletion, points to a possible role for a PI3KC2/Mtm pathway in endocytic recycling at neuromuscular junctions, as well as at MTJs.
Many of the defects observed in mtm mutant muscle parallel those associated with the human disease, XLMTM, demonstrating that the fly offers a tractable model for the cellular basis of centronuclear myopathy. Importantly, the discovery that mtm broadly regulates βPS-integrin turnover through endocytic trafficking led us to uncover a previously untested defect in β1-integrin localization in human XLMTM myofibers. Normal myofiber organization and function rely on integrin adhesions in vertebrate muscle [2], [3], [31], [47]. Thus, disruption of integrin regulation provides a basis for aspects of the severity of myofiber disorganization and dysfunction observed in XLMTM. The conservation between fly mtm and human MTM1 functions brings further significance to the potent interaction demonstrated between mtm and class II Pi3K68D for integrin regulation in flies. Whereas Class I and III PI3-kinases have been the focus of intense study as potential therapeutic targets of specific inhibitory compounds, the Class II PI3-kinases have received little attention. The knowledge of PI3KC2 contributions to specific MTM pathways is significant towards motivating similar studies for potential strategies addressing MTM-related disease.
Mtm is the single fly homolog related to both human MTM1 and MTMR2, and human MTMR2 expression was able to rescue integrin-related defects in mtm-depleted fly myofibers. An mtm pathway function in endocytic trafficking is therefore relevant to a more general understanding of the cell biological functions employed by MTM subfamily members. Mutations in MTMR2 associated with CMT4B neuropathy affect the morphology and function of myelinating Schwann cells [11], which like myofibers, share features of having an extensive plasma membrane and a reliance on integrin adhesions [48]. The regulation of integrin trafficking under the control of a conserved PI3KC2/Mtm pathway may be an important mechanism for controlling cell compartmentalization more broadly in different contexts, and relevant to different MTM-related human disease.
Human samples were obtained and used as per institutional IRB accepted protocol.
Flies were reared at 25°C, unless stated. Stocks used include: mtmΔ77, mtmz2-4747 and RNAi hairpins w; UAS-IR-mtm3-1 and w; UAS-IR-mtm3-5 interchangeably or in combination when “2x IR-mtm” noted, w; UAS-mtm:GFP7 and w; UAS-GFP:MTMR2 [18]; UAS-IR-Pi3K68Dv16240 and UAS-IR-Vps34v100296 (VDRC); UAS-Vps34-KDm8 [49]; How24B-GAL4, DMef2-GAL4, UAS-Dcr2; DMef2-GAL4, UAS-2xeGFPAH2, mys1 and mysts1 (Bloomington); if3 (F. Schöck); Ubi-βPS-integrin:YFP [6]; amph26 [28]; w, dlg1:GFPYC0005 (FlyTrap, [50]); UAS-GFP:LAMP [51]; UAS-GFP:Rab5 and UAS-GFP:myc:2xFYVE [52]; UAS-GFP:Rab7 [53]; UAS-mRFP:PH-FAPP1 (G. Polevoy and J. Brill). UAS-GFP:Atg8a [54] and Ubi-Talin:GFP [55]. To reduce βPS-integrin function, mys1/mysts1 flies raised at permissive 22°C were shifted to 29°C at pupation until lethality at pharate stage. To reduce mtm function specifically during pupation, UAS-EGFP/+; DMef2-GAL4, tubulin-Gal80ts/+ and UAS-EGFP, UAS-IR-mtm3.1/+; DMef2-GAL4, tubulin-Gal80ts/+ flies raised at permissive 18°C were shifted to 29°C at 2 days after pupation until pharate stage. To reduce mtm function in adults, UAS-Dcr2; Mef2-GAL4/UAS-EGFP and UAS-Dcr2; Mef2-GAL4/UAS-EGFP, UAS-IRmtm3.1 flies raised at permissive 18°C were shifted to 29°C at eclosion until adults were analyzed at 6 or 10 days old.
Staged pharate adults were removed from pupal case fastened to double-sided tape and pinned on a sylgard covered petri dish in dissecting buffer (5 mM HEPES, 128 mM NaCl, 2 mM KCl, 4 mM MgCl2, 36 mM sucrose, pH 7.2). Abdomens were opened with longitudinal and two lateral incisions, pinned flat, washed and fixed 30 min. (3.7% formaldehyde, 50 mM EGTA, PBS), washed, unpinned and blocked (0.3% bovine serum albuminum, 2% goat serum, 0.1% Triton, PBS), incubated with primary antibody overnight 4°C, washed (0.1% Triton PBS), reblocked and incubated overnight 4°C with Alexa-secondary antibodies (Molecular Probes), counterstained with phalloidin for F-actin and DNA as needed, and mounted in Fluorsave. Antibodies included Amph Ra29 (C. O'Kane), Dlg 4F3 and βPS-integrin CG.6G11 (Developmental Studies Hybridoma Bank), α-tubulin (Sigma-Aldrich), Zormin B1 (B. Bullard), muscle myosin (D. Kiehart), αPS2-integrin 7A10 [56], anti-Rabenosyn-5 [57] and KDEL (Babraham Institute). Propidium iodide (PI) staining was done on tissue. Pharate adult and adult abdomens were dissected in dissection buffer as above, washed once with PBS 0.2% BSA and maintained in 0.3 mg/ml PI final concentration in PBS 0.2% BSA throughout imaging. Images of abdomen fillets were taken with a Leica DMI 6000B inverted microscope using semi-apochromat 5× objective (N.A. 0.15), of live IOMs with a Zeiss Axiovert 200M using a LD-Plan NeoFluar 20× (N.A. 0.4) objective and of individual myofibers with FV1000 Olympus point scanning confocal using 60× Plan Apo N (N.A. 1.2) and 100× Plan Apo (N.A. 1.45) objectives. Exported TIFFs were handled by Adobe Photoshop or ImageJ software.
White pupae were positioned on double-sided tape on a coverslip, placed in a petri dish with water-soaked filter and incubated at 25°C. At each time point, the coverslip was flipped over for imaging on Leica DMI 6000B, as above.
FRAP was carried out as described [6] in living 3rd instar larvae. Integrin:YFP was heterozygous (Int/+) in all experiments. A total of 21 and 19 individual FRAP experiments from multiple larvae were carried out for int/+ and mtmz2-4747/mtmΔ77;int/+ respectively.
Pharate adults dissected and pinned flat were fixed 1 hr. RT (0.1 M sodium phosphate, 3% paraformaldehyde, 2% glutaraldehyde, 2 mM sodium EGTA, 0.1 M sucrose, pH 7.2), postfixed 1 hr. (1% osmium tetroxide in 0.1 M cacodylate buffer) and stained 1 hr. (1% uranyl acetate). Abdomen fillets were embedded in epoxy resin and 70 nm sections were collected on Formvar and carbon-coated copper grids.
To assay viability, pre-cleared vials were counted for surviving adults, dead adults post-eclosion on food and mid-eclosion, dead pharates and dead pupae at 13 or 17 days after egg laying. Flies 1–6 days old were tested for flight at least 24 hours after CO2 anesthesia, by releasing 12 females in a 2 L cylinder (50.8 cm high). Flies that landed below 0.6 L (14.6 cm) were scored flightless.
8 mM cryosections were obtained from muscle biopsies from 3 genetically confirmed cases of XLMTM and age matched cases without obvious histopathology (Carsten Bonnemann), per institutional IRB accepted protocol. Sections were stained using manufacturers instructions (NovoLink kit, Novocastra) and as previously described [14] using primary antibodies to beta1D integrin (Chemicon; 1∶25) and dystroglycan (Novocastra; 1∶20).
Visual quantification was made for number of IOMs in tergites 3 and 4, number of detached IOMs per abdomen, number of IOMs displaying βPS-integrin costameres or βPS-integrin- or Dlg-marked inclusions, and number of nuclei per IOM. ImageJ software was used to draw and measure nuclei distance to IOM midline and sarcomere length. CellProfiler was used to segment and quantify nuclei morphology. Statistical analysis in Prism software used to determine mean, standard error and Student's t-test, where possible.
Full genotypes used are as shown in Figure Legends and as follows:
Figure 1. (D,G,H,I,J) Control, UAS-EGFP/+; DMef2-GAL4/+. (D′,G′,H′,I,J) RNAi: UAS-EGFP, UAS-IR-mtm3.1/+; DMef2-GAL4/+. (C) 24B-GAL4. Control: UAS-EGFP/24B-GAL4. RNAi: UAS-EGFP, UAS-IR-mtm3.1/24B-GAL4. RNAi with mtm cDNA: UAS-mtm:EGFP, UAS-IR-mtm3.1/24B-GAL4. RNAi with human MTMR2 cDNA: UAS-EGFP:huMTMR2, UAS-IR-mtm3.1/24B-GAL4. DMef2-GAL4. Control: UAS-EGFP/+; DMef2-GAL4, UAS-LacZ/+. RNAi: UAS-EGFP/+; DMef2-GAL4, UAS-IR-mtm3.5/+. (E, F). 24B-GAL4: Control: UAS-EGFP/24B-GAL4. 1x RNAi: UAS-IR-mtm3.1/24B-GAL4. 2x RNAi: UAS-IR-mtm3.1/24B-GAL4; UAS-IR-mtm3.5/+. DMef2-GAL4. Control: UAS-EGFP/+; DMef2-GAL4/+. 1x RNAi: UAS-IR-mtm3.1/+; DMef2-GAL4/+. 2x RNAi: UAS-IR-mtm3.1/+; DMef2-GAL4/UAS-IR-mtm3.5.
Figure 2. (B,C,D,E) Control: UAS-EGFP/+; DMef2-GAL4/+. (B′,C′,D′,E′) RNAi: UAS-EGFP, UAS-IR-mtm3.1/+; DMef2-GAL4/+. (F) Control: Ubi-βPS-integrin:YFP/+. mutants: mtmΔ77/mtmz2-4747/Ubi-βPS-integrin:YFP. (G,H) dlg1:GFP. (I) Control: UAS-Dcr2; Mef2-GAL4/UAS-EGFP. RNAi: UAS-Dcr2; Mef2-GAL4/UAS-EGFP, UAS-IRmtm3.1.
Figure 3. (B,C) Control: UAS-EGFP/+; DMef2-GAL4/+. RNAi: UAS-EGFP, UAS-IR-mtm3.1/+; DMef2-GAL4/+. (D, D′) amph26. (E) amph26/amph26, UAS-IR-mtm3.1/+; DMef2-GAL4/+. (F) mys1/mysts1.
Figure 4. (A) Control: UAS-GFP:LAMP/+; DMef2-GAL4/+. (A′–A″) RNAi: UAS-GFP:LAMP, UAS-IR-mtm3.1/+; DMef2-GAL4/+. (B) Control: UAS-GFP:Rab5/+; DMef2-GAL4, UAS-lacZ/+. (B′–B″) RNAi: UAS-GFP:Rab5/+; DMef2-GAL4, UAS-IRmtm3.5/+. (C,D) Control: UAS-GFP:myc:2XFYVE/+; DMef2-GAL4/+. (C′,D′,C′,D″′) RNAi: UAS-GFP:myc:2XFYVE, UAS-IR-mtm3.1/+; DMef2-GAL4/+.
Figure 5. (B,C,D,E,F) Control: UAS-EGFP/+; DMef2-GAL4, UAS-lacZ/+. RNAi: UAS-EGFP/+; DMef2-GAL4, UAS-IR-mtm3.5/+. (A,A′,B,C′,C″D,E′,E″,F) Single RNAi: DMef2-GAL4, UAS-lacZ/UAS-IR-Pi3K68D or UAS-IR-Vps34/+; DMef2-GAL4, UAS-lacZ/+. mtm co-RNAi: DMef2-GAL4, UAS-IR-mtm3.5/UAS-IR-Pi3K68D or UAS-IR-Vps34/+; DMef2-GAL4, UAS-IR-mtm3.5/+.
|
10.1371/journal.pntd.0005613 | Transmission dynamics and control of Rickettsia rickettsii in populations of Hydrochoerus hydrochaeris and Amblyomma sculptum | Brazilian Spotted Fever (BSF), caused by the bacterium Rickettsia rickettsii, is the tick-borne disease that generates the largest number of human deaths in the world. In Brazil, the current increase of BSF human cases has been associated with the presence and expansion of capybaras Hydrochoerus hydrochaeris, which act as primary hosts for the tick Amblyomma sculptum, vector of the R. rickettsii in this area.
We proposed a semi-discrete-time stochastic model to evaluate the role of capybaras in the transmission dynamics of R. rickettsii. Through a sensitivity analysis, we identified the parameters with significant influence on the R. rickettsii establishment. Afterward, we implemented the Gillespie’s algorithm to simulate the impact of potential public health interventions to prevent BSF human cases.
The introduction of a single infected capybara with at least one infected attached tick is enough to trigger the disease in a non-endemic area. We found that to avoid the formation of new BSF-endemic areas, it is crucial to impede the emigration of capybaras from endemic areas by reducing their birth rate by more than 58%. Model results were corroborated by ex-situ data generated from field studies, and this supports our proposal to prevent BSF human cases by implementing control strategies focused on capybaras.
The proposed stochastic model illustrates how strategies for the control and prevention of vector-borne infectious diseases can be focused on amplifier hosts management practices. This work provides a basis for future prevention strategies for other neglected vector-borne diseases.
| A stochastic model for the spread of R. rickettsii among the Amblyomma sculptum tick vector and the Hydrochoerus hydrochaeris, amplifier host, was formulated. We found that the introduction of a single infected capybara, with at least one infected attached tick, is enough to trigger the disease in a non-endemic area. To avoid disease propagation, it is crucial to impede the emigration of capybaras from endemic areas by keeping the population in a stable state. Otherwise, to eliminate the disease from endemic areas, a reduction by more than 58% in the capybara birth rate is necessary. These results allow improving the planning of public actions to prevent Brazilian Spotted fever based on the control of capybara population and provided a basis for planning prevention programs for other neglected vector-borne diseases.
| Rickettsia rickettsii is the etiological agent of the Brazilian spotted fever (BSF), the deadliest spotted fever in the world. This infection is partially pathogenic to Amblyomma sculptum ticks, main vectors of the R. rickettsii in South America [1, 2], generating a drop in the infection rate with each tick generation [3]. In addition, A. sculptum ticks are unable to maintain the R. rickettsii infection in successive generations by transovarial and transstadial transmissions [4]. Hence, the maintenance of R. rickettsii depends on a constant introduction of new susceptible animals (i.e., newborn of vertebrate hosts), which act as amplifier hosts and guarantee the constant creation of new cohorts of infected ticks [4–7]. This suggests that control strategies focused on vectors [8, 9] are not enough for the prevention of this tick-borne infectious disease.
In Brazil, the capybara Hydrochoerus hydrochaeris acts as the amplifier host of R. rickettsii infection [10, 11]. In southeastern Brazil, both capybaras and BSF occurrences have increased significantly over the last three decades [5, 12]. In turn, these occurrences have been spatiotemporally associated with rising production and spatial expansion of sugarcane crops, the main food source of capybaras [13]. In BSF-endemic areas, population densities of capybaras have reached numbers up to 40 times higher than those recorded in natural environments such as the Amazon and Pantanal [14]. However, the effectiveness of control strategies focused on this amplifier host and their impact on the transmission of R. rickettsii are unknown.
The dynamics of complex transmission cycles, such as tick-borne infectious diseases have been broadly analyzed. Hudson et al. published the first deterministic model representing the dynamics of the Louping-ill disease, an acute viral zoonosis which mainly affects sheep [15]. O’Callaghan et al. put forward tick-borne diseases dynamics considering the potential effect of a vaccination program for Ehrlichia ruminantium [16]. Similarly, Rosà et al. developed a model for the tick-borne encephalitis virus including the tick stages and different transmission routes [17, 18]. Despite their important role in the understanding of vector-borne diseases dynamics, unfortunately they do not include the seasonal behavior of the vector life cycle, neither the onset nor the extinction of these diseases. As it is well established, stochastic models should be used in phenomena that do not satisfy the law of large numbers such as large communities with minor outbreaks [19]. In fact, the extinction of endemic diseases can only be analyzed with stochastic models, since extinction occurs when the epidemic process deviates from the expected level [19].
In this work, we propose a discrete-state semi-discrete-time stochastic framework to evaluate the role of capybaras in the transmission dynamics of R. rickettsii. We identify the parameters with significant influence on the R. rickettsii establishment and subsequently evaluate the impact of potential public health interventions to prevent BSF in humans.
The R. rickettsii dynamics in populations of H. hydrochaeris and A. sculptum is represented in Fig 1. Our model was adjusted to a semi-discrete time dynamics in order to consider the 1-year life cycle of the tick A. sculptum, which is primarily controlled by the larval behavioral diapause [20]. Thus, larvae exclusively quest and feed from April to July for 110 days, nymphs from July to October for 104 days and adults particularly quest, feed and reproduce from October to March for 151 days, as shown in Fig 1.
The tick population is classified according to the life cycle stages as larvae (L), nymphs (N) or adults (A), which could be detached from a capybara or attached to it. When a tick gets infected by an infected capybara, it remains infected until it dies. Thus, each A. sculptum stage is also classified according to whether it is susceptible (S) or infected (I). In this way, the population of each tick stage is represented by three indexes, where the first index denotes the infection status, the second index denotes the tick life cycle stage, and the third index denotes detachment (D) or attachment (A).
Attached ticks at the larval and nymph stages detach at the respective rates of θL and θN. Furthermore, detached ticks at larval, nymph and adult stages can attach at rates αL, αN and αA, and can die at rates δL, δN and δA. The production rate ρ of larvae is assumed to be proportional to the total number of adult attached ticks AA.
On the other hand, capybaras are classified as susceptible SC, infected IC and recovered RC, as shown in Fig 1. They reproduce at a constant rate μC, die at rate δC and recovered at rate γ. All capybaras have the same susceptibility and there is no increased death rate of infected individuals due to disease. Susceptible capybaras can get infected by an attached larva, nymph or adult tick at a rate of λL, λN and λA, respectively. Once capybaras are infected, they keep the R. rickettsii in the bloodstream for 7 to 10 days [11], during which the infection of new susceptible larvae or nymphs that feed on it can occur at rates βL and βN, respectively. After this period, capybaras become immune to the disease. We do not consider a transmission rate from infected capybaras to susceptible adult ticks βA, because of the time of R. rickettsii infection in ticks is greater than the time of laying. Thus, eggs can be infected only if adult ticks become infected during the larval or nymph stages [4]. Additionally, we consider a vertical transmission and hence infected adult ticks can produce infected detached larvae at rate ρII. In this way, there are 25 distinct reactions in the stochastic model which are listed in Table 1. The equivalent deterministic equations associated with the stochastic reactions are presented in the S1 File.
The stochastic process detailed in Table 1 was simulated in the R programming language employing the Gillespie’s algorithm. For each simulation we performed the number of iterations needed to get a stable performance.
We assumed a finite number of individuals distributed over a finite set of discrete states. Given an initial time t0 and an initial population state X(t0), the Gillespie’s algorithm allows us to generate the time-evolution trajectory of the state vector X(t) ≡ (X1(t), …, XN(t)), where Xi(t) is the population size of state i at time t and N is the number of states. Changes in the number of individuals in each state occur due to reactions between interacting states. The states interact through M reactions Rj, where j = 1, …, M denotes the jth reaction. A reaction is defined as any process that instantaneously changes the population size of at least one state [21, 22]. The time step to the next reaction was then determined as τ = 1 α 0 ( x ) l n ( 1 / r 1 ), where α0(x) = Σαj(x) and the index of the next reaction to execute Rj is the smallest integer j satisfying j = Σ i = 1 j α i ( x ) > r 2 α 0 ( x ) [21, 22].
All parameters were estimated using previously generated data from ex situ field works in southeastern Brazil [4, 5, 11, 14, 23]. It is noteworthy that the natural capybara birth rate μ assumes the value of 70% of adults, 64% of females, a litter size mean of 4.2 pups, 1.23 births per female per year and a pregnancy success of 85% [14, 23]. If capybaras die at an exponential rate, then δC is the fraction required to die each day. The birth rate of a susceptible tick was determined assuming a female weight of 500mg [4], CEI (mg egg mass/mg engorged female × 100) of 48.4% [4],18.8 eggs/ 1 mg of eggs [24] and hatching success of 68% [4]. Likewise, the birth rate of an infected tick was determined assuming a transovarial transmission of 42.8% [4], filial infection rate of 50% [4], female weight of 372.20 mg [4], CEI of 39.55% [4], 18.8 eggs/1 mg of eggs, [24] and hatching success of 44.2% [4]. We consider a population of 20 adult female ticks per capybara and the groups of capybaras were restricted to 50 individuals [14, 23, 25]. A full list of the model’s parameters used in the simulations is given in Table 2.
To quantify the impact of the variation of each parameter on the output of the BSF model, we combined uncertainty through the Latin hypercube sampling (LHS) with the robust Partial rank correlation coefficient (PRCC) method [26, 27]. Initially, we obtained a random parameter distribution divided into one hundred equal probability intervals, which were then sampled. Thus, a LHS matrix was generated with one hundred rows for the number of simulations (sample size) and six columns corresponding to the number of varied parameters (α, μC, λ, γC, δC, ϵC). The parameters α and λ were varied according to the life cycle season. BSF model solutions were then simulated using each combination of parameter values through ten year simulation. One thousand model outputs were obtained and the parameter and output values were transformed into their ranks. Subsequently, we computed the PRCCs between each parameter and the average infected population size.
Initially, we simulated the formation of an endemic area by the introduction of infected capybaras and infected attached ticks. For a more realistic simulation of current BSF-endemic areas of southeastern Brazil, we consider a growing population of capybaras (births greater than deaths and high emigration). We assume that this scenario corresponds with the onset of a BSF epidemic. We found that the introduction of a single infected capybara or a single infected tick is unable to trigger the disease in a non-endemic area of 50 susceptible capybaras. However, the introduction of an infected capybara with at least one infected tick attached is enough to establish a new endemic area, as shown in Fig 2 scenario A. This scenario illustrates how the fraction of infected capybaras and ticks converges to a constant from year 2 onward. In this equilibrium state, the average fraction of infected capybaras is 8.9% (95% CI = 0%-28.6%), 18.2% (95% CI = 0%-44.9%) and 17.5% (95% CI = 0%-43.9%) in the larvae, nymphs and adults season respectively. In this scenario, the average fraction of infected detached larvae is 1.25% (95% CI = 0%-7.04%), infected attached larvae 0.7% (95% CI = 0%-6.5%), infected detached nymphs 1.35% (95% CI = 0%-9.36%), infected attached nymphs 0.8% (95% CI = 0%-6.95%), infected detached adults 0.46% (95% CI = 0%-5.17%) and infected attached adults 0.54% (95% CI = 0%-5.64%). These results are consistent with previous observations in BSF-endemic areas in which the fraction of infected A. sculptum adults attached to horses has been reported at 0% [28] and the fraction of infected detached adults A. sculptum at 1% (95% CI = 0.01%-7.8%) [28], 0.2% (95% CI = 0.01%-1.04%) [10] and 1.28% (95% CI = 0.07%-5.59%) [29]. The fraction of infected capybaras and other tick populations remains unreported.
In scenario A, an average of 57.1% (95% CI = 22.8%-91.4%), 78.2% (95% CI = 49.6%-99.9%), and 77.2% (95% CI = 48.1%-99.9%) of capybaras became immune after a primary infection with R. rickettsii in the larvae, nymph and adult seasons, respectively. In practical terms, these immune capybaras include the Rickettsia-seropositive animals that are usually found in serosurvey studies in BSF-endemic areas. These numbers agree with previous serosurvey studies that reported 50-80% of capybaras to be seropositive for R. rickettsii in BSF-endemic areas [10, 30]. In this established endemic area, we observe that 563 capybaras migrated over 10 years.
Sensitivity analysis of scenario A show that the capybaras birth rate variation had the greatest impact on the average infected population size. In the nymph and adult ticks seasons, when the infection of capybaras is greater, the correlation value between the birth rate of capybaras and the average infected population is PRCC>0.6, being significant in both seasons (Fig 3). Since in BSF-endemic areas of southeastern Brazil the population of capybaras is growing [5, 12–14], theoretically, the epidemic will survive forever [31]. Considering μ = δC + ϵ, where μC represents birth, δC death and ϵ emigration rate of capybaras, the population growth generates a high rate of emigration and consequently the spread of the disease. In this way, to better understand the effect of changes in capybara population, we investigate the host-tick-infection dynamics for different values of μC, under three additional scenarios: 1) an endemic area with a stable state of capybaras (births equal deaths, no emigration and no importation of the disease from outside); 2) an endemic area with a decrease of 80% and 3) 90% in the birth rate of capybaras.
Contrary to the expected behavior [32], when the population is in a stable state, the disease does not disappear. Indeed, the proportion of infected individuals remains constant over time from year 2 onward as shown in Fig 2B. In this scenario, the average fraction of infected capybaras is 4.6% (95% CI = 0.01%-19.3%), 6.3% (95% CI = 0.01%-23.1%) and 7.4% (95% CI = 0.01%-25.4%) in the larvae, nymphs and adults seasons, respectively. To ensure births equals deaths (μC = δC) and thereby guarantee no emigration (ϵ = 0), it is necessary to control the capybara’s birth rate in 58%. Thereby, though the disease does not disappear, this birth rate guarantees that the R. rickettsii will not spread from BSF-endemic areas.
Thus, for the elimination of the R. rickettsii from endemic areas, a decrease in the capybara’s birth rate to values lower than 0.0021 is necessary. However, this can lead to a decline in capybaras population over time. When a decrease of 80% (μC = 0.0011) in the birth rate of capybaras is considered, infected individuals tend to disappear in the fourth year along with a decrease in the total population size, as shown in Fig 2C. Otherwise, when a decrease of of 90% (μC = 0.0005) in the birth rate of capybaras is considered, infected ticks and capybaras tend to disappear from the second year as shown in Fig 2D.
Strategies to reduce the birth rate of capybaras include the reduction of the carrying capacity, their removal, either by euthanasia or regulated hunting, and their reproductive control. As capybaras natality depends primarily on the availability of food sources, as is typically the case for rodents [33], the reduction of the carrying capacity in BSF-endemic areas is a plausible strategy to reduce their birth rate [34]. Polo et al. found a spatiotemporal relationship between the occurrence of BSF and the increment and expansion of sugarcane crops, the main food source of capybaras in southeastern Brazil and the most important agricultural product in the region [13]. Furthermore, in this area, there is a constant availability of water sources, which allow the establishment of capybaras, as this is a semiaquatic vertebrate that depends on water sources for thermic regulation, reproduction, and predator protection [33]. Certainly, controlling these aspects is not feasible.
Additionally, because of the constant increment and abundance of vital resources offered by the environment in BSF-endemic areas of southeastern Brazil, it is important to consider that in response to the removal or elimination of recovering capybaras from theses areas, a reintroduction of susceptible animals can occur [13]. This, along with the long survival of unfed A. sculptum in pastures [35], and the fact that just one infected capybara with a single infected tick attached is sufficient to establish an endemic area, can cause a rapid spread of the disease and consequently an increased risk of transmission to humans.
Reproductive control of capybaras through deferentectomy and ligation of fallopian tubes was previously tested in southeastern Brazil [36]. It was observed that the reproductive management did not negatively influence individual or collective behavioral aspects, with the animals defending their territory and not migrating [36]. The sterilization of capybaras has already been authorized as a way to prevent BSF human cases in a small endemic area of southeastern Brazil [37].
Future studies encompassing field data should be performed to evaluate the role of alternative reservoirs in the dynamics of R. rickettsii. This is a limitation of our work since we considered BSF-endemic areas of southeastern Brazil, where capybaras are the major, though not the exclusive, hosts for either larval, nymphal or adult stages of A. sculptum. In most BSF-endemic areas, the only medium-to large-sized animal species is the capybara; therefore it is the only host for the A. sculptum adult stage, and consequently, the only suitable host species to sustain an A. sculptum population in the area [7, 10, 38]. However, there have been some previous reports of A. sculptum immature stages (larvae and nymphs) on small animals (wild mice, marsupials, birds) which usually share the same habitat with capybaras [39, 40]. Nevertheless, comparing to capybaras, the amount of larvae or nymphs that feed on these animals is minimal; i.e., while hundreds to thousands larvae and nymphs are commonly found feeding on a single capybara [41], we usually find less than 10, or exceptionally a few dozen A. sculptum ticks on individual small animals [39–41]. Indeed, this condition is also favored by the fact that all active stages of A. sculptum tend to host-quest on vegetation above 15 cm from the soil, precluding their direct contact with small mammals or ground-feeding passerine birds [20, 42].
This work offers an alternative for the planning of prevention strategies for tick-borne neglected infectious diseases. The planning of these type of interventions is usually performed by public entities through heuristic techniques, such as trial and error methods, without obtaining optimal results. The results of our work accurately match with data previously generated from field studies on the R. rickettsii dynamics in southeastern Brazil and will potentially allow the formulation of public policy to prevent BSF human cases based on the control of capybara population growth. Furthermore, this work will provide a basis for the planning of prevention programs for other neglected vector-borne infectious diseases.
|
10.1371/journal.pcbi.1002066 | A New View of the Bacterial Cytosol Environment | The cytosol is the major environment in all bacterial cells. The true physical and dynamical nature of the cytosol solution is not fully understood and here a modeling approach is applied. Using recent and detailed data on metabolite concentrations, we have created a molecular mechanical model of the prokaryotic cytosol environment of Escherichia coli, containing proteins, metabolites and monatomic ions. We use 200 ns molecular dynamics simulations to compute diffusion rates, the extent of contact between molecules and dielectric constants. Large metabolites spend ∼80% of their time in contact with other molecules while small metabolites vary with some only spending 20% of time in contact. Large non-covalently interacting metabolite structures mediated by hydrogen-bonds, ionic and π stacking interactions are common and often associate with proteins. Mg2+ ions were prominent in NIMS and almost absent free in solution. Κ+ is generally not involved in NIMSs and populates the solvent fairly uniformly, hence its important role as an osmolyte. In simulations containing ubiquitin, to represent a protein component, metabolite diffusion was reduced owing to long lasting protein-metabolite interactions. Hence, it is likely that with larger proteins metabolites would diffuse even more slowly. The dielectric constant of these simulations was found to differ from that of pure water only through a large contribution from ubiquitin as metabolite and monatomic ion effects cancel. These findings suggest regions of influence specific to particular proteins affecting metabolite diffusion and electrostatics. Also some proteins may have a higher propensity for associations with metabolites owing to their larger electrostatic fields. We hope that future studies may be able to accurately predict how binding interactions differ in the cytosol relative to dilute aqueous solution.
| The cytosol is the major cellular environment housing the majority of cellular activity. Although the cytosol is an aqueous environment, it contains high concentrations of ions, metabolites, and proteins, making it very different from dilute aqueous solution, which is frequently used for in vitro biochemistry. Recent advances in metabolomics have provided detailed concentration data for metabolites in E.coli. We used this information to construct accurate atomistic models of the cytosol solution. We find that, unlike the situation in dilute solutions, most metabolites spend the majority of their time in contact with other metabolites, or in contact with proteins. Furthermore, we find large non-covalently interacting metabolite structures are common and often associated with proteins. The presence of proteins reduced metabolite diffusion owing to long lasting correlations of motion. The dielectric constant of these simulations was found to differ from that of pure water only through a large contribution from proteins as metabolite and monatomic ion effects largely cancel. These findings suggest specific protein spheres of influence affecting metabolite diffusion and the electrostatic environment.
| The composition of metabolites, ions and proteins, and processes such as metabolism and signalling which take place in the E.coli cytosol are largely well defined [1], [2]. However, the structure and dynamical nature of the cytosol solution is less well understood whether on the local or cytosol-wide levels. Current perception of the cytosol solution is often of a unstructured mixture with behaviour that differs quantitatively but not qualitatively from an ideal solution. Alternatively, there are theoretical descriptions of a cytosol organised into functionally specific regions and even separated protein and small molecule regions linked by metabolite transit pathways [3], [4].
Cytosolic metabolites are extremely varied, but a large majority of these molecules are negatively charged. Assumed electro-neutrality is maintained by a large concentration of potassium ions and to a lesser degree by magnesium and poly-amines such as putrescine and spermidine. The large amount of charge in the cytosol suggests that electrostatics is a dominant force. However, the Debye length at physiological ionic strength is very short (less than 1 nm) [3], [4]. This electrostatic screening is probably essential for the observed extent of macromolecular crowding [5].
The charge distribution and dynamics of the solution also determines the dielectric constant (), which is reduced by increasing concentration of monatomic ions [6], [7] while Zwitterionic metabolites are thought to increase [8]–[10]. The effect of proteins seems to vary, with some studies suggesting an increment [11]–[13] and others a decrement [14], [15]. Experimental data on of the cytosol is sparse but in general it suggests that cytosolic is significantly larger than that for pure water [16]–[19].
The hydrophobic effect is also significantly modulated by ionic strength. Increasing salt concentration increases the strength of the hydrophobic effect [20] possibly through the weakening of water hydrogen bonding [21]. Almost all theoretical treatments of these issues assume simple solutions of monatomic ions and water, sometimes at infinite dilution. There has been little examination of differences in solutions containing positive monatomic ions and larger, negatively charged solutes.
Given the complexity of the cytosol environment it is very difficult to predict the true nature of structure, dynamics and thermodynamics. With a high level of electrostatic screening and heightened hydrophobicity, is it likely that metabolites and proteins engage in significant and long lasting interactions? A recent theoretical study has attempted to make sense of non-ideal behaviour for two component solutions of some common organic molecules [22]. For some mixtures it was shown that activity can actually decrease with increasing concentration, suggesting a high level of non-ideal behavior. Another study found significantly lower thermodynamic activities between in vivo like and standard conditions for enzyme-inhibitor assays, again suggesting significant non-ideal behaviour [23].
Using a recent extensive list of metabolites and their concentrations from exponentially growing E. coli [24], we have produced two types of atomistic molecular dynamics simulations of a simplified cytosolic model. One included metabolites only and another also included a protein component, for which we used ubiquitin. Although ubiquitin (PDB code 1UBQ) is a eukaryotic protein, it was chosen owing to its small size and large amount of literature dedicated to its study [25]–[27]. Molecular dynamics allowed us to compute several properties of interest, including , amount of contact between cytosolic molecules and diffusion coefficients. The simulations indicate that metabolites spend a large proportion of their time as part of ‘non-covalently interacting metabolite structures’ (NIMS). Our results also indicate that the cytosolic is larger than that of water with monatomic ions. These data allow us to make suggestions about the global structure of the cytosol and the amount of time different metabolites spend free in solution.
This study involved two large cytosol simulations with cubic boxes of 100 Å dimensions, one containing metabolites with monatomic ions (100M) and another with four additional ubiquitin molecules (100MP). Two smaller cytosol simulations (50M and 50MP), a pure water (tip3p) and water + KCl (tip3p+KCl) all with Å dimensions were produced for the dielectric analysis. For a complete list of the simulations of this study and their simplified labels it is instructive to refer to table 1 and the methods section.
The structure of cytosol simulations quickly collapsed from almost equal spacing of metabolites to a series of non-covalently interacting metabolite structures (NIMS) inter-spaced with solvent, ions and fully solvated metabolites. This process was conveniently measured through solvent accessible surface area (SASA) of all metabolites except monatomic ions (Figure 1).
Both 50 Å simulations were deemed equilibrated after 30 ns (Figure 1) while the 100M and 100 MP were equilibrated after 35 and 50 ns respectively. Hence, all analyses were carried out only on this structurally equilibrated data (see supporting information Figure S7). Around 16.7% of SASA is lost within the 50 MP system which is similar to the 100 MP box where around 16.4% is lost. These percentage values were calculated using the running averages shown in red in Figure 1. The effect of the box size on metabolite behaviour and general size of NIMS is difficult to gauge but the fact that there is little relative difference between 100 and 50 Å may suggest that smaller box sizes can be used for computationally expensive calculations.
Figure 2 shows a view through the 100 M box at the beginning of the production simulation and after 200 ns. It is clear that after equilibration there is a significant difference in structure. Within the 200 ns simulation of the 100 Å boxes many NIMS were formed which were stable over relatively long time periods. The most interesting of these NIMS were those with a stacking core of nucleotide base like groups (Figure 3 A). These stacks continuously gain and lose bases and persist as long as 50 ns. Some stack NIMS seem reminiscent of RNA and we speculate that these structures often show similarities with the elongation complex of RNA polymerases [28] in the way phosphates are aligned with ribose rings (Figure S3).
The inclusion of four ubiquitin molecules perturbed the metabolite structures. Many large NIMSs became attached to protein surface areas containing positively charged residues (Figure 4), in many cases for time periods of 50–100 ns. The attachment or detachment of large NIMS from the protein may contribute to the large SASA fluctuations of Figure 1. These protein-connected NIMSs can also form bridges connecting two proteins which correlates their motions (Figure 4).
SASA analysis was used to investigate any propensity for metabolites to interact. For the SASA and diffusion analyses, only the 100 Å boxes are discussed, however the 50 Å boxes were found to follow similar trends. As might be expected metabolites with larger surface areas have more contact with non-water entities. A comparison of the average contact area in the 100 M and 100 MP boxes for each type of metabolite can be found in the supporting information (Figure S4).
Figure 5 displays the amount of time metabolites spend in contact with other molecules and hence are unavailable for any specific interactions. The threshold for our definition of contact is two hydrogen bonds or more (see methods section). Larger metabolites are contacted at least 70% of the time while smaller metabolites show much larger variability with some as low as 20% and other as high as 95%. This analysis gives an indication of metabolites availability for metabolism but of course cannot replace thermodynamic data. A comparison of time in-contact data for the 100 M and 100 MP simulations can be found in supporting information (Figure S5 and Tables S1 and S2). Also further analysis of average and maximum contact events is presented in Figure S2.
A SASA analysis was also applied to ubiquitin, in the 100 MP simulation, to find the metabolite contact area for each ubiquitin residue. Here the residue contact area is defined as the SASA without the environment minus the SASA with the environment and this was reported as a percentage of the average SASA without the environment. Those ubiquitin residues which interact with metabolites most are part of the same patch (nine of the top ten percentage contact area, see supporting information Table S2). Lys 48, becomes covalently attached to the C-terminus of other ubiquitin molecules is part of this patch [25]. This patch was involved in a very close contact event between two ubiquitin molecules in the 100 MP simulation (supporting information Figure S1).
Diffusion coefficients were calculated through the Einstein-Helfand relation. Figure 6 shows the diffusion coefficient against the number of atoms for each type of metabolite in the 100 Å boxes. Recent work has shown a periodic box size dependence for water diffusion in water [29]. Here some diffusion rates were slightly reduced in the 50 Å compared to the 100 Å boxes, however many were identical (Figure S6 of supporting information). We have identified only one literature value for metabolite diffusion of for arginine-phosphate [30], [31], this is within the range of values for molecules with 20 atoms seen in Figure 6. A relation between maximum D and numbers of atoms is clear. However for smaller metabolites ( atoms) D ranges over an order of magnitude. It was not possible to find a clear relation between electrostatic charge or hydrophobicity and D. A comparison of D for the 100 M and 100 MP simulations suggesting metabolites diffuse slightly more slowly in the 100 MP simulation is in supporting information (Figure 7).
The diffusion coefficient of ubiquitin in the 100 Å simulation was , and the average of lateral diffusion in the x, y and z planes was . These values can be compared to experimental values for lateral diffusion of and for green fluorescent protein (GFP) in E.coli [32]–[35]. The order of magnitude difference in these protein diffusion values can be rationalised by the larger size () of GFP and the lack of structural proteins and membranes in our simulations. While this comparison is of limited use it is included as this is the most relevent experimental value available and it shows that our computed values are within a reasonable range. Another relevant comparison is with the large Brownian dynamics models of McGuffee et al.; here a protein of very similar size (CspC) was found to have a diffusion coefficient of with the smallest observation interval used [36]. In the McGuffee et al. study the friction parameter of their Brownian dynamics was adjusted such that the diffusion of green florescent protein matched experimental values. The McGuffee model also differed in that it contained many different types of larger proteins, and so this close agreement may be fortuitous.
The dielectric constant () and conductivity () can give insight into the electrostatic properties of a solution and other associated properties such as hydrophobicity. As suggested in the introduction and for such a complex heterogeneous solution is difficult to estimate. Owing to the necessity for long simulations with extremely frequent data collection (every 10 fs), smaller simulation boxes were used for this analysis (dimensions of 50 Å). and the translational dielectric constant () values were found through an Einstein-Helfand analysis described in the theory section. Regression analyses were applied from 100 to 500 ps for all systems except 50 M which used 100 to 300 ps (supporting information Figure S8).
Table 2 shows the results of the present analysis. is larger in the simulation without ubiquitin compared to that with ubiquitin, probably owing to the increase in ion and metabolite diffusion (Figure 6). for tip3p water is of course zero, while with the addition of 0.3 M KCL it is greater than the cytosol simulations, caused by higher diffusion rates of charge carriers. The tip3p + KCL value of 6.69 compares well with the experimental value of [37].
Unfortunately, direct experimental measurements of cytosolic are not available in the literature. However, spherical or spheroidal models (E. coli is rod shaped) together with various experimental data have been used to give estimates of E. coli cytosolic . Dielectrophoretic analysis gives 0.35 [38], dielectric spectra analysis 0.22 [19] and electrorotation analysis 0.44 [39]. These model-based measurements also predict a cytosolic of , which does not agree with other literature values [16]–[19]. The calculated conductivity with ubiquitin (50 MP) of 3.2 is an order of magnitude greater than these fitted measurements.
Overall, contributions were small compared to total . did not relate well to values for or rates of diffusion. It may be expected that, owing to its large , the 50 M system would have the larger but the 50 MP system contributes far more to from the conductivity. Also, the tip3p+KCl system has a very small contribution. This suggests a strange difference in the dynamics of charge carriers compared with those in the ubiquitin simulation, vibrating more sharply around a similar position than those in the metabolite only simulation.
The rotational component of , , (Table 2) follows trends found in the literature. The pure water system has of 92.5 which is slightly lower than some literature calculated values of around 97 [40]. This is almost certainly related to the use of a longer simulation length in this study (data not shown). The tip3p+KCl system had a reduced which agrees with another literature study of the SPC water model [6]. The metabolite only system has slightly lower than tip3p alone, as the metabolites with large dipoles compensate for the decrementing effect of the salt and those with small dipoles. Finally, the ubiquitin system displays a very large dielectric increment, however, this size of increment is not without precedence [12]. Previous values were similar but used less sampling meaning larger statistical error. Given the relatively small dipole of ubiquitin this increment may be smaller than average.
To the authors knowledge this is the first attempt to produce an atomistic simulation of the cellular cytosol solution. There is relatively little experimental data with which to compare, but comparison with available data on diffusion coefficients was satisfactory.
The stacking NIMS found here (Figure 3) are interesting and possibly important but are they realistic? Studies comparing aromatic stacking interactions show a reasonable agreement between molecular mechanics free energy calculations, high level electronic structure calculations and experiment [41]–[43]. Also there is experimental evidence for self-association of ATP in solution [44]. However, for guanine-cytosine stacked dimers with and without methyl groups, OPLSAA has been shown to produce non-stacked complexes where other force fields found the correct stacked formation. This may suggest that stacked metabolite complexes could be more prevalent with other force fields [45]. The alignment of phosphate and ribose groups in NIMS, such as that in Figure S3, has similarities to the elongation complex of RNA polymerases and may give an indication of how RNA polymers first emerged. Whilst speculative it is possible that highly reactive conditions (high temperatures or levels of radiation) and large amounts of time could do the job of the catalytic conditions found in RNA polymerases.
The analysis presented here suggests that NIMS are mostly mediated through hydrogen bonds, charge-charge, and interactions. A recent study has found good agreement in geometries and energies of a large set of relevant intermolecular complexes with high-level ab initio calculations [45]. Two other studies have demonstrated the high accuracy of OPLSAA in reproducing association constants of relevent small molecules in chloroform and relative free energies of hydration, heats of vaporization and pure liquid densities for 40 mono- and disubstituted benzenes [43], [46]. No parameter set is perfect but on the whole these study add weight to the idea that the metabolite interactions described here are realistic.
It should be no surprise that 2+ ions are found to be important to metabolite interactions. Many metabolites such as ATP require interaction with for enzyme-mediated reactions. ions were found to have two ionic-bonds or more for more than 80% of both 100 Å simulations (Figure 5). is generally not involved in NIMSs and may populate the solvent fairly uniformly, hence its important role as an osmolyte. All larger metabolites were found to spend of their time in contact with other molecules. While smaller metabolites vary in diffusive and contact character with some diffusing quickly and maintaining contact only 20–30% of the time. The presence of ubiquitin does not effect the amount of contact time experienced by metabolites.
There is a small difference in diffusion between the two 100 Å systems (Figure 7) which suggests that proteins have an effect on the dynamics of metabolites. In turn this suggests that with larger protein molecules the metabolites diffusion rates would be further reduced. We can speculate that in regions without proteins metabolite diffusion rates would be increased. Recent Brownian dynamics simulations have modeled many macromolecules in cytosolic volumes [36], [47], [48]. These models have been used to answer questions about macromolecular diffusion and stability outside of the scope of these atomistic models. However, it is possible that effects owing to metabolites could be important in these types of model.
of the cytosol of E. coli has many competing factors. Interestingly, total for the 50 M and tip3p systems are similar as the metabolite increment cancels the decrement of the monatomic ions of the tip3p+KCl 0.3M system. For the cytosol any increment in the rotational contribution due to proteins is an unknown and could have a large effect on , possibly only on a local level. Ubiquitin, used here, clearly has a large increment but can this be said of all proteins? A recent study has analyzed the dipole moments of the protein database [49] and gives an average protein biological unit dipole of 639 D, with ubiquitin having a dipole of 239 D. This suggests that most proteins have a dipole at least twice as big as ubiquitin. However, excluded volume will also have an effect reducing the effect of dipoles due to larger proteins. A higher dielectric compared to pure water will decrease electrostatic screening according to Debye-Huckel theory. A recent study has explored electrostatic screening using molecular dynamics and free energy calculations [50], and suggests that screening at high salt concentration is less than may have been expected from approximate treatments. Hence, the electrostatic screening found in cytosol solution may need further investigation. For the purposes of bio-simulations using implicit solvent it may be that a value closer to the 148 found here will give conditions closer to those found in vivo.
Owing to the diffusive and electrostatic considerations discussed above, it may be possible that proteins have a specific electrostatic and diffusive spheres of influence. If some proteins attract more metabolite ions than others, then this will again affect the local screening of the solution. Hence, proteins may have a locally specific electrostatic environments and propensities for associated metabolites and NIMS. In one example the electrostatic field of a protein is suggested to attract and orient specific metabolites [51], another study suggests that electric fields related to function are very protein specific and conserved through protein families [49].
Recently, kinetic models of cellular metabolism have started to appear in the literature [52]. These studies often attempt to approximate the thermodynamic activity of metabolites through Debye-Huckel theory [53]. Considering the high level of interaction between metabolites found in this study, the use of theory based on infinite dilution may not be sufficient to give realistic thermodynamic activities for these models. A recent experimental study has performed enzyme-inhibitor assays with an in vivo like solution (300 mM potassium, 50 mM phosphate, 245 mM glutamate, 20 mM sodium, 2 mM free magnesium and 0.5 mM calcium, at a pH of 6.8) rather than a standard concentration of the inhibitor [23]. In the in vivo like solution some enzymes have capacities (Vmax) which are less than half those found in optimised conditions. The solutions used are far from the complexity of the real cytosol and so further investigation of more complex solutions may be required. In the future it may be possible to calculate accurate thermodynamic activities using free energy calculations. These ideas may have implications for drug discovery. For example drug candidates predicted to spend significant amounts of time in NIMS and unavailable for binding to enzymes may not be optimal.
The behaviour and effect on the cytosol environment of molecules used by the cell to protect against stresses such as high osmolarity, pressure or anhydrobiotic conditions could be explored with simulations such as those in this study. A molecule which diffuses rapidly and is generally free from NIMS will be more osmotically active, if this molecule does not affect other aspects of the environment, would be a suitable osmolyte protectant. From this study we can predict that , glyceric acid, malate, 3-phosphoglycerate, and phenolpyruvate (metabolite codes are in supporting information Table S1) may be more osmotically active than other metabolites of a similar size. These models represent a specific phase in the cell cycle in optimal external conditions. The constituents of the cytosol can change in response to many factors and inevitably properties such as diffusion rates and molecular associations can be effected. Additionally, understanding the effects of different metabolites, compatible solutes, osmolytes and ions on the properties of the cytosol may allow us to better understand the reactions of the cell to extreme environments such as high salt concentration, high temperature or desiccation [54].
The simulations carried out in this study give an interesting picture of the molecular behavior in the cytosol solution. Metabolites and proteins are seen to have significant level of non-ideal behavior, with metabolites forming large non-covalently interacting metabolite structures (NIMS) and proteins slowing the diffusivity of metabolites. The electrostatic fields of proteins are powerful and control the local dielectric conditions possibly allowing selective filtering of metabolites. In the future these types of simulations may, as part of comparative or thermodynamic analyses, shed light on many poorly understood aspects of cellular environments.
Molecular diffusion coefficients were calculated using the Einstein relation [55],(1)where is the displacement of the atoms of a molecule over time , is the diffusion coefficient and is the number of dimensions of the position data. The Einstein relation was chosen over the velocity correlation function owing to better convergence behavior and the lack of a need to store velocity data. Mean squared displacement (msd) plots were averaged over replicas of the data with 50 ps removed from the start of each successive replica and the linear regression was applied from 1000 to 3000 ps.
Solvent accessible surface area (SASA) was employed to show the amount of time each molecule spends free in solution or as part of a larger non-covalently interacting structures. SASA was calculated, using the “Double Cube Lattice Method” of Eisenhaber et. al. [56], for each molecule with and without the surrounding environment and the difference taken in order that the average molecular surface area in contact with other non-water molecules is found (average contact area). This average contact area was then displayed as a percentage of the average SASA of the metabolite or residue without the surrounding environment, the percentage contact area.
Another analysis calculates percentage of simulation which metabolites are in contact with other non-water molecules. Here only a thermodynamically significant contact was of interest. The average excluded SASA found when two hydrogen bonds were present for all metabolites was calculated from the 100 M simulation. Hence, here contact was defined by an excluded SASA threshold of 0.48 . The use of SASA to define this contact means that other types of interaction such as those involving clouds are also included.
The calculation of using computer simulation was originally reported by Neumann and Steinhauser [57]. The dielectric constant of water models in molecular mechanics simulations has often been calculated in the literature [58], [59]. These studies generally calculate the static dielectric constant via the fluctuations of the system dipole ,(2)(3)where represents molecules and atoms in a molecule, is the Boltzmann constant, is the temperature and is the volume. is generally the origin of the coordinate system or the center of mass of the system.
In the present study the use of equation 3 is difficult due to the presence of molecules with net charge. For a charged molecule the choice of reference position directly affects the molecular dipole. For an overall neutral system these differences are thought to cancel, however convergence can be extremely slow [60]. A recently developed methodology decomposes into rotational () and translational () contributions [61],(4)(5)(6)where is the total charge of a molecule and is the center of mass of a molecule. describes the position of charge centers through the system and is the sum of molecular dipoles with respect to their center of mass. Combining equations 3 and 4 gives an equation for which may overcome some of the problems of equation 3 alone,(7)
This can be further simplified this by assuming that we use enough data such that giving,(8)
For convenience the rotational, translational and cross term contributions to are denoted , and respectively with, . is calculated through a simple ensemble average of . is directly related to the electrical current () and therefore the static conductivity,(9)
This means there are possible alternative routes to finding as is also easily obtainable from molecular simulation. These possibilities have recently been explored in the case of simple ionic liquids [60], [62], [63]. Hence, in the present study is found using the Einstein-Helfand method, as(10)where is the correlation length of current auto-correlation function. A linear regression fit of the resulting curve gives the static conductivity from the slope and from the y-axis intercept. The cross term is certain to be very small. Recent studies have evaluated for a series of ionic liquids made up of molecules which all have both translational and rotational dipoles [62], [63]. All of these studies have found very small . In the present study, a very small minority of molecules have both a translational and rotational dipoles, hence will be very small and has not been calculated.
All simulations used the GROMACS MD package [64], the OPLS force-field [65] was used for Zwitterionic protein residues and parameters for non-standard molecules were generated using hetgrpffgen provided with the Schrödinger Suite (Schrödinger LLC). This parameter generation method has recently been explored using solvation free energies of small, neutral molecules and was generally found to be of a high quality [66]. The development of the OPLSAA force field has focused on reproducing experimental measurements of thermodynamic properties for representative small molecules and was recently found to be the best at reproducing geometries and energies of inter-molecular complexes along with MMFF [45]. The recently developed Bussi et. al. thermostat was used, owing to its good reproduction of real dynamics and diffusive properties [67], [68]. The Parrinello-Rahman barostat was used for all production calculations. Temperature was set to 37 degrees Celsius.
Equation 8 must be applied to a periodic simulation using a long range electrostatic lattice summation and conducting boundary conditions, therefore periodic boundaries and particle mesh Ewald [69] was used throughout this study. Coulombic cutoffs at 1 nm have been shown to give more accurate dielectric calculations and were used throughout this study [57]. Lennard-Jones interactions were truncated with a switching function from 0.8 to 0.9 nm. System configurations were stored every 4 for the longer, 200 ns simulations. Subsequently, shorter 100 ns simulations were carried out storing configurations every 10 for the analysis.
Two box sizes were used, with dimensions of 50 Å and 100 Å, to assess possible size effects and provide a more tractable simulation for the analysis. The numbers of metabolite molecules used in each box was calculated from concentrations measured by Bennett et. al. [24]. Metabolites with concentrations sufficiently low such that less than 0.5 metabolites would be found in a particular box size were not automatically included. However, the total observed intracellular metabolite concentration given by Bennett et. al. was . This total is a higher concentration than that found through automatically included metabolites (0.23 M). We chose to increase the total metabolite concentration to 0.28 M, by randomly selecting from a list of less abundant metabolites with a probability biased by their concentration.
It is not possible to accurately estimate from published metabolomics data the concentrations of free metabolities as opposed to the total metabolite concentration. However, particularly for the most abundant species, Bennett et. al. [24] suggest that the concentrations are well in excess of the Km of enzymes that consume the metabolites, ensuring saturation of the enzymes (which will generally have much lower concentrations), and suggesting that a significant portion of the high-concentration metabolites will be free in solution. Nonetheless, the concentrations we use may overestimate the free concentrations of the various metabolites to unknown and variable extents, which is a limitation of the current study.
All metabolites were protonated according to pKas at pH 7.6 [70] found either though experimental data or calculation with Epik (Schrödinger LLC). The methods used by Bennett et. al. were not able to detect putrescine (JD Rabinowitz, personal communication, 2010). Putrescine has a 2+ charge at pH 7.6 and thus was used to give a neutralising charge along with potassium and magnesium ions (magnesium was used to represent all 2+ mono-atomic ions). Concentrations of putrescine (28 mM), magnesium (40 mM) and potassium (290 mM) ions in line with literature studies [71]–[78] were added such that the system was neutralised. Putrescine and magnesium are often found interacting with DNA, RNA and other large macromolecules [79]–[81] and therefore are less likely to be found free in the cytosol and in our simulation boxes. While potassium may be more likely to be found free in the cytosol and is more osmotically active [82]–[84]. Hence, the amount of potassium ions should be more related to the osmotic strength of the external medium compared to other ions or metabolites.
Larger macromolecules (proteins) were also considered, and to this end 50 and 100 Å boxes containing ubiquitin were also constructed. Ubiquitin (PDB code 1UBQ) is a eukaryotic protein, it was chosen owing to its small size and large amount of literature dedicated to its study [25]–[27]. A protein concentration of was assumed along with possible protein volume of [5], [76], [85]. Table 1 shows the details of the four simulation boxes created for this study.
The effective concentration of the single ubiquitin in the 50 Å is around which is higher than desired, however making this box larger would have prohibited running simulations long enough for the analysis. 50 Å boxes of tip3p water and tip3p with 0.3 M KCl (tip3p+KCl) were also created and equilibrated as part of the dielectric analysis. Types and numbers of metabolites used for each box are listed in supporting information, Table S1.
Model cytosol boxes were constructed through a simple Monte Carlo procedure. Each metabolite to be added to a box was treated as a buffered sphere and random positions were trialled until one was found which did not clash with the edge of the box or any other metabolite. Consequently, the initial structure of the boxes had no contact between any of the constituent metabolites. Owing to these considerations structural equilibration of the boxes was closely monitored before any analysis could be carried out. The use of a barostat throughout the structural equilibration is essential as the actual size of the simulation box reduces slightly.
The authors thanks Dr. Andrew Cossins and Dr. Olga Vasieva for useful discussions over the biological issues discussed in this work.
|
10.1371/journal.pntd.0003726 | Insensitivity to the Spatial Repellent Action of Transfluthrin in Aedes aegypti: A Heritable Trait Associated with Decreased Insecticide Susceptibility | New vector control paradigms expanding the use of spatial repellents are promising, but there are many gaps in our knowledge about how repellents work and how their long-term use might affect vector populations over time. Reported here are findings from a series of in vitro studies that investigated the plasticity and heritability of spatial repellent (SR) behaviors in Aedes aegypti exposed to airborne transfluthrin, including results that indicate a possible link between repellent insensitivity and insecticide resistance.
A dual-choice chamber system was used to observe directional flight behaviors in Aedes aegypti mosquitoes exposed to passively emanating transfluthrin vapors (1.35 mg/m3). Individual SR responder and SR non-responder mosquitoes were identified, collected and maintained separately according to their observed phenotype. Subsequent testing included re-evaluation of behavioral responses in some mosquito cohorts as well as testing the progeny of selectively bred responder and non-responder mosquito strains through nine generations. At baseline (F0 generation), transfluthrin actively repelled mosquitoes in the assay system. F0 mosquitoes repelled upon initial exposure to transfluthrin vapors were no more likely to be repelled again by subsequent exposure 24h later, but repelled mosquitoes allowed to rest for 48h were subsequently repelled at a higher proportion than was observed at baseline. Selective breeding of SR responders for nine generations did not change the proportion of mosquitoes repelled in any generation. However, selective breeding of SR non-responders did produce, after four generations, a strain of mosquitoes that was insensitive to the SR activity of transfluthrin. Compared to the SR responder strain, the SR insensitive strain also demonstrated decreased susceptibility to transfluthrin toxicity in CDC bottle bioassays and a higher frequency of the V1016Ikdr mutation.
SR responses to volatile transfluthrin are complex behaviors with multiple determinants in Ae. aegypti. Results indicate a role for neurotoxic irritation of mosquitoes by sub-lethal doses of airborne chemical as a mechanism by which transfluthrin can produce SR behaviors in mosquitoes. Accordingly, how prolonged exposure to sub-lethal doses of volatile pyrethroids might impact insecticide resistance in natural vector populations, and how already resistant populations might respond to a given repellent in the field, are important considerations that warrant further monitoring and study. Results also highlight the critical need to develop new repellent active ingredients with novel mechanisms of action.
| There is growing evidence to support the expanded use of spatial repellents for vector control, but there are still many uncertainties about how repellents work and how their long term use may impact vector populations over time. Here, we conducted a series of in vitro experiments that investigated spatial repellent (SR) behaviors in Aedes aegypti mosquitoes exposed to airborne transfluthrin, a volatile pyrethroid commonly used in repellent products. We were able to show that repellent insensitivity is linked to reduced insecticide susceptibility and increased knock down resistance allele frequency, indicating that sub-lethal doses of airborne transfluthrin can elicit SR behaviors in mosquitoes by inducing an agitated state via neurotoxic pathways independent of olfactory stimulation. This raises questions about how the use of volatile pyrethroid repellents may impact insecticide resistance in target vectors over time, highlighting the need to further understand all of the physiological drivers of SR behaviors and emphasizing the requirement to develop new repellent active ingredients with novel, non-toxic mechanisms of action.
| New vector control tools and paradigms are desperately needed to complement existing approaches [1–3], and there is growing evidence to support the expanded use of spatial repellents to help address this need [4–9]. The ultimate goal of public health interventions utilizing repellents is to exploit the behavior modifying effects of certain chemicals to prevent human-vector contact and, therefore, reduce disease transmission. Such approaches are among the most promising new strategies under investigation, with much progress already shown towards defining the parameters of spatial repellent-based interventions to control the global arbovirus vector Ae. aegypti [10–13]. However, there are gaps in our knowledge about how repellents work, including the exact molecular and physiological mechanisms by which various chemicals elicit SR behaviors in important vector species [5, 14–17] and the hereditary basis by which SR behavioral traits are maintained in populations of disease vectors [18, 19].
Spatial repellency (SR) is one of several behavior modifying effects of insecticides on mosquitoes that have been recognized for decades [6, 8] and have been shown to contribute to disease reduction in many settings [5, 20, 13]. In outlining a new classification system to more accurately describe the actions of chemicals used for malaria vector control, Grieco et al. (2007) defined SR actions as those that stimulate “movement away from the chemical source without the mosquito making physical contact with the treated surface” [6]. An expanded concept of SR, which also includes chemical actions that interfere with host detection and/or otherwise disrupt the blood-feeding process, was established by WHO in 2013 to help determine guidelines for efficacy testing [9]. Taken together, it is clear that what is casually referred to as spatial repellency is really a set of complex and multifactorial behaviors which can be generally thought of as reactions to air-borne chemical stimuli that deter mosquitoes from entering a space to take a blood meal from an otherwise suitable host.
Despite the complexities inherent in the modification of mosquito behavior, much evidence to date seems to indicate that olfactory mechanisms underlie many repellent behaviors [17, 21, 22]. For example, DEET, which is probably the most widely used and thoroughly studied mosquito repellent [23, 24], is thought to work either through direct olfactory stimulation [16, 25] and/or through interference with normal host cue detection, essentially masking the presence of a potential blood meal [14, 26]. Although DEET is typically found in products labeled for personal protection that are applied directly to the skin and is not, strictly speaking, a spatial repellent able to protect occupants of a defined area, knowledge of its mechanisms of action is likely to inform much of our view of how SR compounds function. Indeed, epidemiological and entomological evidence garnered from the use of indoor residual spraying with DDT for malaria control also supports a model whereby the SR action of the chemical results from a separate mechanism, likely olfaction, from that which produces neurotoxicity: SR activity is preserved in many locations where insecticide resistance is widely reported [27]. Similar observations have also been reported in pyrethroid tolerant mosquitoes that still demonstrate behavioral avoidance to sub-lethal doses of various pyrethroids [28, 29, 15]. Additionally, it has also long been observed that some proportion of mosquitoes continue to locate hosts and feed even in the presence of a repellent [30, 31], and in Ae. aegypti this DEET insensitivity has been shown to be a heritable trait with incomplete penetrance [19] associated with specific odorant receptor polymorphisms [32, 26].
Less clear, however, is whether or not olfactory pathways are the only physiological drivers of SR behaviors in mosquitoes. For instance, Ogoma et al. (2014) have reported that airborne pyrethroids and DDT both elicit multiple behavioral effects on a given mosquito population at the same time, including deterrence (the prevention of mosquito entrance into a structure), irritancy and excito-repellency (eliciting the premature exit of mosquitoes from a structure via physical contact with an insecticide treated surface or with insecticide vapors, respectively), reduced blood feeding, increased 24h mortality and reduced fecundity [7]. Kawada et al. recently reported reduced pyrethroid (permethrin and deltamethrin) contact repellency in a strain of Anopheles gambiae s.s. with the L1014Skdr mutation, but not in strains of An. arabiensis or An. funestus s.s. with cytochrome P450 driven metabolic resistance traits, supporting a role for the non-lethal disruption of neuronal sodium ion channel function in eliciting the observed excito-repellency/irritancy behaviors [15]. While they did not evaluate SR behaviors specifically, these results are in line with previous knowledge that many pyrethroid compounds (i.e., permethrin, deltamethrin and alphacypermethrin) can induce irritant and/or hyperactive responses in mosquitoes at sub-lethal concentrations [33, 34] and this hyperactivity can promote the avoidance of insecticide treated nets [35]. It is clear that physical contact with surfaces treated with these pyrethroid insecticides can produce repellency behaviors through neurologically disruptive mechanisms. It is unknown, however, whether or not a highly active and more volatile pyrethroid insecticide like transfluthrin, which also has SR properties [36, 37, 12, 7], elicits the same physiological responses through airborne exposure. This question is especially important as residual pyrethroids are currently the most commonly used class of public health insecticide worldwide and there are growing concerns about the rapid expansion of pyrethroid resistance in key vector species [5, 38, 39]. Critically, it is unclear how the use of volatile compounds that could act through the same physiological pathways as the most commonly used residual insecticides might complicate the insecticide resistance landscape.
Given the complex and multifactorial nature of SR behaviors in mosquitoes, the molecular and hereditary drivers of the behavior are likely to vary across different active ingredients and target organisms. Nonetheless, elucidating which mechanisms dominate in specific transmission settings is an important step to understanding how to best use spatial repellents in a public health context [40] and how their long-term use might impact vector populations over time [6, 29]. Additionally, this data could be used to guide the rational design of new active ingredients that mitigate resistance driving mechanisms [5]. Here, we report on a series of in vitro experiments that first examined the plasticity and heritability of non-contact SR behaviors in Ae. aegypti that were exposed to airborne transfluthrin, and subsequently explored a link between SR insensitivity and reduced insecticide susceptibility in a selectively bred strain of this important arbovirus vector.
Aedes aegypti (L.) mosquitoes were colonized from wild-caught (P1) larvae collected from discarded automobile tires near the Belize Vector and Ecology Center (BVEC) in Orange Walk Town, Belize (18°04.938’N, 88°33.390’W). The P1—F4 generations were reared and tested at the BVEC field laboratory at ambient light, temperature and humidity. Later generations (F5—F10) and experimental crosses were reared and tested under climate controlled conditions (28°C, 60% RH, and 12L:12D light-dark schedule) at the Uniformed Serviced University of the Health Sciences (USUHS) in Bethesda, MD. Larvae were fed Chiclid Gold fish pellets (Kyorin Co., LTD, Himeji, Japan) and adults were provided 10% sucrose solution from soaked cotton ad libitum. Using CDC bottle bioassay methods, F0 adults exhibited greater than 90% susceptibility to transfluthrin, malathion and DDT at 60 minutes (S1 Fig). SR Behavioral assays were performed using 5–12 day old mosquitoes, which were sorted into cohorts of 20 mosquitoes approximately 24h prior to testing. Female test mosquitoes were unmated, to allow for downstream selective breeding, and were sugar starved (provided only water-soaked cotton) for approximately 24h before testing, following standardized methods [41]. Because high mortality rates were observed in male mosquito populations, they were not sugar starved prior to testing.
SR behavior was evaluated using a high throughput screening system (HITSS-SRA configuration) (Fig 1), previously described by Grieco et al. (2007) [41] and recently adopted by the WHO as a standard procedure for in vitro efficacy testing of spatial repellents [9]. The dual-choice chamber system, which allows the observation of directional mosquito movement in response to a single chemical stimulus outside the context of host cues, consists of a clear Plexiglas central unit connected at opposite ends to one treatment chamber housing repellent-treated netting and one control chamber housing a net treated with acetone only (Fig 1). Tests were conducted to evaluate Ae. aegypti SR responses to passively emanating transfluthrin (2,3,5,6-tetrafluorobenzyl (1R)-trans-3-(2,2-dichlorovinyl)-2,2-dimethyl cyclopropanecarboxylate) (S.C. Johnson and Son, Inc., Racine WI), a volatile synthetic pyrethroid with widely demonstrated SR efficacy against mosquitoes [7, 36, 37, 12]. Briefly, reagent grade (unformulated) transfluthrin was dissolved in 100% acetone (Hofius Ltd./Ace Hardware, Belize City and Fisher Scientific, Waltham MA). This solution was then applied evenly by micropipette across the surface of 11cm x 25cm pieces of nylon organdy netting (No. I10N, G-Street Fabrics, Bethesda MD) and allowed to air dry a minimum of 15 minutes before use. Industry guidelines (M.C. Meier, personal communication, 16 August 2011) and concurrent experimental hut studies using transfluthrin in Belize [42] indicate a standard field application rate (FAR) of 1.35mg active ingredient per cubic meter of airspace to produce indoor SR activity against mosquitoes via passive emanation. Accordingly, HITSS treatment nets delivering 1x the FAR into the assay system were treated with 0.9mL of a 2.2x10-6 M (8.4x10-4 mg/mL) solution. Concentrations tested ranged from 0.5xFAR to 1000xFAR. Control nets were treated with 100% acetone only.
Cohorts of 20 mosquitoes were introduced into the central HITSS chamber and, after a 30 second acclimation period, butterfly valves situated at both ends of the central chamber were opened simultaneously to allow free movement of mosquitoes in either direction into either end chamber. After a ten minute exposure period, the butterfly valves were closed and the numbers of mosquitoes in each chamber were counted. Spatial repellency is measured by considering the number of mosquitoes that have moved into the untreated, control chamber (away from the treated surface) relative to the total number of mosquitoes that have moved in either direction using a weighted spatial activity index (SAI), equal to [(Nc- Nt)/(Nc+ Nt)]x[(Nc+ Nt)]/N] where N is the total number of mosquitoes per replicate and Nc and Nt are the number of mosquitoes in the control and treatment chambers, respectively. Possible values for the weighted SAI range from 1 to -1, with a value of 1 indicating the strongest SR response possible (movement of all mosquitoes away from the chemical source), zero indicating no net response, and a value of -1 indicative of a strong attractive response (movement of all mosquitoes towards the chemical source). To account for mosquito mortality, the total number of mosquitoes tested per each replicate was corrected using Abbott’s formula [43].
A SR dose-response curve was established using unselected (control) females by varying the dose of transfluthrin in the HITSS treatment chamber and measuring differences in corresponding SAI values and overall assay mortality (S2 Fig). The dose corresponding to 1xFAR (1.35 mg/m3) produced the largest SAI value (0.10, significantly greater than zero at P<0.02) and an overall non-contact mortality of only 2.8% and was selected for use in all subsequent HITSS SR replicates.
Male and unmated nulliparous female mosquitoes were tested separately and, after each experimental replicate, were identified as either SRA responders (SRA+) if they had escaped into the untreated control chamber or SRA non-responders (SRA-) if they either stayed in the central chamber or flew into the treatment chamber (S3 Fig). Mosquitoes that were located in the treatment chamber at the end of a replicate (i.e. had made physical contact with the transfluthrin treated netting) were enumerated for statistical purposes but then discarded and not further processed or analyzed. Though both male and female mosquitoes were tested during these experiments, only female behavior was analyzed statistically and only female results are presented here. Typically, males were tested in fewer replicates only to provide sufficient numbers of each behavioral phenotype (SRA+ responders and SRA- non-responders) for selective mating purposes.
To evaluate the plasticity of SR responses in unselected F0 females exposed to transfluthrin, test replicates were performed and mosquitoes were immediately collected and maintained separately based on their observed behavioral phenotype, i.e. SRA+ responders and SRA- non-responders. Mosquitoes were re-assayed on a subsequent day (day 2), after either a 24h or 48h resting period, and the weighted SAI for each phenotype cohort was compared to baseline (day 1) results using Student’s t-test at 95% confidence.
The heritability of SR behavioral responses was evaluated by performing test replicates and collecting mosquitoes based on their SR behavioral phenotype, as described above (S3 Fig). SR responder females were then selectively mated with SR responder males to establish an SRA+ strain of Ae. aegypti, and non-responder females were mated with non-responder males to establish an SRA- strain. Changes in the SAI scores in test populations from each strain were followed for 9 generations and were compared using ANOVA with Dunnett’s test for multiple comparisons at 95% confidence. An additional control strain of Ae. aegypti originating from the same field collected P1 larvae but which was allowed to freely mate was also maintained and tested.
In order to monitor relative changes in transfluthrin insecticide susceptibility over time and across different experimental populations, CDC bottle bioassay tests [43] were performed at various selection points, including the F0, F5 and F8 generations and in progeny from an experimental cross between F9 SRA- females and newly colonized wild type F0 males. A discriminating dose of 94 ng transfluthrin (0.25 nm, approximately 0.125xFAR) per bottle was established using F2 unselected control females (S1 Table). Test replicates lasted one hour, with mosquito knockdown recorded every 15m and final mortality recorded at 24hr.
Using the PCR genotyping approach developed by Linss et al. (2014) [44], Ae. aegypti voltage gated sodium ion channel V1016I and F1534C kdr allele frequencies were estimated using cohorts of 30 mosquitoes each from the F9 Control, SRA+ and SRA- populations and the experimental cross progeny. Both target site mutations have been previously observed in Ae. aegypti populations from Latin America and the Caribbean and have been shown to contribute to pyrethroid resistance [45, 46, 44].
Unless otherwise noted, SAI scores were calculated for each test population at each time point using 180 total mosquitoes, consisting of 9 replicates of 20 mosquitoes each, following established procedures [9]. Herein, the term ‘test population’ is used to refer to a sample of mosquitoes from a unique generation (e.g. F3) of a unique behavioral phenotype ‘strain’ (e.g. SRA-, SRA+ or control). Raw data was organized and descriptive analyses were performed using Excel 2007 (Microsoft Corp., Albuquerque NM). A non-parametric signed rank test (PROC UNIVARIATE) in SAS v8 statistical software (SAS Institute Inc., Cary, NC) was used to determine if mean SAI values were different from zero for each test population. SAI values were compared between populations via Student’s t-test and ANOVA with Dunnett’s test for multiple comparisons using SPSS Statistics 22 software (IBM Corp., Armonk NY). The kdr allele frequencies and herterozygosity were compared using Z-tests on the difference between sample proportions, and a chi-square test with one degree of freedom was used to evaluate deviations from Hardy-Weinberg equilibrium [47]. All analyses were performed at α = 0.05.
Two variations of the behavioral plasticity experiment were performed using F0 mosquitoes, with differing results (Table 1 and Fig 2). During the first experiment, mosquito cohorts (total n = 180 mosquitoes, average baseline SAI = 0.08 ±0.03 SEM) were re-assayed after a 24 hour rest period and results indicated a large degree of plasticity in behavioral responses to the repellent: mosquitoes repelled on day one (n = 29) were not more likely to be repelled again on day two (SAI = 0.03 ± 0.02) (Fig 2). Mosquitoes not repelled on day one (n = 129) were equally unlikely to be repelled on day two
(SAI = 0.03 ±0.04) (Fig 2). For the second experiment, mosquitoes (total n = 280, average baseline SAI = 0.05 ±0.04) were not re-assayed until the second day after the original test (48 hours post exposure). Unlike mosquitoes that were allowed to rest for 24hr, day one repellent responders from this cohort (n = 60) were more likely to be repelled again on day two (SAI = 0.30 ±0.08, P<0.05) (Fig 2). As was observed in the first experiment, non-responding mosquitoes from this experiment (n = 155) were also equally non-responsive on day two (SAI = 0.06±0.04) (Fig 2).
The baseline average SAI value for F0 female mosquitoes, which gave rise to all subsequent SRA+ and SRA- lineages, was 0.14 ±0.06 (significantly greater than zero at P<0.02), confirming that parental mosquitoes were actively repelled by volatile transfluthrin in the assay system. Selective breeding experiments were then carried out through the F9 generation (Table 2 and Fig 3). SAI results from the unselected control strain (S4 Fig) and the SRA+ strain (Fig 3) did not indicate any changes in behavioral responses to volatile transfluthrin at any time point compared to baseline (no significant differences at P = 0.05). Results from the SRA- strain, on the other hand, showed a steady decrease in SAI scores, which reached statistical significance (P<0.05) by the F4 generation (SAI = -0.05 ±0.04) (Fig 3). This SR insensitive phenotype was confirmed in each subsequent SRA- generation, with the exception of the F7 cohort in which the reduced SAI value (0.02 ±0.03) was not significantly different from baseline at P = 0.05 (Fig 3).
Baseline CDC bottle tests indicated greater than 95% susceptibility to transfluthrin toxicity (24hr mortality) at the discriminating dose in the F0 parental mosquitoes that gave rise to all selectively bred strains (Fig 4). Insecticide susceptibility was then reevaluated in the F5 and F8 generations of colony and selectively bred mosquitoes (Fig 4). For the colony (unselected control, S5 Fig) and SRA+ (responder, Fig 4) strains, no significant changes in insecticide susceptibility were noted by either time to knockdown or 24hr mortality. In the selectively bred SRA- repellent insensitive strain there was a moderate but significant (P<0.05) 23% reduction in mortality observed in the F6 generation compared to the control strain (60% ±1% vs. 95% ±6%) while the F8 SRA- test population was highly resistant with a mortality of just 14% ±11%, a significant (P<0.01) 77% reduction in mortality compared to the unselected control (Fig 4A).
An additional round of selective breeding of F8 SRA- non-responders gave rise to F9 SRA- mosquitoes that continued to exhibit repellent insensitivity (SAI = -0.04 ±0.05) (Fig 3) as well as significantly decreased CDC bottle assay knockdown and 24h mortality (13% ±13%) (Fig 4). Mating females from the F8 SRA- population with wild type F0 males newly colonized from the same location in Belize, however, restored both transfluthrin SR sensitivity (SAI = 0.11 ±0.03)(Table 2 and Fig 3) and insecticide susceptibility (24h mortality = 84% ±7%) in the resulting progeny (Fig 4).
Analysis of kdr allele frequencies was performed in the F9 control, F9 SRA+, F9 SRA-, and experimental cross progeny cohorts. Results indicated that the V1016Ikdr allele was more frequent (50%) in the SR insensitive, insecticide resistant SRA- population than in the susceptible SRA+ (16%, P<0.01) or the control (22%, P<0.02) cohorts (Fig 5). Overall V1016Ikdr allele frequency remained high in the experimental cross progeny in which SR sensitivity and insecticide susceptibility were both restored (Fig 5). However, there was a significant (P<0.01) increase in the proportion of heterozygotes, from 27% in the SRA- population to 65% in the experimental cross offspring (Fig 5). The assumption of Hardy-Weinberg equilibrium was rejected in both of the SRA+ (χ2 = 10.25, P<0.01) and SRA- strains (χ2 = 6.53, P<0.02), but not in either the control population or experimental cross progeny. There were no differences or changes in F1534Ckdr allele frequencies observed, with kdr prevalence over 90% for all cohorts tested.
The in vitro SR behaviors observed here were relatively plastic in that individual behavioral responses observed on day one were not consistent with subsequent behaviors observed upon identical chemical exposures at a later time point, reinforcing the notion that spatial repellency is a complex behavior with multiple determinants some of which are likely non-heritable [18]. Despite the overall high degree of variability in repellent behaviors on subsequent days, active SR responses were clearly more reproducible in mosquitoes that were given 48hr rest compared to those given only 24hr rest (Fig 4). This observation is consistent with other field [12] and laboratory [48] experiments that have shown post exposure habituation of mosquito behaviors that gradually resolves after appropriate recovery periods. The specific mechanisms driving these prolonged changes in behavior and their recovery, however, remain untested and in need of further investigation.
In the second set of experiments, SR responders (SRA+) and non-responders (SRA-) were identified and selectively bred for 9 generations. One of the possible outcomes of these experiments was the establishment of an SRA+ strain of Ae. aegypti with increased sensitivity to the SR action of volatile transfluthrin, and it was originally hypothesized that such a strain of super-responders might possess olfactory receptors with a particular affinity for detecting airborne transfluthrin. However, SR responses were not augmented in the selectively bred SRA+ strain at any time point. Conversely, there was a clear reduction in SR behaviors noted in the SRA- strain, ultimately leading to a population of mosquitoes insensitive to the SR activity of volatile transfluthrin. These results do not preclude the possibility that transfluthrin might elicit some SR behaviors by activating and/or interrupting certain olfactory pathways. In fact, the reduction in repellent sensitivity observed in the SRA- strain is in line with previous work by Stanczyk et al. (2010) that similarly demonstrated heritability of a DEET insensitivity trait in mosquitoes and further linked the phenomenon to changes in antennal olfactory reception [19]. Though similar in outcome, the DEET insensitivity trait described by Stanczyk et al. (2010) was clearly dominant, while the transfluthrin insensitivity observed here was restored after a single cross of SRA- females with repellent sensitive wild type males. Additionally, the HITSS SR system used here is unique in that it is designed to permit the observation of directional mosquito movement absent any attractive stimuli, thus allowing for the measurement of active spatial repellency as a distinct entity not confounded by attraction inhibition. Accordingly, it is likely that the transfluthrin insensitive phenotype observed here relies on a different mechanism of action than the DEET insensitive phenotypes, which have been previously linked to changes in antennae sesillum function [26, 19].
As mentioned above, many insecticidal compounds are known to induce irritant and/or hyperactive responses in mosquitoes at sub-lethal doses [33, 12, 34], and this hyperactivity has been observed to promote the avoidance of treated surfaces [35]. These behavior modifying effects are sometimes referred to as excito-repellency, which is defined as the action of irritating a mosquito sufficiently so that it flies away from the source of the chemical before knockdown or death occurs [6, 23]. In this context, the strong correlation between reduced insecticide susceptibility in CDC bottle bioassays and SR insensitivity in HITSS bioassays observed in the selectively bred SRA- strain suggests that the SR behaviors observed here resulted from neurotoxic irritation of mosquitoes by sub-lethal doses of airborne transfluthrin. This view is bolstered by the observed link between the SRA- phenotype and an increase in the frequency of at least one target site mutation, the V1016Ikdr allele, which echoes previous reports of an association between kdr mutations and decreased excito-repellency behaviors in some field populations of Anopheles spp. exposed to pyrethroids [15, 49]. One weakness of the present study is that kdr allele frequencies were not established in the P1 parental population. However, the presence of the V1016I kdr allele in the F9 control (freely mating) population at low but stable frequencies does indicate that the allele was likely present in the parental strain and may have contributed to the less than 100% mortality observed in the baseline CDC bottle bioassays and was likely selected for during these experiments.
In addition to suggesting the neuro-physiological irritation of mosquitoes by active ingredient vapors as a primary mechanism by which transfluthrin can elicit SR behaviors in Ae. aegypti, the results of these selective breeding experiments are also notable for having experimentally reduced insecticide susceptibility in a population of vectors exposed only to sub-lethal doses of an airborne insecticide. This is of particular importance as one of the proposed benefits to the expanded use of spatial repellents in vector control programs is the potential to alleviate much of the selective pressure that encourages the emergence of insecticide resistance from sustained use of toxic interventions in the current vector control paradigm [7, 5, 8, 50]. Our results indicate that if a repellent elicits SR behaviors in the target vector through, at least in part, the same mechanisms that produce toxicity at higher doses, then the potential for selecting resistance traits might remain. Our observation that a single cross of SRA- females with wild type F0 males restored both SR sensitivity and insecticide susceptibility to offspring suggests that the insecticide resistant/SR insensitive phenotypes observed here were predominantly in V1016Ikdr homozygotes. This could indicate dominance of wild type voltage gated ion channel function over V1016kdr, and is predictable given that kdr mutations have been associated with high fitness costs in Ae. aegypti [51, 52]. It should be noted, however, that at least one other kdr allele, F1534C, was present at high frequencies (>90%) throughout this study, suggesting that any single allele represents only one factor contributing to the overall insecticide susceptibility and SR sensitivity profile of an individual mosquito. The relative contributions of various resistance traits, including metabolic mechanisms, to repellent insensitivity and to overall fitness need to be further elucidated.
It is also important to consider that when populations of SR responders and non-responders were allowed to mate freely (control strains), repellent sensitivity and insecticide susceptibility were maintained. The in vitro selective breeding approach used here favored the emergence of repellent insensitivity/decreased insecticide susceptibility only when SR insensitive females were mated exclusively with SR insensitive males. The degree to which natural mosquito populations would experience the same selective pressure in a standalone SR-based system is uncertain. Firstly, it is difficult to imagine a scenario in which repellent insensitive or repellent sensitive individuals that survive exposure to a volatile insecticide would significantly out-compete one another post-exposure, particularly when it has been shown that the use of coils to deliver airborne pyrethroids results in the decreased fitness of all mosquitoes, even those not repelled [7]. Additionally, it is not known how or to what degree chemical exposure to repellents might affect natural male mosquito populations in an operational setting, exposures that are likely to vary significantly according to where the active ingredient source is placed and the typical mating behaviors of the target vector.
Nonetheless, it is essential to consider these results while recognizing that pyrethroids are the most commonly used class of insecticide worldwide [5, 53, 54]. Indeed, for public health applications pyrethroid use constitutes the front line approach for both indoor residual spraying [55] and insecticide treated bed nets [56], resulting in significant and growing concerns over the rapid spread of pyrethroid resistance [39, 57, 38]. Against this backdrop, these findings are potentially more worrisome, as the effects of introducing a volatile pyrethroid repellent in an area where residual pyrethroids are already in use are unknown and require further evaluation and monitoring. As with insecticide resistance in general, the operational relevance of these findings are not known at this time. Clearly, more work must be done to define what these observations mean within the larger landscape of pyrethroid use, including how prolonged exposure to sub-lethal doses of volatile transfluthrin might impact insecticide resistance in natural vector populations and how already resistant populations might respond to a given repellent in the field. Furthermore, given the clear evidence that SR effects can produce beneficial public health outcomes [5–8, 13], these results suggest that an ideal SR compound would not only have a low toxic profile but also be unrelated to the chemical classes currently used in vector control. Acknowledging this highlights the pressing need to identify new insect behavior modifying compounds with novel mechanisms of action [58].
Collectively, these results show that the in vitro SR responses observed here are complex behaviors with a mix of heritable and non-heritable determinants. Based on the link between the SR insensitive phenotype and decreased insecticide susceptibility, evidence also supports a model whereby sub-lethal doses of volatile transfluthrin can elicit SR responses in Ae. aegypti by inducing a hyperactive or agitated state via neurotoxic pathways, likely independent of olfactory stimulation or interruption. Care should be taken before extrapolating these results to other active ingredients or vector species. It should also be emphasized that these results do not indicate that transfluthrin elicits SR behaviors in Ae. aegypti exclusively by disrupting motor-neuron activity: olfactory and/or gustatory pathways may also play a role, whether via active detection and avoidance of odor cues or through the disruption of host detection and/or feeding, possibilities that should continue to be investigated using a variety of methods. Additionally, the appearance of decreased insecticide susceptibility and increased kdr allele frequency in the selectively bred offspring of mosquitoes exposed only to sub-lethal insecticide vapors raises some important questions about how the long-term use of repellents might impact vector populations over time. The answers to these questions will be dependent on several factors including which molecular mechanisms are driving specific repellent behaviors, the hereditary nature of repellent sensitivity and insensitivity, and other physiological effects of using sub-lethal concentrations of compounds that have insecticidal, as well as repellent, properties. Though the story is complex and further research is needed to better understand all of the physiological drivers of SR behaviors, evidence still supports the expanded use of spatial repellents in public health applications to control disease vectors, albeit with continued monitoring of potential changes in target vector repellent sensitivities and/or insecticide susceptibilities and a renewed emphasis on the need to develop new active ingredients with novel, non-toxic mechanisms of action.
|
10.1371/journal.pcbi.1004040 | Rational Design and Adaptive Management of Combination Therapies for Hepatitis C Virus Infection | Recent discoveries of direct acting antivirals against Hepatitis C virus (HCV) have raised hopes of effective treatment via combination therapies. Yet rapid evolution and high diversity of HCV populations, combined with the reality of suboptimal treatment adherence, make drug resistance a clinical and public health concern. We develop a general model incorporating viral dynamics and pharmacokinetics/ pharmacodynamics to assess how suboptimal adherence affects resistance development and clinical outcomes. We derive design principles and adaptive treatment strategies, identifying a high-risk period when missing doses is particularly risky for de novo resistance, and quantifying the number of additional doses needed to compensate when doses are missed. Using data from large-scale resistance assays, we demonstrate that the risk of resistance can be reduced substantially by applying these principles to a combination therapy of daclatasvir and asunaprevir. By providing a mechanistic framework to link patient characteristics to the risk of resistance, these findings show the potential of rational treatment design.
| Hepatitis C virus (HCV) affects approximately 170 million people world-wide and chronic infections can lead to cirrhosis and liver cancer. New combination therapies of direct acting antivirals have achieved remarkably high cure rates in clinical trials. However, high mutation rates and high diversity of HCV populations, combined with the reality of suboptimal treatment adherence, make drug resistance a clinical and public health concern. By constructing a mechanistic framework to assess the risk of drug resistance, we provide guidelines for rational design and adaptive management of these promising new therapies. In particular, we identify a high-risk period when missing doses is particularly risky, and quantify the number of extra doses needed to compensate when doses are missed. This framework is a step towards developing a tool for clinicians to design combination therapies and adaptively manage treatment regimens to achieve favorable clinical outcomes.
| Hepatitis C virus (HCV) affects approximately 170 million people and chronic infections can lead to cirrhosis and hepatocellular carcinoma [1,2]. Recently, development of direct acting antivirals (DAAs) against HCV infection has revolutionized the field of HCV treatment, because of their high potency, broad applicability and mild side effects [3,4]. Combination therapies of DAAs have achieved remarkably high rates of sustained virological response in clinical trials [5–10]. However, most DAAs have relatively low genetic barriers [11–13], with the exceptions of a few pan-genotypic, yet high-cost DAAs [6]. Because of the high intrinsic mutation rate of HCV [14,15] and the resulting high viral diversity [1,16,17], combined with the reality of suboptimal treatment adherence [18,19], viral resistance is still a clinical and public health concern [13,20]. This is especially true for high-risk groups such as patients with psychiatric disorders or depression [21], and in resource-limited settings where patients have limited access to clinical cares and cannot afford the expensive pan-genotypic DAAs with high genetic barriers [22,23].If treatment is not properly managed, resistance could quickly develop to combination therapies and render these new DAAs useless, as observed for other antimicrobial treatments, squandering the potential health gains from these recent breakthroughs [24–26].
Suboptimal patient adherence to dosing regimens is a crucial risk factor for resistance development in both HIV and HCV treatments [18,19,27,28]. Although high rates of sustained virological response have been achieved in clinical trials, adherence levels may vary substantially among the vast population of infected patients, owing to long treatment periods, complicated regimens associated with DAA combination therapies and limited access to health care [18,19,29–31]. Rational design of combination therapy that achieves viral eradication in patients and maximizes the durability of available DAAs in the presence of suboptimal adherence is a research priority [18,32–34]. In addition, theories that guide individualized regimens based on the genetic composition of a patient’s infection and real-time adjustments for missed doses are needed to avoid resistance. Mathematical models are well suited to address this problem. Previous modeling studies for HIV infections have illuminated potential mechanisms underlying treatment failure and explained puzzling clinical observations [35,36]. However, HCV is a curable disease and its infection, goal of treatment and mechanism of resistance differ from HIV in many respects [37], including no known latent reservoir and a finite treatment period to eradicate the virus. Here, by integrating pharmacokinetics/pharmacodynamics (PK/PD) and viral dynamics into mathematical models, we develop the first general theory to assess the impacts of suboptimal adherence on the outcome of DAA-based therapies for HCV infection. We derive design principles that can be generalized to therapies involving different classes and different numbers of drugs. Using large-scale data from in vitro resistance assays and human clinical trials, we apply this framework to a combination therapy of daclatasvir and asunaprevir [38], and derive evidence-based adaptive treatment strategies for treatment protocols over time according to resistance profiles and adherence patterns.
Resistance to antiviral treatments can develop through selection of preexisting mutants or de novo generation of new mutants. A core principle for designing effective combination therapy is that, if patients fully adhere to the treatment regimen, the treatment must suppress all preexisting mutants and de novo resistance should be unlikely [39]. Missing doses, however, can lead to suboptimal drug concentrations, allowing growth of some preexisting mutants with partially resistant phenotypes. Growth of these mutants allows the viral population to survive longer, possibly generating further mutations that contribute de novo resistance against the full combination therapy. For example, consider a combination therapy of two DAAs, A and B, as shown in Fig 1A. If missed doses and pharmacokinetics lead to a drop in the concentration of drug A, this allows growth of the preexisting mutant, AB’, (which is already resistant to drug B), thus opening opportunities to generate the fully resistant mutant, A’B’. Therefore, the dynamics of the subset of preexisting mutants that have a high level of resistance against single DAAs determine resistance evolution and treatment outcomes for combination therapies. In the following, we denote these mutants as ‘partially resistant’ mutants.
The fitness of a particular strain in a treated patient is determined by the PK/PD of the drug, the level of resistance of the strain, and the availability of target cells, i.e. uninfected hepatocytes for HCV (Fig 1B). We can integrate all these factors (for any class of DAA therapy) into a single number, the effective reproductive number under treatment, Reff(t) (Fig 1C). Reff(t) is the average number of cells infected by viruses produced by a single infected cell. It acts as a measure of viral fitness, and can be calculated as:
Reff(t)=(1−ε(t))⋅R0⋅h(t)
(1)
where t is time since treatment starts, τ is the time since last dose, ε(τ) is the efficacy of the drug combination at time τ during the dosing cycle, R0 is the reproductive number of the virus in the absence of treatment, and h(t) is the normalized abundance of target cells (see S1 Text). Under effective treatment, the availability of target cells, h(t), increases quickly to reach the infection-free level [40], and therefore, the overall viral fitness increases over time as h(t) increases under effective treatment (Fig 1B and 1C). When adherence is optimal (i.e. no missing doses), the value of Reff for a partially resistant mutant is always less than 1 (i.e. viral suppression); however, if doses are missed, drug concentration declines exponentially and Reff can become greater than 1 (i.e. viral growth) (Fig 1C). Note that, although we consider the fitness of a single ‘partially resistant’ mutant here, the competition between different mutants is described implicitly in Eq 2 by the target cell availability, h(t): if a ‘partially resistant’ mutant rises to a high abundance due to missing doses, then h(t) will decrease to a low level again, leading to decreases in fitness for all viral mutants.
We now consider how suboptimal adherence impacts the dynamics of partially resistant mutants. As an illustration, we contrast simulations assuming perfect adherence versus suboptimal adherence. Missing doses leads to rapid decreases in drug concentration, and thus increases in Reff of a partially resistant mutant (Fig 2A–2C). This means that extra doses are needed to compensate for the missed doses to suppress the mutant to extinction (Fig 2D), and also that the number of newly infected cells rises substantially, which increases the opportunity for de novo resistance (Fig 2E).
We approximate the time-varying values of Reff(t) during periods when doses are missed, by calculating the average effective reproductive number, Rave,m, as (see Materials and Methods):
Rave,m(t)≈(1−εave,m)⋅R0⋅h(t)
(2)
where t is the time when the patient starts to miss doses, m is the number of consecutive doses missed and εave,m is the average drug inhibition during the period when m consecutive doses are missed. This allows us to generalize our theory to any DAA combinations for which εave,m can be either estimated from pharmacokinetics/pharmacodynamics data or calculated from mutant resistance profiles [36].
We then ask, if m consecutive doses are missed beginning at time t, how many extra doses, Nm, are needed to compensate? This number, which we denote ‘compensatory doses’, can be approximated as (see Materials and Methods):
Nm(t)≈m⋅Rave,m(t)≈m⋅(1−εave,m)⋅R0⋅h(t)
(3)
This allows us to estimate the total duration of treatment needed to clear infection for a given adherence pattern. Furthermore, since h(t) increases over time under effective treatment [40], Eq 3 shows that a higher number of extra doses are needed to eliminate the infection if doses are missed later in treatment.
To assess the risk that a partially resistant lineage will give rise to full resistance, we calculate the expected number of target cells, Φm, that become infected by fully resistant mutant viruses due to de novo mutation during a period when m consecutive doses are missed. This quantity is the product of the cumulative number of cells newly infected by a partially resistant mutant and the effective mutation rate from that mutant to the fully resistant mutant, μeff (see Materials and Methods):
Φm(t)≈μeff•I(t)•Rave,m(t)Rave,m(t)−1•(e(Rave,m(t)−1)•δ•m•T−1)⏞Θ(t)⏟cumulative number of cells newly infected by a partially resistant mutant
(4)
where I(t) is the number of cells infected by the partially resistant mutant at time t when the first dose is missed, and Θ(t) represents the potential to generate new infections. δ is the death rate of infected hepatocytes, and T is the scheduled interval between two doses. Φm quantifies the risk that a fully resistant mutant infects target cells, but whether it emerges and becomes established within the host depends on its fitness and the stochastic dynamics of invasion [41–43].
The strong dependence of Φm on μeff predicts that designing combination therapies to increase the genetic barrier to full resistance, e.g. using DAAs with higher genetic barrier or adding an extra drug into the combination, can reduce Φm by orders of magnitude or more, thus it would lead to drastic reductions in the probability of generating full resistance (compare trajectories a and b in Fig 3A).
Eq 4 also allows us to assess when during treatment it is most risky to miss doses, which can inform treatment guidelines. Changes in two quantities, I(t) and Θ(t), determine changes in Φm over the course of a treatment regimen. For as long as adherence is perfect, I(t) decreases exponentially, while Θ(t) increases over time since Rave,m(t) increases as the abundance of target cells rises over time (Fig 3B). Thus the value of Φm first increases (due to rapid recovery of target cells) and then decreases exponentially (due to decrease of infected cells). This leads to a high-risk window period, during which missing doses is especially risky for generating full resistance (Fig 3A). This qualitative finding is robust to changes in model parameters, though quantitative predictions of the risk of full resistance depend on the fitness of the mutant (R0), the half-life of infected cells (δ), and the rate at which the target hepatocytes become available upon treatment (S1 Fig).
These results suggest principles for designing combination therapies and rational optimization of treatment outcomes. First, the genetic barrier to full resistance to a therapy is an important determinant of the risk of resistance. Assessment of treatment readiness has been a low-cost routine clinical practice for HIV treatment [44]. Similar strategies can be implemented for HCV treatment. Based on the assessment, for patients who are predicted to maintain high adherence, combinations of DAA that ensure the fully resistant mutants are not pre-existing would be sufficient. For patients with risk factors for low adherence, therapies should be designed by selecting drug combinations that impose a higher genetic barrier than required to suppress all pre-existing mutants. Second, we have shown there exists a high-risk window period where the risk of de novo resistance is high. Intervention efforts to ensure a high level of adherence during the high-risk window period (indicated by the value of Φm) would reduce the risk of resistance and treatment failure. Third, because of the exponential growth of ‘partially resistant mutants’ when doses are missed, missing a number of doses consecutively leads to a much higher risk of de novo resistance than missing the same number of doses separately [36]. Thus, missing a block of doses should be avoided.
Adaptive treatment strategies could be developed based on the theoretical findings shown above. If doses are missed during treatment, the patient should be treated with extra doses, computed as the maximum value of the Nm values calculated for all partially resistant mutants. For the lowest risk of de novo resistance, the prescribed number of compensatory doses (Nm) should be taken, uninterrupted, immediately after doses are missed. Otherwise the infected cell population may rebound to a high level, which can make further missed doses very risky for resistance.
To demonstrate the practical applicability of our theory, we consider a recently developed interferon-free combination therapy based on an NS5A inhibitor, daclatasvir, and an NS3 protease inhibitor, asunaprevir [38]. In clinical trials, a large proportion of patients infected with HCV genotype-1b achieved sustained virological response (i.e. viral eradication) when treated with daclatasvir and asunaprevir for 24 weeks, although viral breakthrough and viral relapse occurred in a small fraction of patients [45,46].
We first consider patients with the wild-type virus at baseline, i.e. the wild-type virus is the dominant strain before treatment. Using the PK/PD data for each drug [47–49] and the resistance profiles data measured for genotype-1b HCV [50,51], we predicted which mutants are potentially fully-resistant to this combination therapy and calculated the values of Nm and Φm for each of the partially resistant mutants (Fig 4A and 4B) (see Supplementary Materials for more detail). Choosing the highest values of Nm and Φm among all the partially resistant mutants allows us to project the overall risk arising from missed doses over the course of treatment, and we found required numbers of compensatory doses were modest and the risk of de novo resistance is low (S2A Fig). To demonstrate that the theoretical approximations represent the full viral dynamics accurately, we simulated a multi-strain viral dynamics model (see Materials and Methods), assuming 1–3 day blocks of consecutive doses are missed randomly within a treatment regimen lasting 24 weeks. The model predicts that relapse of L31M+Y93H or L31W would be observed when overall adherence is less than 90% (Fig 4C and 4D). Indeed, the L31M+Y93H mutant has already been detected in one relapse patient in a clinical trial [46]. There is excellent agreement between simulation results and theoretical predictions (based on Eqs 3 and 4) for the number of cells infected by different mutants after 24 weeks of treatment and the cumulative number of cells infected by partially resistant mutants over the treatment period (Figs 4D and S3).
We then simulated outcomes when the doses are guided by the adaptive treatment strategy (guided dosing; see Methods for detailed simulation procedure). Because the risk of de novo resistance when doses are missed is low, there is no high-risk period for de novo resistance in this case (Fig 4B). If patient dosing is guided, i.e. all the required doses and the extra doses to compensate for the missed doses are taken, the infection can be cleared successfully (Fig 4E). Again, we find excellent agreement between simulation results and theoretical predictions (Fig 4F).
Many patients bear the Y93H mutation at baseline and this mutation reduces the genetic barrier to full resistance by one nucleotide[46]. Our theory suggests that reducing the genetic barrier to full resistance will drastically increase the risk of treatment failure. We repeated our analysis for patients with Y93H at baseline, to test how our adaptive treatment strategy works when the risk of resistance is high. As predicted, many more days of treatment are needed to compensate for missed doses, and the risks of generating full resistance de novo are high (>0.01) during the first 3 weeks of effective treatment if 2 consecutive doses are missed (or first 4 weeks if 3 doses are missed; Fig 5A and 5B and S2B Fig). De novo full resistance is likely if doses are missed randomly and adherence is less than 90% (dark red area in Fig 5C). The predicted number of infected cells agrees well with simulation, except when adherence is very low such that viral load rebounds back close to the pre-treatment level (Fig 5D and S4–S6 Figs). In stark contrast, when doses are guided, the risk of de novo resistance becomes much lower (compare Fig 5C with 5E). Again, for patients who do not clear infection after 24-week treatment, extended periods of treatment as predicted by our theory (using Eq 3) can clear infection with low risk of resistance. The efficacy of the adaptive treatment strategy is robust across different parameter values (S7–S12 Figs and S1 Text). Therefore, our treatment strategy can improve clinical outcomes substantially by adjusting on-going treatment based on patient adherence patterns.
In this study, we integrate PK/PD parameters and viral dynamics into a unified framework to assess the impacts of suboptimal treatment adherence on the risk of treatment failure. Using simulations incorporating PK/PD and resistance profile data collected previously [48,50,51], we showed that treatment outcomes of combinations therapies of daclatasvir and asunaprevir can be improved by this adaptive treatment strategy, especially when the Y93H mutant is the dominant strain before treatment begins.
We have identified several factors that influence the risk of de novo resistance to a combination therapy. Among these factors, the genetic barrier to full resistance plays a dominant role. Thus, for patients with risk of low adherence, combinations that impose a higher genetic barrier are recommended. This is especially important in resource-limited settings where patients have limited access to health care and adherence is not closely monitored. The recently developed HCV entry inhibitors [52], which inhibit host factors that are required for viral entry (instead of viral factors), may offer a promising direction for HCV combination therapy, because of their high genetic barriers to resistance, and their synergistic interactions with other classes of DAAs. For situations where therapies with low genetic barriers to resistance are used, we have identified a high-risk window period during which de novo resistance is likely if doses are missed. Intervention efforts should focus on enhancing patients’ adherence during this period. Additional complementary strategies could further reduce the risk of treatment failure. First, if doses are missed during the high-risk window, the immediate addition of another drug with a different mechanism of action from existing drugs may eliminate any low level of fully resistant mutants that has arisen. Alternatively, a patient could be treated preemptively using additional drugs during the entire high-risk period and switched to fewer drugs afterwards. Another important factor is the number of consecutively missed doses as shown previously [36]. Consecutively missed doses lead to exponential growth of ‘partially resistant’ mutants, and thus substantially increase the risk of de novo resistance.
Our theory also predicts the number of compensatory doses (Nm) needed to compensate for missed doses, in order to eliminate preexisting mutants. Interestingly, clinical trials have shown that adherence levels tend to decrease over time [19,31]; we show that more doses are needed to compensate for missed doses that occur later in treatment because of the rebound of target cells. While many previous studies have focused on average adherence [18,19,29–31,36], we emphasize that the timing of the missed doses is also a critical determinant of treatment outcome and the risk of resistance.
There exist substantial heterogeneities among patients owing to variation in HCV genotypes, patient viral loads, death rates of infected cells [40,53] and effectiveness of drug penetration [47]. Our analysis has identified several factors that influence the impact of suboptimal adherence, particularly the rebound rate of target cells under treatment, the half-life of infected cells and the overall viral fitness, R0. We used the best available estimates of these parameters, but further empirical work is needed. If resistance profiles and viral parameters could be measured directly from a specific patient, then our framework linking these factors could be tailored to give patient-specific guidelines.
Certain model assumptions reflect uncertainties in our current knowledge of HCV infection. First, our prediction about time to viral extinction should be treated cautiously. We predict the time of extinction (as in other models [54–56]) by assuming that infected cells decline at a rate set by their death rate, and infection is cleared when the number of infected cells is below one. However, factors such as pressures from the immune system and infections in different tissue compartments may influence the extinction threshold. Furthermore, if DAA treatment causes intracellular viral RNA to decay with negligible replication [57], the decline of infected cells may result from a combination of cell recovery and death of infected cells. Indeed, sustained virological response has been observed in clinical trials of DAA combination therapies with shorter durations of treatment [5,6]. Our model can be adjusted easily once the decay dynamics of infected cells are understood better. Second, our model captures the main features of pharmacodynamics and viral dynamics by assuming quasi-equilibrium for viral populations and drug penetration into liver cells. Further work that incorporates detailed intracellular interactions [57] and different body compartments may improve model accuracy, once pertinent parameters are measured. However, a more detailed model may become analytically intractable.
This quantitative framework is a step towards developing a tool (for example, see Ref. [58]) for clinicians to design combination therapies and adaptively manage treatment regimens to achieve favorable clinical outcomes. It highlights the importance of characterizing resistance profiles of HCV, assessing readiness for treatment, and monitoring adherence patterns during treatment, so that treatment can be designed and adjusted in an evidence-based manner. This framework can be adapted easily to combination therapies based on interferon, entry inhibitors [52] or other DAA candidates, or treatments of other curable diseases without a latent reservoir.
To analyze the dynamics of the virus, we constructed an ordinary differential equation (ODE) model to describe the long-term within-host dynamics of a single HCV strain under drug treatment, based on an established model developed by Neumann et al.[53] (see Supplementary Material). In the model, ε represents the proportion by which the therapy reduces viral growth (ε is in the range of 0 and 1). Then, the fitness of the virus, Reff(t), is the product of the complement of the therapy’s efficacy (1- ε(τ)), the reproductive number of the virus, R0, and the availability of target cells, h(t) (Eq 1).
To approximate the time-varying viral fitness, Reff(t), during the period when m consecutive doses are missed, we assume that the abundance of target cells stays constant. This is a good approximation, because the length of the period when consecutive doses are missed tends to be short compared to the time scale of target cell rebound. Then the only time-varying quantity in Eq 1 is ε(τ). We can calculate the average level of drug inhibition during the period when m doses are missed, εave,m, by incorporating parameters for pharmacokinetics and pharmacodynamics (for example, see Wahl and Nowak[36]). Then the time-average effective reproductive number, Rave,m(t), for a mutant when m consecutive doses are missed starting at time t can be expressed as Eq 2. In practice, because the precise number of target cells at time t is hard to estimate, we can approximate Rave,m(t) by setting h(t) = 1, and then Rave,m(t) becomes Rave,m(t) ≈ (1 – εave,m) ∙ R0. Because h(t)≤1, this always overestimates the viral fitness and thus is a conservative estimate in terms of guiding treatment. Note that the assumption that h(t) = 1 is valid only when the viral load at time t is much lower than it was before treatment, which is the case if adherence is not too low. Otherwise, h(t) would decrease significantly due to large amount of infection.
To calculate Nm for each mutant, we make the simplifying assumption that the dynamics of the viral populations are at quasi-equilibrium, because changes in the viral populations occur much faster than changes in infected hepatocytes. Then, the dynamics of the number of cells infected by mutant viruses, I(t), are described by:
dI(t)d(t)=(Reff(t)−1)⋅δ⋅I(t)
(5)
where δ is the death rate of infected hepatocytes. If we approximate Reff(t) using the constant Rave,m for the period when doses are missed, Eq 5 can be solved analytically. Then, the number of infected cells after missing m consecutive doses starting at time t0 can be expressed as:
I(t0+m+T)≈I(t0)⋅exp((Rave,m(t0)−1)⋅δ⋅m⋅T)
(6)
We now consider the situation when m consecutive doses are missed, and ask how many uninterrupted doses (compensatory doses) must be taken so that the number of cells infected by the mutant is suppressed to a same number as if the m doses had not been missed. We first calculate the number of infected cells if the m consecutive doses are taken, i.e. if no doses is missed:
Ioptimal(t0+m⋅T)≈I(t0)⋅exp((Rave,0(t0)−1)⋅δ⋅m⋅T)
(7)
where I(t0) is the number of cells infected by the mutant at time t0, Rave,0 is the average effective reproductive number of the mutant when all doses are taken, and T is the scheduled interval between doses.
We then analyze the situation where a patient skips m consecutive doses, starting at time to, and then takes Nm compensatory doses immediately afterwards. In this case, assuming the number of target cells does not change much during this period, we can approximate the number of cells infected by the mutant at the end of the Nm doses as:
Isuboptimal(t0+m⋅T+Nm⋅T)≈I(t0)⋅exp((Rave,m(t0)−1)⋅δ⋅m⋅T)⋅I(t0)⋅exp((Rave,0(t0)−1)⋅δ⋅Nm⋅T)
(8)
By equating the right hand sides of Eqs 7 and 8 and solving the equation, we derive the expression for Nm:
Nm(t0)≈Rave,m(t0)−Rave,0(t0)1−Rave,0(t0)⋅m
(9)
For potent therapies, usually Rave,0(t0) ≈ 0. Then we get Eq 3.
In the derivation above, we have assumed that the target cell abundance stays constant during the period under consideration. This would be a good approximation if only a few days of doses are missed or if the target cell has already rebounded to the infection-free level. If the abundance of target cells changes considerably during the period under consideration, an alternative, conservative approach would be to assume h(t) = 1 and take Nm,max(t0) ≈ m ∙ (1 − εave,m) ∙ R0 compensatory doses after missing m consecutive doses of treatment.
One important application of Nm is to predict the number of remaining doses needed to eradicate a mutant, Nerad, in a patient during treatment. This number can be calculated as follows. If adherence is perfect, the number of infected cells declines exponentially at a rate set approximately by the death rate of infected cells, δ: (t) ≈ I0 ∙ exp(−δ ∙ t), where I0 is the number of cells infected by a mutant of interest before treatment. If we assume that a mutant goes extinct if the expected number of infected cells in a patient goes below 1, the number of doses needed to eradicate a mutant before treatment (assuming adherence is perfect), Nerad,0, is calculated as:
Nerad,0≈log(I0)δ∙T
.
When doses are missed during treatment, it is clear from the calculation of Nm above that Nm–m extra doses of treatment are needed to eradicate the virus. Therefore, if a patient has taken a total of x doses and has had k instances of missing doses before time t, with mi days of doses missed in the ith instance (i = 1,2,…,k), then the number of remaining doses needed to eradicate the mutant is calculated as:
Nerad=Nerad,0−x+∑i=1k(Nm,i−mi)
(10)
We can use Eq 10 to predict the number of cells infected by a mutant as: I(t) ≈ exp(δ ∙ Nerad(t) ∙ T). In our model, and a patient is cleared of infection when all mutants are driven to extinction. The accuracy of this approximation is shown in Figs 4D and 4F and 5D and 5F.
To calculate the risk of full resistance during the period when m doses are missed, we first calculate the number of cells newly infected by a partially resistant mutant when m doses are missed, Ωm(t). Again, we use Rave,m(t) to approximate Reff(t), the total number of cells infected by the mutant virus, starting at time t. Ωm(t) can be expressed as an integration of new infections during the period of missing doses (according to Eq 5):
Ωm(t)≈Rave,m(t)⋅δ⋅∫tt+m⋅TI(x)dx=I(t)⋅Rave,m(t)Rave,m(t)−1⋅(e(Rave,m(t)−1)⋅δ⋅m⋅T−1)
(11)
The expected number of target cells that become infected by fully resistant mutant viruses, Φm, is a product of the effective mutation rate from the partially resistant mutant to the fully resistant mutant (μeff) and the total number of cells infected by the partially resistant mutant (Ωm): Φm(t) = μeff ∙ Ωm(t), as shown in Eq 4.
Note that we track the population of newly infected cells to assess the risk of de novo generation of full resistance. This assumes implicitly that the fully resistant mutant is selected only when it enters a cell. This is a good assumption for DAAs that act on intracellular stages of the viral life-cycle, such as viral genome replication or assembly. However, in situations where the drug blocks viral entry into the cell, the mutant virus may have a selective advantage for entering a cell. Then the viral population should be tracked instead, but the results presented here still can be applied to drugs that block cell entry by multiplying with a simple scaling factor [59].
We constructed a simulation model considering the dynamics of the baseline virus and all the potentially partially resistant mutants (see Supplementary Material). This simulation model follows a hybrid approach used previously to simulate the evolutionary dynamics of HIV [60]. It considers the dynamics of multiple strains of HCV deterministically (using ODEs) while treating the extinction and mutation processes as stochastic events (see Supplementary Material for detail).
In the simulation, a patient is treated for a total period of 24 weeks. We generate two types of dosing patterns: random dosing and guided dosing. For the random dosing pattern, doses are missed in blocks of 1–3 days at times chosen randomly with equal probability during the treatment period. This probability is set as a constant in each run, but varied across runs such that different overall levels of adherence are generated. In each simulation, we assume that at least one-day treatment is taken immediately after each dose-skipping event (i.e. 1, 2 or 3 consecutive missed doses), to ensure that two dose-skipping events do not occur consecutively (otherwise, longer blocks of doses would be missed than intended). For guided dosing, we ensure that doses are always taken during the high-risk window period predicted by our theory. After this high-risk window period, we set a constant probability of missing doses in blocks of 1–3 days. Immediately after a block of doses is missed, we ensure a sufficient number of uninterrupted doses (calculated as Nm) are always taken. If the virus is not eradicated after the 24-week treatment period, the patient is treated with an uninterrupted number of doses as predicted by our theory. The outcome of the simulation at the end of the procedure is reported.
|
10.1371/journal.pcbi.1002096 | Interrogation of the Protein-Protein Interactions between Human BRCA2 BRC Repeats and RAD51 Reveals Atomistic Determinants of Affinity | The breast cancer suppressor BRCA2 controls the recombinase RAD51 in the reactions that mediate homologous DNA recombination, an essential cellular process required for the error-free repair of DNA double-stranded breaks. The primary mode of interaction between BRCA2 and RAD51 is through the BRC repeats, which are ∼35 residue peptide motifs that interact directly with RAD51 in vitro. Human BRCA2, like its mammalian orthologues, contains 8 BRC repeats whose sequence and spacing are evolutionarily conserved. Despite their sequence conservation, there is evidence that the different human BRC repeats have distinct capacities to bind RAD51. A previously published crystal structure reports the structural basis of the interaction between human BRC4 and the catalytic core domain of RAD51. However, no structural information is available regarding the binding of the remaining seven BRC repeats to RAD51, nor is it known why the BRC repeats show marked variation in binding affinity to RAD51 despite only subtle sequence variation. To address these issues, we have performed fluorescence polarisation assays to indirectly measure relative binding affinity, and applied computational simulations to interrogate the behaviour of the eight human BRC-RAD51 complexes, as well as a suite of BRC cancer-associated mutations. Our computational approaches encompass a range of techniques designed to link sequence variation with binding free energy. They include MM-PBSA and thermodynamic integration, which are based on classical force fields, and a recently developed approach to computing binding free energies from large-scale quantum mechanical first principles calculations with the linear-scaling density functional code onetep. Our findings not only reveal how sequence variation in the BRC repeats directly affects affinity with RAD51 and provide significant new insights into the control of RAD51 by human BRCA2, but also exemplify a palette of computational and experimental tools for the analysis of protein-protein interactions for chemical biology and molecular therapeutics.
| The atomic scale interactions that occur at the interfaces between proteins are fundamental to all biological processes. One such critical interface is formed between the proteins, human BRCA2 and RAD51. BRCA2 binds to and delivers RAD51 to sites of DNA damage, where RAD51 mediates the error-free repair of double-stranded DNA breaks. Mutations in BRCA2 have been linked to breast cancer predisposition. Therefore, an accurate picture of the interactions between these two proteins is of great importance. BRCA2 interacts with RAD51 via eight “BRC repeats” that are similar, but not identical, in sequence. Due to lack of experimental structural information regarding the binding of seven of the eight BRC repeats to RAD51, it is unknown how subtle sequence variations in the repeats translate to measurable variations in their binding affinity. We have used a range of computational methods, firstly based on classical force fields, and secondly based on first principles quantum mechanical techniques whose computational cost scales linearly with the number of atoms, allowing us to perform calculations on the entire protein complex. This is the first study comparing all eight BRC repeats at the atomic scale and our results provide critical insights into the control of RAD51 by human BRCA2.
| The human breast cancer suppressor protein BRCA2 controls the functions of the RAD51 recombinase, an enzyme conserved in all kingdoms of life, which carries out the strand exchange reaction central to homologous DNA recombination (HDR) [1]. This essential cellular pathway is responsible for the error-free repair of DNA double strand breaks and is central to the maintenance of genome integrity and the prevention of diseases such as cancer [2].
Attempts to understand the role of BRCA2 in the regulation of HDR have been primarily driven by biochemical and cellular biological studies using regions of the full-length protein, amenable to cellular, biochemical and structural analyses. Two regions in the BRCA2 protein have been shown to interact directly with RAD51. The “BRC repeat” is a conserved motif of BRCA2 of approximately 35 amino acids that is thought to be the primary mode of interaction with RAD51. All known BRCA2 orthologues have been shown to contain at least one BRC repeat motif, but curiously the number of BRC repeats present varies between orthologues ranging from one (e.g. Caernorhabditis elegans Brc-2 and Ustilago maydis Brh2) to fifteen (e.g. Trypanosoma brucei) [3]. All vertebrate Brca2 proteins contain eight BRC repeats, clustered into a single large exon located in the central portion of the protein and show significant conservation of sequence and inter-repeat spacing [4]. The interaction between the BRC repeats of human BRCA2 and RAD51 has been characterised predominantly through structural and biochemical approaches and regulates many of RAD51's activities including RAD51 oligomerisation, and its ordered assembly on single-stranded or double-stranded DNA substrates to control the stepwise events of the strand exchange reaction [5], [6].
A second motif, unrelated in sequence to the BRC repeats, is found at the C-terminus of BRCA2 and, uniquely, is capable of interacting only with oligomerised RAD51 species in the presence or absence of DNA [7], [8]. A further major distinction is that this motif has no significant impact on the execution of HDR by RAD51, but rather links the disassembly of RAD51 complexes that form during HDR to the timing of entry into mitosis [9].
Three pieces of evidence suggest that the BRC repeats of human BRCA2 regulate RAD51-mediated strand exchange. Firstly, BRCA2-deficient cells are defective in HDR [10], [11]. Secondly, it has been shown that a region of BRCA2 comprising all eight human BRC repeats, or a subset of repeats fused to a DNA-binding domain, are capable of stimulating RAD51-mediated HDR and additionally, in the latter case, partially rescuing the HDR defect in BRCA2-deficient cells [5], [6], [12]–[14]. Thirdly, recent biochemical characterisation of the BRC repeats in isolation, as well as the intact human Brca2 protein, shows that they can stimulate RAD51 assembly on single-stranded DNA and inhibit its assembly on double-stranded DNA, hence promoting the stepwise DNA transactions required for strand exchange [5], [6], [15]–[17].
The crystal structure of the complex between the fourth human BRC repeat, BRC4, and the catalytic core domain of RAD51, conserved between all RAD51 orthologues (RecA in eubacteria and RadA in the archaea), has provided mechanistic insights into how BRC peptides can interact with RAD51 [18]. Interestingly, the BRC4 repeat binding to RAD51 was shown to antagonise RAD51 oligomerisation by directly binding to the oligomerisation surface of RAD51 found at the protomer∶protomer interface in oligomerised RAD51 assemblies. Intriguingly, this interaction uses precise molecular mimicry, rather than steric obstruction, to bind to RAD51 using an evolutionarily convergent amino acid sequence. BRC4 binds RAD51 using the motif 1524-FHTA-1527 ( Homo sapiens BRCA2 numbering) to establish contacts with RAD51 otherwise utilised by the sequence 86-FTTA-89 in the linker region of an adjacent RAD51 protomer.
A binding mode of BRC repeats antagonistic to RAD51 oligomerisation is not inconsistent with its stimulatory role in controlling RAD51. It has recently been reported that all BRC repeats may harbour a specific motif architecture that allows binding modes with RAD51 that may be permissive for RAD51 oligomerisation [19]. The identification and characterisation of two modules in the BRC repeats highlights an “FxxA” module that antagonises oligomerisation and an “LFDE” module (by BRC4 sequence nomenclature) that does not affect oligomerisation (and is likely to be permissive for oligomerisation), and complementary binding pockets in RAD51. These findings also suggest that binding modes at the BRC repeat-RAD51 interface are conserved across all known BRC repeats, permit differential regulation of RAD51 and are in essence a new example of hotspot-mediated protein-protein interaction. These tetrameric modules, and the corresponding pockets in RAD51, have been demonstrated to harbour the majority of binding capacity of an entire BRC repeat and their integrity is required for cellular viability through a critical mechanistic role in HDR.
Although these experimental studies focused upon BRC4, a known “strong binder” of RAD51, it was also shown that this conserved motif architecture was predicted to be partially intact even in the fifth BRC repeat, BRC5, a “weak binder” of RAD51, as an “LFDE”-like module was present. Indeed, this module was able to reconstitute RAD51 binding and regulation of RAD51 assembly of DNA when fused to a functional “FxxA” module, derived from BRC4.
Despite significant sequence similarity between the BRC repeats of BRCA2, several studies have reported that these motifs display varying affinities for RAD51 [20]–[22]. The functional relevance of having multiple repeats of varying affinities for RAD51 remains unclear, but may engender tighter regulation of RAD51 behaviour in the more complex genomic environment of higher organisms. Indeed, the finding that BRC repeats use two modules to mediate structural and functional associations with RAD51 and the observation that some repeats, such as BRC5, may contain just one of the modules, albeit of high affinity, speak to this idea.
In this study, we have combined experimental determination of the relative affinities of human BRC peptides for RAD51 with an array of computational simulations that address the atomistic determinants of the behaviour of BRC repeat binding to RAD51. We have used classical molecular dynamics (MD) simulations to explore the interface between RAD51 and the different BRC repeats and also their cancer-associated mutations at a critical interaction hotspot. From these simulation trajectories we have obtained the binding free energies of different BRC-RAD51 complexes using not only classical force fields, but also our newly developed QM-PBSA technique [23], which includes in the calculations the first principles quantum mechanical energies of the entire complexes. Furthermore, we have performed computational alanine scanning mutagenesis studies [24] on the repeats in order to pinpoint the energetic hotspots and quantify their strength in terms of the energetic contribution of each residue and used the more rigorous thermodynamic integration approach to verify critical findings. Our calculations confirm previously reported experimental binding behaviour and provide a rationale for observed differential affinities of BRC repeats for RAD51. Encompassing a range of accuracy and computational expense, these approaches to studying this promiscuous interface between RAD51 and, potentially, multiple peptides, provide fresh mechanistic insights into the regulation of RAD51 by multiple BRC repeats and serve as a template for the interrogation of protein-protein interactions of significant biological interest, often not amenable to direct experimental assessment.
Several studies have previously reported the variation in binding affinities of human BRC repeats to RAD51 [20]–[22]. However, a quantitative comparison of these repeats has not been provided and indeed the majority of experimental insights are based upon BRC4, a stronger binder of RAD51 for which a high-resolution crystal structure exists of the complex. Attempts to purify a homogeneous preparation of RAD51 in a monomeric state amenable to biophysical studies of interaction with BRC peptides with a view to providing thermodynamic parameters have not been successful. In order to circumvent this technical challenge, we have developed a fluorescence polarisation (FP) assay that indirectly measures binding by determining the ability of BRC peptides to act as soluble inhibitors of the BRC4-RAD51 interaction in order to gauge the relative binding affinities of each of the repeats.
This assumes that all BRC peptides can bind to the same surface of RAD51 and are, in essence, competing for the interface on RAD51 pre-bound by BRC4. As all known BRC repeats share common sequence fingerprints that are matched by complementary sequence fingerprints in eukaryotic RAD51 orthologues in species with a BRCA2 orthologue [3], and this binding specificity has been confirmed experimentally, this assumption is likely to extend across all known BRC-RAD51 interactions.
RAD51 used for experimental determination of relative binding affinity was the full-length protein that maintains the capacity to oligomerise. However it should be noted that the structure of the BRC4-RAD51 complex is monomeric and comprises only the core catalytic domain, lacking the first 97 residues of RAD51 comprising the N-terminus and linker region [18]. The interactions of the BRC4 peptide with RAD51 extend along the length of the peptide, including the “LFDE”-module at its C-terminus in the partial context of an . BRC binding to this region is likely to alter the N-terminal domain of RAD51 that is located in a conformation likely to sterically clash with the BRC peptide. The N-terminal domain is connected to the core catalytic domain through a flexible linker region and it is thought that this region of RAD51 engenders conformational flexibility in the N-terminus of RAD51 that is stimulated to accommodate, or be displaced by, BRC peptide binding. Indeed this conformational flexibility has been noted in several high resolution crystal structures of RAD51 orthologues and electron microscopic reconstructions of human RAD51 oligomeric assemblies on DNA in the presence of BRC peptides. The absence of the linker region in the construct used for crystallisation also renders the RAD51 species monomeric.
The outline of the FP assay for detection of disruption of the BRC4-RAD51 interaction is shown in Figure 1(a). Briefly, wild-type full length RAD51 was complexed with Alexa488-conjugated BRC4 and incubated with varying concentrations of each of the eight BRC repeats (unconjugated), present as unlabelled soluble competitive peptides.
In accord with the findings of several qualitative analyses [21], [22], the BRC repeats showed a well-defined relative order of competitive inhibition of the BRC4-RAD51 interaction (Figure 1(b)). BRC4 was the most potent competitive inhibitor, followed by BRC2 and BRC1. BRC8 showed a markedly weaker competitive inhibition. BRC7 and BRC3 showed mild competitive behaviour but failed to achieve inhibition even at the highest concentrations of peptide () and BRC5 and BRC6, in accord with previous reports, showed no significant competition of the BRC4-RAD51 interaction. The BRC4 T1526A mutant (a previously reported non-binding mutant identified by sequential mutagenesis) [25] showed weak competitive inhibition relative to wildtype BRC4.
Understanding protein-protein interactions using computational methods is a major goal at the nexus between structural biology, biophysics and computational chemistry, but is often compromised by limitations of accuracy, high computational cost and the inability to simulate large systems. In this study, we combine a variety of computational methods, with a range of accuracy and computational expense, that are able to measure and rationalise protein behaviour in the context of existing macromolecular complexes. Such methods can help us achieve an understanding of a wide variety of problems relevant to the basic biology of all cellular processes reliant on protein-protein interactions to allow, for example, small molecule chemical intervention with therapeutic or chemical biological rationale.
We begin our analysis with a computational alanine-scanning mutagenesis study [24], [26] of BRC4 using the MM-PBSA method [27], [28]. This approach estimates the contribution of each residue to the free energy of binding at a protein-protein interface by mutating each residue in turn to alanine and measuring the effect of the mutation on the overall free energy of binding. This is done while accounting for the dynamical nature of the interactions and the effects of solvation. Such simulations are directly analogous to the experimental technique of alanine scanning mutagenesis [29], [30], which is used to identify “energetic hotspots” on protein-protein interfaces [31], [32].
Figure 2(a) summarises the MD procedure and Figure 2(b, black line) reports the change in binding free energy () resulting from the mutation of the side chain of each residue of BRC4. As previously reported by Rajendra and Venkitaraman [19], this computational mutagenesis approach highlighted both F1524 and L1545/F1546/E1548 via alanine scanning and A1527 via glycine scanning as residues contributing significantly to the binding of BRC4 to RAD51. Thus, these results are both predictive and fully supportive of a model whereby two modules in the BRC repeats are involved in hotspot-mediated interaction with RAD51.
Having established that computational alanine scanning mutagenesis confirms the presence of two modules within BRC4 previously identified experimentally [19] that contribute to its interaction with RAD51, we sought to understand if further analysis could provide insights into the behaviour of the larger regions of the BRC peptides to establish why they displayed different experimental affinities for RAD51. As no high resolution structural information is available for human BRC repeats 1–3 and 5–8, and accurate biophysical interrogation is hindered by technical challenges in the purification of a suitable N-terminally-truncated monomeric RAD51 species, we turned to computational simulations to analyse the interactions of each of the BRC peptides with RAD51.
In order to approach this problem, we used classical MD simulations of the N-terminal 15 residues of the BRC peptides, denoted “BRCnA” with residue sequences shown in Figure 3. We chose this region for two key reasons. Firstly, given that the full length RAD51 is used for FP assays and only the core catalytic domain is used in simulations, a simulation including the C-terminal 18 residues of BRC peptides (“BRCnB”) would be questionable as this region may be sterically interdependent with the N-terminal domain of RAD51, which is missing in the simulated complex. Attempts to simulate the BRC5B peptide region suggested much weaker binding than observed experimentally and binding modes that did not conform to the crystallised RAD51-BRC4 complex (Figure S1). Secondly, the BRCnA region contains the FxxA module that has a defined functional effect (antagonism of RAD51 oligomerisation) that could be later benchmarked against the binding energy between RAD51 protomers in an oligomeric assembly. Care should be taken when comparing the results of MM-PBSA simulations and our FP assays, as any contribution to binding affinity caused by sequence variation outside the truncated BRCnA peptides is neglected in our computational model. However, Figure 2(b) confirms that use of separate RAD51-BRC4A and RAD51-BRC4B trajectories gives very similar binding behaviour to the full RAD51-BRC4 complex around the significant hotspot regions, indicating that our conclusions concerning the effects of sequence variation in the BRCnA half-peptides are unaffected by our choice of truncation of the experimental peptide. The tail regions of the BRC4A peptide show more variation in the alanine scan of Figure 2(b) since they are more mobile than the residues located in the hotspot. As such, they may become trapped in local minima of the free energy landscape for the duration of the simulation, artificially affecting the free energy of binding calculated by MM-PBSA. With this in mind, throughout this study, we used a combination of MM-PBSA to obtain the total free energy of binding and computational alanine scanning to quantify the contribution of each of the BRCnA peptides in the significant hotspot region, and thus discern the effects of sequence variation.
We used the MD protocol outlined in Figure 2(c) to generate structures of the RAD51-BRCnA interfaces starting from the RAD51-BRC4 crystal structure. BRCnA peptides derived from BRC repeats 1–3 and 5–8 were generated by mutating selected side chains of the BRC4A structure as described in the methods. Here, we assume that each BRC repeat folds in the same manner as BRC4A since no changes in secondary structure are expected to occur on the time scales of these simulations, although significant, localised variations in binding behaviour were observed for some BRC repeats.
Given the importance of the BRCnA repeats in antagonism of RAD51-RAD51 oligomerisation, the relative binding free energies at this protein-protein interface is of significant mechanistic interest. Despite sequence similarity in the hotspot region itself (FHTA in BRC4A mimics FTTA in RAD51), the protein sequence used by RAD51 to self oligomerise is known to partially comprise a helical region [33], which is unlikely to form spontaneously from the structure of BRC4 over the time scale of these simulations. Figure 4 summarises the alternative method used here to generate the RAD51-RAD51 interface starting from a dimeric unit from the crystal structure of the Saccharomyces cerevisiae Rad51 (see methods) and retaining 15 residues of the Rad51 ligand, to match the sequence register of the 15 residues of the BRCnA repeats.
We have used the single trajectory classical MM-PBSA technique, with the gas phase binding entropy of the molecules calculated using a normal modes analysis, to compute the relative free energies of binding of each of the eight BRCnA repeats to RAD51, and compared them to the binding free energy of the RAD51-RAD51 interface. With the exception of BRC5A, Figure 5 shows that the relative free energies of binding of the BRC repeats to RAD51 are very similar, which reaffirms the requirement for very precise measurements of their affinities. Interestingly, MM-PBSA predicts the truncated RAD51 ligand to be the strongest binder to the RAD51 oligomerisation interface, although the difference is mostly entropic and this term is usually assumed to carry the greater uncertainty. Three of the BRC repeats (BRC1A, BRC2A and BRC4A, in that order) bind with affinity comparable to RAD51. Our combination of FP assays and MM-PBSA indicates that BRCA2 also contains five more weakly-bound BRC repeats and the sequence differences that give rise to this variation in affinity will be investigated in the following sections.
The relative binding free energies of the BRCnA repeats to RAD51 as determined by MM-PBSA () are in reasonable qualitative agreement with the inhibition order of the BRC repeats derived from FP assays (). Notable discrepancies are the over-estimation of the binding affinity of both BRC1A and BRC6A in the MM-PBSA approach. One reason for this may be limitations of the computational system, such as the neglect of the 18 BRCnB C-terminal residues and the N-terminal domain of RAD51. Another reason may be limitations of the force field used to describe the interactions between receptor and ligand, which on this length scale are inherently quantum mechanical in nature.
To address the limitations in accuracy of classical force fields, caused by their dependence on a large number of parameters and their inherent inability to describe charge transfer and polarisation we have recently developed a new computational approach that allows us to calculate, from first principles quantum mechanics (QM), the binding free energy of biomolecular complexes consisting of thousands of atoms [23]. In this QM-PBSA approach binding energies are obtained with Density Functional Theory (DFT) calculations which do include charge transfer and polarisation effects. Here, we use QM-PBSA calculations to re-assess the free energy of binding of four of the studied complexes, RAD51-BRC4A (reported in a previous work [23]), RAD51-RAD51, and the two discrepancies between experiment and MM-PBSA, RAD51-BRC1A and RAD51-BRC6A.
Figure 6(a) reveals that there is very good correlation between the gas phase binding energies calculated within MM-PBSA and QM-PBSA. The classical force field is very accurate for the RAD51-BRC1A interaction but under-estimates both the RAD51-BRC6A and RAD51-RAD51 gas phase binding energies. The relative free energies of binding are calculated within QM-PBSA by combining these gas phase binding energies with the scaled solvation free energy and the classical relative entropy change of the solutes upon binding. Figure 6(b) reveals that the binding order of the investigated complexes is the same as predicted by MM-PBSA (), although the relative binding free energy of RAD51-BRC1A and RAD51-RAD51 are under-estimated by 2–5 kcal/mol in MM-PBSA.
Despite significantly improving the calculation of the gas phase quantity in the MM-PBSA scheme, the QM-PBSA method is still potentially subject to inaccuracies. Firstly, the error in the entropy contribution, calculated by classical normal modes analysis, may be large and future improvements in this area should concentrate on increasing the precision of this term. Secondly, the binding energy is calculated by sampling snapshots taken from the classical MD trajectory, which assumes adequate sampling of the ligand's conformational space by the classical force field. To demonstrate this latter limitation, in Figure 6(c), the magnitude of the vector difference between the QM and MM forces averaged over the snapshots is plotted for each ligand atom in the hotspot region of the RAD51-BRC4A and RAD51-BRC1A complexes. The differences are generally small indicating that the QM configurational space is well sampled by the classical force field. However some discrepancies exist in polar groups, especially the R07 side chain in BRC1A, and methods to force-match the force field to the QM forces based on the local environment of the proteins [34], [35] are the subject of ongoing work.
Having established, via three complementary methods, that the BRC repeats show a defined relative order of affinity for RAD51, and in concert with the identification of a sequence motif architecture comprising two specific interaction modules across all BRC repeats with complementary binding energy hotspots in RAD51, we sought to understand why each of the different BRC repeats varied in their affinity to RAD51.
We first utilised computational alanine-scanning mutagenesis, which we have already shown to be predictive of the hotspot-mediated interaction of BRC4 with RAD51, to probe the RAD51-RAD51 and RAD51-BRC4A interaction interfaces for differences in binding free energy that could be associated with sequence variation. The RAD51-RAD51 oligomeric interface differs from the RAD51-BRC4A interface in both structure (the hairpin is replaced by a helical segment) and interaction type (dispersion interactions account for approximately of the QM gas phase binding energy, compared to just in BRC4A). Yet the binding free energies and even each residue's individual contribution to binding, revealed by the computational alanine scan, are remarkably similar (Figure 7(a)).
A representative snapshot of the RAD51-RAD51 complex is shown in Figure 7(b). The interactions of the FxxA hotspot motif, namely F06 and A09 hydrophobic interactions and the backbone inter-protein hydrogen bonds of residue T07 (which do not contribute to the alanine scan), are present in RAD51 as well as BRC4. In the RAD51-RAD51 complex, the alanine scan reveals that residues T10 and F12 provide significant additional contributions to binding. T10 forms an intermittent hydrogen bond with RAD51 via D187 with an occupancy of . F12 forms contacts with residues F166, P168 and Y191 of RAD51. Hydrophobic contact is also formed to some extent between H13 and the RAD51 surface.
The major differences between the RAD51-RAD51 self-oligomerisation interface and the RAD51-BRC4A complex are the increased contribution to binding of residue T08 in the latter and the removal of the hydrophobic F12, which is replaced by the charged residue K12 with little change in binding affinity. In order to bind to the RAD51 interface, K12 forms a salt bridge with D187 of RAD51 (Figure 7(c)) and, to accommodate this change in interaction, BRC4 adopts a structure, whose stability was noted in a previous study [36]. In fact, the backbone interactions between residue 08 and residues 11 and 12 that span the hairpin are found here in all eight simulations. Residue T08 appears to contribute further to hairpin stability by forming side chain hydrogen bonds with the side chain of S10 and the backbone of K12 (Figure S2(a)) and hydrophobic contacts with the methylenes of the K12 side chain and the D187 atom (Figure S2(b)). The latter interaction accounts for the higher contribution to binding of T08 in RAD51-BRC4A relative to RAD51-RAD51 and by interacting simultaneously with residues S10, K12 and D187 (Figure 7(c)), we speculate that T08 stabilises the hairpin interaction with RAD51. This hypothesis is supported by the close homology between five of the eight repeats (BRC1A, BRC3A, BRC4A, BRC7A and BRC8A) in the hotspot region, all of which contain the sequence –FxTASxK– and have very similar alanine mutagenesis scans to RAD51-BRC4A (Figure S3).
The information gained from alanine mutagenesis can be used to resolve the discrepancy between MM-PBSA (or QM-PBSA) and our FP assays. The relative free energies of binding of these five BRC4A-like repeats, appear to be determined not by any sequence variation in the hotspot region, but by the strength of the electrostatic attraction between the ligands of varying net positive charge (Figure 3) and the negatively charged receptor. If we rank these repeats in order of increasing charge (), we observe very good agreement with the relative abilities of these repeats to compete the RAD51-BRC4 interaction in FP assays (). The unexpectedly strong affinity of BRC1A for RAD51 in MM-PBSA appears to be caused by its strongly bound C-terminus (Figure S3), which, as in the 1N0W RAD51-BRC4 crystal structure, is expected to point away from RAD51 in the context of the full BRC1 peptide, and is therefore an artefact of our computational model.
As we have shown in the previous section, five of the eight BRC repeats use very similar motifs to bind RAD51. BRC5, however, replaces the sequence –FxTASxK– with –FxTSCxR–, the most notable change in sequence being the replacement of A09 in a hydrophobic pocket in RAD51 by the polar residue S09. Isothermal titration calorimetry measurements have recently shown that a single A09S mutation in BRC4 is sufficient to significantly reduce the rate constant for the RAD51-BRC4 association reaction and, in turn, reduce the capacity of BRC4 to dissociate the RAD51-DNA complex [37]. In Figure 8(a), we compare a computational alanine scan of the RAD51-BRC5A interface with that of the RAD51-BRC4A interface. The most interesting difference between the two curves is at position 09. Although S09 remains bound throughout the MD simulation, its contribution to the binding free energy is 1.3 kcal/mol lower than the A09 contribution in RAD51-BRC4A. This energy difference is sufficient to explain experimental observations of loss of binding affinity upon A09S mutation in BRC4 [37] and may be rationalised by considering the relative solvation free energies of the two residues, which are accounted for naturally in the MM-PBSA scheme.
The binding mechanism of BRC5A to RAD51 is otherwise very similar to the RAD51-BRC4A complex (Figure 8(a)). Cysteine is less polar than serine, which it replaces as residue 10, and forms neither intra-protein hydrogen bonds with T08 nor inter-protein hydrogen bonds with D187. The R12-D187 interaction is present, though is weaker than the K12-D187 interaction that it replaces in the BRC4A-like repeats. Overall, both FP assays and MM-PBSA predict a very weak interaction between RAD51 and BRC5.
The use of computational simulation, in particular MD, allows the investigation of dynamical motion and access to structures that are not amenable to experimental structural determination. This is particularly relevant for the BRC repeat, BRC2A, which differs significantly in sequence from BRC4A in the hotspot region, replacing the sequence –FxTASxK– with –FxSAHxT–. Analysis of the RAD51-BRC2A MD trajectory reveals that H10 does not form hydrogen bonds within the BRC2A hairpin or directly with RAD51, which is the role of S10 in BRC4A. Yet BRC2A is among the most strongly bound repeats according to both FP assays and MM-PBSA.
By using MD simulations to explore the conformational space of the receptor-bound ligand and alanine scans to probe the contribution of each residue to the binding free energy, we are able to rationalise the high affinity of BRC2A for the RAD51 interface. Figure 8(b) reveals a different binding mode to that observed in the crystal structure of the RAD51-BRC4 complex. Firstly, the computational alanine scan reveals contributions to binding from residues F03, which forms a hydrophobic contact with the RAD51 surface, and R04, which is due to longer-ranged electrostatic effects. Secondly, the S10-D187 hydrogen bond is replaced by S08-D187 and the T12-D187 bond is present for a higher proportion of the simulation than the K12-D187 interaction in, for example, RAD51-BRC4A ( vs. ). This change in binding pattern appears to introduce strain into the RAD51-BRC2A hotspot interface. The two hydrogen bonds formed between the backbone of residue 07 and RAD51 are well conserved in the other seven repeats, but here are and longer than at the RAD51-BRC4A interface (Figure S4). Despite a significant variation in residue sequence in the BRC2A hotspot region compared to the other BRC repeats, the similarity of its binding free energy and alanine scan with those of BRC4A is striking.
Armed with this knowledge, we can propose mechanisms for binding of the different BRC repeats to RAD51 with varying affinity, with implications for the regulation of HDR. BRCA2 is mutated in a significant proportion of individuals with familial breast and ovarian cancer [38], [39]. However, of the many sequence alterations in BRCA2 that have been found in cancer samples (Breast Cancer Information Core, http://research.nhgri.nih.gov/bic/) [18], it remains unclear which represent silent genetic variations and which represent pathogenic mutations. This remains a major problem in the field. We have therefore sought to test whether our atomistic simulations of BRCnA-RAD51 complexes might reveal information concerning the ability of cancer-associated BRCA2 alterations to affect the interaction between BRCA2 and RAD51. With this in mind, we have performed a series of additional simulations on carefully selected single residue mutations, which are designed both to test our predictions regarding the amino acid sequences that give rise to differential binding affinities in the BRC repeats and to examine the effects on binding of single residue mutations associated with cancer development.
One such potentially pathogenic mutation is the S08P substitution in BRC2A, at a site which we have predicted to bind directly to RAD51 in the wildtype complex. We have performed an additional MD simulation of the BRC2A S08P mutant in complex with RAD51 and found that, as expected, the S08P mutation significantly reduces the binding free energy of BRC2A by over 10 kcal/mol (Figure S5). Hence, by using atomistic simulations, we are able to directly link pathogenic mutations in the BRC repeats with changes in binding affinity, which may in turn affect the integrity of HDR. Interestingly, the alanine scan (Figure 8(b)) reveals that the source of this reduction in binding free energy is not only the loss of direct interactions between S08 and RAD51, but also the removal of the T12-D187 hydrogen bond. The geometry of the proline mutation does not allow intra-hairpin backbone hydrogen bonds and the idea that destabilisation of the hairpin and loss of co-operativity between residues spanning the hairpin may reduce binding affinity will be investigated further in the next section.
We have already proposed that T08 in the BRC4 repeat plays an important role in stabilising the interactions between the and RAD51, namely the S10-D187 and K12-D187 hydrogen bonds, by forming side chain hydrogen bonds with the side chain of S10 and the backbone of K12 (Figure S2(a)) and hydrophobic contacts with the methylenes of the K12 side chain and the D187 atom (Figure S2(b)). We now investigate this stabilisation further by examining the binding behaviour of BRC6A, which contains the strongly hydrophobic isoleucine residue at position 08.
Figure 9 shows a snapshot of the RAD51-BRC6A interaction and the results of the computational alanine scan. Firstly, the direct hydrophobic interaction between I08 and D187 of RAD51 is increased relative to T08. Interestingly, the T08I substitution has the additional effect of increasing the occupancy of the S10-D187 and K12-D187 hydrogen bonds ( and in BRC6A vs. and in BRC4A), which can be rationalised by the observation of strong hydrophobic contact between D187, I08 and K12 (Figure S2(b)). Very similar behaviour is observed in MD simulations of the cancer-associated T08I mutation in RAD51-BRC7A (Figure S5). Indeed, the overall binding free energy is actually increased relative to wildtype BRC7A.
Although, intra-hairpin hydrogen bonding interactions (T08-S10 and T08-K12) are lost upon T08I substitution, there is no evidence of this causing a decrease in stabilisation in the RAD51-BRC6A or mutant RAD51-BRC7A interactions. The unexpected relative stability of RAD51-BRC6A in MM-PBSA compared with our FP assays may be due to limitations of the computational model, such as the neglect of the BRC6B C-terminus and the N-terminal domains of RAD51. However, a more likely scenario is that the BRC6A (and mutated BRC7A) hairpin will unfold on time scales longer than we can access in our simulations. Indeed, the intra-hairpin hydrogen bond formed between the backbones of residues 08 and 12 undergoes larger fluctuations in simulations of BRC repeats containing I08 than in any repeats containing the highly-conserved residue T08 (Figure S6), which may result in a shorter lifetime of the fold and loss of binding free energy over longer time scales [40].
We have shown above that, on the time scale of these simulations, the T08I substitution enhances the contributions of the hairpin residues, S10 and K12, to binding. To investigate whether this binding contribution may be reduced in some circumstances, we have also investigated the T08A mutation in RAD51-BRC4A (a previously identified structural mutation derived from an equivalent cancer-associated mutation in BRC1). In agreement with our speculation that T08 stabilises the S10-D187 and K12-D187 interactions, the alanine scan (Figure 9) reveals reduced contributions to binding from both S10 and K12 upon T08A mutation (Figure S2(b)), as well as loss of direct interactions from residue T08. Despite this clear loss of binding around the hotspot region, MM-PBSA actually predicts the T08A mutant to have a more favourable binding free energy than wildtype BRC4A (Figure S5).
In order to resolve this discrepancy between MM-PBSA and our analysis of the BRC4A binding hotspot, we have sought to also compute the free energy of the T08A mutation in BRC4A by thermodynamic integration (TI) [41], [42]. The TI technique is one of the most rigorous approaches for calculating free energy changes as it actually connects the start and end states of the mutation along a non-physical but thermodynamically well-defined path of intermediate species ( values) in order to calculate the free energy change associated with the transformation. Provided the sampling is converged with respect to the number of values, it naturally includes all of the entropic contributions from the accessible conformational space, going beyond the harmonic frequencies approximation of the MM-PBSA approach and limited only by the quality of the force field. It is therefore a more rigorous method than the single trajectory MM-PBSA approach, but is also considerably more computationally expensive as one TI calculation will typically need about 100 times more computation than an MM-PBSA calculation. The change in free energy obtained by TI for the T08A mutation in RAD51-BRC4A is . This change in binding free energy is sufficient to explain the weak competitive inhibition of the BRC4 T08A mutant relative to wildtype BRC4 in our FP assay (Figure 1(b)) and confirms that T08 plays an important role in RAD51-BRC4A binding, forming direct contact with RAD51 and maintaining hairpin stability.
We have carried out an investigation of the interactions that determine the stability of a protein-protein system that is essential for normal cellular function (DNA repair) and found to be mis-regulated in cancer. Crucially, this biologically significant protein-protein interaction occurs between a single protein (RAD51) and, mutually exclusively, one of several protein motifs in another (BRCA2) that have measurable variation in affinity despite only subtle changes in sequence. Our approach is based on a combination of computational and experimental techniques and seeks to establish the relative binding affinity of each of the eight human BRC repeats to RAD51.
It has recently been proposed that the BRC repeats interact with RAD51 through two energetic “hotspot” regions, which are in distinct modules of the repeat (termed here, BRCnA and BRCnB) [19]. The reported differences in affinity of a complete BRC repeat for RAD51 are likely to be explained both by the total contributions of both modules to interactions with RAD51, as well as their site accessibility to RAD51 in different functional settings (e.g. monomeric, oligomeric or filamentous forms on DNA). Although each module comprises a discrete tetrameric binding motif with potentially divergent functional effects on RAD51, in this study, we have decided to focus our computational simulations on BRCnA peptides as we have more structural and biochemical insights into the conformations explored by both the BRC repeat and RAD51 components within the confines of this region of the protein-protein interface.
Due to the inherent experimental difficulties with these systems, such as the spontaneous aggregation of purified RAD51 [43], it has not been possible to measure, so far, free energies of binding through approaches such as isothermal titration calorimetry for all eight BRC repeats with a fully monomeric RAD51 core catalytic domain (in accord with the crystallised BRC4-RAD51 complex). We have been able to fill this knowledge gap by using a wide range of computational techniques, not only to measure relative free energies of binding of the different repeats, but also to link these affinities with the variations in sequence observed across the BRC repeats. For systems such as these, for which relatively little experimental structural information is available, we emphasise the need for multiple computational approaches, balancing accuracy with computational expense. Studies employing homology modelling and the computation of in silico interaction energies are able to scan a large number of residues at the BRCn-RAD51 interface with high efficiency and have successfully predicted a number of mutations that enhance RAD51-BRC4 binding (e.g. the L1545F mutation in the “LFDE” hotspot) [37]. However, this approach assumes that each of the BRC repeats interacts in the same manner as BRC4 with RAD51 and neglects both relative free energies of solvation of the ligands and longer time scale dynamics of their interaction with RAD51 (such as the S10-D187 hydrogen bonds that fluctuate on nanosecond time scales).
In this paper, we have approached the system with a more rigorous (and also more computationally expensive) set of methods that are applicable across any protein-protein interaction. In using classical MD to sample the conformational space of the complexes derived from the RAD51-BRC4 crystal structure, we have assumed that the BRC repeats interact with binding modes broadly similar to those of BRC4, but also found small but significant differences such as the model we have proposed for the RAD51-BRC2A interface. By post-processing the resultant MD trajectories using MM-PBSA analysis, we have obtained a relative order of binding that is in reasonable qualitative agreement with our FP assays. Furthermore, the QM-PBSA method we have developed allows us to compute binding free energies from large-scale quantum mechanical first principles calculations, which is an important step towards resolving the affinities of similar repeats which typically vary by a few kcal/mol. In order to link these free energy calculations with variations in sequence, a useful tool is alanine scanning mutagenesis, which estimates the contributions of each residue of the BRC repeats to the total binding energy, can be compared directly with experiment and can be used to reveal binding hotspots and potential sites for small molecule targeting. The alanine scanning derived contributions confirm the hotspot model and show that the majority of the binding energy is concentrated in the FxxA hotspot of BRC4A and its analogues for the other repeats. Also, based on computational alanine scanning and significant to the processes behind the regulation of RAD51 by BRCA2, we have rationalised experimental observations that the A09S mutation in BRC4 reduces the free energy of binding to RAD51 [37].
The sequence variation in the BRCnA region has a significant effect on the stability of the structural environment in which the FxxA hotspot is embedded. The hairpin domain of the BRCnA repeats is critical for maintaining potent interaction with RAD51 and we have found that both T08 and I08 are capable of stabilising this fold, on the time scale of these simulations, via intra-hairpin interactions. Stability of the hairpin can be compromised by mutations that are associated with cancer predisposition and may hence compromise the integrity of HDR. The computational tools we have employed allow us the ability to study the effect of essentially any mutation to the repeats and we have hence used them to interpret the mechanism of crucial cancer-associated mutations. For example, alanine scanning mutagenesis reveals reduced inter-protein interactions between RAD51 and the hairpin of the T08A mutant form of BRC4A. In this case, MM-PBSA fails to recover the relative binding affinities of wildtype BRC4 and its mutant form observed in our FP assays and so we have turned to the more accurate and computationally expensive TI technique to confirm our observations from experiment and computational alanine scanning. Simulations such as these are vital since mutations do not always fall in the BRC repeat for which there is a high-resolution structure or in a region of obvious functional relevance in the BRC repeat (we note that the known cancer-associated mutations are not in the F/A residues of FxxA but they do have an effect on it). This may be true of L/F/D/E as well (no known mutations in hotspot residues but the structural context may be affected).
There is significant therapeutic relevance associated with our insights into the behaviour of the in the context of the natural variation found in different repeats and our understanding of how the interaction with RAD51 can be enhanced or reduced. We envisage that our approach can be used for the rational computer-aided design of peptidomimetic drugs that specifically compete for, and block, the BRC-RAD51 interaction. This could be achieved through mimicry of the BRC-RAD51 interface or potentially through the use of “stabilised hairpins”, in a manner akin to recent developments in the stabilisation of chemical scaffolds [44]. By defining how a promiscuous interface is able to interact with different primary sequences with varying binding capacities, we have the potential to understand the dynamic nature of protein-protein interactions and identify the determinants of molecular discrimination that could be studied with regard to biological consequences of binding mode and mechanism of protein-protein interactions and re-evaluating peptidomimetic insights for the rational design of small molecule targeting of protein-protein interactions.
FP measurements were carried out in a 384-well, low-volume, black, flat bottom polystyrene NBS microplate (Corning 3820) using a PHERAstar Plus plate reader (BMGLabtech). The polarisation values are reported in millipolarisation units (mP) and were measured at an excitation wavelength of 485 nm and an emission wavelength of 520 nm. Following assay optimisation, full length wild type RAD51 protein was used at a final concentration of 135 nM and Alexa488-BRC4 peptide at 10 nM. By varying the concentration of Alexa488-BRC4 it was shown that FP was independent of total fluorescence (data not shown). The Z′ factor for the assay was calculated to be 0.771 (data not shown). To assess the relative ability of the BRC repeats to displace the Alexa488-BRC4 in this assay, unlabelled BRC repeat peptide was added to each well at a final concentration of (serial dilution) and measurements made in quadruplicates. To validate the method, the experiment was repeated for selected BRC repeats using an ELISA assay (Figure S7), as described previously by Rajendra and Venkitaraman [19].
All peptides, listed in Figure 3 but with full sequences as described in Figure S8, were synthesised by the Cancer Research UK Peptide Synthesis Facility with a C-terminal amide except Alexa488-BRC4, synthesised by Cambridge Research Biochemicals Ltd with an additional N-terminal Alexa488 moiety attached by an aminohexanoic acid spacer. Peptides were purified to by HPLC, sequence-verified by time-of-flight mass spectrometry and diluted in water.
MD simulations were performed with the amber10 package [45], using the X-ray crystal structure of the RAD51-BRC4 complex [18] (PDB: 1N0W) as the starting structure. Water molecules were treated using the TIP3P force field and all protein interactions were described by the amber ff99SB biomolecular force field [46]. Coulomb interactions were treated using the Particle Mesh Ewald sum, with a real space cut-off of 10 Å. The cut-off length for Lennard-Jones interactions was also set to 10 Å. A short energy minimisation was performed in vacuum to remove steric contacts, water and sodium counter-ions were added and the system was heated to 300 K with weak harmonic restraints on the complex at constant pressure (NPT ensemble). Finally, all restraints were removed and the system was equilibrated for 2 ns at 300 K, at the end of which the root mean square deviation of the protein backbone atoms was converged and was less than 2 Å relative to the original crystal structure. In addition, three 12 ns production runs (with different initial velocities) were performed to provide structures for a computational alanine scan of the full RAD51-BRC4 complex (Figure 2(b)).
In order to study the relative binding affinities of the eight BRC repeats to RAD51, we have removed all water molecules from the equilibrated structure of RAD51-BRC4 and truncated the BRC4 peptide to include only the N-terminal 15 residues (P1519-K1533) that bind to the RAD51 oligomerisation interface (RAD51-BRC4A). RAD51 was terminated by and groups, which are more than 25 Å from the hotspot region and are, therefore, not expected to affect binding energetics. The BRC4 half peptides were all terminated by and in accordance with our experimental procedures. The choice of terminal groups for the BRC repeats may affect the strength of binding determined by MM-PBSA, as discussed in the results section, but does not affect the contribution of each residue in the hotspot region to binding, on which our discussion concerning sequence variation is based. Starting structures of the remaining seven BRC repeats in complex with RAD51 were obtained by truncating at the atom only residues that differ between the two structures and using the leap module of amber10 to rebuild the mutated side chains. The resulting eight complexes (plus the five mutated ligands detailed in Figure 3 and discussed in Supporting Text S1) were re-solvated and, as above, were heated to 300 K and simulations were performed for times ranging from 26 ns to 53 ns (simulations were stopped when the running average of the MM-PBSA binding free energy did not vary by more than 1 kcal/mol in the final 12 ns). Snapshots were saved every 6 ps for MM-PBSA single trajectory analysis over the final 24 ns. In order to test the reproducibility of alanine scanning mutagenesis of the hotspot regions, we have performed an additional 24 ns simulation of the RAD51-BRC6A complex (Figure S9).
The above procedure predicted BRC2A to be a weak binder, in contrast to our FP measurements. It may be expected that the RAD51-BRC4A complex is a poor starting configuration since homology with the RAD51-BRC2A complex is relatively low. To further explore the configuration space of BRC2A, an additional simulation was performed starting with the RAD51-BRC2 complex (entire BRC2) and following 2 ns of simulation, the final structure was truncated and used as input for the RAD51-BRC2A simulation, which led to the reported dynamics and a favourable binding free energy. To confirm that this approach was not artificially lowering the binding free energy, a similar procedure was applied to the RAD51-BRC3A interaction. No gain in binding free energy was observed.
Precluding computational analyses on the oligomerisation interface between RAD51 monomers (competed by the BRCnA region of the BRC repeats), no high-resolution crystal structure exists for a human RAD51 oligomeric species either in solution or on DNA. In order to address this problem, we performed analyses on a modelled structure based on the interface between Rad51 protomers from the budding yeast Saccharomyces cerevisiae Rad51 orthologue [33] (PDB: 1SZP). Both the “receptor” and “ligand” components of a dimeric unit of the ScRad51 dimer were truncated to match as closely as possible the complexes between the humanised RAD51 receptors and the BRC repeats described above. Namely, the Rad51 receptor N-terminus was removed, keeping only residues E156-P395. In order to compare directly with the BRCnA peptides, a 15 amino acid peptide (L139-R153), which is responsible for binding at the Rad51-Rad51 interface in yeast and contains the FVTA motif (conforming with the FxxA motif conserved across RAD51 orthologues), was retained as the ligand. The truncated complex was minimised in vacuum and equilibrated in water at 300 K, as in Figure 2(a). Finally, the equilibrated complex was removed from water and all residues of both the receptor and ligand were mutated to the human form, leaving a 15 amino acid ligand containing the FTTA motif interacting with the fully “humanised” RAD51 receptor. The root mean square deviation of the backbone atoms of the resulting complex remained below 2.5 Å relative to the equilibrated yeast structure throughout the subsequent production run, indicating that the yeast dimer is a reasonable input model for human RAD51.
Although yeast have no identifiable BRCA2 orthologue and yeast filament structures have been shown to differ slightly from those of human RAD51 by low resolution electronic microscopic reconstruction [47], [48], our method of humanisation and energy minimisation of the yeast structure make this model suitable for our analyses. Furthermore, we have based our model on the closest orthologue of human RAD51 that is currently available with a high-resolution structure, and the I345T mutation used to aid crystallisation of the yeast Rad51 on DNA [33] is not expected to affect the oligomerisation interface studied here.
Free energy calculations of the resulting trajectories were performed using both MM-PBSA [24] and QM-PBSA [23] techniques, retaining 163 residues of the RAD51 receptor (a atom complex). Classical free energy calculations were carried out using the MM-PBSA post-processing module in amber10. In the single trajectory MM-PBSA approach, the relative free energy of binding between a receptor and its ligands is given by:(1)where the gas phase binding energy is split into electrostatic (EL) and van der Waals (vdW) terms, and averaged over the ensemble of snapshots extracted from the MD simulation. Infinite non-bonded cut-offs were used for these molecular mechanics contributions. Similarly, the binding free energy of solvation from the Poisson-Boltzmann continuum solvation model includes electrostatic (PB) and non-polar surface area (SA) terms. For calculating the free energy of solvation, dielectric constants of 1.0 and 80.0 were used for the solute and solvent respectively and the Poisson-Boltzmann equation was solved on a grid of spacing 0.5 Å. A spherical solvent probe of radius 1.4 Å and atomic radii provided by the amber force field were used for the implicit solvent molecules and solute atoms, respectively. The non-polar contribution to the free energy was calculated via , where SA is the solvent-accessible surface area and is . Finally, is the binding entropy of the molecules, arising from changes in the translational, rotational and vibrational degrees of freedom of the solute species, and was estimated by normal mode analysis, using the NAB module of amber10. The trajectory was sampled every 0.75 ns and each snapshot was minimised in the generalised Born implicit solvent model, using initially conjugate gradients and then Newton-Raphson minimisation, until the root mean square of the elements of the gradient vector was less than . The harmonic frequencies of the vibrational modes were then calculated at 300 K for these minimised structures using normal mode analysis.
Trajectories were sampled every 120 ps for computational alanine scanning using the MM-PBSA post-processing module in amber10. Alanine mutant structures were generated by truncating each residue of the ligand in turn at the atom and by replacing the atom with a hydrogen atom at the correct distance along the bond. Glycine scans on alanine residues found in the interaction hotspots were performed in the same way by truncating at the atom, as is standard in alanine scanning experiments [49]. Although glycine scanning cannot be quantitatively compared to alanine scanning, it allows us to compare the contribution to binding of alanine residues on different BRC repeats and qualitatively identifies residues involved in hotspot mediated interactions (Figure S10).
The T08A mutation in RAD51-BRC4A was investigated using TI in amber10. Gaussian quadrature with nine nodes () and soft-core potentials [50] were used to smoothly mutate all side chain atoms from threonine to alanine in three stages. At each value of , the system was minimised for 1000 steps and heated to 300 K over a period of 0.15 ns with restraints on the heavy atoms of the proteins. To avoid large temperature fluctuations in the solute, a Langevin thermostat with a collision frequency of was employed with a time step of 1 fs. All restraints were removed and the systems were equilibrated for periods ranging from 2 to 6 ns. Productions runs lasted from 2.5 to 8 ns, with the vdW transformations requiring longer simulations to reach convergence (Figure S11).
In the QM-PBSA approach [23], the relative free energies of binding are replaced by:(2)where instead of using a classical force field to obtain the gas phase binding energy of each snapshot, we use a full DFT quantum mechanical calculation. Quantum mechanical calculations of total energies were performed with the onetep program [51], using the PBE gradient corrected exchange-correlation functional [52]. Interactions between electrons and nuclei were described by norm-conserving pseudopotentials. The onetep program achieves computational cost that scales linearly with the number of atoms by exploiting the “near-sightedness” of the single-particle density matrix in non-metallic systems [53]. The density matrix is expressed in terms of a set of non-orthogonal generalised Wannier functions (NGWFs) [54] that are localised in real space with radii of 4.0 Å. The NGWFs were expanded in a basis of periodic cardinal sine (psinc) functions [55] with a kinetic energy cut-off of 830 eV. The spherical cut-off approach for Coulomb potentials [56] was used to eliminate all interactions of the molecules with their periodic images. Van der Waals interactions were included by augmenting the DFT energy expression by damped London potentials with parameters optimised specifically for the PBE functional [57] (). The root mean square error in gas phase binding energies of a benchmark set of complexes calculated using the DFT methodology described above has been shown to be approximately 1 kcal/mol when compared to MP2 and CCSD(T) methods extrapolated to the complete basis set limit [57]. is the weighted polar part of the solvation free energy from the MM calculation:(3)where is determined for each complex studied by a best fit power law curve to a plot of against [23]. QM-PBSA is more computationally demanding than MM-PBSA and so the trajectory was sampled every 1.5 ns. Following previous work [23], in order to improve convergence with the number of snapshots sampled, four additional snapshots were chosen so as to minimise the difference between the properties of the sampled set (as calculated by MM) and the high sample limit of the MM distribution. The chosen properties were the mean and standard deviation of the binding free energy and the fractional occupancies of intermittent hydrogen bonds (D187-S10 and D187-K12 for the repeats BRC1A, BRC4A and BRC6A and D187-T10 for RAD51-RAD51). Using this method, total binding free energies were converged to within 0.5 kcal/mol with respect to the number of snapshots sampled.
MM force errors were evaluated as the magnitude of the vector joining the MM and QM forces for each atom, , averaged over snapshots sampled every 1.5 ns.
|
10.1371/journal.pgen.1000816 | Genetic Dissection of Differential Signaling Threshold Requirements for the Wnt/β-Catenin Pathway In Vivo | Contributions of null and hypomorphic alleles of Apc in mice produce both developmental and pathophysiological phenotypes. To ascribe the resulting genotype-to-phenotype relationship unambiguously to the Wnt/β-catenin pathway, we challenged the allele combinations by genetically restricting intracellular β-catenin expression in the corresponding compound mutant mice. Subsequent evaluation of the extent of resulting Tcf4-reporter activity in mouse embryo fibroblasts enabled genetic measurement of Wnt/β-catenin signaling in the form of an allelic series of mouse mutants. Different permissive Wnt signaling thresholds appear to be required for the embryonic development of head structures, adult intestinal polyposis, hepatocellular carcinomas, liver zonation, and the development of natural killer cells. Furthermore, we identify a homozygous Apc allele combination with Wnt/β-catenin signaling capacity similar to that in the germline of the Apcmin mice, where somatic Apc loss-of-heterozygosity triggers intestinal polyposis, to distinguish whether co-morbidities in Apcmin mice arise independently of intestinal tumorigenesis. Together, the present genotype–phenotype analysis suggests tissue-specific response levels for the Wnt/β-catenin pathway that regulate both physiological and pathophysiological conditions.
| Germline or somatic mutations in genes are the underlying cause of many human diseases, most notably cancer. Interestingly though, even in situations where every cell of every tissue of an organism carries the same mutation (as is the case for germline mutations), some tissues are more susceptible to the development of disease over time than others. For example, in familial adenomatous polyposis (FAP), affected persons carry different germline mutations in the APC gene and are prone to developing cancers of the colon and the rectum—and, less frequently, cancers in other tissues such as stomach, liver, and bones. Here we utilize a panel of mutant mice with truncating or hypomorphic mutations in the Apc gene, resulting in different levels of activation of the Wnt/β-catenin pathway. Our results reveal that different pathophysiological outcomes depend on different permissive signaling thresholds in embryonic, intestinal, and liver tissues. Importantly, we demonstrate that reducing Wnt pathway activation by 50% is enough to prevent the manifestation of embryonic abnormalities and disease in the adult mouse. This raises the possibility of developing therapeutic strategies that modulate the activation levels of this pathway rather than trying to “repair” the mutation in the gene itself.
| The evolutionarily conserved Wnt/β-catenin pathway is a critical regulator of proliferation and differentiation and plays a pivotal role during embryonic development and in the maintenance of tissue homeostasis in the adult. A multitude of studies have documented that impaired or excessive activation of the Wnt/β-catenin pathway result in a large number of pathophysiological conditions, including cancer (for review see [1]). Tight regulation of Wnt/β-catenin signaling is ensured by compartmentalized expression of the different Wnt ligands and receptor components and this is complemented by multiple layers of negative regulation. In particular, the tumor suppressor protein Apc provides a platform for the formation of a β-catenin destruction complex, and thereby acts as a negative regulator of activated Wnt signaling. Loss of Apc function leads to ligand-independent accumulation of β-catenin and its nuclear translocation, where it binds to Tcf/Lef family transcription factors and induces expression of target genes such as Axin2, Cyclin D1 and c-Myc that are involved in proliferation and transformation (for review see [2]).
During embryonic development, Wnt/β-catenin signaling plays an important role in the anterior-posterior patterning of the primary embryonic axis in vertebrates. Unregulated activity of the Wnt pathway during embryonic development leads to anterior defects. For example in mice, loss of Dkk1, a Wnt antagonist, results in truncation of head structures anterior to the mid-hindbrain boundary [3] and mice doubly deficient for the Wnt antagonists Sfrp1 and Sfrp2 have a shortened anterior-posterior axis [4]. Ectopic expression of Wnt8C in mice causes axis duplication and severe anterior truncations [5], while embryos lacking functional β-catenin have impaired anterior-posterior axis formation [6]. Embryos homozygous for the mutant Apcmin allele, which results in truncation of the full-length 2843 amino acid protein at residue 850 and in heterozygous mice leads to an intestinal phenotype akin to familial adenomatous polyposis (FAP) in humans, fail to develop past the gastrulation stage due to proximalisation of the epiblast and ectopic activation of several posterior mesendodermal genes [7],[8]. While these observations establish indispensable roles for components of the Wnt pathway in patterning the anterior-posterior axis, recent genetic rescue studies have helped to define signaling threshold requirement(s) for head morphogenesis [9].
Mutations in components of the destruction complex (APC, AXIN, GSK3β etc) are implicated in tumorigenesis and result in aberrant, ligand-independent activation of the WNT/β-CATENIN pathway. For instance truncating nonsense mutations in APC, loss of heterozygosity (LOH) or promoter hypermethylation are most prominently associated with aberrant WNT signaling that is characteristic of more than 90% of sporadic forms of colorectal cancer in humans [10]–[12]. Meanwhile epigenetic and genetic impairment mutations that reduce expression of wild-type AXIN2/CONDUCTIN [13],[14] or amino-terminal missense mutations in CTNNB1 (β-CATENIN) [15] are most commonly associated with aberrant WNT signaling in cancers of the liver (hepatocellular carcinoma and hepatoblastoma), stomach, kidney (Wilms tumor) and ovaries. It remains unclear why in humans the intestinal epithelium is most sensitive to cancer-associated somatic mutations in APC rather than to those in other components of the WNT signaling cascade, and to what extent this may be due to the loss of interaction between APC and actin-regulatory proteins and microtubules that affect cell migration, orientation, polarity, division and apoptosis, rather than the proliferation/differentiation generally associated with WNT/β-CATENIN signaling (for review see [16]). However, at least in the mouse the C–terminal domains of Apc are dispensable for its tumor suppressing functions [17]. In addition, phenotypic changes observed after the conditional deletion of Apc including those on apoptosis, migration, differentiation and proliferation are rescued by concomitant deletion of the Wnt/β-catenin target gene Myc [18].
Signaling threshold levels in vivo have been assessed by various approaches, including administration of (ant-) agonistic compounds, the (inducible) over-expression of transgenes and the creation of haploinsufficiency through the combination of knock-out and hypomorphic alleles. Elegant combinations of different hypomorphic Apc alleles, for instance, have demonstrated that within the context of intestinal tumorigenesis, there is a clear correlation between gene dosage and phenotype severity [17],[19]. In particular, these studies implied an inverse correlation between the level of Apc protein expression and activation of the Wnt/β-catenin pathway, and in turn, proliferation and differentiation of epithelium along the crypt-villus axis as well as cell renewal in the stem cell compartment [20]. Here we genetically identify differences in signaling threshold levels that determine physiological and pathological outcomes during embryonic development and various aspects of tissue homeostasis in adult tissue. Using combinations of epistatically related hypomorphic alleles of components of the Wnt/β-catenin signaling cascade, we identify tissue-specific signaling threshold levels for anterior specification during embryogenesis, intestinal and hepatic homeostasis in the adult. Our observations add further support to the “just-right” model [21] of Wnt/β-catenin signaling activation where distinct dosages are required to perturb the self-renewal of stem cell populations and lead to neoplastic transformation in the intestine and liver.
In order to modulate the activity of the Wnt/β-catenin pathway in the mouse, we took advantage of the Apcmin [22] and Apcfl [23] alleles. The premature stop codon encoded by the Apcmin allele encodes a truncated 850 amino acid Apc protein, which lacks the 15- and 20 aa repeats and Axin binding repeats required for β-catenin regulation [24], while the unrecombined Apcfl allele results in attenuated expression levels of wild-type Apc mRNA [23]. We used Western blot analysis of lysates from mouse embryo fibroblasts (MEFs) to quantitate expression of full-length Apc protein and the capacity to augment Wnt3a-dependent signaling in cells from the corresponding Apc allele combinations. We observed an inverse relationship in the hierarchy of allele combinations between full-length Apc protein expression (Figure 1A), and signaling activity of the Wnt/β-catenin pathway recorded with a Tcf4 reporter plasmid (Figure 1B). Owing to the presence of residual amounts of full-length Apc protein, the two soluble Wnt antagonists Sfrp5 and Dkk1 were able to suppress Wnt3a-mediated reporter activation in cells of all tested allele combinations. However, in the presence of Wnt3a, pSUPERTopFlash reporter activity was inhibited less effectively by Sfrp5 and Dkk1 in cells with impaired expression of full-length Apc protein (Figure 1B). Therefore, genetic modulation of the expression levels of full-length Apc protein enables experimental manipulation of Wnt/β-catenin pathway activation for a given concentration of Wnt ligand or its soluble antagonists.
To assess whether the outcome of incremental modulation of Wnt/β-catenin signaling by genetic means in MEFs would impact differentially during development and in adult tissue homeostasis in vivo, we set out to generate adult mutant mice with genotypes comprising different combinations of Apc alleles. Surprisingly, we were unable to obtain Apcmin/fl mice at term from crossing heterozygous Apc+/fl with Apcmin/+ mice. Since homozygous Apcmin, but not Apcfl, mice die in utero due to gastrulation defects [7], we genotyped 117 embryos at E12 and found that all 30 Apcmin/fl embryos lacked all structures anterior to the hindbrain. Anterior morphological defects first became visible in E8.5-E9.5 Apcmin/fl embryos, and remained restricted to that region throughout embryonic development (Figure 2A and 2B). Histological cross-sections of Apcmin/fl E12 embryos revealed the presence of a prominent cap of neural tissue that formed at the most anterior part of the embryo, in the absence of cranial structures and the mandible (Figure 2C). Next we used the BAT::gal reporter allele to confirm excessive Tcf4-dependent β-galactosidase reporter activity in the neural tissue cap of Apcmin/fl E15 embryos. As predicted from the Tcf-reporter analysis in MEFs, we also observed BAT::gal reporter activity around the fronto-nasal region with a gradual increase from Apc+/+ to Apc+/fl and Apcmin/+ embryos. This was further extended to most abnormal anterior structures in the Apcmin/fl embryos (Figure 2E). Furthermore, analysis of E5.5-E7.5 embryos by wholemount confocal immunohistochemistry revealed anterior extension of β-catenin expression in the anterior visceral endoderm, an axial signaling centre in the outer endoderm layer of early embryos [8],[25], of Apcmin/fl embryos when compared to their Apc+/fl counterparts (Figure 2D). However, “headless” Apcmin/fl embryos were present at the expected Mendelian ratios until E15.5 (Table S1A) and live embryos could still be detected at E17.5 (Theiler stage 25–26) (Figure 2A) but at less than the expected Mendelian ratio. Our observations therefore support a role for limiting Apc-dependent signaling) functions during the development and patterning of the most anterior structures of the embryo similar to that proposed for excessive Wnt3 signaling in Dkk1-deficient or compound mutant Dkk1+/−;Wnt3+/− mice [9],[20], and reminiscent of the function played by Otx2 [26].
To establish that the “headless” phenotype in Apcmin/fl mice arose from altering the extent of Wnt/β-catenin signaling rather than arising from other potentially dominant-negative activities mediated by the truncated Apcmin protein, we conducted three further genetic experiments. First, we created a more severely 580 amino acid truncated Apc protein by excising exon 14 in Apc+/fl mice that were crossed with the CMV:Cre deletor strain to induce a germline nonsense frame-shift mutation in the corresponding recombined Apc580Δ allele. Subsequent matings of Cre-transgene-negative Apc580Δ/+ mice with Apc+/fl mice failed to yield Apc580Δ/fl pups at birth (Table S1B). Meanwhile, inspection of E9.5, E12 and E16 litters revealed that approximately 25% of all embryos displayed a “headless” phenotype indistinguishable from that observed in stage-matched Apcmin/fl mice (data not shown).
Second, we attempted to rescue the “headless” phenotype in Apcmin/fl mice by genetically limiting expression of β-catenin in corresponding Apcmin/fl;Ctnnb1+/− compound mutant mice. Resulting Apcmin/fl;Ctnnb1+/− MEFs revealed an approximately 50% reduction of Wnt/β-catenin signaling when compared to their Apcmin/fl;Ctnnb1+/+ counterparts (see below). When mating Apcfl/fl;Ctnnb1+/− with Apcmin/+;Ctnnb1+/+ mice, we recovered Apc+/fl;Ctnnb1+/+, Apc+/fl;Ctnnb1+/− and Apcmin/fl;Ctnnb1+/− mice at weaning age at a similar ratio, while among E13.5 embryos, all four possible genotypes were represented at comparable frequencies (Table S1C and Figure S1). Importantly, Apcmin/fl;Ctnnb1+/− mice developed normally into fecund adults (Figure 2B and data not shown), suggesting that limiting Wnt/β-catenin signaling corrected the development of detrimental phenotypes observed in Apcmin/fl mice.
Since the atypical Wnt receptor component Ryk has recently been suggested to amplify Wnt signaling during cortical neurogenesis through β-catenin-dependent as well as independent pathways [27], we also tested whether the “headless” phenotype was promoted by Ryk activity. However, and in contrast to β-catenin, the embryonic lethality of Apcmin/fl mice was not rescued by genetically limiting the expression of the atypical tyrosine kinase Ryk, because we failed to recover either Apcmin/fl;Ryk+/− or Apcmin/fl;Ryk-/- compound mutant mice at weaning (Table S1D), suggesting that Ryk expression was not contributing to the Wnt/β-catenin induced phenotype.
Collectively, our observations extend previous reports that identified a Wnt signaling gradient along the anterior-posterior axis and a requirement for Dkk1 and other Wnt antagonists at the anterior end to prevent posteriorization [3]–[6],[28],[29]. In particular, our experiments clarify genetically that the tight signaling requirements for head morphogenesis previously attributed to Apc or the extracellular components Dkk1 [3], Sfrp [4],[30], Wnt3a [9] and Wnt8a [5] occur exclusively through the Wnt/β-catenin pathway.
Unlike Apcmin/fl embryos, Apcmin/min embryos die around the time of gastrulation [7], consistent with our observation that Apc580Δ/min MEFs, which serve as a model for unavailable Apcmin/min counterparts, reveal higher Tcf4 reporter activity than Apcmin/fl MEFs (see below). Since the morphological defects in E4.75 Apcmin/min embryos correlate with excessive nuclear β-catenin in the epiblast and primitive ectoderm [8], we also examined the effect of genetically limiting β-catenin in these embryos. Unlike the phenotypic rescue observed in Apcmin/fl;Ctnnb1+/− mice, we detected Apcmin/min;Ctnnb1+/− embryos only at E4.5 and E5.5 but not at later stages (E6 and E7). This finding is reminiscent of the time points of embryonic death of Apcmin/min embryos [7] and suggested that reduction of Wnt/β-catenin signaling was insufficient to rescue their death immediately after gastrulation (data not shown). Therefore, higher threshold levels of Wnt/β-catenin signaling selectively inhibit development at an earlier stage (i.e. gastrulation) and genetic reduction of Wnt/β-catenin signaling through ablation of one Ctnnb1 allele reduces signaling only below the threshold that is tolerated during later stages of development. However, we cannot formally exclude other essential function(s) of the full-length Apc protein, which could be provided by residual full-length protein encoded by the Apcfl allele, and which may be required around the time of gastrulation.
Apcmin/+ mice develop intestinal polyposis upon spontaneous LOH of the wild-type Apc allele which arises from centromeric somatic recombination [31],[32]. Meanwhile, genetic studies estimated the polyposis threshold level to correspond to 10–15% of the full-length protein produced from biallelic Apc expression [19]. We therefore established aging cohorts of mice harbouring different Apc allele combinations to constitute an allelic series for Wnt/β-catenin signaling based on the results in Figure 1. As observed previously, Apcfl/fl mice on a mixed 129Sv x C57BL/6 background remained free of intestinal polyps (>18 month, n = 24), while all Apcmin/+ mice (n = 22) developed macroscopic lesions primarily within the proximal portion of the small intestine. Although tumor multiplicity and incidence was reduced in Apcmin/fl;Ctnnb1+/− mice, leaving 6 of 15 mice (40%) free of polyps (Figure 3A), the remaining macroscopic lesions were of tubulo-villous structure and of similar size to those observed in age-matched Apcmin mice (Figure 3B). The similar latency of disease onset between Apcmin/+ and Apcmin/fl;Ctnnb1+/− mice suggests a common requirement for LOH. We therefore amplified exon 14 from polyps which contain the min allele-specific A>T transition to confirm LOH in all polyps from Apcmin (n = 12) and Apcmin/fl;Ctnnb1+/− mice (n = 4) (Figure 3C and data not shown). Based on our in vitro analysis (Figure 1), these results are similar to observations by Oshima et al. showing a requirement of less than 30% of wild-type Apc to prevent Wnt signaling from reaching the permissive threshold for intestinal polyps to form [33]. Surprisingly, restricting the pool of available cellular β-catenin in Apcmin/fl;Ctnnb1+/− mice selectively reduced tumor multiplicity rather than tumor size when compared to Apcmin mice. This suggests that, once LOH has occurred, Wnt/β-catenin signaling exceeds the permissive threshold level, even in light of a 50% reduction in β-catenin and fuels maximal tumor growth, which indeed may be mediated most effectively by submaximal Wnt activity [34].
In humans, de-regulated WNT/β-catenin signaling plays an important role during onset and progression of hepatocellular carcinomas (HCC) and frequently arises from either dominant mutations in the CTNNB1 (β-catenin) gene, or biallelic inactivation of the AXIN1 and AXIN2 genes that involves LOH associated with somatic (epi-)mutation [35]–[37]. Somatic APC mutations, by contrast, are rarely associated with liver carcinogenesis, but FAP patients with germline APC mutations frequently develop hepatoblastomas as well as colonic adenocarcinomas [38]. In addition, adenovirally transduced, complete Apc gene inactivation in the murine liver resulted in hepatomegaly-associated mortality [39], while its sporadic inactivation triggered the development of HCC [40]. We therefore assessed the incidence of liver tumors in moribund mice of the different Apc allele combinations. We found that all Apcfl/fl mice (n = 15), but none of their Apcfl/fl;Ctnnb1+/− littermates (n = 8), had developed HCC by 450 days of age (Figure 4A and 4B), but remained free of intestinal polyps (Figure 3A). We also used PCR analysis to exclude Cre-independent, spontaneous recombination of the Apcfl allele(s) in these tumors (Figure S2). Taken together with our observation of a reduced (but not complete loss) of Apc protein, this argues that tumors are formed with low level Apc and not in the absence of Apc. Therefore, our results suggest not only that HCC formation can occur due to excessive Wnt/β-catenin signaling but importantly that the permissive signaling threshold for hepatic tumorigenesis is lower than that for intestinal tumorigenesis consistently associated with LOH. Surprisingly, we observed HCC in 47% of Apcmin/fl;Ctnnb1+/− mice (n = 15) including 20% that showed intestinal co-morbidity. Survival analysis of mice from this cohort, where disease was confined either to the intestine (n = 6) or the liver (n = 4; Figure 3D), suggested the requirement for a stochastic secondary event to occur akin to intestinal Apc LOH. However, our genomic analysis of hepatic biopsies from Apcmin/fl;Ctnnb1+/− mice confirmed the absence of Apc LOH (Figure 4A), while qPCR and Western blot analysis revealed similar Apc expression between hepatic lesions and adjacent unaffected tissue from Apcfl/fl mice (Figure 4C). As expected, expression of Wnt target genes in unaffected livers from Apcfl/fl mice was elevated compared to livers from wt mice (Figure 4D). Meanwhile, in Apcfl/fl mice we found further, tumor-specific overexpression of some Wnt-target genes (incl. Cd44) that coincided with attenuation of others (notably encoding the negative regulators Axin2, Dkk2 and Wif1).
In order to clarify the nature of potential additional somatic mutations that may affect or cooperate with Wnt/β-catenin signaling, we excluded the presence of activating mutations in Ctnnb1-exon3 that would ablate the negative regulatory phosphorylation sites in β-catenin (Table S2A). We also failed to identify aberrant hypermethylation of the proximal Axin2 promoter (Figure 4E) and also excluded activating mutations in codons 12, 13 or 61 of H-Ras (Table S2B), although Harada et al. previously observed that simultaneous introduction of H-Ras and a constitutively active form of β-catenin by adenoviral gene transfer conferred HCC, while introduction of β-catenin alone did not [41],[42]. Indeed, exposure of Apcfl/fl mice to the liver-specific carcinogen diethylnitrosamine (DEN), which is known to promote mutations in H-Ras, resulted in a higher tumor incidence than in Apc+/fl mice (Figure 4F). Contrary to the observation with intestinal lesions collected from Apcmin/fl;Ctnnb1+/− and Apcmin mice, we found that hepatic tumor volumes in Apcfl/fl mice were larger than in Apc+/fl mice (261 mm3±167 mm3 (n = 12) vs. 193 mm3±414 mm3 (n = 9), p = 0.036; Mann-Whitney test; mean ± SEM; n = 9) suggesting that the extent of aberrant Wnt/β-catenin activity may control both initiation and progression of lesions in the liver.
Collectively, these data suggest differential signaling threshold requirements for intestinal and hepatic tumorigenesis and likely differences in the molecular mechanisms by which Wnt/β-catenin signaling promotes tumorigenesis in these two tissues. The relatively low proliferative activity of the hepatic stem cell compartment, for instance, may provide protection from Apc LOH, even when facilitated by haploinsufficient expression of a recQ-like DNA helicase in Apcmin/+;BlmCin/+ compound mutant mice which remain free of HCC [43]. In light of the lack of Axin2 promoter hypermethylation, the reduction of tumor-specific Axin2 expression may arise from other stochastic events. For instance, the AXIN2 locus contributes to some cancers by LOH or rearrangements in humans [44]. On the other hand, chronic inflammation and the associated excessive activation of the Interleukin-6 pathway may cooperate with activating mutations in CTNNB1 during malignant transformation of human HCC [45]. Despite similar Tcf4 reporter activity recorded between Apcfl/fl and Apcmin MEFs, Apcmin mice remained free of HCC. This observation may be explained by the premature death of Apcmin relative to Apcfl/fl mice (Figure 3D) together with the late onset of liver tumorigenesis. Indeed, we observe hepatic tumors in the Apcmin/fl;Ctnnb1+/− mice which live longer than Apcmin mice. On the other hand, hepatic tissue shows exquisite sensitivity to differential threshold levels of Wnt/β-catenin signaling, whereby the resulting signaling gradient provides a mechanism for metabolic liver zonation [39]. Indeed, we observed here that partial attenuation of full-length Apc expression in Apcfl/fl mice not only increased the number of cells with nuclear β-catenin (Figure 5A and 5B), but also altered expression of Wnt target genes and liver zonation. In particular, and in agreement with our previous findings [46], we observed that attenuation of full-length Apc favored expansion of a perivenous gene expression program (incl. GS, Glt1 and RHBG) at the expense of a periportal signature (incl. CPS, Arg1 and Glut2) (Figure 5A and 5C). Our observation that aberrant Wnt signaling in Apcfl/fl mice in the absence of additional somatic mutations in H-Ras bias towards tumors with perivenous characteristics is consistent with the finding that H-Ras mutated HCCs favor a periportal gene expression program [47].
To gain biochemical insights into the extent to which Wnt signaling thresholds are related to the tumorigenic response in mice, we generated MEFs of genotypes similar to those of cells having undergone Apc LOH in Apcmin mice. In particular, we inactivated the latent Apcfl allele by Cre-mediated recombination in MEFs following infection with an AdCre-GFP adenovirus that expressed the Cre-recombinase as a GFP-fusion protein (Figure S3). Western blot analysis confirmed expression of the 580 amino acid truncated protein encoded by the recombined Apc580Δ allele, in the presence of the 850 amino acid Apcmin protein (Figure 6A). To prevent our analysis from being affected by potential “plateau effects”, we stimulated MEFs with submaximal concentrations of Wnt3a and found a ∼3-fold increase in Tcf-reporter activity between cells harboring the unrecombined Apcfl or recombined Apc580Δ allele, respectively (Figure 6B, compare Apcmin Δ vs. Apcmin/fl and Apc min/Δ;Ctnnb1+/− vs. Apcmin/fl;Ctnnb1+/−). Furthermore, we confirmed that ablation of one Ctnnb1 allele reduced reporter activity by approximately 50% (compare Apcmin/Δ vs. Apcmin/Δ;Ctnnb1+/−; Apcmin/fl vs. Apcmin/fl;Ctnnb1+/− and Apc fl/fl vs. Apcfl/fl;Ctnnb1+/−), and the comparison suggested similar Tcf4-reponsivenes between Apcmin/+ and Apcfl/fl cells. As predicted from the extent of the activating Apc mutations, we also observed a gradual increase of Tcf reporter activity in the absence of Wnt3a ligand (Figure S4).
Since systemic effects observed in adult Apcmin/+ mice may arise secondary to LOH-dependent intestinal tumorigenesis, we next used Apcfl/fl mice to explore this in the context of the Wnt/β-catenin signaling requirement for the maintenance of the hematopoietic cell population [48]. Specifically, Apcmin/+ mice develop lymphodepletion around the time when intestinal tumors are observed [49], and this is associated with a progressive loss of immature and mature thymocytes, and the depletion of splenic natural killer (NK) cells. Comparison of 17 week old wild-type, Apcfl/fl;Ctnnb1+/−, Apcfl/fl and Apcmin/+ mice revealed a strong reduction of mature single positive CD4+ and CD8+ cells in the spleen of Apcmin mice and a less pronounced reduction in immature double positive CD4+,CD8+ cells (Figure 7A and Figure S5A). Moreover, this was reflected by a reduction in splenic CD3+ thymocytes and DX5+, CD3- NK-cells (Figure 7B and Figure S5B) in Apcmin/+ mice when compared to Apcfl/fl mice. Since we did not observe lymphodepletion as a consequence of incremental increases in Wnt/β-catenin signaling from wild-type to Apcfl/fl;Ctnnb1+/− and Apcfl/fl mice (where the latter allele combination generates comparable signaling to that of Apcmin/+ cells), we conclude that this phenotype in aging Apcmin/+ mice is likely to be secondary to LOH-induced intestinal tumorigenesis. This conclusion is consistent with the lymphodepletion phenotype persisting in tumor-bearing irradiated Apcmin/+ mice that have been reconstituted with wild-type bone marrow [49] and our observation that thymic atrophy and associated T-cell depletion reported by Coletta et al., [49] in their tumor bearing 14 week old Apcmin/+ mice is not a reproducible finding at 17 weeks in our Apcmin/+ colony (data not shown) which displays a relative delay in polyposis onset.
The present study underscores the power of hypomorphic alleles in the mouse to understand mechanisms that help to explain at the molecular level the specificity of pleiotropic signaling cascades. Here, we propose the existence of differential permissive Wnt/β-catenin signaling threshold levels during development and tissue homeostasis, and how they relate to each other with respect to specific pathophysiological outcomes. Combining biochemical assessment of different Apc allele combinations in MEFs with the corresponding mouse phenotype genetically defines threshold levels that are lower for liver tumorigenesis than for influencing cellular identity along the anterior-posterior axis, which in turn are lower than that required for intestinal tumorigenesis (Figure 6B). Our data complement those by Ishikawa et al [50] who observed a more severe head morphogenesis defect in mice homozygous for the hypomorphic ApcneoR allele which showed an 80% attenuation of full-length Apc protein (compared to ∼70% in Apcmin/fl cells, Figure 1A) and a 7-fold increase in Tcf4-reporter activity (compared to ∼5.5-fold in Apcmin/fl cells, Figure 1B and Figure 6B). Previously, we have shown that functional cooperation between individually insufficient (epi-) genetic alterations induced sufficient aberrant Wnt/β-catenin signaling to trigger intestinal tumorigenesis in compound A33Dnmt3a;Apcmin mice, with polyps characterized by retention of the wild-type Apc allele and epigenetic silencing of the Sfrp5 gene [51]. Together with the findings presented here, these observations add further support to the “just-right” signaling model which predicts cellular transformation to require specific and distinct dosages of Wnt/β-catenin signaling in intestinal, mammary or hepatic cells, and which was based on the observation that LOH in mice carrying the hypomorphic Apc1572T allele predisposed to metastatic mammary adenocarcinomas rather than intestinal or hepatic tumorigenesis [52]. Indeed, analysis of somatic mutations found in polyps of FAP patients indicates an active selection process favoring APC genotypes that provide residual levels of β-catenin regulation over its complete loss, which would trigger maximal activation of the Wnt/β-catenin pathway [21]. Furthermore, our results demonstrate a lower requirement of Wnt/β-catenin activation levels for neoplastic transformation of hepatocytes than of intestinal epithelium. Meanwhile, human HCC are frequently associated with somatic mutations in AXIN1 or AXIN2 rather than with those in APC [36],[37] suggesting that APC truncation mutations may be selected against during the process of hepatocyte transformation. Our findings that the frequency of HCC is higher in Apcfl/fl mice than in Apcmin/fl;Ctnnb1+/− mice (despite the higher Tcf reporter activity in MEFs of the latter genotype) may not only be accounted for by the shorter overall survival of Apcmin/fl;Ctnnb1+/− mice, but also predicted from the “just-right” signaling model [21],.
Our data also implies that Wnt/β-catenin signaling is likely to conform to cell type-specific bistable switches, where the input stimulus must exceed a threshold to change from one cellular state (and associated response) to another. In the context of Apc LOH-dependent intestinal polyposis, for instance, the predicted two-fold increase of Wnt/β-catenin signaling between Apcmin/Δ;Ctnnb1+/− cells (corresponding to Apcmin/LOH;Ctnnb1+/− lesions in Apcmin/fl;Ctnnb1+/− mice) and Apcmin/Δ cells (corresponding to Apcmin/LOH lesions in Apcmin/+ mice, Figure 6B), has no further detrimental effect on polyposis-associated survival of Apcmin/+ compared to Apcmin/fl;Ctnnb1+/− mice (Figure 3D). Indeed, a recent report delineates a nested feedback-loop that may include a Wnt signaling-associated MAPK cascade [53] as one of the components which provides the non-linear input-output relationship for GSK3β and associated Wnt/β-catenin activity [54] to generate the dramatic threshold responses that characterize a bistable system.
Differential sensitivity to genetic dosage provides the basis for establishing therapeutic windows when targeting non-mutated components in diseased tissue. Indeed, for instance, the notion of therapeutic exploitation of non-oncogene addiction is based on the difference in signaling thresholds tolerated between normal and neoplastic cells. Based on our hitherto limited capacity to target and/or compartmentalize drug delivery, global single-allele inactivation models may provide a convenient first screen to identify potential drug targets. Here, we extend this concept from our previous findings for Stat3 in the context of inflammation-associated gastric cancer [55] to Ctnnb1 in tumors of the liver and intestine and associated aberrant Wnt signaling.
All animals were handled in strict accordance with good animal practice as defined by the relevant national and/or local animal welfare bodies, and all animal work was approved by the appropriate committee.
Heterozygous Ctnnb1+/− mice were generated by excising exons 3–6 from the germline following the mating Ctnnb1fl/fl males with female C57Bl/6 E2a:Cre mice [56]. Ryk+/−, the Apc mutant Apcmin/+ and Apcfl/fl mice and the BAT-gal transgenic reporter mice have been described previously [22],[23],[57],[58]. All experimental mice were on a mixed genetic 129Sv x C57BL/6 background.
qPCR analysis from liver was performed as described [59]. Following extraction of total RNA with TRIzol reagent (Sigma), first strand complementary DNA was synthesized using the Omniscript RT kit (Qiagen). The PCR reactions were carried out under the following conditions: 94°C for 2 min, denaturation at 92°C for 30 s, annealing at 56°C for 30 s and extension at 72°C for 45 s. Primers were obtained from Invitrogen. The number of cycles was 20 for GAPDH, 25 for Arginase1, Glut2 and RHGB, and 30 for β-catenin and Apc. The calculation of relative expression ratios was carried out with the Relative Expression Software Tool (REST) Multiple Condition Solver (MCS) (http://www.gene-quantification.com/) using the pairwise fixed reallocation randomization test. Primers used are listed in Table S3.
Dissected liver tissue was fixed for 1 h in 4% paraformaldehyde or overnight in 10% formalin (Sigma) at 4°C depending on the antibody used (see below). After fixation, tissue samples were transferred to 70% ethanol and embedded in paraffin wax.
Samples were prepared as described previously [46]. Immunoperoxidase staining for GS, CPS I and CYP2E1 (4% PFA) and β-catenin (formalin) was carried out as follows. Sections were dewaxed in Histoclear for 7 min. Sections were washed in PBS and blocked for 30 min in 2% Roche blocking buffer (Roche) before addition of the following antibodies: anti-mouse GS (1∶400; BD Transduction Laboratories), anti rabbit CPS (1∶1,000; a kind gift of Wouter Lamers), and CYP2E1 (1∶500; a kind gift of Magnus Ingelman-Sundberg) in blocking buffer overnight at 4°C. Immunostaining for β-catenin (1∶50; BD Transduction Laboratories) was carried out as previously described [60]. Excess primary antibody was removed by washing 3 times in PBS for 10 min each. Sections were incubated with the DAKO Envision peroxidase-labeled anti-mouse or rabbit secondary antibody polymer for 30 min. The DAB substrate–chromogen mixture was added to the sections and allowed to develop for 10 min. The reaction was terminated in dH2O and the sections counterstained with hematoxylin where appropriate. Specimens were observed using a Leica DMRB microscope. Image collection from the Leica was made with a Spot camera and images collated into figures in Photoshop.
Mouse embryo fibroblasts (MEFs) were derived from E13 embryos and propagated in DMEM supplemented with 10% FBS. The day before transfection, cells were seeded at 5×104 cells/well into 24-well plates. Wnt3a-conditioned medium was a gift from Liz Vincan (Peter MacCallum Cancer Institute, Melbourne) and Nicole Church (JPSL, Ludwig Institute for Cancer Research, Melbourne) and the recombinant human Dkk1- was from R&D Systems (#1090-Dk). Transfections were carried out using either FuGENE 6 transfection reagent (Roche) or nucleofector (Amaxa), 200 ng pSuperTOPflash, 4 ng pRL-CMV and 200 ng of pCMV-HA-SFRP5 expression construct. Two days later, cultures were processed using the Dual-Luciferase Reporter Assay kit (Promega) and luminescence was measured using a Lumistar Galaxy luminometer (Dynatech Laboratories).
Mice were injected intraperitoneally with a single dose of diethylnitrosamine (DEN) (10 mg/ml) at 40 mg/kg at 14 days of age. Mice were sacrificed 6–8 months later and livers were scored for the presence of macroscopic tumors.
Single cell suspensions from spleens were prepared by passing organs through a 40 µm mesh. Cell suspensions were treated with NH4Cl to lyse red blood cells, and then nonspecific binding was blocked by incubating with mouse Fc block (2.4G2). The cells were incubated for 30 min at RT with the relevant fluorochrome-conjugated antibodies to CD3 (clone 2C11), CD4 (GK1.5), CD8 (53–6.7) and DX5 (#558295). All antibodies and Fc Block for flow cytometry were purchased from BD Biosciences, San Jose, CA. Expression of surface markers on cells was detected using a FACSCalibur flow cytometer (BD Biosciences) and analyzed using the FlowJo software (Tree Star, Inc.) Forward scatter/side scatter (FCS/SSC) gating was used to exclude debris and doublets and dead cells were gated out on the basis of PI positivity measured on the FL-3 channel.
Embryos are killed by submerging in ice-cold PBS for a few minutes and fixed by rocking for 45 min in ice-cold 4% PFA in PBS. Specimens are washed 3×5 min in PBS and subsequently incubated o/n at 30°C in X-gal staining solution. After washing in PBS for a few minutes, stained embryos were photographed.
Cells were lysed using Triton-X based lysis buffer (30 mM Hepes), 150 mM NaCl, 1% Triton-X-100, 2 mM MgCl2), with Complete EDTA-free protease and phosphatase inhibitor cocktail (Roche). This was followed by centrifugation at 13000 g for 5 min at 4°C and denaturing at 95°C for 5 min. Protein concentration was determined using a BIO-RAD assay kit. Proteins were then separated by SDS-PAGE (Invitrogen), blotted onto nitrocellulose and incubated with the appropriate antibody overnight. After incubation with the secondary antibody, proteins were visualized using ECL chemiluminescence detection kit (GE Healthcare). For detection of APC, cell lysates were prepared by resuspending cells in ice-cold Lysis buffer [20 mM HEPES, pH 7.4, 150 mM NaCl, 5 mM EDTA, 1% TritonX-100, 1% deoxycholate and Complete EDTA-free protease inhibitor cocktail] and incubation on ice for 15 min. Lysates were clarified by microcentrifugation at 16,060 g for 30 min at 4°C. Total cell lysates were then analysed by SDS-PAGE (3–8% NuPAGE) and detected using the Odyssey infrared imaging system (Odyssey). Quantification of Western blots was performed by using Image J pixel analysis (NIH Image software). Data from Western blots is presented as band density normalized to the loading control, and is representative of three independent experiments. Anti-Active-β-Catenin (anti-ABC), clone 8E7, was from Upstate (#05-665), rabbit polyclonal antibody to the N-terminus of APC (H-290) was obtained from Santa Cruz Biotechnology (Santa Cruz) and anti-mouse Actin (AC-40) was from Sigma-Aldrich.
Parts of exon 16 containing the Min allele specific T>A substitution was PCR-amplified and the gel-purified amplicons were sequenced on an ABIprism377 DNA sequencer (Applied Biosystems). Apc (ex16) forward primer 5′-TCACCGGAGTAAGCAGAGACAC-3′, reverse primer 5′-TTTGGCATAAGGCATAGAGCAT-3′. Bisulfite treatment of genomic DNA and methylation specific PCR was carried out as described [61].
Adenovirus expressing Cre Recombinase fused to enhanced green fluorescent protein (GFP; Cre-GFP) was produced by cloning a cDNA encoding Cre-GFP into pShuttle, the adenoviral transfer vector (Q-BIOgene). Linearised plasmid was then co-transformed into Escherichia coli with pAdEasy1 (Ad5ΔE1/ΔE3) (Q-BIOgene). The pAdCreGFP was linearised and transfected into Q-HEK293A cells (Q-BIOgene) using the calcium phosphate method (Promega). 10 days after transfection, adenoviral infected cells were collected and the adenovirus was released by three rounds of freeze/thawing, and amplification in Q-HEK293A cells, as described in the protocol (Q-BIOgene). For Tcf4 reporter assays MEFs were plated at 5×104 cells/well and were transfected with pSuperTOPflash, and pRenilla-luc. After 24 h, cells were infected with either Ad-LacZ (control virus) or Ad-CreGFP (20 µl/well, TCID50 1.995×108/ml). 48 h after infection, cells were lysed and assayed for luciferase activity. For Western blot analysis, MEFs were plated at 1.5×105 cells/well in 6 well plates and infected with AdCreGFP (20 and 50 µl/well, TCID50 1.995×108/ml) or Ad-LacZ for 48 h. For microscopy, MEFs were plated on glass coverslips, infected with virus, and after 48 h, infected cells were washed twice with PBS and fixed in 4% formaldehyde/PBS for 5 min. DIC and fluorescent images were produced using a Nikon 90i microscope.
Statistical significance was determined by unpaired t-test or, where indicated, using Mann-Whitney analysis.
|
10.1371/journal.pgen.1000223 | Variations in Stress Sensitivity and Genomic Expression in Diverse
S. cerevisiae Isolates | Interactions between an organism and its environment can significantly influence
phenotypic evolution. A first step toward understanding this process is to
characterize phenotypic diversity within and between populations. We explored
the phenotypic variation in stress sensitivity and genomic expression in a large
panel of Saccharomyces strains collected from diverse
environments. We measured the sensitivity of 52 strains to 14 environmental
conditions, compared genomic expression in 18 strains, and identified gene
copy-number variations in six of these isolates. Our results demonstrate a large
degree of phenotypic variation in stress sensitivity and gene expression.
Analysis of these datasets reveals relationships between strains from similar
niches, suggests common and unique features of yeast habitats, and implicates
genes whose variable expression is linked to stress resistance. Using a simple
metric to suggest cases of selection, we found that strains collected from oak
exudates are phenotypically more similar than expected based on their genetic
diversity, while sake and vineyard isolates display more diverse phenotypes than
expected under a neutral model. We also show that the laboratory strain S288c is
phenotypically distinct from all of the other strains studied here, in terms of
stress sensitivity, gene expression, Ty copy number, mitochondrial content, and
gene-dosage control. These results highlight the value of understanding the
genetic basis of phenotypic variation and raise caution about using laboratory
strains for comparative genomics.
| Much attention has been given to the ways in which organisms evolve new
phenotypes and the influence of the environment on this process. A major focus
of study is defining the genetic basis for phenotypes important for organismal
fitness. As a first step toward this goal, we surveyed phenotypic variation in
diverse yeast strains collected from different environments by characterizing
variations in stress resistance and genomic expression. We uncovered many
phenotypic differences across yeast strains, both in stress tolerance and gene
expression. The similarities and differences of the strains analyzed uncovered
phenotypes shared by strains that live in similar environments, suggesting
common features of yeast niches as well as mechanisms that different strains use
to thrive in those conditions. We provide evidence that some characteristics of
strains isolated from oak tree soil have been selected for, perhaps because of
the shared selective pressures imposed by their environment. One theme emerging
from our studies is that the laboratory strain of yeast, long used as a model
for yeast physiology and basic biology, is aberrant compared to all other
strains. This result raises caution about making general conclusions about yeast
biology based on a single strain with a specific genetic makeup.
| A major focus of genetic study is to elucidate the effects of genetic variation on
phenotypic diversity. The evolution of phenotypes is often driven by environmental
factors and the interactions between each organism and its environment. Recently,
there has been a renewed interest in characterizing the diversity and ecology of
organisms long used in the laboratory as models for biological study. Yeast, worms,
flies, and mice have been studied on a molecular level for decades and have provided
many insights into basic biology. However, most of our knowledge base exists for
only a handful of domesticated lines. Little is known about the natural ecology of
these organisms or the degree to which individuals of each species vary within and
between natural populations.
The budding yeast Saccharomyces cerevisiae exists in diverse niches
across the world and can be found in natural habitats associated with fruits, tree
soil, and insects, in connection with human societies (namely through brewing and
baking), and in facultative infections of immuno-compromised individuals [1]. These
yeasts are transported by insect vectors and likely through association with human
societies. Recent population-genetic studies have begun to explore the genetic
diversity of S. cerevisiae strains [2]–[5]. These
studies have demonstrated little geographic structure in natural yeast populations
and relatively low sequence diversity, particularly within vineyard strains. It has
been proposed that low sequence diversity in this species may be due to a more
recent common ancestor compared to other yeasts [6]. Genomic comparisons
also suggest low rates of outcrossing between strains [7], which may limit the
fixation of genetic differences under selection by reducing effective population
sizes [8].
Although the genetic diversity of S. cerevisiae populations is
emerging from large-scale sequencing projects, the phenotypic diversity within and
between yeast populations has been less systematically studied. Myriad studies have
characterized strain-specific differences in specific phenotypes to identify the
genetic basis for phenotypes of interest (for example, those related to wine making
[9],
thermotolerance [10]–[12], sporulation efficiency
[13]–[16], drug sensitivity [17]–[19], and others [20]–[25]). The degree to which
these phenotypes vary across diverse strains has not been systematically explored.
Other genomic studies have investigated variation in genomic expression across
strains, with the goal of investigating the mode and consequence of gene-expression
evolution [26]–[30]. These studies
demonstrated significant variation in gene expression between strains, and in some
cases pointed to the genetic basis for those differences [27], [31]–[35].
However, each study investigated only a few strains, typically vineyard strains. The
broader phenotypic variation across diverse yeast strains and populations,
particularly natural isolates, is largely uncharacterized.
Here we investigated the variation in stress sensitivity and genomic expression in a
large panel of Saccharomyces strains. We quantified the sensitivity
of 52 strains collected from diverse niches to 14 environmental conditions and
measured genomic expression in 18 of these strains growing in standard medium. We
observe a large amount of phenotypic variation, both in terms of stress sensitivity
and gene expression. Associations among phenotypes revealed relationships between
environmental conditions and among yeast strains. One case in particular suggests
that genetically diverse strains collected from oak soil have undergone selection
for growth in a common niche. This study provides a representative description of
expression variation and stress sensitivity within and across yeast populations,
particularly non-laboratory strains, setting the stage for elucidating the genetic
basis of this variation.
Fay and Benavides conducted a population-genetic study of 81
Saccharomyces strains by analyzing ∼7 kb of coding
and non-coding sequence from each isolate [2]. We characterized the
phenotypic diversity of 52 of these strains, shown in Figure 1. This set included natural isolates
from European vineyards, yeasts collected from African palm-wine fermentations,
commercial wine- and sake-producing strains, clinical yeasts, natural isolates
collected from African and Asian fruit substrates, strains from oak-tree soil
and exudates from the Northeastern United States, three common lab strains, and
other isolates (see Table S1 and [2] for references). We
also characterized two haploid S. cerevisiae strains (RM11-1a
and YJM789) and three other Saccharomyces species (S.
paradoxus, S. mikatae, and S.
bayanus) for which whole-genome sequence is available [36],[37]. Each strain was grown under 31 different
conditions representing 14 unique environments, chosen to provoke diverse
physiological responses. These environments varied in nutrient composition,
growth temperature, and presence of toxic drugs, heavy metals, oxidizing agents,
and osmotic/ionic stress. Cells were grown on solid medium in the presence of
each environmental variable, and viability was scored relative to a no-stress
control for each strain (see Materials and
Methods for details).
The results reveal a tremendous amount of phenotypic diversity in environmental
sensitivity (Figure 2).
Although there were similarities between strains, no two strains were exactly
alike in phenotypic profile. Each displayed a propensity for growth under at
least one environment and sensitivity to one or more conditions. Some strains
were generally tolerant to stressful environments across the board. For example,
strain Y2, originally collected from a Trinidadian rum distillery, and clinical
isolates YJM454 and YJM440 were tolerant of most of these conditions, while the
S. bayanus strain used in our study was sensitive to nearly
all stresses tested. Several strains, including commercial sake-producing
strains, showed a wide standard deviation of growth scores across the stresses,
reflecting that they were either highly sensitive or highly resistant to
different stresses. In contrast, most vineyard isolates grew moderately well in
most of the environments examined (see Discussion).
Exploration of the range of strain sensitivities measured for each environment
also suggested common and unique features of
Saccharomyces' habitats. Collectively, this set of
strains showed the greatest variation in copper sulfate tolerance, sodium
chloride resistance, and freeze-thaw survival, implicating these as
niche-specific features not generally experienced by yeast. In contrast, strains
showed the least variation (but some variability nonetheless) for growth on
non-fermentable acetate, in minimal medium lacking supplemental amino acids, and
at 37°C. Presumably, defects in respiration, prototrophy, and growth at
physiological temperature represent a significant selective disadvantage,
regardless of the particular niche.
Hierarchical clustering of the phenotype data revealed interesting relationships
between groups of strains. In particular, several groups of strains displayed
similar profiles of stress sensitivity across the environments tested (Figure 2). As a group, the
sake-producing strains were extremely resistant to lithium chloride but
sensitive to copper sulfate, calcium chloride, cadmium chloride, and SDS
detergent (p<0.005 based on 10,000 permutations, see Materials and Methods); indeed, this group was slightly more
sensitive to stress in general. Many of the vineyard strains shared specific
phenotypes, including resistance to copper sulfate, as previously noted for
other vineyard strains [26],[38],[29]. The
group of laboratory strains was also highly resistant to copper sulfate as well
as sodium and lithium chloride. In contrast, strains collected from oak soil
were particularly sensitive to copper sulfate and sodium chloride but highly
resistant to freeze-thaw stress (p<0.005, 10,000 permutations).
The similarities in phenotypic profiles could arise through selection (either
directional or purifying) due to shared selective pressures across strains
living in the same environment. Alternatively, phenotypic similarity could
result simply if the strains are genetically related due to a recent common
ancestor. For example, many of the lab strains are closely related, since a
large fraction of their genomes is derived from a common progenitor [39],[40]. We wished to
distinguish between these possibilities for other strain groups. Natural
selection can be inferred by comparing the population genetic structure
(FST) to an analogous measure of phenotypic structure
(QST) [41],[42]. A deviation from
unity suggests that either divergent (QST/FST>1) or
purifying (QST/FST<1) selection has occurred across
populations. We wished to analyze each subpopulation separately, and therefore
we devised a simple alternative approach to identify deviations from neutral
phenotypic variation. We calculated the average pairwise phenotypic distance
over the average pairwise genetic distance for pairs of strains collected from
the same environment (‘sake’,
‘vineyard’, ‘oak’,
‘clinical’, ‘natural’ or
‘other fermentation’). This ratio was compared to the ratio
of distances calculated for pairs of strains between niche groups, generating
the parameter P/G. A P/G ratio = 1 is expected
under neutrality, where the phenotypic to genetic distance is equal for
within-group versus between-group comparisons. In contrast, a value of
P/G<1 suggests that the strains within the group are more similar in
phenotype than would be expected under the neutral model, whereas a ratio
>1 indicates that the strains are phenotypically more variable than
expected based on their genetic relatedness.
The results provide evidence of both selection and shared ancestry for different
groups of strains. First, the P/G ratio did not deviate significantly from unity
for strains in the ‘clinical’,
‘natural’, or ‘other fermentation’
groups (average
P/G = 1.02+/−0.22), nor did
it deviate significantly for randomized simulations (data not shown). In
contrast, P/G was 4.2 and 3.0 for sake strains and vineyard strains,
respectively. Thus, the similarity in their phenotypes likely arises due to
their recent divergence from a common ancestor. Interestingly, these P/G values
were significantly higher than expected by chance (p<0.0001 from 10,000
permutations), suggesting that the strains show more phenotypic
variation than expected. This could arise if strains have experienced
diversifying selection for disparate phenotypes, although it could also result
if genetic distances are underrepresented or skewed due to limited sequence
data.
In contrast, strains collected from oak-tree exudates and soil are phenotypically
more similar than would be expected under a neutral model. We observed a P/G
ratio of 0.31 (p = 0.0013 from 10,000
permutations), indicating that phenotypic variation within this group is lower
than expected based on the strains' genetic relatedness. This suggests
that the strains have undergone selection for growth in a common environment
(see Discussion). Consistent with this model, the S. paradoxus
strain YPS125, also collected from Northeastern oak flux [6], is
phenotypically more similar to S. cerevisiae strains collected
from that environment (pairwise R of 0.61, 0.66, and 0.77 to YPS1000, YPS1009,
and YPS163, respectively) than to the other S. paradoxus strain
in our collection (R = 0.51). At least some of
the phenotypes shared by these strains are likely important for their ability to
thrive in their niche (see Discussion).
Numerous studies have characterized differences in genomic expression between
individual strains of yeast, typically vineyard and lab strains [13],
[26]–[31],[34],[43]. To more broadly
survey the variation in genomic expression across populations, we measured
whole-genome expression in 17 non-laboratory strains compared to that in the
diploid S288c-derived strain DBY8268, using 70mer oligonucleotide arrays
designed against the S288c genome. The long oligos used to probe each gene
minimize hybridization defects due to sequence differences from S288c. We
verified this by hybridizing genomic DNA from 6 strains of varying genetic
distance from S288c: indeed, fewer than 5% of the observed gene
expression differences described below could be explained by defective
hybridization to the arrays (see Materials and
Methods). Therefore the vast majority of measured expression
differences are due to differences in transcript abundance.
A striking number of yeast genes showed differential expression from the
laboratory strain in at least one other strain (Figure 3A). Of the ∼5,700 predicted
S. cerevisiae open reading frames, 2680
(∼47%) were statistically significantly altered in expression
(false discovery rate, FDR = 0.01) in at least
one non-laboratory strain compared to S288c, with an average of 480 genes per
strain. At an FDR of 0.05, over 70% of genes were significantly
altered in expression in at least one non-lab strain (Table 1). The number of expression
differences is comparable to that observed by Brem et al., who reported over
half of yeast genes differentially expressed between the vineyard strain RM11-1a
and S288c [27].
However, closer inspection revealed that many of these expression differences
were common to all of the non-laboratory strains (Figure 3A), revealing that these expression
patterns were unique to S288c. This group was enriched for functionally related
genes, including those involved in ergosterol synthesis, mitochondrial function,
respiration, cell wall synthesis, transposition, and other functions (Table 2). Many of these
functional groups were also reported by Brem et al., who noted that multiple
categories (including ergosterol synthesis and mitochondrial function) can be
linked to a known polymorphism in the Hap1p transcription factor [44].
Indeed, the expression differences specific to S288c were enriched for targets
of Hap1p (p<10−11, hypergeometric distribution) as
well as targets of Hap4p (p<10−6) [45],
which regulates genes involved in respiration. Hence, many of the observed
expression differences may result because of S288c-specific physiology (see
Discussion).
For a more representative description of expression variation in non-laboratory
strains, we sought to represent the expression differences in a way that was not
obscured by S288c. First, we identified genes whose expression varied
significantly from the oak strain YPS163. Second, we identified transcripts
whose abundance varied from the mean of all non-laboratory strains (see Materials and Methods). Although the mean
expression value of each gene is merely an arbitrary reference point, this data
transformation serves to remove the effect of S288c from each array while
maintaining the statistical power to identify expression differences.
Roughly 1330 (23%) of yeast genes varied in expression in at least one
non-laboratory strain relative to the mean of all strains, while 953
(17%) of genes varied significantly from YPS163
(FDR = 0.01). In both cases, two thirds of
significant expression differences were specific to only one strain (Figure 3B and 3C). The number
of genes with statistically significant expression differences from the mean
ranged from 30 (in vineyard strain I14) to nearly 600 (in clinical isolate
YJM789), with a median of 88 expression differences per strain. The number of
expression differences did not correlate strongly with the genetic distances of
the strains (R2 = 0.16). However,
this is not surprising since many of the observed expression differences are
likely linked in trans to the same genetic loci [27],[31],[34],[35],[43]. Consistent with this
interpretation, we found that the genes affected in each strain were enriched
for specific functional categories (Table S4), revealing that altered expression
of pathways of genes was a common occurrence in our study.
We noticed that some functional categories were repeatedly affected in different
strains. To further explore this, we identified individual genes whose
expression differed from the mean in at least 3 of the 17 non-laboratory
strains. This group of 219 genes was strongly enriched for genes involved in
amino acid metabolism (p<10−14), sulfur metabolism
(p<10−14), and transposition
(p<10−47), revealing that genes involved in
these functions had a higher frequency of expression variation. Differential
expression of some of these categories was also observed for a different set of
vineyard strains [26],[28], and the genetic
basis for differential expression of amino acid biosynthetic genes in one
vineyard strain has recently been linked to a polymorphism in an amino acid
sensory protein [35]. We also noted that the 1330 genes with
statistically variable expression in at least one non-laboratory strain were
enriched for genes that contained upstream TATA elements [46]
(p = 10−16) and genes
with paralogs (p = 10−6)
but under-enriched for essential genes [47]
(p = 10−25). The
trends and statistical significance were similar using 953 genes that varied
significantly from YPS163. Thus, genes with specific functional and regulatory
features are more likely to vary in expression under the conditions examined
here, consistent with reports of other recent studies [30],[43],[48],[49] (see
Discussion).
Expression from transposable Ty elements was highly variable across strains.
However, Ty copy number is known to vary widely in different genetic backgrounds
[50],[51], suggesting that
these and other observed expression differences could be due to copy number
variations in particular strains. Indeed, numerous expression differences could
be linked to known gene amplifications in S288c, such as ASP3,
ENA1, CUP1, and hexose transporters [52],[51]. We quantified the
contribution of increased copy number to the observed increases in gene
expression relative to S288c in 6 of our strains. In general,
∼2–5% of expression differences could be wholly or
partially explained by differences in gene copy number (see Materials and Methods). YPS1009 was an exception to the
trend, since nearly 20% of genes with higher expression could be
attributed to increased copy number - most of these genes reside on Chromosome
XII. In fact, more than 80% of genes on Chromosome XII met our
criteria for increased copy number (Figure S1A), indicating that the entire
chromosome is duplicated in this strain. Another example of chromosomal
aneuploidy is evident in strain K9, for which Chromosome IX appears amplified
(Figure
S1B). Whole-chromosome aneuploidy has been frequently observed in strains
growing under severe selective pressure (for example [53]–[56].
Interestingly, the majority of genes on these duplicated chromosomes do not show
elevated transcript abundance in the respective strains.
In fact, only ∼25% of genes with increased copy number in each
strain showed elevated expression (defined at
FDR = 0.01 or as genes whose expression is
>1.5× over S288c). This is in stark contrast to previous
studies demonstrating little dosage compensation in S288c in response to gene
amplification and chromosomal aneuploidy, leading to the conclusion that yeast
does not have a mechanism for dosage compensation. [53],[54],[57]. Instead, our
results suggest that some form of feedback control acts to normalize the dosage
of most genes in non-laboratory yeast strains. The remaining quarter of
amplified genes may be inherently exempt from this feedback mechanism.
Alternatively, relaxed feedback may occur for specific amplifications if the
resulting transcript increase provides a selective advantage to the strain in
question. Indeed, 15–40% (depending on the strain) of genes
lacking feedback control show at least 1.5× higher expression beyond
what can be accounted for by gene amplification alone, indicating that the
expression differences are affected by both gene dosage and regulatory
variation. These genes are excellent candidates for future studies of adaptive
changes.
As observed for gene expression, we found that some genomic amplifications were
common across all 6 strains compared to S288c. All strains showed decreased Ty1
copy number, ranging from 2–15× lower than S288c. This is
consistent with previous studies that showed higher Ty1 copy number (including
active and partial Ty elements) in S288c compared to wine strains and natural
isolates [50],[51],[58]. Most strains also
showed even lower Ty1 transcript abundance, beyond what could be explained by
copy number variations. Thus, in addition to a higher Ty content, S288c also
shows higher expression from Ty genes, perhaps reflecting elevated rates of
retrotransposition under the conditions studied here. In contrast, all strains
showed higher copy number of the mitochondrial genome compared to S288c,
typically elevated 2–3× but nearly 7× higher in
clinical strain YJM789. The most likely explanation is that these strains harbor
more mitochondria than S288c, a fact confirmed in vineyard strain RM11-1a by
mitochondrial staining [25].
In addition to revealing phenotypic diversity within and between yeast
populations, natural variation can also uncover new insights into the effects of
each environment on cellular physiology. For example, we noted correlations
between environments based on the distribution of strain-sensitivity scores. The
most likely explanation is that these stresses have similar effects on cellular
function, and thus strains display similar sensitivities to them. Resistance to
sodium chloride and lithium chloride or tolerance of ethanol and elevated
temperature were highly correlated (R = 0.66 at
p<0.0001 and R = 0.51 at
p<0.0006, respectively, based on 10,000 permutations), consistent with
the known effects of these stress pairs on ion concentrations or membrane
fluidity/protein structure, respectively. Other relationships were not
previously known, including the correlation between sensitivity to SDS detergent
and the heavy metal cadmium (R = 0.64,
p<0.0001) and between ethanol and caffeine tolerance
(R = 0.59, p<0.0001). In contrast,
resistance to freeze-thaw stress was anticorrelated to sodium chloride
resistance (R = −0.35,
p = 0.006), suggesting antagonistic outcomes of
the same underlying physiology. These relationships point to commonalities in
the cellular consequences inflicted by these environments that will be the
subject of future investigations of stress-defense mechanisms.
We also conducted an associative study to identify gene expression patterns
correlated with environmental sensitivity across the 17 non-laboratory strains
(see Materials and Methods for details). As
basal expression differences could significantly contribute to the inherent
ability of cells to survive a sudden dose of stress, the results point to genes
whose expression is related to, and perhaps causes, the phenotypes in question.
Among the top genes associated with copper sulfate resistance was the
metallotheionein CUP1, important for copper resistance and
known to have undergone tandem duplications in copper-resistant strains [59],[60]. Of the genes whose
expression was correlated to sodium chloride tolerance, nearly 20%
are known to function in Na+ homeostasis and/or osmolarity maintenance
(including RHR2, COS3, SIS2
identified through genetic studies [61]–[63] and
JHD2, SRO7, YML079W,
YOL159C, TPO4, UTH1
implicated in high-throughput fitness experiments in S288c [64]). Thus, these and
likely other genes whose expression is highly correlated with each
stress-sensitivity profile play a functional role in surviving that condition.
Other correlations were not expected. Ethanol and caffeine tolerance were both
correlated to the expression of genes encoding transmembrane proteins
(p<0.003, hypergeometric distribution), perhaps related to the effect of
these drugs on membrane fluidity. Sensitivity to the cell-wall damaging drug
Congo Red was significantly correlated to the expression of genes involved in
mitochondrial function and translation, respiration, and ATP synthesis
(p<10−13), revealing a link between
mitochondria/respiration and the cell wall. Although these connections will
require further characterization, they demonstrate the power of using natural
diversity to uncover previously unknown relationships between stresses and
cellular processes.
This study demonstrates the vast amount of phenotypic variation in
Saccharomyces strains collected from diverse natural habitats,
used in industrial processes, and associated with human illness. Considering the
phenotypic responses to the conditions studied here provides insights into the
relationships between specific strains and their niches. For example, the wide
variance in growth scores of sake-producing strains indicates that they are either
highly resistant or sensitive to the different environments studied here, suggesting
that they may be specialized for growth in the defined conditions of sake
fermentation. In contrast, many of the vineyard isolates survived relatively well in
most of the conditions tested. This may reflect their ability to thrive in more
variable, natural environments and may also have facilitated their dispersal into
new environments in a manner associated with human interactions [5].
Geographic dispersal might also explain the higher-than-expected phenotypic
diversity of vineyard strains, which might be driven by diversifying selection
(suggested by our analysis) due to unique pressures imposed after expansion into new
environments.
Although many of the phenotypic differences we observed are probably neutral,
providing no benefit or disadvantage to the strains in question, some are likely to
provide a selective advantage. Copper-sulfate resistance in European vineyard
strains may have arisen through positive selection, since copper has long been used
as an antimicrobial agent in vineyards and orchards [1],[65]. Another example may
apply to the oak strains studied here. Our simple metric comparing phenotypic to
genetic diversity in strains collected from similar environments suggests that oak
strains are phenotypically more similar than expected based on their genetic
relationship. Formally, this could arise if multiple traits are evolving neutrally
(but slower than the genetic drift represented by the sequences used here) since the
strains diverged from a distant, common progenitor. However, the fact that
S. paradoxus oak isolate YPS125 is phenotypically more similar
to S. cerevisiae oak strains than the other S.
paradoxus isolate in our analysis instead supports that these strains have
undergone selection for growth in a common environment. One intriguing phenotype is
freeze-thaw resistance, which may be important to survive the wintry niche from
where these strains were collected. Consistent with this hypothesis, we have
recently isolated numerous Sacharomycete strains (including
S. cerevisiae) from Wisconsin oak exudates, of which
86% (19/22) are freeze-thaw tolerant (DJK and APG, unpublished data).
Ongoing studies in our lab are dissecting the genetic basis for this phenotypic
difference.
In addition to stress sensitivity, gene expression also varies significantly across
yeast populations. More than a quarter of yeast genes varied in expression in at
least one non-laboratory strain under the conditions studied here. Consistent with
other recent reports [30],[48],[49],[66], we find that genes with specific structural or
functional characteristics (including nonessential genes and those with upstream
TATA elements and paralogs) show higher levels of expression variation across
strains. This has previously been interpreted as a higher rate of regulatory
divergence for genes with these features, either in response to selection [48] or
mutation accumulation [49]. However, these features are also common to genes
whose expression is highly variable within the S288c lab strain grown under
different conditions ([67] and data not shown), particularly those induced
by stressful conditions [46],[68]. It is also notable that genes with TATA elements
show higher ‘noise’ in gene expression within cultures of the
same strain [69],[70]. Thus, an alternative, but not necessarily
mutually exclusive, hypothesis is that the expression of these genes is more
responsive to environmental or genetic perturbations, again consistent with previous
studies [66],[30],[48],[49]. We have conducted our experiments under
‘common garden’ lab conditions in attempt to minimize
environmental contributions to expression phenotypes. However, because each strain
may have evolved for growth in a unique environment, each may in fact respond
differently to the same growth conditions used here. Indeed, this may explain the
prevalence of metabolic genes in our set of genes showing variable expression in
multiple strains, since many of these strains have not evolved for growth in highly
artificial laboratory media.
Emerging from our analysis is the fact that S288c is phenotypically distinct from the
other non-laboratory strains studied here. This strain displays extreme resistance
to specific stresses, harbors fewer mitochondria, contains more transposable
elements, and shows unique expression of many genes compared to all other strains
investigated (a direct comparison of the number of differentially expressed genes in
S288c is difficult due to the different statistical power in calling these genes).
We have also found that this strain has an aberrant response to ethanol, since it is
unable to acquire alcohol tolerance after a mild ethanol pretreatment, unlike
natural strains [71]. It is likely that additional responses found in
natural strains have been lost or altered in this domesticated line. The progenitor
of S288c was originally isolated from a fallen fig in Merced, California, and
sequence analysis indicates that S288c is genetically similar to other natural
isolates [1]–[3]. A recent study by Ronald et
al. counters the proposal that S288c has undergone accelerated divergence during its
time in the laboratory [72]. Instead, our results suggest that the strain has
evolved unique characteristics through inadvertent selection for specific traits
(such as growth on artificial media) and population bottlenecks. Thus, the
laboratory strain of yeast may not present an accurate depiction of natural yeast
physiology. Indeed, no single strain can be used to accurately represent the
species, a note especially important for comparing phenotypes across species.
Complete exploration of an organism's biology necessitates the study of
multiple genetic backgrounds to survey physiology across populations.
Despite its limitations, the lab strain offers nearly a century of detailed
characterization, along with powerful genetic and genomic tools. A useful approach
is to complement studies on laboratory strains with investigations of natural
variation. By characterizing stress sensitivity in a large set of strains, we have
leveraged the power of natural diversity to uncover new relationships between
stresses and to reveal previously unknown connections between genes, stresses, and
cellular processes. These connections lead to hypotheses about stress defense
mechanisms that can often be dissected using the valuable tools provided by the lab
strain. Application of genomic techniques to characterize natural yeast strains will
foster such studies while revealing additional insights into genetic and phenotypic
variation in Saccharomyces.
Strains used in this study and references are found in Table S1.
In addition to sequence data from [2], an additional 5,305 bp of noncoding DNA was
sequenced for 41 S. cerevisiae strains over 8 intergenic
sequences (GENBANK accession numbers EU845779 - EU846095) for a total of 13,016
bp over 13 loci. Phylogenetic analysis shown in Figure 1 was performed on the combined
sequence set using the program MrBayes [73]. Evolutionary
distances were estimated using the Jukes-Cantor (JC) model based on 2,056 bp
noncoding sequence data present in all strains; results and significance were
very similar when the distance was based on 9,334 bp of noncoding sequence
excluding only pairwise-deletion data [74]. Strains with
evolutionary distances equal to zero over this subsequence (but clearly non-zero
when all sequence was assayed) were set to 0.00001 to facilitate permutation
calculations. Paralogs were defined as genes with a BLAST E-value score
<10−100.
Yeast strains were grown in YPD medium at 30°C to an optical density of
∼0.3 in 96-well plates. Three 10-fold serial dilutions were spotted onto
YPD agar plates containing the appropriate stress, as well as a YPD plate for a
no-stress control. Cells were also plated onto minimal medium [75] or
YP-acetate. In the case of freeze-thaw stress, 200 µl cells was frozen
in a dry ice/ethanol bath for two hours or left on ice as a control before
spotting onto YPD plates. Cells were grown for 2–3 days at
30°C unless otherwise noted, and viability of each dilution was scored
relative to the no-stress control for each strain. All experiments were done in
at least duplicate over 2–3 doses of most stresses (see Table S2
for raw data and stress doses). Final resistance scores were summed over the 3
serial dilutions then averaged over replicates and stress doses, providing a
single score ranging from 0 (no growth) to 6 (complete growth) for each strain
and each stress condition.
For Figure 2, strains were
clustered based on phenotypes using the Pearson correlation and UPGMA clustering
[76]. Correlations between stresses were calculated
based on the Pearson correlation between strains, excluding 14 strains of highly
similar genetic distance (JC<0.0008). Phenotypes specific to groups of
strains collected from similar environments (see Table S1
for groupings) were calculated based on the median growth score of strains in
that group. Significance was estimated by 10,000 permutations of strain-group
labels, scoring the frequency of observing a median growth score equal to or
greater than that observed.
A parameter, P/G, was calculated to compare the similarity in phenotype to the
similarity in genotype for strains within and between niche groups. The average
pairwise phenotypic distance, taken as the Pearson distance (1 –
Pearson correlation) between phenotype vectors, was divided by the average
pairwise JC distance for strains within a niche group. This value was divided by
the same ratio calculated for all pairs of strains between niche groups (see
Table
S1 for niche groupings). Significance was estimated based on 10,000
random permutations of strain-group labels. The distribution of P/G ratios from
randomized trials was centered on 0.99; furthermore P/G was ∼1.0 for
strains in the ‘clinical’, ‘natural’,
and ‘other fermentation’ groups, reflecting either neutral
drift for these groups or that these strains were inappropriately grouped
together into somewhat amorphous categories.
Seventeen strains (including B1, I14, M22, M8, PR, RM11-1a, K1, K9, YJM308,
YJM789, YJM269, Y12, SB, Y1, Y10, YPS1009, and YPS163) were chosen for
whole-genome expression analysis. Cells were grown 2–3 doublings in
YPD medium to early log-phase in at least biological triplicate. Cell
collection, RNA isolation, and microarray labeling and scanning were done as
previously described [77], using cyanine dyes (Flownamics, Madison, WI)
and spotted DNA microarrays consisting of 70mer oligos representing each yeast
ORF (Qiagen). For all arrays, RNA collected from the denoted strain was compared
directly to that collected from the diploid S288c lab strain DBY8268, with
inverse dye labeling used in replicates to control for dye-specific effects. At
least three biological replicates were performed for all comparisons. Data were
filtered (retaining unflagged spots with R2>0.1) and
normalized by regional mean-centering [78]. Genes with
significant expression differences (compared to the S288c control, strain
YPS163, or the mean expression across all strains) were identified separately
for each strain with a paired t-test (or unpaired t-test in reference to YPS163)
using the BioConductor package Limma v. 2.9.8 [79] and FDR correction
[80], taking p<0.01 as significant unless
otherwise noted (see Table S3 for limma output and Figure S2
for a comparison of the statistical power for each strain). All microarray data
are available through the NIH Gene Expression Omnibus (GEO) database under
accession number GSE10269.
Array-based comparative genomic hybridization (aCGH) was performed in duplicate
on six strains (K9, M22, RM11-1a, Y10, YJM789, and YPS1009) relative to the
DBY8268 control as previously described [81], using amino-allyl
dUTP (Ambion), Klenow exo-polymerase (New England Biolabs), and random hexamers.
Post-synthesis coupling to cyanine dyes (Flownamics) was performed using inverse
dye labeling in replicate experiments. Technical variation in hybridization was
defined as the mean+2 standard deviations (a log2 value of 0.3) of all
spot ratios, based on triplicate comparisons of DBY8268 to DBY8268 genomic DNA.
For non-lab strains compared to DBY8268, genes with negative aCGH ratios outside
the range of technical variation on both duplicates were defined as those
affected by copy number and/or hybridization defects. Transcript levels within
0.45 (3 standard deviations of technical variation) of the aCGH ratio were
identified as those largely explained by copy number and/or hybridization
defects – on average, fewer than 5% of genes with
statistically significant (FDR = 0.01)
differential expression compared to DBY8268 fell into this class. Genes with a
positive aCGH ratio >0.7 in log2 space were defined as genes with
increased copy number in each non-lab strain. All microarray data are available
through the NIH Gene Expression Omnibus (GEO) database under accession number
GSE10269.
A vector of relative phenotype scores was generated by dividing scores from Figure 2 by the score measured
for DBY8268. The Pearson correlation between this vector and the measured
expression vector for each strain relative to DBY8268 was calculated for all
genes in the dataset. Genes whose expression was correlated above or below what
was expected by chance (p<0.01) were defined based on
100 permutations of each of the ∼6,000 expression vectors.
|
10.1371/journal.pbio.1000550 | Co-Evolution of Transcriptional Silencing Proteins and the DNA Elements Specifying Their Assembly | Co-evolution of transcriptional regulatory proteins and their sites of action has been often hypothesized but rarely demonstrated. Here we provide experimental evidence of such co-evolution in yeast silent chromatin, a finding that emerged from studies of hybrids formed between two closely related Saccharomyces species. A unidirectional silencing incompatibility between S. cerevisiae and S. bayanus led to a key discovery: asymmetrical complementation of divergent orthologs of the silent chromatin component Sir4. In S. cerevisiae/S. bayanus interspecies hybrids, ChIP-Seq analysis revealed a restriction against S. cerevisiae Sir4 associating with most S. bayanus silenced regions; in contrast, S. bayanus Sir4 associated with S. cerevisiae silenced loci to an even greater degree than did S. cerevisiae's own Sir4. Functional changes in silencer sequences paralleled changes in Sir4 sequence and a reduction in Sir1 family members in S. cerevisiae. Critically, species-specific silencing of the S. bayanus HMR locus could be reconstituted in S. cerevisiae by co-transfer of the S. bayanus Sir4 and Kos3 (the ancestral relative of Sir1) proteins. As Sir1/Kos3 and Sir4 bind conserved silencer-binding proteins, but not specific DNA sequences, these rapidly evolving proteins served to interpret differences in the two species' silencers presumably involving emergent features created by the regulatory proteins that bind sequences within silencers. The results presented here, and in particular the high resolution ChIP-Seq localization of the Sir4 protein, provided unanticipated insights into the mechanism of silent chromatin assembly in yeast.
| As eukaryotic species evolve, transcriptionally silent portions of their genomes—termed “heterochromatin”—mutate rapidly. To maintain the “off” state of certain genes in silenced regions, regulatory DNA sequences called silencers, which reside within a rapidly mutating region, must co-evolve with the regulatory proteins that bind these sequences to turn off transcription. Although hypothesized to occur widely in nature, such “molecular co-evolution” of genetic regulators has been demonstrated in only a few cases. Unlike previous examples of gene regulatory co-evolution, we found that the transcription factors that bind silencers in two budding yeast species are, in fact, functionally interchangeable, even though the silencers are not. Surprisingly, the Sir1 and Sir4 silencing proteins, which are heterochromatin components that bind the transcription factors rather than the silencer DNA sequences per se, are the proteins engaged in rapid co-evolution with the silencers. Silencer sequences therefore contain additional, evolutionarily labile information directing the assembly of heterochromatin. As mutations in Sir1 and Sir4 over evolutionary time can compensate for changes in the silencers, this “extra information” likely involves cooperative assembly of the transcription factors with the Sir1 and Sir4 “adaptor” proteins. The localization patterns of two species' Sir4 proteins across both species' genomes in interspecies yeast hybrids illuminate unexpected features of heterochromatin structure and assembly.
| Among all specialized chromatin structures, the difference between heterochromatin and euchromatin is perhaps the most fundamental, motivating intense study of the differences between these two structures. DNA sequences within heterochromatic regions evolve rapidly in animals [1],[2], plants [3],[4], and fungi [5], presenting a paradox of how the specification of heterochromatin structure persists despite rapid changes in the underlying sequence [6]. In Saccharomyces the biology of heterochromatin has proven eminently accessible to genetic studies through its role in gene silencing [7], and comparative studies of silencing now seem poised to illuminate key processes underlying heterochromatin evolution.
Molecular co-evolution of transcriptional regulatory proteins with their sites of action has been proposed to maintain regulatory functions across species divergence [8],[9]. In this context, “co-evolution” is typically understood as compensatory changes in a DNA sequence motif and the DNA-binding domain of the cognate transcription factor. Although it has been suggested that such co-evolution is prevalent in nature [8], in only a few instances has it been directly tested [10]–[12]. In Dipteran insects, for example, co-evolution of bicoid binding sites in the hunchback promoter and the bicoid homeodomain has been proposed to maintain hunchback-mediated developmental patterning along the anterior/posterior axis in Musca and Drosophila [13],[14]. However, the large size and complexity of animal regulatory elements, and the difficulty of performing cross-species complementation tests in animals, have precluded clear distinction between regulatory divergence and bona fide co-evolution.
Transcriptional silencing by Sir (Silent Information Regulator) proteins is necessary for the specialized haploid mating-type system found in Saccharomyces [15],[16]. DNA regulatory elements termed “silencers” contain binding sites for the Origin Recognition Complex (ORC), Rap1, and Abf1, which in turn direct the assembly of silent chromatin structures at the HML and HMR loci. The current model for the establishment of silencing holds that a Sir2/Sir3/Sir4 complex is brought to silencers by protein-protein interactions between ORC and Sir1, and between Rap1 and Sir4 [7]. Upon nucleation of these complexes, silent chromatin formation is catalyzed by the histone deacetylase activity of Sir2, and propagated, at least in part, through interactions between Sir3 and newly deacetylated histone tails [17]–[19]. Sir proteins are not thought to bind specific DNA sites; instead, efficient nucleation of silencing complexes at silencers requires interactions between Sir1 and Sir4, bridging the ORC-Sir1 and Rap1-Sir4 interactions [20]. Silencing also occurs at telomeres, which recruit Sir proteins primarily through arrays of Rap1 binding sites within the terminal repeats (TG1–3) [21]. However, more complex regulatory elements reside near the terminal repeats, and at some telomeres these may also serve to recruit Sir proteins [32],[23].
We have recently shown that silencer elements are among the most rapidly evolving regulatory sequences in Saccharomyces genomes [5]; however, the regulatory proteins that directly bind silencers are highly conserved, essential proteins. Intriguingly, the Sir1 and Sir4 proteins parallel the silencers in their rapid evolution, but these proteins show distinct patterns of evolution. SIR1-related genes have undergone multiple duplication and loss events: for example, S. bayanus has four functional paralogs of the single S. cerevisiae SIR1 gene, including the ancestral KOS3 (Kin Of Sir1) paralog, which S. cerevisiae has lost along with two other paralogs (Figure 1A) [24]. In contrast, the Sir4 protein is among the 40 most diverged proteins between S. cerevisiae and S. bayanus (Figure 1B), with 45% identity between its orthologs relative to a genome-wide average of 83% identity [25],[26]. Although SIR2 and SIR3 each have a paralog resulting from the whole-genome duplication (SIR2 has three additional, more ancient paralogs), neither gene has experienced subsequent duplication or loss events [27],[28].
S. cerevisiae and S. bayanus are post-zygotically isolated—haploids of these two species can mate to form mitotically stable hybrid diploids, but meiotic spores derived from these diploids are usually inviable [29],[30]. The rapid evolution of the silencers, the Sir4 protein sequence, and the elaboration of Sir1 paralogs make these two species an excellent phylogenetic context for comparative studies of silencing. Here, we describe functional studies in S. cerevisiae/S. bayanus interspecies hybrids that demonstrated how co-evolution among two heterochromatin proteins, Sir1 and Sir4, and multiple silencer DNA elements allowed two divergent lineages to maintain robust silencing despite these rapid genetic changes. This example of regulatory co-evolution is of particular interest because the co-evolving proteins are not the agents that directly bind to the divergent regulatory DNA sites.
In the course of a genetic screen for S. bayanus silencing mutants, we discovered that S. cerevisiae SIR4 failed to complement S. bayanus sir4Δ mutants for silencing of both HML and HMR, but S. cerevisiae SIR2 and SIR3 complemented mutations in S. bayanus orthologs (Figure S1; Zill et al. in preparation). This result was unanticipated as there are many cases of human proteins that can replace their yeast counterparts, even for proteins that function in large complexes and have considerably more sequence divergence than that seen between S. cerevisiae and S. bayanus proteins [31]–[34]. The incompatibility was unidirectional as S. bayanus SIR2, SIR3, and SIR4 complemented S. cerevisiae sir2Δ, sir3Δ, and sir4Δ, respectively.
Importantly, SIR4 functional divergence was due to one or more coding changes, as the level of expression of the two Sir4 orthologs, measured at either the RNA or protein level, was equivalent (Figure S2). To assay the function of both species' silencing machineries in the same cellular milieu, we developed a highly sensitive transcriptional reporter assay in S. cerevisiae/S. bayanus interspecies hybrid diploids that allowed us to monitor silencing of each species' HMR locus (hereafter referred to as Sc-HMR or Sb-HMR). The reporter consisted of the K. lactis URA3 open reading frame placed under the control of the endogenous HMRa1 promoter of each species, in two separate, but otherwise isogenic, hybrid strains (Figure 2A).
In these hybrids the S. cerevisiae SIR4 (Sc-SIR4) allele could not, on its own, silence Sb-HMR (Figure 2A, row 2). Reduced dosage of Sir4 per se did not cause loss of silencing at Sb-HMR, as S. bayanus diploids with only one copy of Sb-SIR4 showed no silencing defect (Figure 2B, row 2), nor did S. cerevisiae diploids with only one copy of Sc-SIR4 (unpublished data). Furthermore, a hybrid diploid containing two copies of Sc-SIR4 (the Sb-SIR4 gene was replaced by Sc-SIR4) also failed to silence Sb-HMR (Figure 2A, row 5). In contrast, one Sc-SIR4 gene was able to silence Sc-HMR in all hybrid strains tested (Figure 2A, bottom panel; Figure 3A). Thus, the hybrid cellular environment did not interfere with Sc-Sir4 function, and within a species, SIR4 was not haplo-insufficient. It appeared that Sc-Sir4 was either inhibited at Sb-HMR by something encoded by the S. bayanus genome specifically or somehow failed to interact with proteins that promoted Sb-HMR silencing.
Transcription analysis of a critical set of the hybrid strains showed good correspondence between expression of the HMR::URA3 reporter and growth patterns observed on FOA and CSM/-Ura media (Figure 3B).
We note that in the interspecies hybrids with both species' SIR4 alleles (Sc-SIR4/Sb-SIR4), Sb-HMR silencing appeared weakly defective relative to the complete silencing of Sb-HMR in S. bayanus diploids by both the reporter assay and direct RNA measurement (compare Figure 2A, row 1 with Figure 2B, row 1; Figure 3B). In addition, Sb-HMR silencing was further weakened in hybrids lacking Sc-Sir4 (Figure 2A, compare row 3 with row 1). This result was paradoxical because Sc-Sir4 appeared to have very little ability to silence Sb-HMR in hybrids lacking Sb-Sir4. As explained below, these weak Sb-HMR silencing defects were likely due to an emergent property of the hybrids, resulting from unusually strong interactions between Sb-Sir4 and S. cerevisiae silent loci that effectively reduced Sb-Sir4 associations with Sb-HMR. The presence of Sc-Sir4 limited the competition for Sb-Sir4.
The inability of Sc-Sir4 to function at Sb-HML and Sb-HMR could have been manifested either during its recruitment or after its assembly into chromatin [35]. To determine where in the assembly of S. bayanus silenced chromatin Sc-Sir4 protein was blocked, we compared the ability of Sc-Sir4 and Sb-Sir4 proteins to associate with all silent loci of both species at high resolution using chromatin-immunoprecipitation followed by deep-sequencing of the precipitate (ChIP-Seq). Sir4 ChIP-Seq was performed using hybrid diploids expressing Sc-Sir4 only, Sb-Sir4 only, or both Sc-Sir4 and Sb-Sir4. Because of the sequence divergence between HML and HMR of the two species, the occupancy of each species' HML and HMR loci could be evaluated simultaneously. In each strain, only one SIR4 allele carried a 13xMyc epitope tag [36]. In hybrids expressing Sc-Sir4 only, robust enrichment of Sc-HML and Sc-HMR silencers was observed as expected, with very weak enrichment of Sb-HML and Sb-HMR silencers (Figure 4A, Table 1). Strikingly, Sc-Sir4 association with an internal region of Sb-HMR was indistinguishable from non-silenced regions. In contrast, as predicted from the genetic results, Sb-Sir4 associated robustly with HML and HMR loci from both species, and did so most robustly at S. cerevisiae silencers (Figure 4A, Table 1). The ChIP-Seq results were validated at Sc-HMR, Sb-HMR, and control loci using standard ChIP-qPCR analysis (Figure 5A). Thus, Sc-Sir4 showed strongly reduced association with Sb-HML and Sb-HMR silencers and no detectable association with their internal regions. The relative absence of Sc-Sir4 from these normally silenced regions of the S. bayanus genome was consistent with two possibilities. Perhaps Sc-Sir4 could not interact properly with Rap1, ORC, or the S. bayanus Sir1 paralogs assembling on their silencers, or perhaps an S. bayanus protein was preventing stable association between Sc-Sir4 and S. bayanus silencers.
The comparative Sir4 ChIP-Seq data provided a surprising insight into the mechanism of Sir4 incorporation into silent chromatin. Although Sc-Sir4 binding to Sb-HML and Sb-HMR loci was barely detectable in hybrids expressing Sc-Sir4 only (Figure 4A, center and right panels; Table 1), in hybrids expressing both Sc-Sir4 and Sb-Sir4, Sc-Sir4 binding increased substantially at Sb-HML and Sb-HMR silencers and internal regions (Figure 4A; Table 1). Thus, despite the poor ability of Sc-Sir4 to associate with Sb-HML and Sb-HMR on its own, Sb-Sir4 somehow provided Sc-Sir4 access to them. It appeared that Sir4 association with S. bayanus HML and HMR involved two distinguishable modes of interaction, but Sc-Sir4 was capable of only one (Sb-Sir4-dependent). Moreover, the divergent mode was apparently critical only for the initial association of Sir4 with a silencer, and not for subsequent associations with the silenced region. The ability of Sc-Sir4 to form dimers suggested one straightforward explanation for the Sb-Sir4-dependent chromatin association: inter-specific Sc-Sir4/Sb-Sir4 dimerization through a conserved coiled-coil domain [37],[38].
Sb-Sir4-assisted incorporation of Sc-Sir4 into Sb-HML and Sb-HMR was consistent with Sc-Sir4 contributing to silencing at these loci, as suggested by the decreased Sb-HMR silencing in hybrids lacking Sc-Sir4 (Figure 2A, row 3). However, this hypothesis per se could not explain the sensitivity of Sb-HMR silencing to reduced Sb-SIR4 dosage that was observed in interspecies hybrids, but not in S. bayanus diploids (compare Figure 2A, rows 1 and 3, with Figure 2B, row 2; Figure 3B). Further analysis of Sir4 localization on the S. cerevisiae/S. bayanus hybrid genome by ChIP-Seq provided an explanation of this hybrid-specific Sb-SIR4 dosage sensitivity, as described next.
Given the differential association of Sc-Sir4 and Sb-Sir4 with the two species' HML and HMR loci, we asked if any other loci, genome-wide, also showed a dramatic discrepancy. In S. cerevisiae, silencing by Sir proteins occurs at telomeres and subtelomeres, in addition to HML and HMR [18],[39],[40]. A comparison of the interspecies hybrids expressing Sc-Sir4 only versus Sb-Sir4 only showed that all S. cerevisiae TG1–3 terminal repeats (which contain embedded Rap1 binding sites), including those present on the centromere-proximal side of some Y′ elements, were comparably occupied by both species' Sir4 proteins (Figure 4B). (Y′ elements are helicase-encoding repetitive sequences of unknown origin and function that are found in some subtelomeric regions immediately adjacent to the terminal repeats [22].) This result was not surprising as the telomerase-replicated repeated sequence, templated by the TLC1 RNA, is identical in the two species (our unpublished observations). Thus, it appeared that Sir4 association with the S. cerevisiae genome, as promoted by Rap1, was not substantially different between Sc-Sir4 and Sb-Sir4. Indeed, the C-terminal residues of Sc-Sir4 critical for its interaction with Rap1 are conserved in Sb-Sir4 (our unpublished observations). We note that the smaller ChIP-Seq peaks observed in these regions in the “No tag” control strain (Figure 4B, yellow shading) are likely due to non-specific DNA binding to the anti-myc beads.
Unexpectedly, S. cerevisiae subtelomeres had two types of regions notably more enriched by Sb-Sir4 ChIP than by Sc-Sir4 ChIP. These regions corresponded to X elements, which are regulatory sequences near telomere ends that contain ORC and Abf1 binding sites [22], and the ORFs within Y′ elements. For X elements, ChIP-Seq of Sc-Sir4 showed an average of 7-fold enrichment, whereas Sb-Sir4 showed an average of 14-fold enrichment, with even greater disparity often evident immediately adjacent to X elements (Figure 4B). Therefore, Sb-Sir4 either associated more robustly with factors bound to X elements than did Sc-Sir4 or conceivably was excluded less effectively. X element core sequences (containing the ORC and Abf1 binding sites) are bordered on the telomere-proximal side by X-element combinatorial repeats (formerly known as subtelomeric repeats or STRs; [22]) and the terminal repeats (see http://www.yeastgenome.org/images/yeastendsfigure.html for schematics depicting X-only and X-Y′ telomere ends). The differential pattern of Sir4 association with X elements was consistent with Sb-Sir4 associating more robustly than Sc-Sir4 with sequences at, and immediately adjacent to, the ORC binding sites, presumably via ORC-mediated interactions (Figure 4B). Other S. bayanus proteins produced in the hybrids, such as the Sir1 paralogs, may contribute to the enhanced association of Sb-Sir4 with X elements, as discussed below.
We observed weak Sc-Sir4 association with Y′ elements despite its strong association with neighboring terminal repeats (Figure 4B, right panel), consistent with earlier observations using ChIP-chip and transcription reporter analyses [23],[41]. Surprisingly, Sb-Sir4 associated considerably better than Sc-Sir4 with all Y′ elements, showing an average of 5-fold enrichment across their coding regions by Sb-Sir4 ChIP versus 1.2-fold enrichment by Sc-Sir4 ChIP. We note that the S. bayanus genome lacks Y′ elements, and thus S. bayanus subtelomeres may have reduced Sir4 recruitment potential relative to S. cerevisiae subtelomeres [42],[43]. Thus, the enhanced associations of the Sb-Sir4 protein with X and Y′ elements suggested that, in the hybrid strains, S. cerevisiae telomeres might have competed with Sb-Sir4 association with Sb-HML and Sb-HMR, leading to the somewhat weakened Sb-HMR silencing observed in hybrids with only one copy of Sb-Sir4 (Figure 2A, rows 1 and 3; Figure 3B). Sb-Sir4 association was indeed reduced at Sb-HMR and Sb-HML silencers in a hybrid expressing only one copy of Sb-SIR4, relative to a hybrid with two copies of Sb-SIR4 (Table 1, compare columns 2 and 4). Thus, Sc-Sir4 may have, in effect, protected Sb-HMR silencing in hybrids when Sb-Sir4 was present (Figure 2A, compare rows 1 and 3) by occupying sites at S. cerevisiae telomeres that would otherwise have been bound by Sb-Sir4. (Although the S. cerevisiae Y′ elements are bound by Sb-Sir4 and not by Sc-Sir4 in cells with only a single species' Sir4, in the Sc-SIR4/Sb-SIR4 hybrids, Sc-Sir4 and Sb-Sir4 both occupy Y′ elements (unpublished data). However, the extent of occupancy by Sb-Sir4 is less than in cells with Sb-Sir4 only, consistent with Sc-Sir4's ability to spare Sb-Sir4 binding to Y′ elements in the hybrids.)
The ChIP-Seq data allowed us to determine whether the species restriction to Sc-Sir4 association, evident at Sb-HML and HMR, also applied to S. bayanus telomeres. Although subtelomeric regions of the S. bayanus genome are presently incompletely assembled and annotated (see Saccharomyces Genome Database, www.yeastgenome.org), we identified several candidate subtelomeric contigs based on homology to S. cerevisiae subtelomeric genes and X elements. Contigs from the S. bayanus genome assembly that contained regions bound by both Sc-Sir4 and Sb-Sir4 (as determined by peak-calling software, see Methods) and putative subtelomeric sequence were further examined for Sir4 ChIP enrichment (an example is shown in Figure 5B). Sb-Sir4 associated with one end of each of these contigs and usually with an internal region as well, typically within 10 kb of the contig end. Interestingly, in the Sc-Sir4-only hybrid, Sc-Sir4 association was observed at the contigs' ends, but not at the internal regions that bound Sb-Sir4. This result suggested that Sc-Sir4, even in the absence of Sb-Sir4, was capable of associating with S. bayanus telomere ends, presumably via the conserved Rap1 protein, but could not make some additional contacts necessary to associate with internal sequences.
To test whether the Sc-Sir4 molecules bound to S. bayanus telomeres were capable of silencing S. bayanus subtelomeric genes, we measured the transcription of candidate subtelomeric ORFs in S. bayanus wild-type, Sb-sir4Δ::Sc-SIR4, and Sb-sir4Δ strains. Importantly, the expression of all three putative subtelomeric genes increased in Sb-sir4Δ cells (Figure 5C). Although Sc-Sir4 was capable of silencing Sb-YIR039c and an ORF located on Contig_626, it could not repress the transcription of an ORF on Contig_511 located almost 9 kb from the main peak of Sc-Sir4 ChIP. Thus, Sc-Sir4 could bind to and silence at least a subset of S. bayanus telomeric regions. It was possible that S. bayanus had subtelomeric regulatory elements that promoted silencing, in addition to the Rap1-binding terminal repeats. Depending on the sequence of a particular element, or its proximity to the telomere end, Sc-Sir4 may or may not have been capable of binding and silencing.
The cross-species complementation and ChIP analyses suggested that the incompatibility between Sc-SIR4 and Sb-HML and HMR was caused by the failure of one or more physical interactions occurring at S. bayanus silencers. In principle, the lack of productive Sc-Sir4 association with Sb-HML and Sb-HMR could have resulted either from an S. bayanus-specific inhibitor of silencing that Sc-Sir4 could not overcome or an S. bayanus-specific positive regulator of silencing (e.g., Sb-Rap1 or Sb-Sir1) with which Sc-Sir4 could not interact. To distinguish between these models, in an S. cerevisiae strain, we replaced the Sc-HMR locus with Sb-HMR containing the URA3 reporter, including the flanking silencer elements (Figure 6A). If S. bayanus encoded an inhibitor of silencing that Sc-Sir4 could not overcome, Sb-HMR should be silenced in S. cerevisiae, given the strong conservation of ORC, Rap1, and Abf1 proteins and the Rap1 and Abf1-binding sites in the HMR-E silencer [5]. If, however, Sc-Sir4 failed to be recruited to S. bayanus silencers, we would expect little or no silencing of Sb-HMR in S. cerevisiae.
Upon transfer into S. cerevisiae, Sb-HMR was silenced extremely poorly (Figure 6B, row 1). However, the transplanted Sb-HMR locus could still be silenced in the context of the S. cerevisiae chromosome in hybrids made by mating the S. cerevisiae Sb-HMR strain to wild-type S. bayanus. The transplanted Sb-HMR locus was silenced to approximately the same degree as the native Sb-HMR locus in hybrids (Figure 6B, row 2, compare with 6C rows 1 and 2). The slightly incomplete silencing of the transplanted Sb-HMR was largely due to the Sb-SIR4 dosage sensitivity observed in the original set of hybrids (Figure 2A, row 3), as silencing was strengthened in Sb-SIR4/Sb-SIR4 hybrids (Figure 6B, row 3). Thus, the lack of silencing of Sb-HMR in hybrids expressing only Sc-Sir4 (Figure 2A, rows 2 and 5) was not due to an inhibitor of silencing encoded elsewhere in the S. bayanus genome. Rather, the incompatibility was encoded in the Sb-HMR locus itself, requiring S. bayanus-specific silencing proteins to interpret Sb-HMR-specific sequence information. These “interpreter” proteins potentially included DNA-binding proteins such as ORC, Rap1, and Abf1 or proteins indirectly associated with silencers, such as Sir1, Sir4, or both.
Alignments of Sc-HMR and Sb-HMR suggested that their functional divergence was due to changes in the silencer sequences between the two species (Figure 7). The HMRa1 gene was 83% identical between S. cerevisiae and S. bayanus (the promoter was 93% identical), well above the genome-wide average of 62% identity for all intergenic regions, and the mating-type cassette-homology sequences (shared with MAT and HML) approached 100% identity (Figures 6A and 7). Notably, the silencer sequences share well below the genome-wide average identity for intergenic regions and are difficult to align outside of the conserved Rap1 and Abf1 sites [5].
The simplest model consistent with the results so far was that the silencing incompatibility was limited to Sir4, with Sc-Sir4 having a more restricted range of interactions than Sb-Sir4. To test this possibility, we replaced Sc-SIR4 with Sb-SIR4 in the S. cerevisiae strain bearing Sb-HMR. If the incompatibility involved only SIR4 and silencers, Sb-SIR4 should restore silencing to Sb-HMR. Indeed, the S. cerevisiae strain with Sb-SIR4 and Sb-HMR indeed showed a modest increase in silencing relative to the Sc-SIR4 Sb-HMR strain, confirming that changes in Sir4 itself contributed to the silencing incompatibility. However, this silencing increase—a 5-fold change—was detectable only as an increase in FOA resistance, and was still at least 100-fold below the level of HMR silencing seen in the hybrids (Figure 8A, row 2; compare with Figure 6B, row 2). Thus, although a portion of the incompatibility could be explained by SIR4 and silencer co-evolution, one or more additional S. bayanus proteins were likely required to recruit Sb-Sir4 efficiently or to stabilize its association with S. bayanus silencers.
Interestingly, Sc-Sir4's very weak ability to silence the transplanted Sb-HMR locus resulted in the low-frequency appearance of FOA-resistant colonies (occurring at an approximate frequency of 5×10−5; Figure 8A, row 1). Within these colonies, which grew at nearly the rate expected of Ura- strains, the cells were able to grow under conditions that killed the majority of cells that did not form colonies. Hence this silencing occurred at low frequency, but was nonetheless heritable. Indeed, Sb-HMR silencing by either Sb-Sir4 or Sc-Sir4 was fully dependent on S. cerevisiae Sir1 (Figure 8A), whose role is to promote the establishment of heritable silencing. That Sb-HMR could be silenced at all in S. cerevisiae suggested that a critical subset of Sc-ORC, Rap1, and Abf1 bound productively to Sb-HMR silencers. It was therefore possible that providing additional S. bayanus silencing proteins could stabilize interactions between the S. cerevisiae DNA-binding proteins and S. bayanus silencers. Likely candidates to provide this presumptive function were the S. bayanus Sir1 paralogs, Kos1, Kos2, and Kos3, with Kos3 being the most structurally distinct from Sir1, yet the most similar to the ancestral member of the Sir1 family [24]. Interestingly, Sb-KOS3 enhanced Sb-HMR silencing synergistically with Sb-SIR4, but not with Sc-SIR4 (Figure 8B; compare rows 1, 5, and 10). None of the other Sir1 paralogs of S. bayanus provided a dramatic enhancement of Sb-HMR silencing. The Sb-HMR Sb-SIR4 +Sb-KOS3 strain showed 100-fold better silencing than the Sb-HMR Sc-SIR4 strain (Figure 8B, compare rows 1 and 10). This result was particularly interesting because Sir4 interacts weakly and non-specifically with DNA [44], and Kos3 is not thought to bind DNA at all. Thus, the “interpretation” of differences between the Sb-HMR and Sc-HMR silencers by Sb-Kos3 and Sb-Sir4 presumably required some sort of HMR-allele-specific collaboration with silencer-binding proteins that could be interpreted by Sb-Kos3 and Sb-Sir4 in a species-specific way.
By sequence conservation, Rap1 and Abf1 binding sites can be detected in the Sb-HMR-E silencer, but the ORC binding site is not readily identified (Figure 7) [5]. Given Sb-Sir4's dependence on Sir1 and Kos3, and their dependence on ORC [20],[45],[46], our results suggested two likely explanations for why Sb-HMR was not silenced in S. cerevisiae: either Sc-ORC bound S. bayanus silencers less well than S. cerevisiae silencers, or Sc-ORC bound equivalently but failed to promote silencing because it was in a suboptimal conformation or context with respect to other silencer binding proteins. In either case, the subsequent interactions with Sc-Sir1 and Sc-Sir4 might suffer. To test whether Sc-ORC indeed bound to S. bayanus silencers, we performed ChIP analysis on HA-tagged Sc-Orc5 in S. cerevisiae bearing Sb-HMR. Sc-Orc5 associated with the Sb-HMR-E silencer, albeit at a level several-fold below its association with Sc-HMR-E (Figure 9A, left panel; note log scale on y-axis). A parallel analysis with Sc-Abf1 ChIP showed robust association of this protein with both Sc-HMR-E and Sb-HMR-E silencers (Figure 9A, right panel). We note that both Sc-Orc5 and Sc-Abf1 associations with Sb-HMR-E showed small alterations in the Sb-SIR4 strain relative to the Sc-SIR4 strain. However, these changes did not correlate with Sb-HMR silencing levels (Figure 8A, rows 1 and 2). These ChIP data were consistent with Sb-HMR silencers having conserved functional binding sites for ORC and Abf1.
To test whether Sc-ORC, Rap1, and Abf1 indeed participated in S. bayanus silencing, we monitored silencing of Sb-HMR in hybrids lacking either species' complement of each of these proteins (out of the six ORC subunits, we focused on Orc1 because it directly interacts with Sir1). Because RAP1, ABF1, and ORC1 are essential, we assayed silencing in hybrids heterozygous for each gene. S. cerevisiae diploids sensitized to detect silencing defects at HMR show strong silencing defects if either SIR1 or SIR4 dosage is also reduced [47]. Similarly, Sb-HMR silencing was weakly compromised in hybrids whereas Sc-HMR was not (Figure 2A), potentially providing a sensitized background to uncover similar types of genetic interactions. For this reason, any such silencing defects in heterozygous hybrids were expected to affect silencing of Sb-HMR but not Sc-HMR. Indeed, Sb-HMR, but not Sc-HMR, was further derepressed in hybrids lacking either Sc-RAP1 or Sb-RAP1 (Figure 9B, top panel; Figure S3). Note that Sb-HMR was fully silenced in S. bayanus RAP1/rap1Δ diploids; therefore, reduced RAP1 dosage per se did not cause the loss of silencing observed in the hybrid (Figure 9B, bottom panel). Thus, Sc-Rap1 participated in Sb-HMR silencing in hybrids, likely by direct binding to S. bayanus silencers. In contrast to the analysis with RAP1, Sb-HMR was derepressed to a greater extent in hybrids lacking Sb-ORC1 but not in hybrids lacking Sc-ORC1 (Figure 9B, top panel). Again, Sb-HMR was fully silenced in S. bayanus ORC1/orc1Δ diploids (Figure 9B, bottom panel), ruling out simple dosage explanations. Hence, Sb-Orc1 was more important for Sb-HMR silencing in hybrids than Sc-Orc1, suggesting that a partial species restriction existed with respect to ORC binding or activity at Sb-HMR silencers. Heterozygosity of ABF1 had no effect on either Sb-HMR or Sc-HMR silencing (Figure 9B, top panel; Figure S3).
The rapid sequence and functional divergence of SIR4 between closely related species suggested that an interesting evolutionary force may have contributed to the functional divergence of this gene. To test whether a specific function of the Sir4 protein had been under positive selection within the sensu stricto clade, we aligned SIR4 coding sequence from all five species and computed the ratio of nonsynonymous to synonymous divergence (henceforth ω, also known as dN/dS) across the whole gene. The value of ω for SIR4 was 0.44, substantially higher than the genomic average of 0.10. Only 16 of 4,894 loci we analyzed had a higher ω, indicating that SIR4 was indeed one of the most rapidly evolving genes in the budding yeast genome.
A value of ω significantly greater than 1 is evidence of positive selection [48]. Therefore, a value of 0.44 might suggest that the SIR4 coding region did not evolve under positive selection. However, because Sir4 is a large protein we investigated whether sub-regions or individual codons might have ω>1. To determine whether rapidly evolving Sir4 residues might lie within known functional regions of the protein, we computed ω in 102 bp (34-codon) windows throughout the SIR4 open reading frame (Figure 10A). Consistent with our previous whole-gene estimate, the median ω value for all windows in SIR4 was 0.43 (Figure 10A, solid horizontal line) with a range from 0.02 to 1.87. Because ω estimates calculated in short windows are subject to stochastic noise, we compared the results of this analysis to ∼1,500, 102 bp windows drawn from other S. cerevisiae coding regions. The median of these ω values was 0.05, and 95% of windows lie between 0.0001 and 0.42 (Figure 10A, dashed lines). These comparisons supported two conclusions. First, because the median ω for SIR4 was comparable to the most extreme values in other genes, the unusual molecular evolution of this gene extended over a large fraction of its length. Second, the non-random distribution of windows with high ω suggested that the rapid evolution of certain residues was connected to functional changes within specific regions of the Sir4 protein. In support of this suggestion, simulations of SIR4 evolution indicated that 1.3% of 102 bp windows are expected to have ω>1 by chance compared to 6.7% observed in sensu stricto SIR4 (unpublished data). The high-ω windows in SIR4 were therefore unlikely to reflect noise and instead indicated that the most rapidly evolving codons are concentrated in particular regions of SIR4. Indeed, although the Rap1- and Sir3-binding coiled-coil domain was largely protected from the rapid evolution of SIR4, residues within the PAD (Partioning and Anchoring of plasmids) and the putative N-terminal regulatory domains [37] showed striking signatures of rapid evolution (Figure 10A).
To provide an independent, statistically robust analysis of SIR4 evolution in this clade, we used a likelihood-ratio test to compare nested models of sequence evolution that either allowed or did not allow a subset of codons to have a value of ω>1. The model allowing ω>1 (M8) fit the data significantly better than the alternative model (M7; p = 5×10−4), indicating that some codons were likely to be evolving under positive selection (Figure 10B). The posterior probability of ω≥1 exceeded 0.75 for 11 codons (Figure 10A and C; ω≥1.5 in all cases), however for no single codon did the posterior probability exceed the nominal significance level of 0.95. Inclusion of SIR4 sequences from species outside the sensu stricto was not possible because of poor alignment quality. In summary, although we were not able to identify specific codons that were unambiguously under positive selection, these data suggested that multiple codons within SIR4, including some within the PAD and N-terminal regulatory domains, exhibit signatures of extremely rapid sequence evolution in the Saccharomyces sensu stricto clade.
To examine whether the rapid sequence evolution of SIR4 showed a phylogenetic correlation with the functional divergence we observed, we fit models that allowed different branches of the SIR4 tree to have different values of ω. If such a correlation were observed (i.e., more rapid evolution of SIR4 along the S. cerevisiae lineage than along the S. bayanus lineage), then positive selection on a specific silencing function of SIR4 during the evolution of S. cerevisiae would be likely. Although increased estimates of ω were obtained for some branches (notably the shared S. cerevisiae/S. paradoxus branch; ω = 0.55), none were statistically supported, suggesting that there have been no dramatic shifts in the selection pressures operating on SIR4 since the divergence of the sensu stricto (Figure S4). We note that a change in selection pressure that affected only a subset of codons could easily have gone undetected.
Using interspecies hybrids, we have shown by three functional criteria—cross-species complementation assays, cross-species cis-trans tests, and genome-wide localization by ChIP-Seq—that the functions of both the Sir4 protein and multiple silencer elements have strikingly diverged over the short divergence time between closely related yeast species. Cross-species complementation assays revealed an incompatibility between Sc-Sir4 and Sb-HML and Sb-HMR (Figure 2A). The inability of Sc-Sir4 to silence Sb-HML and Sb-HMR was due to a difference in the protein sequence of Sir4 between the two species rather than a difference in expression level (Figure 10, Figure S2). This incompatibility likely resulted from the coordinated divergence of multiple heterochromatin determinants: Sir1, Sir4, and silencers. Two pieces of evidence implicated cis-acting changes in silencer sequences as being key to the incompatibility. First, comparative ChIP-Seq analysis of Sir4 pinpointed an inability of Sc-Sir4 to associate stably with S. bayanus silencers (Figure 4A, Table 1). Second, and more definitively, transfer of the Sb-HMR locus into S. cerevisiae demonstrated that this locus was inherently unrecognizable to Sc-Sir4 (Figure 6B). This result established that S. bayanus did not produce an inhibitor of Sc-Sir4 function, and mapped the locus of the Sir4 species specificity to the Sb-HMR silencers. As silencing of the transplanted Sb-HMR locus was largely restored in an S. cerevisiae/S. bayanus hybrid (Figure 6B), S. bayanus-specific proteins were required to assemble silent chromatin at Sb-HMR in the manner dictated by the Sb-HMR silencers, with Sb-Sir4 and the Sb-Sir1 paralogs being the most likely candidates for species-specific “interpreter” proteins.
The Sir4 protein and silencers diverged rapidly in concert, a process that was accompanied by loss of three Sir1 paralogs in the S. cerevisiae lineage [24]. As silencing was robustly maintained in each species, it was likely that these factors had co-evolved such that coding changes in Sir4 and a reduction in Sir1 family members led to compensatory changes in silencers, or vice versa. The asymmetrical complementation of SIR4 alleles (Figure 2A), and the enhanced ability of Sb-Sir4 to bind S. cerevisiae silent loci compared to its own silent loci (Figure 4B), suggested that S. cerevisiae silencer elements had become stronger than those of S. bayanus, while S. cerevisiae Sir1 and Sir4 proteins had become weaker (operationally defined) than S. bayanus Sir4 and its four Sir1 paralogs. The intra-species combinations of Sir1 and Sir4 proteins and silencers allowed efficient nucleation of silencing complexes at HML and HMR in each species.
Broadly speaking, we imagine two possible evolutionary paths for this co-evolution, with variations on either path possible. In an “adaptive” model, hypothetical selective pressure(s) induced coding changes in Sir4 and reduction in Sir1 family members (Zill et al. in preparation), which then required “strengthening” mutations (for example, a change that increased the affinity of ORC for a silencer) in the silencers to maintain robust silencing. In a “constructive neutral” model [49], strengthening mutations accumulated in silencers at random, thus relaxing the selective constraints to maintain Sir1 paralogs and certain Sir4 residues. Once Sir1 paralogs were lost, the “stronger” silencers would need to be maintained by purifying selection. Our evolutionary analyses using the PAML software supported a role for positive selection acting on multiple sites within SIR4 during the evolution of the sensu stricto species (Figure 10). However, the SIR4 gene showed an unusually high evolutionary rate across most of its length. Furthermore, although we identified 11 rapidly evolving residues that may suggest specific regions' contributions to the functional divergence of Sir4, none of these crossed the 0.95 threshold for statistical significance. (We note that 4 of the 11 fastest evolving sites were localized within the PAD domain, which mediates interactions between Sir4 and Esc1 at the nuclear periphery, and between Sir4 and the Ty5 retrotransposon integrase protein. Interestingly, Esc1 is also one of the most rapidly diverging proteins in Saccharomyces species [O. Zill, unpublished observations].) Extensive directed mutational analyses will be necessary to test whether the sites under selection are responsible for the functional divergence between Sc-Sir4 and Sb-Sir4.
An important question relevant to these models is, in which lineage did the observed changes in Sir4 and silencer function occur relative to the common ancestor of S. cerevisiae and S. bayanus? Although accurate determination of the ancestral state of the silencing mechanism will require extensive evolutionary analyses, it appears that S. bayanus has retained at least two ancestral characters that S. cerevisiae has lost. First, Kos3, the ancestral Sir1-related protein, has been lost in S. cerevisiae. Second, the SIR4 gene from K. lactis, an outgroup to the Saccharomyces clade, was able to complement silencing function in S. bayanus sir4Δ mutants (Zill et al. in preparation). That a Sir4 protein from a species outside of Saccharomyces is compatible with S. bayanus silencers suggests that these elements did not “gain” a restrictive property in the S. bayanus lineage. The more likely scenario is that Sir4 changed in the S. cerevisiae lineage such that its range of interactions with other species' silencers has become restricted, consistent with earlier observations of cross-species function of Sir4 [50]. It will therefore be of interest to understand in detail the mechanism of silencing in S. bayanus and to determine what forces caused the dramatic shift in Sir1 and Sir4 functionality in the S. cerevisiae lineage. We measured the rates of SIR4 evolution (ω) along all branches in the sensu stricto clade but did not observe a notable asymmetry in these rates (Figure S4). Thus, the functional asymmetry between Sc-Sir4 and Sb-Sir4 was probably localized to a few sites and may not be related to the broad evolutionary forces that have acted on SIR4 across all five species in this clade.
Perhaps the most striking finding of this study was that the heterochromatin proteins that showed the most dramatic evidence of co-evolution with silencers, Sir1 and Sir4, were not the ones that bind specific DNA sites. Rather, these proteins associate with DNA indirectly via the conserved regulatory proteins Rap1, Abf1, and ORC (Figure 11). The key evidence demonstrating functional co-evolution between Sir4 and the Sir1 family and silencers came from attempts to reconstitute Sb-HMR silencing in S. cerevisiae. The changes in Sir4 sequence were not sufficient to explain the inability of Sc-Sir4 to function at S. bayanus silencers: expression of Sb-Sir4 in an S. cerevisiae strain was only modestly effective in silencing an Sb-HMR locus transplanted into that strain (Figure 8A). The Sir1-dependence of the rare, but heritable, silencing events mediated by Sb-Sir4 at Sb-HMR in S. cerevisiae suggested that the limitation involved proteins dedicated to establishing silencing. Indeed, adding Sb-Kos3, the ancestral member of the Sir1 family, together with Sb-Sir4 enhanced silencing of Sb-HMR in S. cerevisiae by 100-fold (Figure 8B), although not completely. It was possible that the site-specific DNA-binding proteins ORC, Rap1, and Abf1 had also co-evolved with silencer sequences. If this were the case, we would expect hybrids lacking the Sb-ORC, Sb-Rap1, or Sb-Abf1 proteins to have shown defective Sb-HMR silencing. However, only Sb-Orc1 inactivation (and by inference, inactivation of the entire Sb-ORC complex) showed the expected S. bayanus allele-specific effect on Sb-HMR silencing (Figure 9B). This effect of Sb-ORC1 deletion on Sb-HMR silencing was relatively modest, and the addition of Sb-ORC1 (together with Sb-SIR4) had no effect on Sb-HMR silencing in S. cerevisiae reconstitution experiments (unpublished data). Because Sc-Orc1, Sc-Rap1, and Sc-Abf1 were capable of supporting Sb-HMR silencing in hybrids (Figure 9B) and in S. cerevisiae (Figure 8B), their DNA-binding domains' interactions with silencers were largely conserved across species and hence were not engaged in notable co-evolution with silencers or with Sir4. Indeed, we were able to ChIP Sc-Orc5 and Sc-Abf1 on the Sb-HMR-E silencer in S. cerevisiae (Figure 9A). Together, these results suggested that the cis-acting differences between the two species' silencers were interpreted largely indirectly, via interactions between ORC, Sir1/Kos3, and Sir4, with a somewhat lesser contribution of differences in ORC-silencer DNA interactions.
Why did Sc-Sir4 not bind efficiently to S. bayanus silencers? Simple explanations such as sequence divergence between S. cerevisiae and S. bayanus silencers precluding sequence-specific contacts with Sir4 are unlikely because biochemical data on Sir4 point to a lack of sequence-specific binding to DNA [44]. Instead, Sir4 is recruited to silencers predominantly via protein-protein interactions [20],[37],[51]. It is unlikely that different proteins bind the silencers in the two species as the preponderance of evidence points to ORC, Rap1, and Abf1 as the critical silencer-binding proteins in both species (Figures 5, 7, and 9). Further, the residues mediating Sc-Orc1 interaction with Sc-Sir1 [52],[53] are conserved in Sb-Orc1 (our unpublished observations). Hence we are forced to consider models in which something special about how ORC, Rap1, and Abf1 bind S. bayanus silencers prevents Sc-Sir4 from interacting with Rap1 or creates a requirement for specific interactions with the Sir1 paralogs that can only be made by Sb-Sir4. Perhaps the precise juxtaposition or conformation of these site-specific DNA-binding proteins allow or restrict interactions with a particular species of Sir4. Alternatively, perhaps a reduced affinity of S. bayanus silencers for ORC or Rap1, or the ensemble of nucleation proteins, is compensated by binding energy provided by Sb-Kos3 (and possibly additional Sir1 paralogs) and Sb-Sir4, but not by Sc-Sir4. Indeed, complete silencing of both Sb-HML and Sb-HMR requires Sb-Sir1, Sb-Kos1, and Sb-Kos2 [24]. Therefore, it is possible that a relatively weak binding site (such as for Rap1) in the Sb-HMR-E silencer could be compensated by increased binding energy provided to the nucleation complex in trans by the combination of Sb-Sir4 and the Sir1 paralogs. Additionally, we note that a requirement for multivalent interactions may help explain why Sir4 fails to interact stably with Rap1 at the many Rap1 binding sites throughout the genome.
An unexpected finding of the Sir4 comparative ChIP-Seq experiment provided insight into the mechanism of silent chromatin assembly. The Sb-Sir4-assisted Sc-Sir4 incorporation into Sb-HML and HMR (Figure 4A) suggested two distinct types of interactions made by Sir4 proteins at these loci: only Sb-Sir4 was capable of making stable contacts either with the Sir1 paralogs, or perhaps with Rap1. However, as Sc-Sir4 was capable of mediating telomeric silencing at some S. bayanus telomeres (Figure 5C), it appeared that Sc-Sir4 could interact productively with the Sb-Rap1 protein. In addition, there was a second and qualitatively distinct mode of Sir4 protein association that was species-independent but occurred only if the species-specific interaction occurred. Three types of interactions might account for the secondary mode of Sc-Sir4 association with Sb-HML and Sb-HMR: direct Sb-Sir4-Sc-Sir4 interaction via a conserved dimerization surface [38], Sc-Sir4 binding to Rap1 via the conserved C-terminal coiled-coil domain, or Sc-Sir4 interaction with deacetylated histone tails [18]. We note that Sc-Sir4 association with the Sb-HMR-E silencer increased in the presence of Sb-Sir4 at least as much as did its association with internal regions of Sb-HMR (Figure 4A). Thus, this secondary mode of Sir4 interaction did not appear to be restricted to regions of Sb-HMR where the deacetylated histones reside (silencers are nucleosome-free regions). Further studies will resolve whether Sb-Sir4-assisted Sc-Sir4 incorporation involves contacts with multiple silencing proteins versus simple Sir4-Sir4 dimerization and whether it requires Sir2 catalytic activity.
Additionally, the enhanced interaction of Sb-Sir4 across Y′ elements at S. cerevisiae telomeres (Figure 4B) suggested that novel or changed interactions in the hybrids somehow led to enhanced Sir4 occupancy of these regions. This differential long-range occupancy by Sir complexes presents an opportunity to ask whether Sir1 and Sir4-mediated interactions during Sir complex nucleation regulate the “strength” of silent chromatin over a distance. Alternatively, Sb-Sir4 (and potentially other S. bayanus silencing proteins) may have been less sensitive to factors that exclude Sc-Sir4 from the Y′ elements. The species-specific Sir4 distributions occurring in these interspecies hybrids should be further dissected to understand the determinants limiting silent chromatin formation across subtelomeric regions.
Another unusual property of the interspecies hybrids led to a weak silencing defect affecting Sb-HMR but not Sc-HMR (Figure 2A, row 1; Figure 3B). In hybrids lacking Sc-Sir4 this defect was more evident (Figure 2A, row 3), which paradoxically suggested that Sc-Sir4 protected Sb-HMR silencing in the presence of Sb-Sir4, despite having no ability to silence Sb-HMR on its own. How might Sc-Sir4 have “enhanced” Sb-Sir4 function at S. bayanus silent loci in hybrids? Strong evidence compatible with Sc-Sir4 protecting Sb-Sir4 from being titrated by Sc-specific sequences was the ability of both Sb-Sir4 and Sc-Sir4 to bind extensively to S. cerevisiae telomeres, as described in the Results. Hence, the hybrid state may result in a dosage sensitivity to Sb-SIR4 not evident in S. bayanus SIR4/sir4Δ intra-species diploids due to additional binding sites provided by the Sc-X and Y′ elements, and potentially other elements. We note the resemblance of this “Sb-Sir4 sequestration” model to the “Circe effect” proposed to explain Sc-Sir4-mediated clustering of S. cerevisiae telomeres [54].
Gregor Mendel's studies were motivated by a desire to understand the emergent properties of interspecies hybrids, such as hybrid vigor, that were of great practical significance at the time. Although he became famously distracted by discovering two fundamental laws of genetics, his original interest in the processes by which hybrid species are not necessarily the “average” of the two parental species remains as interesting today as it was practically important in Mendel's day. Indeed, the striking asymmetry in the ability of Sb-Sir4 to silence Sc-HMR, but inability of Sc-Sir4 to silence Sb-HMR (Figure 2A), was the seminal observation that inspired this study. By and large, however, in interspecies hybrids of S. cerevisiae and S. bayanus, a protein from either species was fully capable of providing all of that protein's function to hybrids. Although this result could be anticipated from the ability to “clone by complementation” genes of one species by their function in another, this study established that symmetry of complementation is an important general consideration. For example, the essential proteins Rap1 and Abf1 from either species had all the functions necessary to support viability of the hybrids, and we established that Sir2 and Sir3 of both species were fully interchangeable (Figure S1), despite being members of a complex in which another member of that complex, Sir4, has extraordinary divergence. By extrapolation, asymmetrical deviations from a general expectation of cross-species compatibility, such as in the case of Sir4, may signal situations of uncommon interest.
The studies presented here capitalized on the extraordinary genetic properties of interspecies hybrids to tease out important dimensions to the evolution and structure of silent chromatin in yeast. Although silencing behavior in these yeast hybrids was rather unusual, some type of defect might have been anticipated from recent studies of hybrid sterility or lethality genes in Drosophila, which have implicated rapidly evolving heterochromatin proteins as key factors contributing to interspecies genetic incompatibility [55],[56]. There is presently no reason to believe that SIR1 or SIR4 play roles in the post-zygotic genetic incompatibility between budding yeast species. It is notable, however, that in budding yeast multiple regulatory sites mediating silencing have rapidly evolved in a phylogenetically asymmetrical fashion along with a set of divergent silencing proteins, paralleling observations of rapid evolution in Drosophila heterochromatin [55],[57],[58]. It will be of great interest to determine whether the similar patterns of heterochromatin evolution in these distant taxa reflect similar underlying evolutionary processes.
The unprecedented resolution of Sir4 distribution provided by ChIP-Seq methods calls into question earlier models for silenced chromatin assembly, and in particular the so-called mechanism of spreading (reviewed in [7],[59]). In the common view, Sir protein recruitment to the silencers or to telomeres allows the deacetylation of H4K16-Ac on adjacent histones, creating new binding sites for additional Sir protein complexes, with sequential cycles of deacetylation and binding leading to spreading of Sir-protein complexes across all nucleosomes in silenced chromatin. The strikingly uneven distribution of Sir4 at HML and HMR (Figure 4A, note y-axes), and at the telomeres (Figure 4B), as shown here and in K. lactis [60], is not entirely inconsistent with the common view of heterochromatin spreading, but is in no way anticipated by it. Clearly, high-resolution characterization of all Sir proteins by ChIP-Seq has the potential to force substantial revision or replacement of the current view.
All S. cerevisiae strains were of the W303 background. Generation of marked S. bayanus strains from type strain CBS 7001 has been described [61]. All yeast strains were cultured at 25°C in standard yeast media. One-step gene replacement and C-terminal 13xMyc tag integration have been described previously [36],[62], and these genetic manipulations were performed identically for S. bayanus, S. cerevisiae, and S. cerevisiae/S. bayanus hybrids. The HMR::URA3 reporter strains were constructed independently in S. cerevisiae sir4Δ and S. bayanus sir4Δ haploid strains, wherein the HMRa1 ORF was replaced with the K. lactis URA3 ORF by PCR-based gene targeting, leaving the HMRa1 promoter intact. For most experiments, interspecies hybrids were made by crossing S. bayanus MATα HMR::URA3 strains (wild-type, sir4Δ, or SIR4-13xMyc) to S. cerevisiae MATa strains (wild-type, sir4Δ, or SIR4-13xMyc). For ORC1, RAP1, and ABF1 heterozygote analysis, gene targeting was performed directly in hybrid diploids or S. bayanus diploids. Three independent transformants were analyzed in all cases. Sc-SIR4-13xMyc and Sb-SIR4-13xMyc alleles were shown to be functional by two independent silencing assays in each case: by mating ability in S. cerevisiae SIR4-13xMyc and S. bayanus SIR4-13xMyc haploid strains and by FOA resistance in hybrid diploids bearing the appropriate HMR::URA3 reporter (unpublished data).
The Sc::(Sb-HMR::URA3) replacement allele (Figure 6A) was generated in two steps. The Sb-HMR::URA3 cassette plus 1 kb of leftward-flanking sequence was PCR-amplified out of the S. bayanus genome, and the PCR product was used to replace the syntenic portion of Sc-HMR (including the E silencer) in S. cerevisiae sir4Δ strains. A HygMX marker was then targeted into the S. bayanus genome 3 kb to the right of Sb-HMR. The entire rightward-flanking 3 kb region plus the HygMX marker was PCR-amplified out of the S. bayanus genome, and the PCR product was used to replace the syntenic portion of Sc-HMR in the S. cerevisiae genome (including the I silencer). The Sc::(Sb-HMR::URA3) replacement allele therefore included a total of 5.5 kb of Sb-HMR sequence, plus the 1.7 kb HygMX marker.
To construct the SIR4 replacement alleles, the Sc-SIR4 and Sb-SIR4 genes were separately cloned into the yeast plasmid pRS315 [63] such that the LEU2 marker was 5′ of, and in opposite orientation to, each SIR4 gene. Each SIR4 gene plus the LEU2 marker was PCR-amplified from each plasmid. The LEU2-Sc-SIR4 PCR product was used to replace the URA3 marker at the Sb-SIR4 locus in an S. bayanus sir4Δ::URA3 leu2 strain; likewise, the LEU2-Sb-SIR4 PCR product was targeted into the Sc-SIR4 locus in an S. cerevisiae sir4Δ::URA3 leu2 strain. The integrated Sc-SIR4 gene was shown to silence Sc-HMR::URA3 in hybrids (Figure 3A), and the integrated Sb-SIR4 gene was shown to silence Sb-HMR::URA3 in hybrids (Figure 6C) and Sc-HML and Sc-HMR in S. cerevisiae strains (Zill et al. in preparation). The expression level of each SIR4 replacement allele was determined by quantitative RT-PCR (Figure S2A).
Assays of yeast strain growth on FOA and CSM/-Ura media were performed using standard “frogging” techniques. Briefly, for each strain, a 10-fold dilution series of yeast cells at an approximate density of 4×107/mL was spotted onto each plate. For Figures 2, 3A, 6, and 8A plates were photographed after 2 d for YPD, and after 3 d for FOA and CSM/-Ura. For Figure 8B, plates were photographed after 3 d for all media. For Figure 9B, plates were photographed after 3 d for FOA and YPD, and after 5 d for CSM/-Ura. We note that some changes in silencing could be seen only on FOA and not on CSM/-Ura. Incomplete silencing of the HMRa1 promoter likely led to heterogeneous expression states within the population of cells, with some remaining silent while others were expressed [64].
RNA isolation was performed using the hot-phenol method [65]. Total RNA was digested with Amplification grade DNase I (Invitrogen) and purified using the RNeasy MinElute kit (Qiagen). cDNA was synthesized using the SuperScript III First-Strand Synthesis System for RT-PCR and oligo(dT) primer (Invitrogen). Quantitative PCR on cDNA was performed using an MX3000P machine (Stratagene) and the DyNAmo HS SYBR Green qPCR kit (NEB). Amplification values for all primer sets were normalized to actin (ACT1) or SEN1 cDNA amplification values. Samples were analyzed in triplicate from three independent RNA preparations.
Yeast whole cell extracts were prepared using 20% TCA and solubilized in SDS loading buffer plus 100 mM Tris base. SDS-PAGE and immunoblotting were performed using standard procedures and the LiCOR imaging system. Anti-c-Myc antibody from rabbit (Sigma, Cat. No. C3956) was used to detect Myc-tagged Sir4 and Abf1 proteins. Mouse anti-Pgk1 antibody (Invitrogen, Cat. No. 459250) was used to verify equal loading. The S. cerevisiae Orc5-HA strain derivation has been described [66].
All chromatin immunoprecipitations (Sir4-Myc, Orc5-HA, Abf1-Myc) were performed as described [67], using formaldehyde cross-linking of log phase cultures for 1 h at room temperature. IPs were performed overnight at 4°C using Anti-c-Myc-Agarose (Sigma, Cat. No. A7470) and Anti-HA-Agarose (Sigma, Cat. No. A2095). Quantitative PCR was performed as described above.
Orthologous S. cerevisiae and S. bayanus genes were identified on the basis of sequence similarity and syntenic context (D. Scannell and M. B. Eisen, unpublished). Percent identities between 4,981 orthologous S. cerevisiae and S. bayanus proteins were then obtained by running BLASTP [73] with default parameters, imposing an E-value cutoff of 1×10−5, harvesting percent identities for each HSP and calculating a length-weighted average. This will necessarily lead to some underestimation of the true divergence between protein pairs, but it is unlikely that the rank order of divergences among pairs would be significantly affected.
For PAML and sliding window analyses, protein alignments were produced using FSA with default parameters [74], and DNA alignments were obtained by back translation with RevTrans [75]. All site and branch models were fit using codeml in the PAML package [76]. To test for positive selection we compared model M8 to M7 using a χ2 test with two degrees-of-freedom. Posterior probabilities of ω>1 for individual codons were obtained from the Bayes Empirical Bayes output of M8 only. For the sliding window (dN/dS, or ω) analyses, a window size of 102 bp and a step-size of 3 bp were used. Only alignment windows without gaps were analyzed. For each window we used codeml to estimate a single ω using model M0 implemented in codeml. All other parameters were estimated from the data.
We estimated the level of selective constraint operating on SIR4 on each branch of the Saccharomyces sensu stricto phylogeny by computing branch-specific ratios of non-synonymous to synonymous substitutions (dN/dS, or ω). Briefly, we performed protein-space alignments of orthologous SIR4 coding sequences with FSA [74] and then used codeml in the PAML package [76] to fit a “free-ratio” model (model = 1, NSSites = 0) to the alignment and obtain independent estimates of ω for each branch.
|
10.1371/journal.pgen.0030186 | αADα Hybrids of Cryptococcus neoformans: Evidence of Same-Sex Mating in Nature and Hybrid Fitness | Cryptococcus neoformans is a ubiquitous human fungal pathogen that causes meningoencephalitis in predominantly immunocompromised hosts. The fungus is typically haploid, and sexual reproduction involves two individuals with opposite mating types/sexes, α and a. However, the overwhelming predominance of mating type (MAT) α over a in C. neoformans populations limits α–a mating in nature. Recently it was discovered that C. neoformans can undergo same-sex mating under laboratory conditions, especially between α isolates. Whether same-sex mating occurs in nature and contributes to the current population structure was unknown. In this study, natural αADα hybrids that arose by fusion between two α cells of different serotypes (A and D) were identified and characterized, providing definitive evidence that same-sex mating occurs naturally. A novel truncated allele of the mating-type-specific cell identity determinant SXI1α was also identified as a genetic factor likely involved in this process. In addition, laboratory-constructed αADα strains exhibited hybrid vigor both in vitro and in vivo, providing a plausible explanation for their relative abundance in nature despite the fact that AD hybrids are inefficient in meiosis/sporulation and are trapped in the diploid state. These findings provide insights on the origins, genetic mechanisms, and fitness impact of unisexual hybridization in the Cryptococcus population.
| Cryptococcus neoformans is a major cause of fungal meningitis, predominantly in immunocompromised individuals. This fungus has two mating types/sexes, a and α, and mating typically requires two individuals with opposite mating types. It is mysterious why the α mating type is overwhelmingly predominant in nature and how the capacity for sexual reproduction is maintained in a largely unisexual population. We postulated that same-sex mating between α isolates may occur naturally, as it does under laboratory conditions. By analyzing natural Cryptococcus diploid hybrid isolates containing two α alleles of different serotypic origins, this study demonstrates that same-sex mating transpires in nature. The observations that Sxi1α, a sex regulator encoded by the mating type locus, is frequently altered in C. neoformans hybrids but rarely in the haploid population, and that Sxi1α is also altered in the fertile VGIII group of the sibling species C. gattii by a different mutation support the hypothesis that these SXI1α mutations may enhance fertility, possibly in concert with other genomic changes. Our study provides insights on the genetic and environmental factors that play important roles in the evolution of the current population structure of this pathogenic fungus.
| The level of genetic variation within a species is correlated with evolutionary potential [1]. Hybridization can provide genetic variation within and between populations by yielding progeny more fit in novel or changing environments, and both intra- and interspecies hybridization are a driving force for evolution [2,3]. Hybridization is observed in animals, and is especially common in plants [4–8]. Hybrids also occur in microorganisms. For example, Trypanosoma cruzi, the cause of Chagas disease, descends from two ancestral hybridization events [9,10]; influenza viruses undergo antigenic variations and host range shifts through hybridization and reassortment [11]; and in the parasite Leishmania, which has no known sexual cycle and a largely clonal population structure, recombinant strains can be generated through interspecific hybridization [12–15].
Because of their morphological and genomic plasticity, fungi are subject to profound genetic changes, including those resulting from hybridization. Indeed, hybridization is one of the most significant biological forces driving fungal evolution, as illustrated by the Saccharomyces sensu stricto complex [16]. This species complex descends from an ancient whole genome duplication event in which two related yeast species hybridized ∼100 million years ago [17–20]. Hybridization can confer novel features; for example, S. cerevisiae–S. paradoxus hybrids exhibit thermal stress vigor [21]. In plant fungal pathogens, hybridization produces novel physiological traits including enhanced virulence [22–24]. By comparison, less is known about the impact of hybridization on the virulence of human pathogenic fungi.
Cryptococcus neoformans is a cosmopolitan human fungal pathogen that causes meningoencephalitis in predominantly immunocompromised hosts [25]. Cryptococcal meningitis is the most common fungal infection of the central nervous system and is considered an AIDS defining condition [25–29]. This species is classified into three serotypes based on capsular agglutination reactions [30]: serotype A (C. neoformans var. grubii, mostly haploid), serotype D (C. neoformans var. neoformans, mostly haploid), and AD hybrids (mostly diploid). Serotype A is responsible for the vast majority of human infection (95% worldwide) [25], but AD hybrids can be fairly common, especially in Europe (∼5%–30%) [31–37], and are likely more common than currently appreciated [32,37–39].
Because this fungus is ubiquitous in nature, and humans are infected through inhalation of infectious propagules from the environment [40–42], it is important to understand the natural life cycle and its impact on population structure. The bipolar mating system of C. neoformans has been well-defined under laboratory conditions. Mating involves cell–cell fusion of haploids of opposite mating type, a and α, to produce dikaryotic filaments. Nuclear fusion occurs at the tip of the filaments in a fruiting body (the basidium) resulting in a transient a/α diploid that immediately undergoes meiosis and sporulation [43,44] (Figure 1). Because clinical and environmental isolates of C. neoformans are predominantly of α mating type (>98%–99.9%) [25,45], it is difficult to envision that a–α mating is the only significant means by which genetic diversity is generated in nature. C. neoformans serotype D strains undergo monokaryotic fruiting to produce filaments and basidiospores under laboratory conditions [46,47]. Fruiting was recently recognized to be a modified form of sexual reproduction occurring between strains of the same mating type [48] (Figure 1). Because monokaryotic fruiting is commonly observed in serotype D α isolates [46–48], and the MATα allele enhances fruiting under laboratory conditions [49], same-sex mating could significantly impact the population structure of this pathogenic fungus in nature.
Although the global population of C. neoformans is largely clonal, recombination does occur at a low level [50–54]. Recently, phylogenetic analysis of the sibling species C. gattii has shown that same-sex mating between two different α strains may have given rise to a more virulent strain occupying a new environmental niche and causing the Vancouver Island outbreak [55]. Population genetic studies of C. neoformans serotype A veterinary isolates in Sydney, Australia, also reveal evidence of recombination in a unisexual α population, providing further indirect support for the occurrence of same-sex mating in natural populations (D. Carter, personal communication). In this study, characterization of natural diploid αADα hybrids provides definitive evidence for same-sex mating occurring in nature.
Under laboratory conditions, intervarietal matings between strains of serotype A and D lead to cell–cell fusion, but genetic differences between these divergent serotypes (∼5%–10% nucleotide polymorphisms) severely limits meiosis and thus few, if any, viable haploid basidiospores are produced [56,57]. Consequently, most natural AD hybrids remain in the diploid (or aneuploid) state (Figure 1), and analysis of these AD hybrids can reveal the genomic nature of their parental strains. For example, most reported AD hybrids are αADa or aADα (mating type/serotype–serotype/mating type) [32–36,57,58], reflecting their origin from traditional a–α mating between serotype A and D strains. All extant aADα hybrids appear to derive from a cross between African serotype A strains (Aa) and serotype D strains (Dα) followed by clonal expansion and emigration from sub-Saharan Africa, the only region where serotype A isolates of a mating type are common [53,59].
In this study, we identified and characterized natural αADα hybrids that arose from same-sex mating between two α strains of A and D serotypes, providing definitive evidence that the laboratory-defined same-sex mating process occurs in nature. In addition, our analysis reveals a common feature in all aADα and αADα hybrids tested: a C-terminal deletion in the serotype D SXI1α gene located in the MAT locus, which encodes a homeodomain transcription factor regulating mating [60]. Characterization of populations containing the C-terminally truncated SXI1α serotype D (SXI1Dα) allele suggests that this mutation may have contributed to the origin of AD hybrids.
The common presence of AD hybrids in both clinical and environmental samples may be indicative of hybrid vigor [33,35,61]. However, unlike clearly documented cases of increased fitness and epidemiological success of plant-pathogenic fungal hybrids [23,62–65], examples of hybrid fitness in human pathogenic fungi have not been well-documented. Previous studies of C. neoformans AD hybrids revealed variable virulence potential [57,58,66,67]. This ambiguity may be due to the analysis of genetically diverse αADa and aADα isolates, which exhibit considerable phenotypic and genotypic variation. The presence of both a and α mating types in a diploid strain may also complicate virulence studies if pheromone sensing occurs during infection [68–70]. Here, αADα hybrids were constructed in defined genetic backgrounds and analyzed for hybrid fitness and virulence. In vitro, laboratory-constructed αADα hybrids exhibited hybrid vigor, and were more UV- and temperature-resistant than either parent. Other virulence attributes of the αADα hybrid were similar to (e.g., capsule) or intermediate between (e.g., melanin) those of the parents. In a murine inhalation model, the laboratory-constructed αADα hybrid exhibited virulence similar to that of the serotype A parent. These observations demonstrate benefits of hybridization in C. neoformans, which may enable less robust serotype D strains to survive both during infection and in the environment.
A report from Litvintseva et al. in 2005 indicated the potential existence of environmental αADα hybrids isolated from North Carolina, USA [35]. To ensure these were indeed AD hybrids, three such isolates and control strains were analyzed by amplified fragment length polymorphism (AFLP) analysis. AFLP results using two primer pairs showed that all three isolates generated a banding pattern representing a composite between those of serotype A and D strains, indicative of an AD hybrid (Figure 2A and 2B). These strains also contained twice the cellular DNA content of haploid controls based on fluorescence flow cytometry analysis (Figure 2C), and are therefore diploid.
Based on serotype- and mating-type-specific PCR, all three isolates have serotype A– and serotype D–specific genes, both within the mating type locus (MAT) and in other genomic regions (Table 1), further confirming their AD hybrid nature. Sequence analysis suggested the three isolates could be clonal, as PCR-amplified gene sequences were identical (data not shown). The combined sequences for five different serotype A–specific genes (STE20α, SXI1α, GPA1, CNA1, and PAK1) were 99.9% identical to those of the sequenced serotype A reference strain H99 (4,226/4,230 bp) [71]. The sequences for four different serotype D–specific genes (STE20α, GPA1, CNA1, and PAK1) were 99.86% identical to those of JEC21 (2,886/2,890 bp), a sequenced serotype D reference strain [72]. Because these AD hybrid isolates contain α mating type genes from both serotype A (STE20α and SXI1α) and D (STE20α) and lack a mating type genes of either serotype (Table 1), they are αADα strains that originated from two α parental strains of serotype A and D. This provides the first direct evidence, to our knowledge, of the cell–cell fusion step of same-sex mating occurring in nature.
The mating behavior of the natural αADα hybrids was examined in crosses with the reference strains JEC20 (a) and JEC21 (α). The αADα hybrids mated with the a reference strain JEC20 to produce mating dikaryotic hyphae with clamp connections (Figure 3), and did not mate with the α reference strain (data not shown). The two parental nuclei (diploid α/α and haploid a) alternated positions in adjacent hyphal cells, a hallmark of compatible matings in basidiomycetous fungi [48,73]. Basidial fruiting bodies were also observed in different developmental stages, they contained one or multiple nuclei, and some were decorated with four long chains of spores (Figure 3). These morphological characteristics are similar to those of matings between haploid α and a cells. However, despite apparently normal morphological differentiation, the spores generated were not viable, and all dissected spores from a cross between the diploid αADα hybrid 6–20 and the a haploid strain JEC20 failed to germinate (n = 105), indicative of abnormal meiosis, as expected from a triploid. Our observations indicate that the αADα hybrid mates as α, but is unable to complete the final stages of sexual reproduction, including spore germination.
Because the SXI1Dα allele could not be amplified from the αADα hybrids with the primers tested (Table 1), the mating type locus of the αADα hybrids was further analyzed to ascertain whether any genetic alterations were apparent. The MAT locus of C. neoformans is unusually large (>100 kb) compared to most fungi and encodes more than 20 proteins [74]. Because of the complex nature of the C. neoformans MAT locus, all genes within the MAT locus of the natural αADα hybrid were examined by comparative genome hybridization (CGH).
Mating-type- (a and α) and serotype-specific (serotype A and D and C. gattii) 70-mer probes for all genes in the MAT locus (Aα, Aa, Dα, and Da alleles for each MAT gene) were designed previously for microarray analysis [49]. Here genomic DNA was labeled and hybridized to microarray slides to characterize the mating type locus gene content. Genomic DNA of the natural αADα hybrid 6–20 and the control (a mixture of H99 [Aα] and JEC21 [Dα]) were labeled with fluorescent dyes and competitively hybridized to a genomic microarray slide containing the mating-type- and serotype-specific 70-mers. The log2 ratio of fluorescence intensity between the hybrid and the control for all a genes was close to zero regardless of serotypes (the average log2 ratio of fluorescence intensity was within ± 0.4, meaning that the fold difference between hybrid and control fell into the range of 0.76∼1.32; data not shown), indicating the genetic contents of the control, and sample were similar. Because there were no a genes in the control, this showed that a genes were also absent in the hybrid strain, consistent with the PCR analysis (Table 1). To ensure that lack of hybridization to Aa or Da probes was not due to failure of the a 70-mers on the microarray slides, hybridizations of Aα/Da, Aa/Dα, and Aa/Da samples using genomic DNA from reference strains were performed. The Aa and Da probes were functional based on this analysis (Figure S1). As shown in Figure 4, the overall fluorescence intensity of α genes in the MAT locus from the natural hybrid isolate 6–20 and the control (Aα + Dα) was similar for both the serotype A and D alleles (log2 ratio of fluorescence intensity was within ± 0.5, meaning that the fold difference between hybrid and control fell in the range of 0.71∼1.41). The only exception was that the SXI1Dα allele appeared to be missing in the natural αADα hybrid, as the fluorescence intensity of the hybrid SXI1Dα was much lower than that of the control (log2 hybrid/control = −3.24, which means hybrid/control ≈ 0.1). This CGH result is consistent with the SXI1α PCR analysis (Table 1), indicating that hybrid 6–20 contains all α genes from both serotype A and D with the apparent exception of the SXI1Dα allele.
However, because the array used was not a tiling array, other potential mutations in the mating type locus, such as indels in regions not covered by the probes and single nucleotide alterations, might not be detected.
Because the SXI1Dα gene in the MAT locus of the αADα hybrids did not amplify using SXI1Dα-gene-specific primers (Table 1), or yield a hybridization signal during CGH analysis (Figure 4), the structure of the SXI1Dα locus in the natural αADα hybrids was examined by Southern hybridization. Surprisingly, hybridization to the SXI1Dα ORF probe was observed, but the size of the hybridizing band was decreased for the natural αADα hybrids compared to the wild-type serotype D control, suggesting that a shorter version of the SXI1Dα gene was present (Figure 5). Sequencing of the SXI1Dα allele from the three αADα hybrids revealed a C-terminal truncation of the ORF (119 bp) and a partial deletion of the 3′ untranslated region (301 bp). Thus, the genomic locus is 420 bp shorter in the αADα hybrids (Figure 5). The 70-mer oligonucleotide on the microarray slide used to detect the SXI1Dα gene lies within the C-terminal deletion interval, and the sequence of one of the primers (JOHE15636) used to PCR amplify the SXI1Dα-specific gene was also within the missing region (Figure 5), explaining the apparent absence of the SXI1Dα gene in the PCR and CGH analyses (Table 1; Figure 4).
To test whether the C-terminal deletion in the SXI1Dα gene is unique to the αADα isolates from North Carolina, or is a uniform feature of the aADα and αADα hybrids with a Dα parental origin, additional hybrids were analyzed. Interestingly, all of the aADα and αADα hybrid strains tested share precisely the same C-terminal truncated version of SXI1Dα (Table 2). Four hypotheses could explain the presence of the truncated SXI1Dα allele in hybrid strains. (1) The “pre-fusion fitness” model: the truncated SXI1Dα allele may confer an advantage to haploid serotype D strains, and selection for the shorter version of SXI1Dα occurred prior to cell fusion. In this model, the truncated version of SXI1Dα is prevalent in aADα and αADα hybrids simply because it is common in the Dα population. (2) The “pre-fusion fertility” model: selection for this C-terminally truncated SXI1Dα was prior to cell fusion. This version of SXI1Dα may enhance the fertility of Dα strains and therefore is common in aADα and αADα hybrid strains that result from fusion between Aa or Aα strains and Dα strains with this allele. (3) The “post-fusion fitness” model: the SXI1Dα truncated version may confer an advantage to AD hybrids such that selection for this allele occurred after hybrid formation. This advantage could involve limiting sporulation, leading to fewer inviable spores in AD hybrid strains. (4) The “natural variant” model: this SXI1Dα allele is a neutral variant that confers no selective benefit.
To test these hypotheses, the prevalence of the C-terminally truncated version of the SXI1Dα allele was investigated in natural Dα isolates. If selection for this SXI1Dα allele occurred prior to the cell fusion events that produced AD hybrids (“pre-fusion fitness” and “pre-fusion fertility” models), this allele should be present in the serotype D α haploid population. If selection for this allele occurred after cell fusion, then it would be expected to be absent in the Dα population (“post-fusion fitness” model). If there was no selection, then this allele need not be common in either the hybrid or the Dα population (“natural variant” model).
Twenty-four isolates recorded as serotype D α strains were screened by PCR to detect SXI1Dα size polymorphisms, and four isolates were found to contain the truncated allele, while the remaining 20 isolates contain the wild-type allele (Table 2). The truncated version of SXI1Dα in these four isolates was sequenced, and the deletion site was identical to that found in the aADα and αADα hybrids. Interestingly, one of five North Carolina Dα isolates, each representing a different genotype [35], harbors the C-terminally truncated SXI1Dα allele. These five Dα strains were isolated together with the natural αADα hybrids in a previous study [35]. The North Carolina αADα hybrids bearing the C-terminally truncated SXI1Dα allele represent the common AD genotype (76%, or 41/54) in this region [35], further supporting the hypothesis that selection for this allele could have occurred. To ensure that these isolates are indeed haploid Dα strains and not unrecognized hybrids, ploidy was analyzed by fluorescence flow cytometry. As shown in Table 2, with one exception (isolate 713), all of the serotype D isolates tested were haploid. Isolate 713 showed a diploid nuclear DNA content and was found to be an unrecognized αADα hybrid isolate from Italy (see below).
Thus, the truncated SXI1Dα allele is present in the global natural serotype D α isolates, albeit at a relatively low level (∼13%, or 3/23), which does not support the “pre-fusion fitness” or “post-fusion fitness” models. The truncated SXI1Dα allele is uniformly present in the aADα and αADα hybrid population (100%, or 10/10) (Table 2), which supports the “pre-fusion fertility” or “post-fusion fitness” models. All strains with the truncated SXI1Dα allele harbor an identical SXI1Dα allele, while those strains without the truncation harbor distinct SXI1Dα alleles based on the sequence of the SXI1Dα 5′ region preceding the deletion site. Thus, the novel truncated allele likely arose once in the haploid progenitor population, arguing against the “post-fusion fitness” selection model. These findings support the “pre-fusion fertility” model, in which the SXI1α truncation allele enhances fertility of the serotype D α haploid parental progenitors and, as a result, increases fusion with an Aa or Aα partner to yield the aADα and αADα hybrid populations.
During this screening, another αADα hybrid isolate (713) was identified. This diploid isolate contains α mating type genes from both serotype A (STE20Aα and SXI1Aα) and D (STE20Dα), lacks a mating type gene of either serotype type based on mating-type- and serotype-specific PCR (data not shown), and is an unrecognized αADα hybrid. To confirm this, the MAT locus was characterized by CGH. Isolate 713 displayed a CGH MAT profile similar to that of the αADα hybrids from North Carolina (Figure S2). All of the mating type genes of both the Aα and the Dα alleles were similar to the control (Aα + Dα) with the only exception being the SXI1Dα gene, which was also truncated in this natural αADα hybrid. The discovery of an independent αADα isolate from Italy suggests that same-sex mating is not restricted geographically, consistent with the fact that Aα and Dα isolates are globally distributed worldwide in nature and are often sympatric. However, we cannot exclude that an ancestral αADα isolate clonally expanded to distinct locations.
To investigate if altered SXI1α alleles also occur in other members of the Cryptococcus species complex, known sequences of the SXI1α gene in the sibling species C. gattii were analyzed [55]. C. gattii and C. neoformans diverged from a common ancestor ∼37 million years ago and are recognized to be separate species [75]. C. gattii is divided into four molecular types: VGI, VGII, VGIII, and VGIV [40]. Although the majority of C. gattii strains are sterile, a significant proportion of VGIII isolates are fertile [55,76]. The SXI1α gene sequences in strains of the three molecular types, VGI (10/10), VGIV (4/4), and VGII (10/11) appeared wild-type (data not shown) with the exception of one VGII strain (WM178, 1/11) in which the SXI1α gene contains a frameshift mutation (Figure S3). Interestingly, a premature stop codon is present in the C-terminus of the SXI1α gene in the majority of strains of the VGIII molecular type (7/8, or 87.5%) (Figure S3). This stop codon truncates the C-terminus of Sxi1α (corresponding to residue 358 in Sxi1Dα) seven amino acids N-terminal to the deletion site found in the truncated SXI1Dα allele in the aADα and αADα populations (365 aa) (Figure S3). Importantly, the homeodomain (aa 144–205 in Sxi1Dα) [60] is intact in both truncated SXI1α alleles. The observation of two different mechanisms of C-terminal truncation in the SXI1α gene occurring in subgroups of two different species (C-terminal deletion in C. neoformans serotype D and AD hybrid strains and premature stop codon in VGIII C. gattii isolates) indicates that C-terminally truncated versions of the SXI1α gene have arisen independently at least twice in the Cryptococcus species complex.
The SXI1α gene is a master regulator of sexual reproduction [60]. Deletion of this gene does not prevent cell–cell fusion, but blocks further sexual morphological differentiation into dikaryotic hyphae, meiosis, and development of basidiospores during a–α mating [60]. The C-terminal deletion of the SXI1Dα gene does not prevent sexual differentiation during mating based on the fact that the two natural Dα strains (431 and 434) with a C-terminal deletion in the SXI1α gene still produce mating hyphae and abundant basidiospores when crossed with the reference strain JEC20 (Figure 6). Differences in filamentation and sporulation observed between the wild-type strain JEC21 and the nonisogenic natural strain 431 could be attributable to other genetic differences. Spores dissected from a cross between strain 431 and JEC20 were viable (germination rate = 83%, n = 72) and showed a typical 1:1 Mendelian segregation of mating types (a:α = 31:29), indicative of normal meiosis. An engineered strain in the JEC21 background with the C-terminally truncated allele of SXI1Dα replacing the wild-type SXI1α allele also mated like wild-type, indicating the truncated SXI1α allele is functional (Figure S4). These observations indicate that the C-terminal deletion of the SXI1α gene does not impair morphological development or meiosis during mating.
The C. gattii SXI1α gene with the premature stop codon at the C-terminus is also functional. Five out of seven VGIII C. gattii strains that contain this SXI1α variant mated with the reference strain JEC20 to form mating hyphae and spores (Figure 7). Two VGIII isolates were sterile and likely harbor other unlinked mutations that impair fertility (Figure 7). The C. gattii isolate NIH836 likely harbors a nonfunctional SXI1α gene as an early stop codon occurs after one-third of the coding sequence (Figure S3); this isolate was sterile, consistent with the known essential role of SXI1α in mating [60]. Many natural C. gattii strains are sterile under laboratory conditions [77], whereas the VGIII molecular type contains many of the known fertile C. gattii isolates. Isolate NIH312, the most fertile C. gattii strain identified thus far [77], is a member of this group and harbors the SXI1α premature stop codon allele. These findings provide further evidence that changes in the C-terminus of the SXI1α gene may enhance fertility.
Previous reports on the virulence of AD hybrids present differing results [57,58,67]. Reduced virulence of AD hybrid isolates compared to the Aα H99 reference strain was observed by Lengeler et al. [57], virulence of AD hybrids similar to that of H99 was reported by Chaturvedi et al. [58], and virulence of AD hybrids intermediate between Aα H99 and Da JEC20 reference strains was presented by Barchiesi et al. [67]. This variation is likely due to both different experimental models and analysis of divergent αADa and aADα isolates, as these isolates differ genotypically and phenotypically. The presence of opposite mating types, a and α, in diploid strains may also have complicated earlier virulence studies, as pheromone production and sensing may occur during infection [68–70].
To avoid these potential complications in virulence studies, AD hybrid strains of only α mating type were constructed based on the H99 (haploid Aα) and JEC21 (haploid Dα) backgrounds (see Materials and Methods for details). Both parental strains have completed genome sequences and are widely used for genetic and pathogenesis studies [72,78,79] (http://cneo.genetics.duke.edu/; http://www.broad.mit.edu/annotation/genome/cryptococcus_neoformans/Home.html).
The laboratory-generated αADα hybrid was first tested in vitro. As an environmental pathogen, C. neoformans may have evolved and maintained virulence traits through selective pressure in the environment [25,56,80,81]. Defined C. neoformans virulence factors include melanization, capsule production, and the ability to grow at high temperature, all of which confer survival advantages in both animal hosts and the environment. The ability to grow at high temperature (37–39 °C) enables human infection [82–84]; production of a polysaccharide capsule inhibits host immune responses during infection and protects cells from dehydration in the environment [85–88]; production of melanin provides protection from toxic free radicals generated by host defenses during infection and from UV irradiation in the environment [89,90]. These virulence properties enable C. neoformans and its sibling species C. gattii to be the only two highly successful mammalian pathogens in the genus Cryptococcus [40,56,91].
In vitro virulence attributes of the laboratory-constructed hybrid strain were compared to those of the parental strains. Haploid Aα (H99), haploid Dα (JEC21), and the laboratory-constructed hybrid αADα (XL1462) strains were examined for sensitivity to UV irradiation, growth at high temperature (39 °C), capsule production, and melanization (see Materials and Methods for details). Each cell type was capable of capsule production based on microscopic observations (Figure 8A). The diploid αADα hybrid cells were larger than those of the parental Aα and Dα strains, and this was confirmed by forward scatter flow cytometry (data not shown). An association of higher ploidy with larger cell size has also been observed in other organisms [92–94]. The Aα strain H99 was more resistant to UV irradiation than the Dα strain JEC21, and the αADα hybrid strain was even more resistant to UV irradiation than the Aα parental strain (Figure 8B). Both higher ploidy, which resulted from hybridization, and the interaction of the serotype A and D genomes independently contribute to this enhanced resistance of AD hybrids to UV irradiation, based on the observation that diploid cells (αAAα or αDDα) were modestly more UV-resistant than haploid cells (Aα or Dα), but less UV-resistant than αADα hybrids (Figure S5). The αADα hybrid strain also grew significantly better at 39 °C than the Aα and Dα haploid parental strains, again displaying hybrid vigor (Figure 8B).
C. neoformans can produce melanin by oxidizing a variety of diphenolic substrates, including the neurotransmitter L-dihydroxyphenylalanine (L-DOPA) [89]. Variation in the rate of melanization yields pigmentation differences. At 22 °C, both the Aα strain and the αADα hybrid were heavily melanized compared to the Dα strain (Figure 8B). At 37 °C, melanization of the hybrid αADα was drastically reduced and was comparable to that of the less melanized Dα parental strain (Figure 8B). This observation indicates a complicated interaction of different virulence attributes (temperature and melanization) in the αADα hybrid.
In conclusion, the αADα hybrid strain displays hybrid vigor for some virulence factors under defined in vitro conditions, but the effect of hybridization on other virulence factors is complex.
As the effects of hybridization on in vitro virulence attributes are complex, the virulence potential of the hybrid was assayed in a murine inhalation model. Animals were intranasally infected with haploid Aα (H99), haploid Dα (JEC21), and the laboratory-constructed hybrid αADα (XL1462) strains. Animal survival and fungal burden in the lungs and brains were monitored. The αADα hybrid strain is as virulent as the highly virulent Aα parental strain, based on both survival rate (Figure 9A) (p = 0.371) and organ burden of fungal cells at the time of sacrifice (Figure 9B). Animals infected with the Dα strain remained viable and showed no symptoms at the conclusion of the study (day 100). Fungal burden in animals infected with the Dα strain was considerably lower than that of animals infected with the Aα or the αADα hybrid strains. This assay indicates that the Aα and αADα strains are both much more virulent than the Dα strain, and thus hybridization with an Aα partner confers a clear benefit to the less virulent serotype D α strain. Enhanced virulence in animals is not likely to be the selective pressure that gives rise to AD hybrids, as mammalian infection is not an obligate part of the normal life cycle of this environmental pathogen, but it may reflect evolved traits that contribute to the common presence of AD hybrids in nature [80,95].
The same-sex mating process has been observed under laboratory conditions [48] and is hypothesized to occur in nature given that C. neoformans has a largely unisexual population and the α mating type predominates in both clinical and environmental isolates. Population genetic studies also provide evidence that same-sex mating occurs in nature. For example, the Vancouver Island outbreak C. gattii strains are hypothesized to descend from two α parental strains [55], and serotype A strains from Sydney, Australia, show evidence of recombination in a unisexual α population (D. Carter, personal communication). However, direct evidence for naturally occurring same-sex mating is lacking, probably because of the difficulty of observing this process in nature.
By characterizing naturally occurring αADα hybrid strains, we present here conclusive evidence for the cell–cell fusion step in the same-sex mating process. Because of genetic divergence, hybrids have an impaired ability to undergo meiosis and remain in a diploid state where both parental genomes, including the MAT locus, are largely intact. These natural αADα hybrids have α mating type alleles from two parents of different serotypes that can be distinguished by serotype- and mating-type-specific PCR, CGH, and sequencing. All mating type genes (>20) of both serotype A α and serotype D α alleles are present in the AD hybrid, based on CGH, with the exception of the SXI1Dα gene, which bears a unique C-terminal deletion. The fact that αADα hybrids have been found in both the US and Italy suggests either that the same-sex mating process is not restricted to a specific geographic location or that αADα strains clonally expanded and dispersed. Additional AD hybrids of this nature likely remain to be recognized, as the Italian αADα hybrid strain was originally classified as a haploid Dα strain. It can be difficult to recognize αADα strains because (1) ploidy analysis of strains is not a common laboratory practice, (2) αADα hybrid strains mate as α strains in mating assays and thus do not behave like aADα or αADa hybrids, which are sterile or self-fertile, and (3) many AD hybrids are not recognized as hybrid strains by the serotype agglutination test commonly used in ecological and epidemiological studies [32,37–39].
Evidence has been presented to advance the hypothesis that some MAT homozygous isolates (α/α or a/a diploids) arise via a post-meiotic nuclear fusion event following a–α mating [96]. It is possible that a post-meiotic nuclear fusion event could generate a/a, a/α, and α/α diploid nuclei that are packaged into spores, generating MAT homozygous and MAT heterozygous diploid isolates, as originally proposed by Sia et al. [97]. However, post-meiotic nuclear fusion following a–α mating seems an unlikely explanation for the αADα isolates described here. First, only αADα, and no aADa, MAT homozygous strains were observed. Second, the αADα isolates descend from two α parents of divergent lineages and as a consequence inherited two very divergent alleles of the MATα locus, in contrast to what would be expected for the post-meiotic fusion model, in which the MATα locus alleles would be strictly identical by descent. Third, the genetic distance between serotype A and D isolates precludes efficient meiosis and sporulation, limiting the routes by which the unusual αADα isolates could have arisen. The most parsimonious hypothesis as to the origin of the αADα diploids is same-sex mating between haploid Aα and Dα parents, and further study of the origins of other MAT homozygous strains (αAAα and αDDα) is warranted. We hypothesize that such isolates may also have arisen via same-sex mating, based on the findings presented here with respect to αADα isolates.
This study provides evidence for the first step in same-sex mating: cell–cell fusion. The natural conditions that stimulate cell–cell fusion events during same-sex mating are still unknown and require further investigation. Furthermore, the current study could not address meiotic reduction of the α/α diploids because none of these isolates was self-fertile under laboratory conditions. Meiosis is similarly precluded in many aADα and αADa hybrids. Only a minority of aADα and αADa hybrids were reported to be self-fertile in a previous study, and only one was observed to produce spores, which germinated poorly (<5%), reflecting a meiotic defect [57]. The extensive DNA divergence between the two serotypes likely triggers a mismatch-repair-system-evoked block to recombination, similar to that in interspecies hybrids in bacteria and budding yeasts [98–101]. In this sense, AD hybrids likely represent a genetic dead end as they cannot complete a normal sexual cycle. They are therefore a source of diversity, but not the source of diversity for the haploid population. While providing direct evidence for α–α same-sex mating in nature, the challenge remains to provide evidence for completion of the α–α sexual cycle, including meiotic reduction and sporulation. This will necessarily entail further studies with natural α/α diploid strains of one serotype (αAAα, αDDα, or αBBα), as the molecular differences are more subtle within each serotype, allowing meiosis. Detailed investigation of such isolates, as has been conducted for laboratory-generated αDDα hybrids [48], will provide insights on the complete same-sex mating cycle as it may occur in nature.
A common feature of the aADα and αADα hybrid isolates is that they all bear a C-terminal deletion in the SXI1Dα gene. Selection for this allele likely occurred prior to the cell–cell fusion events that produced these hybrid strains. Because all of the aADα and αADα hybrids tested bear the same truncated SXI1Dα allele whereas it is uncommon in haploid serotype D α isolates (∼13%), we favor the hypothesis that this allele enhances the fertility of Dα isolates. This interpretation is further supported by the observation that, unlike a complete deletion of the SXI1α gene, the C-terminally truncated SXI1α is still functional and Dα strains with this allele mate robustly and undergo meiosis normally. This hypothesis is also supported by the observation that C. gattii VGIII strains with a similar shortened version of SXI1α caused by a premature C-terminal stop codon also mate robustly. Because the VGIII group includes most of the fertile C. gattii strains characterized thus far, this shortened allele of SXI1α may also be associated with increased fertility. However, cell fusion between a transgenic strain with only the C-terminally truncated SXI1Dα allele in the JEC21 background and Da or Aa partners was not enhanced compared to wild-type under the laboratory conditions tested thus far (data not shown). It is thus not clear if this truncated allele of SXI1Dα directly promotes the cell fusion step of mating, or is linked to another causative mutation in the MAT locus that was not detected in our study. Another possibility is that the effect of C-terminal truncation of SXI1α is genotype specific and mediated in concert with other unlinked mutations, similar to the observation that the role of the mating type locus in virulence is dependent on genetic background and functions as a quantitative trait locus [49,102]. Alternatively, laboratory conditions may not recapitulate the natural environment where cell fusion and mating occur (pH, temperature, nitrogen source, nutrient, and presence or absence of small molecules such as inositol and auxin indole-3-acetic acid [103]). The efficiency of cell fusion varies considerably depending on the isolate and mating medium (unpublished data). The last and, in our view, least likely possibility is that these alterations in the SXI1α gene are neutral variants, and by chance C-terminal truncation and the premature stop codon arose independently in the original ancestors of both the aADα and αADα hybrid populations (the founder Dα strains) and the C. gattii VGIII strains. Our study demonstrates the complexity and diversity of the life cycles of C. neoformans and indicates that hybridization is influenced by both environmental and genetic factors.
Hybridization between two serotypes may have consequences for pathogenesis, as new strains with altered virulence may arise. The fact that AD hybrids occur at a reasonable frequency in both clinical and environmental samples is possibly indicative of hybrid fitness and an impact of hybridization on C. neoformans infection [33,35,61]. To test the effect of hybridization on virulence, yet avoid variations caused by natural genotypic differences and potential complications from the presence of both mating types, αADα hybrid strains were constructed in defined genetic backgrounds (H99 and JEC21), for which complete genome sequences are available and which are widely used in genetic and pathogenesis studies. The constructed αADα hybrid exhibited hybrid vigor under defined conditions, such as growth at high temperature (39 °C) and resistance to UV irradiation. The hybridization effect on melanization is complex and is affected by growth temperature. In most aspects tested in vitro, the αADα hybrid and Aα strains exhibited enhanced fitness compared to the less virulent Dα parental strain. Virulence tests in a murine inhalation model showed that the constructed αADα hybrid is similar in virulence to the Aα parental strain, while the Dα parental strain is almost avirulent. Overall, these observations support the hypothesis that hybridization between serotype A and D enhances the ability of the less virulent serotype D strains to survive both in the environment and in the host. Similar hybrid vigor (UV resistance and tolerance to high temperature) has also been observed in natural aADα hybrids, and the increased fitness of these hybrids is hypothesized to have contributed to their worldwide distribution, whereas the parental Aa strains are geographically restricted to Africa [59].
Our findings provide definitive evidence that C. neoformans can undergo same-sex mating in nature. However, a limitation is that natural αADα hybrids have an impaired ability to undergo meiosis and fail to produce haploid progeny, precluding further evaluation with these isolates of the impact of this life style on the haploid population structure and evolution of the C. neoformans species complex. The hybrid vigor displayed by the laboratory-constructed αADα strain, both in vitro and in vivo, offers a plausible explanation for the common presence of hybrids in clinical and environmental isolates. Whether AD hybrids are a source of diversity, are en route to speciation, or are a genetic dead end requires further investigation.
The unique α–α unisexual mating cycle that C. neoformans can adopt reflects either an adaptation to the sharply skewed distribution of mating types, or a route by which this disparity arose. It may maximize the advantages of both outcrossing and selfing in this heterothallic fungus that has a largely unisexual population. Similar strategies may also occur in other fungal species. For example, the obligate human fungal pathogen, Pneumocystis carinii, may share a similar life cycle. P. carinii is hypothesized to undergo both asexual and sexual cycles, based on cytological studies [104–106]. Only one mating-type-like region is known in this fungus, and there is no evidence of mating type switching [107]. The life cycle of the filamentous hemiascomycetous fungus Ashbya gossypii may also involve fusion of cells or nuclei of like mating type that then undergo meiosis and sporulation, as only the a allele of the MAT locus has been identified thus far for this species [18,108]. Similar inbreeding/selfing reproductive strategies have evolved in other kingdoms. For example, in plants that normally outcross, pseudo-self-compatibility in older flowers allows self-pollenization by a breakdown of self-incompatiblity barriers [109,110], which is conceptually similar to the ability of a heterothallic fungus to engage in same-sex mating. The first gene underlying pseudo-self-compatibility, the S-locus-linked gene PUB8 (the S locus in plants is functionally similar to the mating type locus in fungi or the sex chromosomes in animals), was recently identified [111]. In many insects, parthenogenesis is also conceptually reminiscent of same-sex mating in fungi. Reproduction in parthenogenic strains of the bisexual species Drosophila mercatorum is also analogous to same-sex mating in the heterothallic species C. neoformans [112]. It has been shown in grasshoppers that parthenogenic species can generate a level of variability similar to that in closely related sexual species [113]. Similarly, same-sex mating in C. neoformans could potentially contribute significantly to genetic variation in the largely unisexual fungal population. Elucidating how this life cycle occurs in the genetically tractable fungus C. neoformans, its underlying molecular mechanisms, and its impact on population structure will shed light on similar reproductive strategies occurring in other species.
The congenic strains JEC21 (α), JEC20 (a), H99, and KN99a were used as mating reference strains. Other strains used in this study are CHY621 (ura−, NATR) [114], sxi1Δ mutants CHY610 [60] and CHY618 (ura−, NATR) [114], XL1462 (αADα), XL1501 (αAAα), XL1620 (SXI1DαΔC), and those listed in Table 2. Cells were grown on YPD (1% yeast extract, 2% BactoPeptone, and 2% dextrose) or YNB medium (Difco). Mating or cell fusion was conducted on V8 medium (pH 7.0) in the dark at 22 °C.
To determine mating type, isolates were grown on YPD medium for 1 d at 30 °C and separately cocultured with the reference tester strains, JEC20 (MATa) and JEC21 (MATα), on V8 medium in the dark at 22 °C [78]. The isolate and tester strains were cultured alone on the same plate as controls. The mating reactions were examined after a week for mating hyphae formation, which signaled the initiation of sexual reproduction. Mating type was also determined by PCR with SXI1α, SXI2a, and STE20α/a gene primers that yield mating-type- and serotype-specific amplicons. Primers used are listed in Table 3.
Cells were processed for flow cytometry as described previously [48,97]. Briefly, cells were harvested from YPD medium, washed once in PBS buffer, and fixed in 1 ml of 70% ethanol overnight at 4 °C. Fixed cells were washed once with 1 ml of NS buffer (10 mM Tris-HCl [pH 7.6], 250 mM Sucrose, 1 mM EDTA [pH 8.0], 1 mM MgCl2, 0.1 mM CaCl2, 0.1 mM ZnCl2) and then stained with propidium iodide (10 mg/ml) in 0.2 ml of NS buffer containing RNaseA (1 mg/ml) at 4 °C for 4–16 h. Then 0.05 ml of stained cells was diluted into 2 ml of 50 mM Tris-HCl (pH 8.0) and sonicated for 1 min. Flow cytometry was performed on 10,000 cells and analyzed on the FL1 channel with a Becton-Dickinson FACScan.
Strains were grown in 50 ml of YPD medium at 30 °C overnight with shaking. The cells were washed three times with distilled water and harvested by centrifugation at 4,000g for 8 min. The cell pellet was frozen immediately at −80 °C, lyophilized overnight, and stored at −20 °C until genomic DNA was prepared using the CTAB protocol as described previously [115]. The quality of the purified DNA was examined on an agarose gel.
AFLPs were generated and analyzed as previously described [53]. Two different EcoRI primer combinations (EAC and ETG) were used for the selective PCR, as described previously [68]. Only intense and reproducible bands were scored to identify differences between strains.
Genomic DNA was extracted as described above. DNA was digested with restriction enzymes, separated in agarose gels, and blotted to nitrocellulose (Zeta-Probe, Bio-Rad) by standard methods. Probes were generated with a Prime-It II kit (Amersham). Hybridization was performed using Ultrahyb (Ambion) according to the manufacturer's instructions.
Genomic DNA was sonicated to generate ∼500-bp fragments and purified with a DNA Clean and Concentrator kit (Zymo Research). Five micrograms of DNA was used for Cy-3 dUTP or Cy-5 dUTP labeling reactions using the Random Primer/Reaction Buffer mix (BioPrime Array CGH Genomic Labeling System, Invitrogen). Hybridization conditions were as described previously [82] except that the slides contained a C. neoformans whole genome 70-mer oligonucleotide array and serotype- and mating-type-specific 70-mer oligonucleotides for genes in the MAT locus [49]. After hybridization, arrays were scanned with a GenePix 4000B scanner (Axon Instruments) and analyzed using GenePix Pro version 4.0 and BRB ArrayTools (developed by R. Simon and A. Peng Lam at the National Cancer Institute; http://linus.nci.nih.gov/BRB-ArrayTools.html).
Cells were grown on V8 medium in the dark at 22 °C. Hyphae were fixed in 3.7% formaldehyde and permeablized with 1% Triton in PBS. Nuclei were visualized by staining with DAPI (4′,6-diamidino-2-phenylindole, Sigma) as described previously [47].
PCR products of the SXI1Dα gene were generated using primers JOHE17409 (GCCGTGCAAGGGTGTAGG) and JOHE14895 (GGGCCATTGGAGGAAGCTG) and template genomic DNA from the strains tested. The PCR products were subjected to agarose gel electrophoresis to reveal different sizes of the SXI1Dα alleles in these strains.
To construct the αADα hybrid strain, the auxotrophic strains F99 (Aα ura5) and XL342 (Dα ade2) were cocultured together with strain JEC169 (Da ade2 lys1 ura5) as the pheromone donor on V8 agar medium (pH 5.0) in the dark at 22 °C. These three strains are unable to grow on minimal medium without supplementation of uracil, adenine, and uracil + adenine + lysine, respectively. After 24 h of coculture on V8 medium, cells were collected and spread on YNB minimal medium at 37 °C to select for prototrophic fusion products. Two types of fusion products were obtained: the desired diploid αADα hybrid strains and triploid Da/Aα/Dα strains, which were distinguished by mating behavior and ploidy analyzed by flow cytometry analysis. The chosen diploid αADα hybrid strains were further confirmed by mating-type- and serotype-specific PCR analyses using primers listed in Table 3.
Yeast cells were grown in YPD liquid medium overnight at 30 °C. Cells were collected by centrifugation and washed three times with sterile distilled water. Cell density was determined by absorption at 600 nm and cells were 10× serially diluted with sterile water. To examine melanin production, 3 μl of serial dilutions of cells were spotted on melanin-inducing medium containing L-DOPA (100 mg/l) [116] and incubated at 22 °C and 37 °C in the dark for 2 to 4 d. Melanization was observed as the colony developed a brown color. To analyze growth at different temperatures, cells were spotted on YPD medium and incubated at the indicated temperatures. Cell growth was assessed on days 2, 3, and 4. To determine sensitivity to UV irradiation, cells were spotted on YPD medium, air dried for 15 min, and then exposed to UV irradiation (∼48 mJ/cm2) in a Stratalinker (Stratagene) for 0, 6, or 12 s. Cells were then incubated at 22 °C, and cell growth was monitored daily from day 2 to 4. To characterize capsule production, equal numbers of C. neoformans cells were spotted on DMEM (Invitrogen) and incubated at 37 °C for 3 d. Cells were scraped from the plates, suspended in India ink, and observed microscopically. The capsule was visualized with light microscopy as a white halo surrounding the yeast cell due to exclusion of the dark ink particles.
Mice were infected essentially as previously described [117]. Groups of 4- to 8-wk-old female A/J mice (ten mice per strain) were anesthetized by intraperitoneal injection of phenobarbital (∼0.035 mg/g). Animals were infected intranasally with 5 × 104 fungal cells in 50 μl of PBS. The inocula of yeast cells were confirmed by CFU after serial dilutions. To verify strain identity for the inoculation, 100 colonies for the controls and 200 colonies for the hybrid were tested for auxotrophic markers and mating type. Three colonies for each strain were randomly chosen and checked for ploidy by fluorescent flow cytometry. Mice were monitored twice daily, and those showing signs of severe morbidity (weight loss, extension of the cerebral portion of the cranium, abnormal gait, paralysis, seizures, convulsions, or coma) were sacrificed by CO2 inhalation. The survival rates of animals were plotted against time, and p-values were calculated with the Mann–Whitney test. The lungs and brains from two animals from each group were removed, weighed, and homogenized in 2 ml of sterile PBS. Serial dilutions of the organ samples were plated on YPD agar plates containing 100 μg/ml chloramphenicol and incubated at 37 °C overnight. Randomly chosen colonies (100 for the controls and 200 for the hybrid) were tested for auxotrophic markers and mating type. Three colonies from each organ were randomly picked and checked for ploidy by fluorescent flow cytometry. Results of auxotrophic marker, mating type, and ploidy analysis of recovered strains were congruent with those for the infecting strains.
The GenBank (http://www.ncbi.nlm.nih.gov/Genbank/index.html) accession numbers for the Sxi1α sequences discussed in this paper are as follows: C. neoformans αADα hybrid (EF471284), C. neoformans wild-type strain JEC21 (AAN75718), C. neoformans wild-type strain H99 (AAN75175), C. gattii wild-type VGI strain WM276 (AAV28797), C. gattii VGII strain WM178 (DQ096309), C. gattii VGIII strain V28 (AY973651), C. gattii VGIII strain DUMC140.97 (DQ096306), C. gattii VGIII strain NIH312 (DQ096307), C. gattii VGIII strain 97/426 (DQ198312), C. gattii VGIII strain 97/433 (DQ198313), C. gattii VGIII strain 97/428 (DQ198314), C. gattii VGIII strain ICB88 (DQ198315), and C. gattii VGIII strain NIH836 (DQ198305). |
10.1371/journal.ppat.1002499 | Temporal Expression of Bacterial Proteins Instructs Host CD4 T Cell Expansion and Th17 Development | Pathogens can substantially alter gene expression within an infected host depending on metabolic or virulence requirements in different tissues, however, the effect of these alterations on host immunity are unclear. Here we visualized multiple CD4 T cell responses to temporally expressed proteins in Salmonella-infected mice. Flagellin-specific CD4 T cells expanded and contracted early, differentiated into Th1 and Th17 lineages, and were enriched in mucosal tissues after oral infection. In contrast, CD4 T cells responding to Salmonella Type-III Secretion System (TTSS) effectors steadily accumulated until bacterial clearance was achieved, primarily differentiated into Th1 cells, and were predominantly detected in systemic tissues. Thus, pathogen regulation of antigen expression plays a major role in orchestrating the expansion, differentiation, and location of antigen-specific CD4 T cells in vivo.
| Pathogens alter protein expression in an infected host, depending on metabolic or virulence requirements, but the effect of these changes on the immune response is unclear. We identified new class-II epitopes within Salmonella type-III secretion system effector proteins and generated a methodology to visualize endogenous T cells responding to these epitopes. Our study shows that Salmonella flagellin generates a mixed Th1 and Th17 response that contracts early and is enriched in mucosal tissues. In contrast, we found that Salmonella T3SS effectors generate a sustained Th1 response that requires a persisting infection and is enriched in systemic tissues. These data demonstrate that in vivo antigen regulation substantially alters the antigen specificity, helper differentiation, and anatomical location of pathogen-specific CD4 T cells.
| Generating vaccines for current and emerging infectious diseases remains an important goal of immunological research [1], [2]. An ideal vaccine will induce the expansion and maturation of naïve pathogen-specific lymphocyte clones that exist at low frequency in uninfected or unimmunized individuals [3]. During infection or immunization, pathogen-specific T cells expand within secondary lymphoid tissues after recognition of foreign peptides in the context of host MHC molecules. These dividing pathogen-specific T cells acquire effector capabilities that are tailored to combat different classes of microbial pathogen [4], [5]. Following the resolution of primary infection, a cohort of these expanded pathogen-specific T cells persists to provide robust secondary immunity against another encounter with the same pathogen [6].
Naïve CD4 T cells can differentiate into T helper 1 (Th1), Th2, or Th17 effector lineages depending on the instructional cues delivered during initial activation [7]. Th1 cells express the transcription factor T-bet, secrete IFN-γ, and protect the host against infection with intra-macrophage pathogens [8], [9]. In contrast, Th2 cells express GATA-3, secrete IL-4, and combat large extracellular pathogens [10], [11]. Th17 cells are a recently described effector cell lineage that express RORγt, secrete IL-17, and contribute to clearance of extracellular bacterial and fungal infections [5], [12]. A number of variables can influence CD4 T helper cell differentiation, including TCR affinity for peptide/MHC, antigen dose, costimulatory signals, and the local concentration of inflammatory cytokines [7].
In vivo analysis of CD4 T cell differentiation during infection has often involved visualization of adoptively transferred TCR transgenic T cells [13]–[19]. Although this experimental approach allows detection of pathogen-specific T cells using conventional methodologies, it can also introduce experimental variables that are not present in a natural host [20]–[22]. The development of peptide-MHC tetramer enrichment methodologies now allows direct visualization of low frequency antigen-specific CD4 T cell populations without requiring adoptive transfer of TCR transgenic cells [23]. Using this approach, infection via the intra-nasal route was shown to enhance development of Listeria-specific Th17 cells while systemic infection encouraged Th1 development [24]. Most pathogens differentially regulate their protein expression in mucosal or systemic tissues depending on local metabolic or virulence requirements [25], [26], but it is not yet clear how this might affect CD4 T cell expansion and differentiation to stage-specific antigens expressed in different tissues.
Here, we examined this issue using peptide-MHC tetramers and ELISPOTs that allow simultaneous tracking of CD4 T cell responses to Salmonella flagellin and type-III-secretion system (TTSS) effector proteins. We found that flagellin-specific CD4 T cells undergo early expansion and contraction and after oral infection generate Th17 cells that localize to infected mucosal tissues. In contrast, CD4 T cells responding to TTSS effector proteins expanded and accumulated over several weeks and primarily differentiated to a Th1 lineage localized in systemic tissues. Thus, an infected host simultaneously develops distinct populations of CD4 effector lineages that detect stage-specific antigens and have divergent anatomical localization.
I-Ab epitopes have been identified within Salmonella flagellin [27], [28], and these remain the only confirmed class-II epitopes for examining CD4 response to Salmonella in C57BL/6 mice [29]. However, flagellin is an unusual antigen that can be directly detected by the innate immune system, and is also rapidly down-regulated as bacteria transition to intracellular growth [30]–[32]. In order to uncover antigens that are recognized by CD4 T cells during intracellular bacterial growth, we used a bioinformatic approach to identify novel Salmonella I-Ab epitopes. Based on the published structure of I-Ab in complex with human class-II invariant chain peptide (CLIP) [33], we used a positional-specific scoring matrix (PSSM) and Hidden Markov Model (HMM) to interrogate all open reading frames in the Salmonella genome for I-Ab epitopes. Although this yielded numerous epitopes that were then confirmed by peptide immunization (Table 1 and Figure 1A), none were found to be natural epitopes during murine Salmonella infection or after immunization with heat-killed Salmonella (data not shown). Since many of these candidate peptides were derived from cytoplasmic proteins that are unlikely to enter the class-II presentation pathway, we modified our approach to focus on outer membrane or secreted Salmonella proteins, (Table 2), many of which are required for bacterial persistence in vivo [34]. Using this strategy we identified natural I-Ab epitopes in two effector proteins that are encoded by Salmonella Pathogenicity Island 2 (SPI2) TTSS, SseI and SseJ (Figure 1B and Table 3).
In marked contrast to flagellin, SseI and SseJ are though to be expressed during intra-macrophage replication [35]. We confirmed the differential regulation of flagellin (FliC) and SseJ by examining bacterial mRNA expression in vitro. While FliC mRNA was highly expressed under SPI1-inducing conditions, it was down-regulated under SPI2-inducing conditions which simulate the intra-macrophage environment (Figure 2). In contrast, SseJ was expressed under SPI2-inducing conditions but down-regulated under SPI1-inducing conditions (Figure 2). Thus, Salmonella differentially regulate the production of flagellin and SseJ, depending on local environmental cues. Although it was technically challenging to measure Salmonella mRNA expression in vivo, we were able to confirm the differential regulation of FliC and SseJ at early time points after infection. FliC mRNA was highly expressed at 30 minutes after infection but significantly reduced at 5 hours, while in contrast, SseJ mRNA was low at 5 minutes but expression increased at 30 minutes (Figure 2). Thus, flagellin (FliC) and SseJ represent antigens that are differentially regulated during the transition from extracellular to intracellular growth.
We constructed class-II I-Ab tetramers containing flagellin427–441 and SseJ329–341 epitopes and visualized endogenous flagellin427–441-, SseJ329–341-specific CD4 T cells using a sensitive tetramer enrichment methodology [36]. C57BL/6 mice contained low numbers of naive flagellin- and SseJ-specific CD4 T cells, but a population of expanded CD44Hi tetramer positive cells was readily detected in the draining lymph nodes after immunization with flagellin427–441 or SseJ329–341 and CFA (Figure 3). Similarly, infection with Salmonella (BRD509) allowed detection of an expanded population of flagellin- and SseJ-specific CD4 T cells (Figure 3). Salmonella-specific T cells were not detected in unbound column fractions or in CD8 T cells bound to enrichment columns (Figure 3), demonstrating the specificity of these detection regents and the efficiency of enrichment. We chose to focus on visualizing CD4 T cells responding to Salmonella strain BRD509 since this strain has been widely studied previously [37]–[39], and use of more virulent strains precludes analysis at late time points after infection of C57BL/6 mice.
We examined the kinetics of flagellin427–441-specific and SseJ329–341-specific CD4 T cell expansion after intravenous (IV) infection with Salmonella. The pooled secondary lymphoid tissues of a C57BL/6 mouse contained approximately 32 flagellin427–441-specific and 30 SseJ329–341-specific CD4 T cells (Figure 4A). After Salmonella infection, flagellin427–441-specific CD4 T cells expanded to a peak of 410 cells at day 7 (12.9 fold expansion over naïve frequency), before contracting to 88 cells by day 160 (Figure 4A). At the peak of clonal expansion, flagellin427–441-specific CD4 T cells decreased surface expression of CCR7 and CD27, indicating the development of T effector cells (Figure 4B and C). Flagellin427–441-specific T cells started to contract as early as day 10 following infection, and this coincided with a gradual increase in the percentage of cells expressing CCR7 and CD27 (Figure 4A–C). Importantly, this CD4 contraction phase occurred while bacterial burdens remained high in vivo (Figure 4D). In marked contrast, SseJ329–341-specific CD4 T cells expanded after infection and continued to accumulate in secondary lymphoid tissues until day 52 post-infection, eventually reaching a peak of 5,900 cells (210-fold expansion over naïve frequency), before decreasing to 1,800 cells by day 160 (Figure 4A). In contrast to the flagellin-specific response, SseJ329–341-specific CD4 T cells maintained low expression of CCR7 and CD27 until bacterial clearance had been achieved (Figure 4B–D). We also examined intracellular expression of transcription factors associated with T helper lineage commitment in these two populations at time points close to peak expansion. The vast majority of expanded flagellin427–441-specific and SseJ329–341-specific CD4 T cells expressed T-bet, while no staining above background levels was detected using antibodies specific for GATA-3, FoxP3, or RORγt (Figure 4E and data not shown).
In order to confirm the distinct tempo of CD4 responses to Salmonella flagellin and TTSS effector proteins, we also examined CD4 T cell responses by ELISPOT. IFN-γ-producing CD4 T cells were detected seven days after Salmonella infection and were found to predominantly focus on flagellin, rather than TTSS effector epitopes (Figure 5). In marked contrast, 40 days after infection, larger pools of IFN-γ-producing CD4 T cells were detected responding to SseI and SseJ epitopes rather than flagellin (Figure 5). No Th2 or Th17 CD4 T cell response was detected at either of these time points after intravenous infection (data not shown). Together, these data demonstrate distinct temporal differences in the targeting of flagellin and TTSS effector proteins by Salmonella-specific CD4 T cells.
We hypothesized that the temporal difference in the CD4 T cell response to flagellin and SseJ was due to the maintenance of SseJ expression in vivo by persistent bacteria. However, it was also possible that these different kinetics reflected T cell-intrinsic variables that were completely unrelated to antigen persistence. We therefore examined the clonal expansion of SseJ-specific T cells in mice that were administered antibiotics to clear bacteria, beginning five days after infection. Although SseJ-specific T cells were clearly detected in both treated and untreated mice, clonal expansion was significantly reduced in mice that had been treated with antibiotics (Figure 6). Thus, the continued accumulation of SseJ-specific CD4 T cells at later time points after infection required bacterial persistence.
Next, we used ELISPOTs to examine the CD4 T cell response to oral Salmonella infection and closely monitored Th1, Th2, and Th17 responses to flagellin and TTSS effector proteins in mucosal and systemic tissues. We used three oral doses of Salmonella (BRD509) as this was required for optimal priming of effector CD4 T cell responses. Consistent with the absence of Th2 responses after IV infection, IL-4 production was not detected in either mucosal or systemic tissues (data not shown). In contrast, IFN-γ-producing CD4 T cells were detected in intestinal (mesenteric lymph nodes, MLNs, and lamina propria, LP) and systemic (spleen and liver) tissues (Figure 7A). SseI and J-specific CD4 Th1 cells were more numerous in the spleen and liver, while flagellin-specific CD4 Th1 cells were predominantly detected in intestinal tissues (Figure 7A). Furthermore, IL-17A-producing CD4 T cells were also detected and this response was notably focused on the flagellin epitopes and occurred in mucosal tissues (Figure 7B). Indeed, flagellin-specific CD4 T cells in intestinal tissues had a higher ratio of Th17:Th1 cells, while in contrast SseI and J-specific CD4 T cells in systemic tissues had a higher ratio of Th1:Th17 cells (Figure 7C). In order to confirm the preferential generation of flagellin-specific Th17 cells in the intestine, we also examined IL-22 production by CD4 T cells recovered from the spleen or MLN of Salmonella-infected mice. Consistent with the ELISPOT data, IL-22 was detected after stimulation of CD4 T cells from the MLN but not the spleen of infected mice, and was directed towards flagellin427–441 rather than T3SS effectors (Figure 8).
This study describes the first examination of endogenous CD4 T cell responses to multiple Salmonella epitopes using tetramer and ELISPOT assays. Previous analysis of Salmonella-specific CD4 T cells has been severely limited by the lack of defined epitopes and has largely focused on the response to a single I-Ab epitope within flagellin [13], [40], [41]. Our ability to track multiple epitope-specific CD4 responses in this study has revealed that in vivo regulation of bacterial antigen expression is a critical factor in orchestrating both the tempo and effector maturation of Salmonella-specific CD4 T cells at mucosal and systemic sites. High expression of flagellin is reported from extracellular bacteria but this is rapidly decreased after infection of macrophages, and flagelin cannot be detected in the spleen of infected mice [32], [42]. In contrast, expression of SPI-2 genes is initiated within the phagolysosome of infected macrophages and is required for bacterial survival in vivo [35]. Our data confirm these findings and demonstrate differential expression of FliC and SseJ mRNA in response to environmental cues, both in vitro and in vivo. Thus, naïve Salmonella-specific CD4 T cells will encounter flagellin only during the early stage of infection within the intestine, whereas SPI-2 TTSS effector proteins accumulate as more intracellular bacteria replicate in the systemic tissues of the spleen and liver.
Interestingly, our tetramer data reveal a pattern of CD4 T cell expansion and contraction to flagellin and SseJ that corresponds closely to their relative temporal abundance during in vivo growth. In particular, we detected early contraction of flagellin-specific T cells, while SseJ-specific T cell responses remained elevated for several weeks. There may be T cell-intrinsic factors that determine the peak clonal expansion of FliC- and SseJ-specific T cells since SseJ-specific T cells exhibit greater clonal expansion. However, this issue does not explain the maintenance of the SseJ response for several weeks after infection. Furthermore, the elimination of bacteria using antibiotics severely dampened the CD4 T cell response to SseJ, indicating that antigen persistence is essential for sustained clonal expansion of this population. Although antibiotics can affect adaptive response indirectly via effects on commensal flora, it is more likely that the reduction in the SseJ response is due to a direct effect on bacterial persistence. Together, these data suggest that CD4 T cells are surprisingly dependent upon local antigen availability even after clonal expansion has occurred and indicate that clearer definition of in vivo protein expression by pathogens might be critical for identification of protective antigens, especially for pathogens that reside in multiple tissues or have distinct life-cycle stages. Indeed, a correlation between temporal antigen expression and in vivo activation of protective CD4 T cell responses has recently been observed during M. tuberculosis infection [43], [44]. Furthermore, a requirement for sustained antigen expression for the generation of optimal CD4 T cell responses has also been identified in other model systems [45], [46].
Our data also demonstrate divergent CD4 T cell helper development to flagellin and TTSS effector proteins. After intravenous infection, CD4 T cells responding to each of these antigens expressed T-bet and produced IFN-γ following in vitro restimulation, thus Salmonella generates distinct antigen-specific Th1 cell populations that appear at different stages of infection. This finding may help explain why Th1 cells can be detected within days of Salmonella infection but do not appear to actively participate in bacterial clearance until several week later [47]. If the majority of the early Salmonella-specific Th1 response is largely focused on antigens that are transiently expressed, such as flagellin or SPI-1 gene products, then intracellular bacteria can evade detection by the initial wave of Th1 effector cells. In contrast, a second wave of Th1 cells that are specific for highly expressed intracellular antigens such as T3SS effectors would be more likely to participate in the later stages of bacterial clearance from systemic tissues. A similar model of bacterial antigen regulation and effector T cell evasion has been proposed in mycobacterial infection [43], [48], and may be a common feature of persistent bacterial infection.
After oral infection with Salmonella, we detected Th17 cells specific for flagellin epitopes and these were particularly enriched in the intestine. TGF-β is abundant in mucosal tissues and plays an important role in the generation of both Treg and Th17 cells [5]. Intestinal IL-6 production is induced by Salmonella infection and is likely to drive Salmonella-specific differentiation towards a Th17 lineage in the intestine [49], [50]. Indeed, flagellin is a ligand for TLR5 and can directly induce IL-6 production in vitro and in vivo [38], [51]. This may explain why Th17 cells specific for flagellin epitopes found only in infected intestinal tissues while Th17 responses to T3SS effector proteins are largely absent. Th17 cells are thought to provide protective immunity during extracellular bacterial infections via upregulation of anti-bacterial mediators and neutrophil recruitment [5], [52]. Since Salmonella grow intracellularly in vivo [53], a protective role for Salmonella-specific Th17 CD4 T cells during primary infection is not immediately obvious. However, Th17 cells may be vital for mobilizing local innate defenses against secondary infection after bacteria penetrate intestinal epithelium. Indeed, it may be particularly important that these mucosal Th17 cells recognize bacterial antigens such as flagellin that are expressed by the initial bacterial burden.
In summary, this study has identified new Salmonella I-Ab epitopes in T3SS effectors and used peptide-MHC tetramers and ELIPOT assays to simultaneously track multiple CD4 T cell responses during Salmonella infection for the first time. These data show that an infected host develops distinct CD4 effector lineages that are tightly regulated by bacterial antigen expression patterns in vivo, a finding that may assist the generation of vaccines directed against microbial pathogens that replicate in mucosal and systemic tissues.
This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The University of Minnesota and University of California Davis are accredited by the Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC). All animal experiments were approved by University of Minnesota Institutional Animal Care and Use Committee (IACUC) (Protocol numbers: 1011A93173 and 1004A80157) or UC Davis IACUC Protocol 16612.
C57BL/6 and 129Sv mice were purchased from the National Cancer Institute (Frederick, MD) and used at 6–16 weeks of age. All mice were maintained in accordance with University of Minnesota or University of California Davis Research Animal Resource guidelines. Salmonella enterica serovar Typhimurium strains SL1344 and BRD509 (AroA−D−) were kindly provided by Dr. D. Xu (University of Glasgow, Glasgow, U.K).
S. Typhimurium were cultured overnight in Luria-Bertani (LB) broth without shaking, and diluted in PBS after estimation of bacterial concentration using a spectrophotometer. Mice were infected intravenously in the lateral tail vein with 5X105 BRD509. For oral infection, mice were administered 0.1ml of 5% sodium bicarbonate to neutralize stomach pH before oral infection with 5x109 BRD509. The actual bacterial dose administered was confirmed by plating serial dilutions of the original culture onto MacConkey agar plates counting the number of colonies that grew after overnight culture at 37°C. In some experiments mice were administered Enrofloxacin (Baytril) at 2mg/ml in their drinking water, beginning on 5 days post-infection, as previously described [54].
The crystal structure of the murine I-Ab molecule in complex with human class II invariant chain-associated peptide (CLIP) [33], was used to identify potential I-Ab epitopes from S. Typhimurium. Based on the proposed alignment, the likelihood of a particular amino acid residue appearing in a specific position in the MHC class-II binding groove was assigned a numerical score. We focused on binding positions 1, 4, 6, 7, and 9 as these were the most likely to contribute to MHC class-II binding. For example, a tyrosine at position 1 was found in 30 out of 75 core residues examined and was assigned a score of 30. An overall score total was calculated by adding individual scores from each potential binding residue, with 109 being the highest possible score. This scoring approach was applied bioinformatically to each open reading frame of Salmonella using 9-mers overlapping by one (a total of 995,592 peptides). Two algorithms were then applied to this data; the first was a position-specific scoring matrix (PSSM) [55], and the second a hidden Markov model (HMM) [56]. The epitopes with highest PSSM (25 epitopes from 105 to 84) and HMM scores (19 epitopes from 82 to 43) were tested for immunogenicity using an IFN-γ ELISPOT assay. Briefly, four groups of eleven peptides were mixed in equal amounts (5 mg/peptide) emulsified in complete Freund's adjuvant (CFA) and injected subcutaneously into groups of C57BL/6 mice. Draining lymph nodes were harvested eight days later, CD4 T cells enriched by negative selection, and mixed 1∶1 with irradiated splenocytes in IFN-γ ELISPOT plates. Duplicate wells were stimulated for 36 hours with individual peptides and spot forming cells (SFC) per 1x106 total cells enumerated. Although 88% (PSSM) and 74% (HMM) of peptides induced a CD4 T cell response in HKST/CFA-immunized mice, no response to these epitopes was detected after infection with Salmonella. The approach was therefore modified using PSSM to identify the two highest scoring candidate peptides from a combination of 13 outer membrane and secreted proteins. These epitopes were tested using purified CD4 T cells from the spleen and mesenteric lymph nodes of 129Sv mice orally infected 11 weeks previously with 5x107 Salmonella (SL1344). In these peptide identification experiments, genetically resistant 129Sv mice were used so that CD4 responses to virulent (SL1344) bacteria could be examined at late time points when C57BL/6 mice would normally succumb to infection.
For SPI1 or SPI2-inducing cultures, Salmonella BRD509 were grown overnight at 37°C in LB broth, diluted in either LB broth for SPI1-inducing conditions or modified N minimal medium [57] for SPI2-inducing conditions, and grown for either 3 hours for SPI1 or 6 hours for SPI2 conditions. For in vivo detection of mRNA expression, C57BL/6 mice were infected intravenously with 5X105 BRD509 and spleens were harvested at either 30 minutes or 5 hours after infection. Total RNA was isolated using Aurum Total RNA Mini Kit (Bio-Rad, Hercules, CA) or TRI reagent (Applied Biosystems/Ambion, Austin, TX) according to the manufacturer's protocol. To enrich bacterial RNA from infected splenocytes, the MICROBEnrich kit (Ambion) was used, and traces of genomic DNA eliminated using DNA-free DNase treatment kit (Ambion) according to manufacturer's recommendations. To quantitate expression of Salmonella Flagellin (FliC) or SseJ by real-time qPCR, 1 µg of bacterial RNA was used as a template and cDNA synthesis by random hexamers was performed using TaqMan reverse transcription reagents (Applied Biosystems). SYBR Green (Applied Biosystems) based real-time qPCR was carried out based on published procedures [58]. The following primers were used in this study; 16S rRNA, 5′-tgttgtggttaataaccgca-3′ and 5′-gactaccagggtatctaatcc-3′ [59]; FliC, 5′-gtaacgctaacgacggtatc-3′ and 5′-atttcagcctggatggagtc-3′ [58]; SseJ, 5′-tattacgagactgccgatgc-3′ and 5′-gcccgtggtgagtataagggt-3′ (GeneScript's real-time PCR primer design tool). Data was acquired using a ViiA 7 Real-time PCR system (Applied Biosystems) and analyzed using comparative Ct method (Applied Biosystems). Transcripts of Flagellin (FliC) or SseJ were normalized to the respective amounts of 16S rRNA in each sample.
Endogenous Salmonella specific flagellin427–441 - or SseJ329–341 -CD4 T cell responses were monitored using a previously described methodology. Biotinylated I-Ab monomers with an attached peptide derived from Salmonella flagellin427–441 (Flagellin:I-Ab) or SseJ329–341 (SseJ:I-Ab) were produced in S2 Drosophila insect cells cultured using a Wave Bioreactor (GE Healthcare Biosciences, Pittsburgh, PA). Purified monomers were tetramerized using fluorochrome-conjugated streptavidin and batch-tested for optimal binding to CD4 T cells from day 7 peptide-immunized mice, as previously described [36]. Briefly, C57BL/6 mice were immunized with either 100 µg of flagellin427–441 or SseJ329–341 peptide in the presence of CFA (Sigma-Aldrich, St. Louis, MO). On day 7, the inguinal, axillary, and brachial lymph nodes were harvested from naïve or peptide-immunized mice and stained with respective class-II tetramers in the presence of Fc block (culture supernatant from the 24G2 hybridoma, 2% mouse serum, 2% rat serum, and 0.01% sodium azide). Cells were incubated with anti-fluorochrome microbeads (Stem cell technologies, Vancouver, Canada) and tetramer-specific cells enriched using a magnet. Bound and unbound fractions were stained with fluorochrome-conjugated Abs specific for CD3, CD4, CD8, CD11b, CD11c, B220, F4/80, CCR7, and CD27 (eBioscience, or BD Bioscience, San Diego, CA). For transcription factor staining, cells were surface stained, and then treated with Foxp3 staining buffer set (eBioscience). Permeabilized cells were stained with fluorochrome-conjugated Abs specific for T-bet, GATA3, RORγT, or Foxp3. Cells were then analyzed by flow cytometry using an LSR II (BD Bioscience) or FACS Canto (BD Bioscience). Endogenous tetramer specific CD4 T cells were identified using a previously described gating strategy [36]. Tetramer specificity was confirmed by the absence of tetramer binding CD8 T cells in the bound fraction and in CD4 T cells within the unbound fraction. All the data were analyzed using FlowJo software (Tree Star, San Carlos, CA).
Spleen and MLN, including Peyer's patches, were crushed through nylon mesh, and then RBC in spleen was lysed with ACK lysis buffer (Lonza, Walkersville, MD). For nonlymphoid tissue processing, we followed published methods [60]. Livers were perfused, crushed through a cell strainer (BD Biosciences), and resuspended in 35% Percoll (Sigma-Aldrich). After centrifugation, pelleted cells were treated with ACK lysis buffer and then washed with 2% FBS/PBS. For Lamina propria (LP) lymphocytes, small intestine was dissected, and washed with 2% FBS/CMF (Ca2+ and Mg2+ free HBSS with 1mM HEPES, and 2.5mM NaHCO3). The intestine pieces were then stirred at 37°C for 30 min in CMF containing 10% FBS and 1mM dithioerythritol. The epithelial cells were removed by stirring the pieces in 1.3 mM EDTA/HBSS solution at 37°C for 30 min. The intestinal tissues were then treated with collagenase (Invitrogen, Carlsbad, CA) in RPMI 1640 (with 1mM CaCl2, 1mM MgCl2, and 5% FBS) at 37°C for 1 h. Cells were washed in 2% FBS/PBS, and then centrifugated on a 44%/67% Percoll gradient (GE Healthcare Biosciences). Viable cells at the interface were collected, and these cells are LP lymphocytes.
MultiscreenHTS ELISPOT plate (Millipore, Billerica, MA) was coated with either purified anti-mouse IFN-γ (BD Bioscience) or IL-17A (eBioscience). Total CD4 T cells from each tissue were isolated using EasySep Mouse CD4 T cell enrichment kit (Stem cell technologies) according to the manufacturer's protocol. The purity of CD4 T cells were measured by flow cytometry and was typically >90% for splenocytes and MLN, and >70% for non-lymphoid tissues. Purified CD4 T cells (105 cells/well) were added to ELISPOT plates in the presence of 4×105 cells irradiated splenocytes. Cells were restimulated with 10 µM peptides (Table 3) at 37°C overnight. Bound IFN-γ or IL-17A was detected by biotin-conjugated anti-mouse IFN-γ (BD Bioscience) or IL-17A (eBioscience), followed by AP-conjugated streptavidin (BD Bioscience). Bound antibodies were visualized using One-step NBT-BCIP substrate (Thermo, Rockford, IL), and plates analyzed using an ELISPOT reader ImmunoSpot (Cellular Technology, Shaker Heights, OH).
Purified CD4 T cells (1×105 cells/well) from spleens and MLN, including Peyer's patches, were restimulated with 10 µM peptides (Table 3) in the presence of 4×105 irradiated splenocytes for 16 h. IL-22 secretion was measured from culture supernatant using mouse IL-22 ELISA Ready-SET-Go kit (eBioscience). Briefly, High-protein binding plates (Costar, Corning, NY) were coated overnight with capture IL-22 antibody. After incubation in Assay Diluent for 1 hour at room temperature, plates were washed twice before culture supernatant was added. After incubation for 2 hours, plates were washed and biotin-conjugated detection IL-22 antibody added. After 1 hour, plates were washed and incubated for 30 min with avidin-HRP. Bound antibody was visualized using TMB substrate and after 15 min, the reaction stopped by adding 50 µl of 2N H2SO4. Plates were analyzed using a spectrophotometer (SpectraMax M2, Molecular Devices, Sunnyvale, CA).
Statistical differences between groups of normally distributed data were examined using Prism (GraphPad Software, La Jolla, CA). Data in each group were compared using an unpaired t test and were considered significantly different with a p<0.05.
|
10.1371/journal.pgen.1006657 | Novel genes involved in severe early-onset obesity revealed by rare copy number and sequence variants | Obesity is a multifactorial disorder with high heritability (50–75%), which is probably higher in early-onset and severe cases. Although rare monogenic forms and several genes and regions of susceptibility, including copy number variants (CNVs), have been described, the genetic causes underlying the disease still remain largely unknown. We searched for rare CNVs (>100kb in size, altering genes and present in <1/2000 population controls) in 157 Spanish children with non-syndromic early-onset obesity (EOO: body mass index >3 standard deviations above the mean at <3 years of age) using SNP array molecular karyotypes. We then performed case control studies (480 EOO cases/480 non-obese controls) with the validated CNVs and rare sequence variants (RSVs) detected by targeted resequencing of selected CNV genes (n = 14), and also studied the inheritance patterns in available first-degree relatives. A higher burden of gain-type CNVs was detected in EOO cases versus controls (OR = 1.71, p-value = 0.0358). In addition to a gain of the NPY gene in a familial case with EOO and attention deficit hyperactivity disorder, likely pathogenic CNVs included gains of glutamate receptors (GRIK1, GRM7) and the X-linked gastrin-peptide receptor (GRPR), all inherited from obese parents. Putatively functional RSVs absent in controls were also identified in EOO cases at NPY, GRIK1 and GRPR. A patient with a heterozygous deletion disrupting two contiguous and related genes, SLCO4C1 and SLCO6A1, also had a missense RSV at SLCO4C1 on the other allele, suggestive of a recessive model. The genes identified showed a clear enrichment of shared co-expression partners with known genes strongly related to obesity, reinforcing their role in the pathophysiology of the disease. Our data reveal a higher burden of rare CNVs and RSVs in several related genes in patients with EOO compared to controls, and implicate NPY, GRPR, two glutamate receptors and SLCO4C1 in highly penetrant forms of familial obesity.
| Although there is strong evidence for a high genetic component of obesity, the underlying genetic causes are largely unknown, mostly due to the highly heterogeneous nature of the disorder. In this work, we have focused on the most severe end of the spectrum, severe obesity with early-onset in childhood, which is more likely due to genetic alterations. We screened for rare copy number variation (CNV) a sample of 157 Spanish children with early-onset obesity using molecular karyotypes and then studied the genes altered by CNVs in 480 cases and 480 non-obese controls. We identified a higher burden of gain-type CNVs in cases as well as several CNVs and sequence variants that were specific of the obese population. Interestingly, the genes identified shared co-expression partners with known obesity genes. Among those, the genes encoding the neuropeptide Y (NPY), two glutamate receptors (GRIK1, GRM7), the X-linked gastrin-peptide receptor (GRPR), and the organic anion transporter (SLCO4C1) are novel obesity candidate genes that may contribute to highly penetrant forms of familial obesity.
| Early-onset overweight (body mass index [BMI] ≥ 85th percentile for age and sex) and obesity (BMI ≥ 95th percentile for age and sex) currently affects 27.8% of children in Spain (Spanish National Health Survey, 2011–2012), being the most prevalent chronic disorder in childhood and adolescence. In the United States, 17.3% of children aged 2 to 19 years are obese, 5.9% meet criteria for class 2 obesity (BMI ≥ 120% of the 95th percentile or BMI ≥ 35), and 2.1% have class 3 obesity (BMI ≥ 140% of the 95th percentile or BMI ≥ 40) [1]. Early-onset obesity (EOO) entails several comorbidities and predisposes to obesity and related diseases during adulthood, being one of the most important health problems in developed countries.
Single gene alterations with Mendelian inheritance account for less than 5% of non-syndromic cases of severe EOO [2], including mutations in the LEP (MIM 164160) or LEPR (MIM 601007) genes [3–5], as well as in MC4R (MIM 155541) [6,7] which are the most common cause of monogenic obesity. Genetic, genomic and epigenetic alterations have also been identified in syndromic forms of obesity, such as Bardet-Biedl syndrome (MIM 209900) [8], Prader-Willi syndrome (MIM 176270) [9], Beckwith-Wiedemann syndrome (MIM 130650) [10] and other rare diseases. However, obesity is generally considered a multifactorial disorder with high heritability (50–75%), probably higher in early-onset cases [11]. To date multiple studies have tried to elucidate genetic factors contributing to the etiopathogenesis of obesity, and relevant SNPs in more than 100 loci have been identified by Genome Wide Association Studies (GWAS), including those near genes such as FTO (MIM 610966), MC4R, NEGR1 (MIM 613173) or TMEM18 (MIM 613220) [12–15]. Nevertheless, the fraction of BMI variance explained by these GWAS top hits is estimated to be only around 2% [16]. Even the infinitesimal model, that combines the effect of all common autosomal SNPs, only explains ∼17% of the variance in BMI [17]. Gene-based meta-analysis of GWAS allowed the identification of regions with high allelic heterogeneity and new loci involved in obesity [18,19].
In addition, several common and rare copy number variants (CNV) contributing to the heritability of BMI and obesity have been reported, including deletions upstream of the NEGR1 gene [13], proximal and distal deletions at 16p11.2 [20], gains at 10q26.6 containing the CYP2E1 gene (MIM 124040) [21], and homozygous deletions at 11q11 encompassing olfactory receptor genes [22], among others. While several studies in large datasets led to the conclusion that common CNVs are not a major contributor [22], a significantly increased burden of rare CNVs was documented in cases of severe obesity with and without associated developmental delay [15,23]. Specifically, a significant enrichment for CNVs larger than 100 Kb and with a population frequency lower than 1% was identified in subjects with isolated severe EOO when compared to controls [15].
In the present study, we have analyzed the contribution to the phenotype of rare and common CNVs as well rare sequence variants (RSVs) in CNV-related genes in a large Spanish sample of patients with isolated severe EOO using case-control and family-based approaches, with the goal to identify novel genes involved in the pathophysiology of severe obesity.
We used a sequential strategy to identify genes potentially related to EOO through the analysis of CNVs by molecular karyotyping and subsequent mutation screening using a DNA pooled approach in a subset of selected genes. The strategy, including the samples used for each step, is summarized in Fig 1.
A total of 42 autosomal CNVs fulfilling the established criteria (>100 kb, gene containing and present in <1/2000 population individuals) were identified in 36 cases (22.9%). We detected 7 deletions and 35 gains (100.1–3,590kb in length), with 5 samples harboring more than one rearrangement (Table 1). MLPA was used for validation (42/42) and determination of the inheritance pattern (41/42): all tested rearrangements were inherited except for the larger deletion. We also detected and validated 2 additional CNVs on the X-chromosome for a total of 44 CNVs.
Clinical data about the parental phenotype was available in all but two families with CNVs. The progenitor harboring the CNV was obese (defined as BMI >30) in 21 cases (53.8%), was overweight (BMI 25–30) in 7 cases (17.9%) and had a BMI in the normal range in 11 cases (28.2%). More than half of the rare CNVs (25 of 44), 21 gains and 4 deletions, were not found in any of the 9,820 adult population controls.
In order to analyze the global burden of rare CNVs in EOO, we compared the amount, type and length of autosomal CNVs in patients (157) with respect to 500 Spanish population controls (Table 2). Rare CNVs were found in 15.8% of controls with respect to the 22.9% frequency found in patients (p = 0.053). Rare CNVs were predominantly gains in both cohorts (83.3% in EOO patients and 77.6% in controls). When the frequency of deletions and gains was analyzed separately, no differences were observed in deletion-type CNVs (3.8% in controls and 4.5% in patients), while a statistically significant difference in the frequency of gains was detected (p = 0.0358). Thus, there is a higher burden of CNVs in EOO patients due to rare gain-type CNVs.
If we consider specific CNVs as those not described in the initial 9,820 subjects used to establish the frequency of each alteration in the population, 8.2% control individuals carried a CNV fulfilling this criteria while the frequency was 14.0% in EOO patients, with this difference being statistically significant (p = 0.0422).
Regarding the co-occurrence of more than one CNV in the same subject, two or three hits were present in 3.2% patients and 1% controls. This difference was not statistically significant (p = 0.0644) likely due to the small sample size. The inheritance pattern of these alterations was established in patients; in two cases each alteration was inherited from a different parent and in the remaining three both rearrangements were inherited from the same progenitor. Case Ob_15 presented a third de novo event additionally to the two rearrangements inherited from her obese mother.
CNVs were considered to have a higher probability to be pathogenic when they were exclusive of the EOO population, co-segregated with the phenotype in the family, disrupted known genes for the disorder and/or were found in more than one case.
Nine duplications and two deletions were absent in 9,820 population controls and co-segregated with the phenotype in the family (Table 1). One of them was a 137kb gain in 7p15.3 containing a single coding gene, NPY (MIM 162640), identified in a male case (Ob_12) presenting with EOO and attention deficit hyperactivity disorder (ADHD) (Fig 2A). The CNV was inherited from the also obese mother (Fig 2B). Additional cases of severe EOO and ADHD were identified in the maternal branch of this family by report (Fig 2C), but unfortunately no additional samples or clinical data could be obtained.
Some CNVs overlapped with previously described microdeletion/microduplication syndromes (Table 1). Rearrangements partially overlapping with the critical region of the 22q11.2 distal deletion syndrome were identified in two patients. Ob_35 carried a 676kb deletion encompassing several genes including RSPH14 (MIM 605663) and GNAZ (MIM 139160), while a more proximal deletion including TOP3B (MIM 603582) was detected in Ob_34. A gain of 348kb at 1q21.1 overlapping with the region of Thrombocytopenia-Absent Radius syndrome (MIM 274000) was detected in case Ob_39; as its frequency in controls was 1/1,720 it was not included in the subset of selected CNVs.
Two CNVs fulfilling the established criteria were identified in more than one patient (Table 1). A gain of 139kb in 9q34.3 only including C9orf62 was found in three cases (Ob_1, Ob_2, Ob_3). However, the parents also carrying the CNV had either overweight or normal weight. Another gain of 106kb in 7p22.1 encompassing RNF216 (MIM 609948) and ZNF815P was identified in two cases (Ob_4, Ob_5), inherited from obese parents. We then completed the analysis of the CNVs identified in the entire sample of obese individuals (n = 480) and the Spanish adult non-obese controls (n = 480) by MLPA. All rare CNVs were patient-specific except for a second patient with a deletion at 11p15.4. None of the rare CNVs were identified among controls except for the 106kb gain at 7p22.1 that was found in 5 controls. The re-analysis of SNP array data unraveled the complexity of mapping this region due to small segmental duplications and was used to determine the real frequency of the rearrangement, which was above the established threshold of the study (1/2,000).
We also explored more common CNVs already described in association with obesity. The gain in 10q26.3 including CYP2E1 was more common in patients than in controls (6.4% vs 3.6%) as previously described [21], but did not reach significance (OR: 2.01, CI95% 0.93–4.36, p-value = 0.075). The frequency of the homozygous deletion encompassing olfactory receptors in 11q11 was 5.1% in cases, which was slightly lower than the frequency in the control cohort (6.7%). Therefore, our data did not replicate the previous findings that indicate a preferable transmission of the 11q11 deletion to obese children [22]. In addition, in this study we did not detect alterations in the 16p11.2 region, including or next to the SH2B1 (MIM 608937) gene.
All genes included and/or disrupted by CNVs found in more than one patient and/or co-segregating in a familial case were selected for sequence analysis (n = 14): SCFD2, NPY, ISM1 (MIM 615793), TASP1 (MIM 608270), GRM7, LOC401164, TRIML1, SLCO4C1, SLCO6A1, C11orf40, TRIM68 (MIM 613184), GRIK1, TOP3B and GRPR. In order to sequence the total number of patients (480) and controls (480) in a cost-efficient manner, pools of 20 DNA samples were sequenced with each DNA sample located in two pools.
We first validated the suitability and specificity of the pipeline to detect real variants among the pools. RSVs were considered when they had a frequency below 1/1.000 in the public database of the Exome Sequencing Consortium (ExAC) representing more than 60,000 exomes [24]. We selected 23 alterations predicted to be in a single sample and reanalyzed the same sample by Sanger sequencing. All 23 RSVs were validated in the specific samples.
We then compared the total burden of RSVs per gene between patients and controls. Significant differences were identified in a few loci, namely NPY, GRIK1 (MIM 138245) and GRPR (MIM 305670) (Table 3). A single missense RSV in NPY (p.V86D) was identified in patient Ob_158, while no RSVs of this gene were found in controls. Although the residue is not evolutionarily conserved and is located outside the main functional domain, the change is likely to affect the shape and the affinity of the NPY protein and has not been described in ExAC. The study of parental samples revealed that the RSV was inherited from the obese father (BMI 34.3 kg/m2). The low frequency of missense variants in this gene in the ExAC database (only 27 among 118.884 alleles) further reinforces its functional relevance.
A nonsense mutation (p.R897X) was identified in GRIK1, encoding the ionotropic glutamate receptor 1, in patient Ob_163. This nonsense variant has a frequency below 1/15000 alleles and, generally, nonsense and frameshift variants at the GRIK1 gene are rare, representing less than 1/3000 alleles in the ExAC database. Additional missense mutations were identified in both glutamate receptors (GRIK1 and GRM7 (MIM 604101)), but with no significant differences between cases and controls.
Four different missense mutations were detected in GRPR in five patients with obesity, while no mutations in this gene were found in controls (5/662 alleles vs 0/726 alleles; p = 0.0245). One of the mutations has never been found in ExAC, while the remaining three had frequencies <1/1000 alleles and all are predicted to result in significant functional consequences.
Finally, a RSV in SLCO4C1 (MIM 609013) was identified in a patient harboring a deletion encompassing the same gene previously identified by CMA. The RSV (p.I233L) has not been described previously and affects a highly conserved amino acid (phylo P = 0.975). The frequency of the deletion in the control cohort is 0/9,820. Parental studies confirmed that each progenitor had transmitted one of the alterations; the deletion was inherited from the obese father and the RSV from the non-obese mother. These findings are compatible with a recessive pattern of inheritance or a two-hit mechanism, with a major contribution of the CNV (inherited from an obese progenitor) and an additional and milder effect of the RSVs (inherited from a non-obese progenitor).
We focused our subsequent analysis on four novel candidate genes considering our CNVs and RSVs findings: GRIK1, GRM7, GRPR and SLCO4C1. To explore their possible role in obesity, we looked for co-expressions with a stringent list of 15 genes previously related with obesity. We selected a total 10 genes with described highly penetrant mutations in severely obese patients, all of them coding for proteins of the leptin-melanocortin pathway: LEP[4], LEPR[5], MC4R[6], POMC (MIM 176830) [25], PCSK1 (MIM 162150) [26], MC3R (MIM 155540) [27,28], BDNF (MIM 113505) [29], NTRK2 (MIM600456) [30], PPARG (MIM 601487) [31] and SIM1(MIM 603128) [20,32]. We also included 5 additional genes with a relevant intermediary role in the same pathway: ADRB3 (MIM 109691) [33], PCSK2 (MIM 162151) [34,35], NPY[36], NPY1R (MIM 162641) [37], AGRP (MIM 602311) [36,38] (Fig 3A).
A total of 10 shared co-expressed partners were identified in the analysis between our 4 novel strongest candidate genes (GRIK1, GRM7, GRPR and SLCO4C1) and the set of 15 obesity-related genes described in the literature. The maximum number of shared co-expressed partners between our 4 genes and 500 genes sets randomly selected was 7, being the empirical p-value of this difference 0.002 (Fig 3B and 3C).
Our results reveal a relevant contribution of rare CNVs to the etiology of severe EOO with a significantly higher burden of gain-type CNVs in patients compared to controls (p = 0.0358). Among relatively common CNVs we only detected a non-significant higher frequency of the gain in 10q26.3 containing CYP2E1 [21]. Previous studies reported a higher frequency of deletion-type CNVs in patients with severe EOO with and without developmental delay [23]. The sample studied here was stringently selected based on clinical exam and targeted genetic testing in order to exclude subjects with syndromic obesity. Thus, all patients presented isolated EOO without comorbid phenotypes such as developmental delay. This difference in the range of phenotype severity could explain the difference in the type of rearrangements found enriched in these cohorts. Although only one of the CNVs had occurred de novo, the progenitor carrying the alteration also presented overweight or obesity in 71.8% of cases, reinforcing the potential role of many of these genetic alterations in the pathophysiology of the disorder. The only de novo alteration (a 3.6Mb deletion encompassing only 3 genes) was detected in a girl with two additional rearrangements inherited from her obese mother.
We have also screened for point mutations using DNA pools in a subset of selected genes located in obese-specific CNVs [39]. Interestingly, we found additional RSVs in patients in 4 of the selected genes (NPY, GRPR, SLCO4C1 and GRIK1) reinforcing their putative role in the pathophysiology of obesity. Although the pooled DNA strategy might have some limitations such as underdetection of relatively common variants, the complete validation rate (100%) demonstrates its high specificity.
A maternally inherited gain in 7p15.3 only encompassing the NPY gene was identified in a patient with EOO and ADHD. The mother also presented severe obesity, as did several relatives from the maternal branch including two male cousins with associated ADHD. Additionally, a missense RSV also inherited from an obese progenitor was identified in another patient, while no alterations were identified in controls (480). A larger gain of approximately 3Mb on chromosome 7p15.2–15.3 encompassing NPY and other genes was previously described in all affected individuals of an extended pedigree presenting ADHD, increased BMI, and elevated NPY levels in blood [40]. Therefore, the gain encompassing only the NPY gene in patient Ob_12 and his obese mother, the point mutation in patient Ob_158 and her obese father while only 27 missense variants have been described among 118.884 alleles in ExAC are strong evidences supporting that gain of function mutations of NPY can cause severe obesity and ADHD.
NPY is a hypothalamic orexigenic peptide with neuromodulator functions in the control of energy balance and food intake. NPY is overproduced in the hypothalamus of leptin deficient ob/ob mice [41]; when depleted by genetic manipulation, ob/ob mice showed reduced food intake, increased energy expenditure and less obesity [42]. On the other hand, overexpression of NPY in noradrenergic neurons caused diet- and stress-induced gain in fat mass in a gene-dose-dependent fashion [43].
In humans, despite some conflictive reports, NPY gene variants have been significantly associated with weight changes from young adulthood to middle age and with risk of obesity [44]. NPY is widely expressed throughout the central nervous system (CNS) and a systematic review and meta-analyses of drug naïve case-control studies also suggested its implication in ADHD [45]. In addition, increased central availability of NPY by intracerebroventricular administration in male rats resulted in a shift of metabolism towards lipid storage and increased carbohydrate use, along with enhanced locomotor activity and body temperature [46].
Among other genes altered by the CNVs identified, we considered as probably pathogenic those exclusive of the EOO population that also presented exclusive RSVs co-segregating with the phenotype in the family. To further assess the possible implication of these strong candidate genes (GRIK1, GRM7, GRPR and SLCO4C1), we determined the co-expression patterns between them and 15 well-defined genes from the leptin-melanocortin pathway previously related to obesity. This analysis consistently identified a significant enrichment of co-expression shared partners among our genes and the subset of obesity related genes when compared to 500 randomly generated gene sets, reinforcing the possible role of those genes in the pathophysiology of EOO.
The alterations affecting glutamate receptors identified in two EOO patients were a partial gain of the gene encoding the ionotropic glutamate receptor GRIK1 (Ob_33) and a gain partially encompassing the gene encoding the metabotropic glutamate receptor GRM7 (Ob_22). L-glutamate is one of the main excitatory neurotransmitter in the CNS and activates both ionotropic and metabotropic glutamate receptors. A nonsense mutation was found in an additional patient in GRIK1. The metabotropic glutamate receptor 5 (mGluR5) plays a relevant role in energy balance and feeding. Adult mice lacking mGlu5 weighed significantly less than littermate controls and resisted diet-induced obesity [47]. Pharmacological approaches have described a reduction of food intake in response to antagonists of mGluR5 in a baboon model of binge-eating disorder [48] and in mGluR5+/+, but not mGluR5-/- mice [47]. On the contrary, dose-dependent stimulation of food intake has been described in rodents after injection of a mGluR5 agonist [49]. Moreover, the metabolic status and leptin can modify astrocyte-specific glutamate and glucose transporters, indicating that metabolic signals influence glutamatergic synaptic efficacy and glucose uptake [50]. Interestingly, GRM7 is likely a loss of function intolerant gene given the difference between expected and observed frequency of loss of function variants in ExAC (25 expected, 1 observed). Partial gains, depending on the location, might act as loss of function alterations when disrupting the gene. Considering these data, glutamate receptors are promising candidates in the pathophysiology of obesity.
Several alterations affecting GRPR gene were identified, including a gain encompassing the whole gene and 4 point mutations (present in 5 subjects, two males and three females) while none were found in controls. The male patients with hemizygous GRPR RSVs had inherited the variant from heterozygous mothers. Both patients had very early onset obesity in infancy presenting a quite severe phenotype at diagnosis (+5SD and +9SD respectively). One of the mothers (patient Ob_162) had a BMI within the normal range while the other presented adult-onset obesity. Thus, the phenotype of both males is more severe than the phenotype of their mothers, consistent with X-linked inheritance. GRPR encodes the receptor of gastrin-releasing peptide. Gastrin is a hormone secreted by the gastric antrum and duodenum in response to gastric distension and the presence of food in the stomach. This hormone increases the production of hydrochloric acid, pepsinogen, pancreatic secretions and bile to facilitate food digestion and also promotes satiety [51]. It is a hormone directly implicated in the regulation of food ingestion and satiety and, thus, a candidate to be associated with obesity (directly or by an alteration of a gene included in the pathway, such as GRPR).
A possible recessive pattern of inheritance or a double hit mechanism was identified in a patient who harbors a deletion partially encompassing SLCO4C1 and SLCO6A1 (MIM 613365) and a RSV in SLCO4C1, each alteration inherited from one of the progenitors. Considering that the CNV was inherited from an obese progenitor and the RSV from the non-obese mother, we postulate a major contribution of the CNV and an additional but likely milder effect of the RSVs. The SLCO4C1 belongs to the organic anion transporter family and is involved in the membrane transport of thyroid hormones, among others. Interestingly, none homozygote subjects for loss of function variants has been described in ExAC.
Other rearrangements were found in single EOO patients, including those in regions previously associated to disease, such as 22q11.2 or 1q21.1. However, the evidence to link these genomic regions to obesity susceptibility is still weak and further data will be needed.
Except for the patient with a biallelic alteration in SLCO4C1 and the de novo deletion, all CNVs and RSVs identified are heterozygous in the patients and inherited from one of the parents. Parents carrying the allele also showed an obese phenotype as well in most cases. Thus, a dominant effect (either hypo or hypermorphic) for these rare genetic variants with additive effects is suggested, leading to a more severe phenotype in the younger generation. This effect has also been found in other studies [52] and can be due to the more “obesogenic” environment that has developed in industrialized societies during the last two decades. Our results, along with previous genetic, family-based and epidemiologic studies, further indicate that EOO etiology is complex and mostly multifactorial, with the presence of some alleles that can behave as highly penetrant susceptibility variants or monogenic forms of obesity.
In summary, our findings reveal a higher burden of rare CNVs in patients with EOO compared to controls, including novel CNVs likely associated with familial obesity. Dosage sensitive genes altered by these CNVs are candidates for contributing to the pathogenesis of EOO. Some of these genes also harbor patient-specific RSVs, reinforcing their putative role in the pathophysiology of obesity. NPY, GRPR, SLCO4C1 and glutamate receptors emerge as novel candidate genes involved in monogenic familial obesity.
Criteria for severe EOO was a BMI more than three standard deviation measures above the mean for age and gender with onset earlier than 3 years of age. All cases underwent a detailed clinical examination as well as family history in search of syndromic forms of obesity, which were discarded. All studies were performed as part of a research project approved by the Medical Ethical Committee of the Hospital Infantil Universitario Niño Jesús, after receiving written informed consent from the family.
Blood samples from patients were collected. Parental blood samples were also collected in cases in which an alteration was identified. DNA from patients and parents was isolated from total blood using the Gentra Puregene Blood kit (Qiagen) according to manufacturer’s instructions. We excluded genomic and epigenetic alterations associated with pseudohypoparathyroidism (MIM 103580), Prader-Willi, Temple (MIM 616222) and Beckwith-Wiedemann syndromes with a custom-made panel (S2 Table) of Methylation Specific Multiplex Ligation Dependent-Probe Amplification (MS-MLPA) [53].
A total of 480 unrelated subjects with severe EOO were included in the study. As controls for CNV and RSV association analyses, we studied 480 adult individuals of Spanish origin with a current BMI lower than 25 and no known history of childhood obesity, obtained from the National DNA Bank from the University of Salamanca (Spain).
An initial sample of 157 probands was studied by using Omni1-Quad (64 subjects) or Omni Express SNP (93 subjects) platforms, Illumina. Copy number changes were identified using the PennCNV software with stringent filtering, as previously described [54]. CNVs encompassing known genes (RefSeq hg19), longer than 100 kb and with a frequency in control samples lower than 1/2,000 were selected. The frequency of each CNV in the control population was determined using 1M Illumina SNP array data of a total of 9,820 samples from two databases: 1) 8,329 individuals previously used as population controls for developmental anomalies [55] (81.2% of European descent, 2% African, and 16.5% other/mixed ancestry), and 2) 1,491 Spanish adult individuals from the Spanish Bladder Cancer/EPICURO study, which includes 1034 patients with urothelial cell carcinoma of the bladder and 457 hospital-based generally healthy controls with a mean age of 63.7 years [54]. To determine the frequency of CNV in the X chromosome, only the Spanish controls were considered, as data from the other cohort was not available. Given the size of the Spanish control sample (1,491), alterations in the X-chromosome absent in controls or only present in one subject were considered as rare. Briefly, a Hidden Markov Model (HMM) based on both allele frequencies and total intensity values was used to identify putative alterations, followed by manual inspection in conjunction with user guided merging of nearby (<1 Mbp between for arrays with <1 million probes and <200 kbp for arrays with >1 million probes) calls, which represent a single region broken up by the HMM, or gaps. All samples on arrays with densities <1M probes were filtered by a maximal genome-wide LogR ratio standard deviation of 0.25, while the high density 1.2 million probe WTCCC2 data was filtered using an increased standard deviation cut-off of 0.37. Mosaic alterations were excluded. For the two datasets where the Illumina array mapping corresponded to build35 (NHGRI), we utilized the autosomal calls generated previously [40] and mapped the coordinates to build36 using the UCSC LiftOver tool [56].
In order to compare the global burden of rare CNVs in patients and controls, data from 500 individuals randomly selected from the Spanish Bladder Cancer/EPICURO study and not included as controls for the CNVs frequency determination [54] were used. For the comparison, only CNVs in autosomal chromosomes with a minimum length of 100 kb, altering genes, and a frequency in control samples lower than 1/2,000 were considered (S3 Table). Alterations totally overlapping with segmental duplications were excluded to minimize biases due to the different probe coverage among microarray platforms.
An MLPA assay was designed to validate genetic alterations detected by SNP platforms and to study inheritance in families (available upon request). A total of 100 ng of genomic DNA from each sample was subject to MLPA using specific synthetic probes (sequence available upon request) designed to target the specific CNV detected. All MLPA reactions were analyzed on an ABI PRISM 3100 Genetic analyzer according to manufacturers' instructions. Each MLPA signal was normalized and compared to the corresponding peak height obtained in control samples [57]. The MLPA assay was also used to analyze the frequency of the CNVs identified in the entire cohort (480 subjects) and in the control population (480 individuals).
To study RSVs in the genes included in the CNVs, an enrichment kit was designed to capture all the coding regions of the selected genes (n = 14). The targeted enrichment was done with SeqCap EZ Choice Enrichment Kits (Roche Sequencing) and the massive sequencing with MiSeq (Illumina).
In order to sequence a high number of patients and controls (960 in total) in a cost-efficient manner, a pooled DNA approach was used [39]. Each sample was included in two different pools, and each pool contained 20 samples, avoiding two samples sharing both pools. A priori, any heterozygous RSV should be present in approximately 1 every 40 reads (2.5%). Thus, to ensure the identification of all RSVs (expected to be found in just one or few individuals) a high coverage was required.
To discriminate real variants from false positives due to extremely high coverage, we optimized the analysis pipeline. Variant calling was done with MuTect [58] to detect variants in a low proportion of reads. We also considered the quality of reads (base quality >15 in each pool) and the absence of strand bias (between 0.2 and 0.8) to define potential real variants from false positives.
To analyze the results and compare patients and controls, we focused on RSVs. We first established the frequency of each variant in the general control population using EXAC as the reference database, composed of 60,706 unrelated individuals sequenced as part of various disease-specific and population genetic studies. All alterations present in more than 1/1,000 alleles in EXAC in any of the populations included in the dataset were excluded. We specially focused on sequence changes with potential functional consequences, including loss of function variants (nonsense, frameshift and splice sites), missense variants predicted as pathogenic and changes in highly conserves residues. To search for recessive patterns of inheritance, we explored biallelic changes and RSVs that might act as second-hits in patients with previously identified CNVs.
To validate the RSVs detected by NGS and to define the segregation in each family, we designed primers to amplify an amplicon encompassing the variant and sequenced the amplicon by Sanger technology (available under request).
Using Genemania, we explored the co-expression between our candidate genes and the selected subset. To test if there was an enrichment of shared co-expressed partners, 500 sets of 15 genes with expression data available were randomly selected with Molbiotools (http://www.molbiotools.com/). For each set of genes the number of shared co-expressed partners was determined and compared with the interactions between our candidates and the set of obesity-related genes. The empirical p-value was calculated based on the fraction of shared co-expressed partners.
|
10.1371/journal.pntd.0005609 | Rabies surveillance in dogs in Lao PDR from 2010-2016 | Rabies is a fatal viral disease that continues to threaten both human and animal health in endemic countries. The Lao People’s Democratic Republic (Lao PDR) is a rabies-endemic country in which dogs are the main reservoir and continue to present health risks for both human and animals throughout the country.
Passive, laboratory–based rabies surveillance was performed for suspected cases of dog rabies in Vientiane Capital during 2010–2016 and eight additional provinces between 2015–2016 using the Direct Fluorescent Antibody Test (DFAT).
There were 284 rabies positive cases from 415 dog samples submitted for diagnosis. 257 cases were from Vientiane Capital (2010–2016) and the remaining 27 cases were submitted during 2015–2016 from Champassak (16 cases), Vientiane Province (4 cases), Xieng Kuang (3 cases), Luang Prabang (2 cases), Saravan (1 case), Saisomboun (1 case) and Bokeo (1 case). There was a significant increase in rabies cases during the dry season (p = 0.004) (November to April; i.e., <100mm of rainfall per month). No significant differences were noted between age, sex, locality of rabies cases.
The use of laboratory-based rabies surveillance is a useful method of monitoring rabies in Lao PDR and should be expanded to other provincial centers, particularly where there are active rabies control programs.
| Rabies is a viral disease that continues to threaten both human and animal health in endemic countries and almost always results in death of the infected individual. In most areas of Southeast Asia, rabies continues to be a major zoonosis with the main source of human cases being dog bites. In this study, we examined the brains of 415 rabies-suspected dogs from Vientiane Capital and some other provinces of Lao PDR during 2010–2016. Overall, rabies was confirmed in 284 cases. We found that there was a significant increase in rabies cases during the dry season (November to April) when monthly rainfall was less than <100mm. The laboratory-based surveillance of rabies in the domestic animal population is an important component of a rabies control program, however it should be used in conjunction with an effective vaccine program.
| Rabies is an acute viral encephalomyelitis caused by the rabies virus (genus Lyssavirus; family Rhabdoviridae) which can affect all warm-blooded animals including humans. Humans and animals may be infected with rabies virus via saliva through a bite or scratch [1] or butchering [2] of a rabid animal, and if left untreated it almost invariably leads to a fatal outcome. The World Health Organization estimates approximately 30,000 human deaths per year due to canine rabies in Asia out of a total 61,000 deaths per year due to rabies worldwide [3].
The Lao People’s Democratic Republic (Lao PDR) is a rabies-endemic country in which dogs are the main reservoir and continue to present health risks for both human and animals throughout the country [4, 5]. Historical reports of rabies cases in Lao PDR have been recorded in humans since 1963 [6] and rabies cases in dogs have been reported since the late 1960s [7]. It has been estimated that an average of 8,528 dog bites are reported annually in Lao PDR [8] underlining the importance of rabies surveillance and control. More recently, studies have identified the increasing importance of dog rabies in Lao PDR [4], Cambodia [9], and southern China [10], examining diagnostic issues, case numbers, epidemiology and genetic relatedness of contemporary regional strains.
Here we present results that 1) describe the number of dog rabies cases confirmed in Lao PDR from 2010–2016 using laboratory-based surveillance in Vientiane Capital, and including samples submitted from eight central, northern and southern provinces; 2) investigate aspects of dog rabies cases in Lao PDR including age, sex, location and season; 3) discuss opportunities for control of rabies in domestic animals using vaccination and identify constraints in the context of a low-resource environment.
Diagnostic testing for the presence of rabies virus was performed at the National Animal Health Laboratory (NAHL), Ban Khunta, Sikhottabong district, Vientiane Capital, Lao PDR (see Fig 1 for location).
Rabies-suspected animals originated from the nine districts of Vientiane Capital (Fig 1) during 2010–2016. From 2015–2016, samples were also submitted from Champassak, Luang Prabang, Saisomboun, Vientiane Province, Xieng Kuang, Bokeo, Saravan and Sekong provinces (Fig 2).
Samples (carcass or head) presented for diagnosis came from animals that were strongly suspected of having rabies, exhibiting furious or comatose clinical signs, with the former often resulting in a human bite incident requiring rabies virus exclusion. In these circumstances, the bite-victim, and in many cases the animal owner, delivered the animals for urgent diagnostic testing. Rabies diagnostics on carcass material was performed following verbal agreement and consent of the animal owner and the bite victim so that compensation and treatment could be determined. Necropsy procedures were performed to allow sampling of the brain and collection of hippocampal tissue for diagnostic evaluation using the rabies virus direct fluorescent antibody test (DFAT) [11]. The DFAT requires a tissue impression or smear on a microscope slide stained with a FITC anti-rabies monoclonal antibody conjugate that is specific for rabies virus (Fujirebio Diagnostics Inc., USA) which is visualized by fluorescent microscopy.
All analysis was performed using STATA 14.0 (StataCorp, College Station, TX USA). Descriptive statistics were used to characterise the number of rabies cases according to sex, age group, temporal and spatial distribution of rabies positive cases from 2010 to 2016. Significant differences (p<0.05) in rabies positive cases was determined using Pearson’s Chi-Square (Chi2) analysis. Age was compared using the groupings of <3, 3–12, >12–24, >24–36 and >36 months. Factors compared included the seasonal influence on rabies positive cases (wet season = >100mm of monthly rainfall i.e. May to October, compared to the dry season from November to April = <100mm of monthly rainfall [12]). Differences in the proportion of rabies positive samples from urban (Chanthabouly, Sisttanak, Sikhottabong and Saysettha) and rural (Naxaithong, Xaythany, Hatxaifong, Pakngum and Sangthong) districts of Vientiane Capital were also examined. Trends in the number of annual rabies cases and dog age were tested for significance (p<0.05) using the ptrend command which calculates a Chi2 statistic for trend.
Historical rabies case data from 1988–2011 was collected for comparative purposes. Data was derived from the same source being directly, or quoted as, provided by the Department of Livestock and Fisheries (DLF) of the Ministry of Agriculture, Lao PDR or by the Livestock and Veterinary Department (former title of DLF) from conference proceedings [13, 14], and a peer reviewed publication [4](Table 1). Data presented in Table 1 varies in terms of the number and provinces and species from where samples were submitted for diagnosis.
During 2010–2016, 415 dogs or specimens from dogs were submitted to NAHL for rabies diagnosis. The overall number of rabies positive cases was 284 (68.4%).
The numbers of samples submitted for rabies testing annually and the numbers of confirmed rabies cases is presented in Table 1. The median number of samples submitted annually was 57 (IQR: 51–73), with a range from 42 (2014) to 75 (2010). The median annual number of confirmed rabies cases was 36 (IQR: 32–53), range 29 (2014) to 57 (2015), and the annual proportion of rabies positive samples ranged from 56.5% (2012) to 78.1% (2015). There was no significant difference in rabies positivity rates (Chi2 = 10.96; p = 0.009) with no significant trend for the seven-year sampling period (Chi2 = 0.497; p = 0.481). The numbers of confirmed rabies cases by month from 2010–2016 are presented in Fig 3.
During the entire 2010–2016 sampling period, the preponderance of dog rabies cases came from Vientiane Capital (90.5%; 257/284) (Table 2) and of these majority were from Xanythany (37.0%; 105/284), Sikhottabong (16.2%; 46/284), Saysettha (15.1%; 43/28) and Naxaithong (7.8%; 22/284) districts. Fewer rabies cases were recorded from dogs in Hatxaifong (4.9%), Sisattanak (4.9%), Chantabouly (3.9%) and Pakngum (0.7%) districts, with no cases from Sangthong district (Table 2 and Fig 1). From 2015–2016, samples submitted from 8 provinces outside Vientiane Capital had at least one rabies case with the majority coming from the southern province of Champassak (6.6%; 16/284) and the centrally located Vientiane Province (1.4%; 4/284) (Table 2 and Fig 2).
The number of rural rabies cases was higher than that of urban districts (i.e., 166 cf 118) although this was not statistically significant (Chi2 = 0.044; p = 0.834). Focusing on rabies cases in Vientiane Capital, despite the larger number of samples from Xanythany, Sikhottabong, Saysettha and Naxaithong districts there was no statistical difference in the proportion of rabies cases between the districts (Chi2 = 12.4; p = 0.133) (Table 3). Fig 3 demonstrates annual frequency of rabies cases for the Vientiane Capital districts from 2010–2016.
There was a significant seasonal difference in the number of rabies cases (Chi2 = 8.32; p = 0.004) with a higher number of cases and higher proportion of positive cases during the dry season (n = 173; 74.3%) when compared to the wet season (n = 111; 61.0%) (Table 3). There was a trend in the increase of monthly rabies cases from the start of the dry season (November) through to the start of the wet season (May) cases over the sampling period (Fig 4). Despite no significant monthly trend in rabies positive cases from January to December (Chi2 = 1.38; p = 0.240) there was a significant decrease in rabies cases in the 12-month period May to through to April (Chi2 = 8.87; p = 0.003) (Fig 4).
There was no significant difference in the proportion of rabies positive cases between the sexes (males 62.3%; Chi2 = 2.12; p = 0.145) or age groups (Chi2 = 4.32; p = 0.363) (Table 3). There was no significant trend in rabies cases with increasing age (Chi2 = 2.17; p = 0.141). The median age of rabies-positive animals (12 months; interquartile range (IQR) 4–24) was, however, more than double the median age of the negative animals (5 months (IQR: 3–12 months).
The results presented here clearly demonstrate that dog rabies remains a significant public health problem in Lao PDR. During the period from 2010 through 2016, the number of dog rabies cases were reasonably stable, ranging from 29 to 57 cases annually. Previous studies [4, 13, 14] have reported somewhat higher annual numbers of dog rabies cases in Lao PDR between 1988 and 2009 and there was an apparent increase in the number of rabies cases during 1992 and 1996 (Table 1) although the proportion of rabies-positive cases were similar to those reported here. In the period from 2004–2011 there were 635 laboratory-confirmed animal rabies cases (51.1% of submissions) [4] in Lao PDR, and Kamsing et al. [15] (cited by [16]) stated that from 1993 to April 2008, a total of 2813 dog brain samples from central and southern Lao PDR were examined, of which 1308 (46.5%) were rabies positive. There is no clear reason why the number of samples submitted for diagnosis should have decreased during the period of the study and it is interesting to note that the proportion of rabies-positive cases was similar or slightly higher than those previously reported. However, historical results, especially those based on tests conducted in Lao PDR prior to 2000, should be treated with caution due to lower accuracy of the Sellers stain microscopy to detect Negri bodies in comparison with the more sensitive DFAT [17, 18]. The numbers and provincial distribution of rabies cases presented here are similar to those reported by Ahmed et al. [4], who examined the epidemiology of rabies in Lao PDR from 2004–2011 and also found that the majority of cases were located in Vientiane Capital, Champassak and Vientiane Province. However, it is likely this observation may be influenced by sampling bias due the closer proximity of these provinces to diagnostic facilities and knowledge regarding rabies control and dog-bite treatment.
This study has demonstrated a significantly higher proportion of rabies-confirmed cases during the dry season (<100 mm monthly rainfall) compared to the wet season. A possible explanation for this difference may be that animals are less likely to socialize during periods of high rainfall leading to a reduction in rabies transmission. Seasonality in dog rabies has previously been reported in Peru [19], Chile [20], Bolivia [21] and the United States [22], where the springtime excess of cases has been associated with greater contact between animals during the mating season, factors relating to herd immunity and an increased number of susceptible animals [19].
Male dogs accounted for a higher proportion of rabies-positive animals than females, although the association was not significant. This observation has been made previously in Florida in the USA [23], Nigeria [24] and Ghana [25] and is thought to be due to an association with the breeding season when it was observed that male dogs were fighting each other for the same bitch resulting in wounds that may have led to rabies infections.
This study has a number of limitations. A major constraint in determining the true burden of animal rabies in Lao PDR is the passive nature of surveillance. At present, the only source of samples for rabies diagnosis are those submitted following an animal bite incident, of which the majority are caused by dogs. This is likely to greatly underestimate the true number of rabies cases and may only represent the “tip of the iceberg”. Furthermore there is little information on the human:dog ratio for comparison of rabies risks. Another limitation is the fact that surveillance was confined to the increasingly urbanized Vientiane Capital, where reference diagnostics such as the DFAT are available, albeit only at NAHL, which severely limits the capacity to estimate the true number of rabies cases throughout the Lao PDR as there is comparatively limited capacity for rabies diagnosis outside of Vientiane Capital. Interprovincial submission of specimens for rabies diagnosis rarely occurs due to financial constraints or lack of awareness of the need to submit samples for reference diagnosis, further compounding the underreporting of both animal and human rabies. Another limitation of the study is that it has exclusively focused on canine rabies however the occurrence of rabies and other lyssavirus infections in wildlife, livestock and other domestic animals is a distinct possibility and the lack of samples from other species is likely to be due to the restricted capacity of the diagnostic service and the low level of awareness of rabies infections in these species. In this study, majority of samples came from dogs with suspected rabies infection, as dog bite was the primary reason for presentation at NAHL.
The availability of simple, accurate and affordable alternatives to the DFAT, such as rabies antigen rapid diagnostic tests (RDTs) [26–29], would be a useful adjunct to current surveillance methods in rural Lao PDR. However, a recent study has highlighted diagnostic problems with these tests [30] and therefore rabies antigen RDTs results should be interpreted with caution and may only be useful for surveillance purposes rather than for acute diagnosis and patient management (i.e., post-exposure prophylaxis).
Successful control and eradication of rabies is best achieved by an effective and targeted dog vaccination programs. Achieving acceptable levels of herd immunity to control rabies in the dog population is dependent on effective and consistent vaccination rates. Regarding rabies-control in Lao PDR, the government has an active program with the aim of rabies elimination, including advocacy by the authorities for the prevention and control of rabies, mobilization of community involvement, dog vaccination, allocation of resources, promotion of public awareness, provision of post-exposure prophylaxis and facilitation of the coordination between human and animal health sectors [5]. Rabies vaccination is largely focused on the major population centers of Lao PDR encouraging dog owners to vaccinate their pets which has led to the emergence of local veterinary practitioners in small animal clinics who provide vaccination services. Unfortunately, there is limited and unreliable data on the types of vaccines that are being administered. International partners have supported vaccination programs in Lao PDR in the recent past, with 50,000 doses of vaccine from the OIE Rabies Regional Vaccine Bank funded by the European Union in September 2012 followed by a further delivery of 120,000 additional doses [5] Such vaccinations are normally administered during mass vaccination campaigns such as World Rabies Day (28 September) or World Immunization Week and comprise a substantial proportion of the vaccines recorded as administered during these campaigns. However, maintaining the momentum of such programs in resource limited settings without continued external resources remains an issue. The challenge remains to find effective means to secure effective and safe vaccines via national and international disease control programs and then apply them in a strategic manner to achieve consistent and effective rabies control.
Other factors to control rabies in Lao PDR are also required. Stray dogs remain a problem throughout rural and urban areas of Lao PDR and may significantly contribute to the spread of rabies however the influence of stray dogs on rabies epidemiology in Lao PDR is not well understood and requires further investigation. Transboundary movement of dogs has the potential to impact the epidemiology of rabies in Lao PDR given that the country shares borders with China, Thailand, Cambodia and Vietnam. However, the official transboundary movement of dogs is strictly regulated with the transit of dogs for trade purpose prohibited and pets imported and exported must be rabies vaccinated. Wildlife and domesticated non-canine species are also potential reservoirs and vectors for rabies however their influence on rabies spread in Lao PDR is poorly understood and further studies are required to understand the influence of these animals on rabies epidemiology in Lao PDR.
Dog bite remains the main source of rabies transmission in the Lao PDR and dog vaccination remains the best method for rabies control. Dog owners can play an important role in rabies-control activities by ensuring that their dogs are vaccinated and restrained to reduce the opportunity for bites by rabies-infected animals. Governments and donors have an equally important role to play in providing a conducive environment and by facilitating increased awareness through mass media campaigns, effective delivery of vaccination and surveillance so that the situation can be actively monitored nationwide.
|
10.1371/journal.pbio.1001013 | Development of Axon-Target Specificity of Ponto-Cerebellar Afferents | The function of neuronal networks relies on selective assembly of synaptic connections during development. We examined how synaptic specificity emerges in the pontocerebellar projection. Analysis of axon-target interactions with correlated light-electron microscopy revealed that developing pontine mossy fibers elaborate extensive cell-cell contacts and synaptic connections with Purkinje cells, an inappropriate target. Subsequently, mossy fiber–Purkinje cell connections are eliminated resulting in granule cell-specific mossy fiber connectivity as observed in mature cerebellar circuits. Formation of mossy fiber-Purkinje cell contacts is negatively regulated by Purkinje cell-derived BMP4. BMP4 limits mossy fiber growth in vitro and Purkinje cell-specific ablation of BMP4 in mice results in exuberant mossy fiber–Purkinje cell interactions. These findings demonstrate that synaptic specificity in the pontocerebellar projection is achieved through a stepwise mechanism that entails transient innervation of Purkinje cells, followed by synapse elimination. Moreover, this work establishes BMP4 as a retrograde signal that regulates the axon-target interactions during development.
| Brain functions rely on highly selective neuronal networks which are assembled during development. Network assembly involves targeted neuronal growth followed by recognition of the appropriate target cells and selective synapse formation. How neuronal processes select their appropriate target cells from an array of interaction partners is poorly understood. In this study, we have addressed this question for the axons emerging from the pontine gray nucleus, a major brainstem nucleus that relays information between the cortex and the cerebellum, a brain area responsible for the control of skilled movements but also emotional processing. Using advanced microscopy techniques, we find that developing mossy fibers establish synaptic contacts rather promiscuously, and elaborate extensive synapses with Purkinje cells, an inappropriate target. These contacts are subsequently eliminated, and proper synaptic connectivity is then restricted to granule and Golgi neurons. We identify bone morphogenetic protein 4 (BMP4) as a regulator of these inappropriate mossy fiber-Purkinje cell contacts. BMP growth factors are best known for their functions in cell specification during embryonic development, and our results support an additional retrograde signaling function between axons and their target cells in early postnatal stages. In summary, we show that the specificity of the synaptic connections in the ponto-cerebellar circuit emerges through extensive elimination of transient synapses.
| The specificity of synaptic connectivity in the central nervous system is a prerequisite for brain function. The neuronal circuits in the vertebrate cerebellum represent a remarkable example of wiring specificity. This was first recognized by Santiago Ramón y Cajal when he chose cerebellar circuits as revealed by the Golgi method for his early studies on brain organization (discussed in [1]). In its simplest form, the cerebellar microcircuit integrates input from two afferent classes—climbing and mossy fibers. Climbing fibers selectively innervate Purkinje cells. By contrast, mossy fiber afferent activity is relayed to Purkinje cells via granule cells in the inner granular layer of the cerebellum (IGL) [2]–[4]. In the IGL, mossy fibers also form synapses on Golgi cells, a class of inhibitory interneurons that provide feed-forward inhibition in the cerebellar circuit. Climbing and mossy fiber information is then integrated in Purkinje cells and transduced via cerebellar efferent projection neurons in the deep cerebellar nuclei. Despite the apparent simplicity of the cerebellar circuit, it is unknown how the specificity of synapse formation emerges during development for each of the principal cerebellar afferent systems. Indeed, the molecular mechanisms regulating synapse specificity for most circuits in the mammalian brain have remained obscure.
Two key steps determining the incipient pattern of synaptic connectivity during development are axon-target contact formation and synaptic differentiation. Ultrastructural reconstruction of mature neuronal circuits suggests that only a subset of contacts differentiates into bona fide synapses [5]. The fraction of actual synapses compared to cellular contacts (potential synapses) has been termed “filling fraction”, with a filling fraction of 1.0 representing a case where all contacts are synaptic structures [6]. In vertebrate and invertebrate systems several attractive and repulsive factors have been identified that contribute to synaptic specificity [7]–[13]. However, pinpointing whether these specificity factors regulate primarily selective contact formation, synaptic differentiation, or both has been challenging, given the limited resolution of light microscopy in assessing direct cellular contacts in vivo. One possibility is that some signaling pathways regulate primarily contact formation, whereas other factors drive the synaptic differentiation process after axon-target contacts are established.
The ponto-cerebellar projection represents an excellent model system to explore mechanisms of synaptic specificity in the mammalian brain [14]. Mossy fiber axons emerging from the basilar pons (PGN) in the ventral brain stem form a major projection to the cerebellar cortex which relays information from sensory and motor cortex. Structurally, mossy fiber afferents exhibit synaptic specificity at two levels: Mossy fiber axons elaborate synapses exclusively with granule and Golgi cells but not Purkinje cells. At the subcellular level, mossy fiber synapses are restricted to the proximal regions of Golgi cells within the IGL but are excluded from the molecular layer where distal Golgi cell dendrites arborize. Cell culture studies indicated that immature granule cells provide a stop-signal for mossy fiber growth [15]. Maturing granule cells, in contrast, contribute positive signals for the differentiation of mossy fiber synapses and the elaboration of mossy fiber glomeruli [16]–[19]. However, it is unknown how contact and synapse specificity emerges for mossy fibers and their granule and Golgi cell targets. Moreover, negative signals that suppress contact of mossy fibers with inappropriate target cells in vivo have not been identified.
Previous anatomical studies indicated that axons with the appearance of mossy fibers do not exhibit strict targeting specificity but form broader projection patterns during early postnatal stages [20]–[22]. Fibers with mixed mossy fiber and climbing fiber morphology (“combination fibers”) were observed to contact Purkinje cells during development though the extent of these interactions remained unknown. A developmental gene expression analysis in pontine nuclei revealed distinct transcriptional programs for axonal growth and synaptic differentiation of pontine mossy fibers [23]. Surprisingly, the termination of the axonal growth program did not require the presence of granule cells in the cerebellar cortex but was perturbed in mutant mice with Purkinje cell degeneration accompanied by granule cell death [23]. In combination, these studies raised the question of whether Purkinje cells may provide signals for the development of mossy fiber projections.
To define the cellular nature of mossy fiber–Purkinje cell interactions and the rearrangements resulting in the specific synaptic wiring pattern, we undertook a systematic analysis of axon-target interactions in the mouse ponto-cerebellar system. Using correlated light-electron microscopy analysis we quantitatively mapped physical contacts and synaptic structures formed between identified pontine mossy fibers and Purkinje cells. Using this methodology, we observed extensive transient mossy fiber contacts and synapses on Purkinje cells that are subsequently eliminated. Patterning molecules such as WNTs, FGFs, and BMPs have been shown to exert novel neuronal signaling functions at the Drosophila neuromuscular junction and in the mammalian central nervous system [24]–[26]. Of this class of molecules, the BMPs have been extensively studied as retrograde signals at the Drosophila neuromuscular junction [27]–[32] and as trophic factors in mammalian neurons [33]–[35]. However, their signaling functions in vertebrate axon–target interactions have not been determined. We explored a role for BMP signaling in mossy fiber transient target interactions in the developing cerebellum. Using expression analysis, in vitro assays, and conditional knock-out mice we identify BMP4 as Purkinje cell-derived signal that specifically controls mossy fiber–target contact selectivity during development.
To examine the emergence of synaptic specificity of ponto-cerebellar mossy fibers we adopted an in utero electroporation approach [36],[37]. Pontine precursor cells are selectively electroporated by injection of DNA constructs into the 4th ventricle at embryonic day 14.5 (Figure S1). Following differentiation and migration, these cells settle into the pontine gray nucleus (Figure 1A,B) [36],[37]. Pontine axons labeled by electroporation of an EGFP expression plasmid project to the cerebellar cortex assume typical mossy fiber morphology and are restricted to the IGL at postnatal day 21 (P21), consistent with the selective elaboration of mossy fiber–granule cells synapses in the adult cerebellar circuitry (Figure S1). To examine how this target specificity emerges, we examined earlier developmental time points. At P7, we identified a significant number of GFP-positive mossy fiber extensions projecting beyond the IGL into the Purkinje cell layer (PCL) (Figure 1C–E). Using three-dimensional analysis of high resolution confocal stacks, mossy fiber varicosities were found in direct proximity with calbindin-positive Purkinje cell somata and axons, suggesting direct mossy fiber–Purkinje cell contacts (Figure 1F,G). These contacts contained synaptic markers, as they concentrated the endogenous synaptic vesicle protein VAMP2 or a synaptophysin-fluorescent protein fusion that was introduced by electroporation into the pontine projection neurons (Figure 1H,I).
Mossy fibers emerge not only from the PGN but multiple pre-cerebellar nuclei. The elaboration of mossy fiber–Purkinje contacts was not unique to PGN-derived mossy fibers as it was also observed in GFP-O transgenic mice where multiple other mossy fiber populations are marked by EGFP (Figure S1) [38]. Some of these contacts concentrated the postsynaptic scaffolding protein Shank1a (Shank1a-positive: 16% of somatic contacts, n = 50 contacts; 44% of contacts with proximal PC axon segments, n = 167; Figure 1J). In sum, these findings suggest that mossy fiber afferents establish transient synapse-like contacts with Purkinje cells during postnatal development.
In order to visualize the developmental progression of mossy fiber–Purkinje cell contacts and their differentiation into synapses we undertook a systematic light-electron microscopy analysis of PGN-derived mossy fiber axons. Pontine projections were labeled by DiI tracing followed by photoconversion of the dye (Figure 2). Mossy fiber projection patterns and Purkinje cell interactions in the cerebellar hemispheres (Crus1, Crus 2, and Simplex lobules) were quantitatively examined by light microscopy and apparent contacts were subsequently analyzed by electron microscopy. By light microscopy, mossy fiber rosettes and the thin protrusions extending from them are seen in great detail, and somata of Purkinje and granule cells of the cerebellar cortex can be clearly identified by DIC microscopy (Figures 2B,C, S2A–D). To further confirm the identity of the Purkinje cell territory, some sections were additionally labeled with antibodies to calbindin (Figure S2I,J). Camera lucida drawings of mossy fiber terminals from 50 µm sections at P0, P7, P14, and P21 revealed that the mature mossy fiber projection pattern emerges from a series of transformations during postnatal development that includes an extensive invasion of and eventual withdrawal from the Purkinje cell territory (Figure 2D,E).
At P0 mossy fibers extend far into the developing cerebellar cortex where Purkinje cells are unevenly distributed and intermixed with migratory and maturing granule cells of the emerging IGL (Figures 2D, S2). At P7, Purkinje cells form a recognizable monolayer above the IGL. However, pontine mossy fibers substantially invade this Purkinje cell territory with close to 30% of all labeled segments in the IGL penetrating into the PCL (Figure 2E,D, see Methods and figure legends for details on quantitative analysis). This invasion of the PCL was significantly reduced at postnatal days 14 and 21, yielding the mature mossy fiber projection pattern. At postnatal days 7 and 14 mossy fibers in close apposition to Purkinje cell somata often exhibited marked varicosities, resembling the presumptive mossy fiber–Purkinje cell synapses identified using the in utero electroporation approach (Figure 2C,D).
Fifty-five putative contacts identified at the light microscopy level in P0, P7, P14, and P21 tissues were examined by electron microscopy. Tissue sections (50 µm) were re-sectioned into 7 µm semithin sections and re-examined again by light microscopy. Sections encompassing the putative mossy fiber–Purkinje cell contacts were then thin-sectioned (70 nm) and processed for ultrastructural analysis (Figure S2E–H). Cytological characteristics defined in previous studies allowed unambiguous identification of Purkinje and granule cell somata in electron micrographs, as well as other relevant cellular components of the cerebellar cortex (Figure S3) [3],[39]. Over 90% of putative contacts between mossy fibers and Purkinje cell somata identified by light microscopy at P0 and P7 indeed represent direct cellular appositions (Figure 3A,B,E). At P0, none of the mossy fiber–Purkinje cell (or mossy fiber–granule cell) contacts in the developing IGL/PCL had synaptic features, representing a filling fraction (synapses per contacts [6]) of 0.0 (Figure 3A,F). However, at P7 a substantial number of mossy fiber–Purkinje cell contacts exhibited ultrastructural characteristics of synapses (Figure 3B,E,F; several consecutive sections shown in Figure S4A, filling fraction = 0.37). At P14, direct contacts were still observed (5 direct contacts verified by EM of 13 putative contacts analyzed) but only one of them was synaptic (Figure 3C,F, filling fraction = 0.2). Finally, at P21 no direct mossy fiber–Purkinje cell contacts or synapses could be identified (Figure 3D).
During the apparent removal of contacts and the elimination of synapses between P7 and P14 we frequently observed mossy fibers separated from the Purkinje cell soma and/or ensheathed by glial processes (Figure S4C–F), reminiscent of pruning processes with axosome shedding observed in peripheral axons [40]–[42]. In addition, we observed instances where mossy fiber axons were engulfed by Purkinje cells (Figure S4C,D). Glial process ensheathing of Purkinje cells became even more prominent at P21. Some mossy fibers were positioned as close as 150 nm from the Purkinje cell soma (Figure 3D) but glial processes separated mossy fiber endings and the Purkinje cell soma. In summary, during the first 10 postnatal days approximately 30% of all labeled IGL mossy fibers derived from the PGN establish direct contacts with Purkinje cells. The quantitative analysis uncovers remarkable developmental changes in the “filling fraction”, i.e. the differentiation of direct mossy fiber–Purkinje cell contacts into synapses, rising from 0.0 at birth to 0.37 at postnatal day 7 (Figure 3F). In the second to third postnatal week, these synapses are eliminated and the contacts withdrawn, resulting in the selective innervation of granule and Golgi cells in the IGL. Notably, not all mossy fiber axons contact Purkinje cells. Therefore, specific signaling mechanisms must exist, first, to limit the invasion of pontine mossy fiber axons into the Purkinje cell territory during the first postnatal days and, second, to promote the removal of pontine mossy fiber–Purkinje cell synapses in the second postnatal week of development.
In Drosophila melanogaster, growth factors of the bone morphogenetic protein (BMP) family regulate synaptic growth, axon arborization, and synaptic homeostasis [27],[28],[43]–[45]. To explore whether a comparable signaling function might be conserved in the mouse cerebellum, we surveyed the expression of BMP signaling molecules in the developing ponto-cerebellar projection system. Using in situ hybridization, we detected mRNAs for BMP receptor 1A (BMPR1A), BMP receptor 1B (BMPR1B), and BMP receptor type 2 (BMPR2) in the PGN at P0, the time when pontine mossy fiber axons extend into the cerebellar cortex. By P14, detection of BMPR1B mRNA was reduced, while signals for BMPR1A and BMPR2 expression persist (Figure 4A). Within the cerebellar cortex significant expression of several BMP ligands was observed consistent with previous reports ([46]–[48] and unpublished data). We focused our analysis on BMP4 as it is highly expressed in Purkinje cells and dynamically regulated during the refinement of mossy fiber connectivity (Figure 4B). At P0, BMP4 mRNA is abundant in proliferating and premigratory granule cells of the EGL, and in scattered Purkinje cells (identified by their large diameter). BMP4 expression in Purkinje cells was reduced at P7, the time when mossy fiber–Purkinje cell synapses are most common, and expression was strongly up-regulated in Purkinje cells by postnatal day 14 (Figure 4B). At P21, BMP4 was highly expressed in Purkinje cells. In addition a subset of large diameter cells in the IGL (presumably Golgi cells) expressed BMP4. In summary, BMP4 and its signaling receptors are appropriately positioned to regulate mossy fiber target selection during postnatal development.
BMP-receptor activation results in phosphorylation of cytoplasmic SMAD proteins that translocate to the cell nucleus and activate transcription [49],[50]. Classical morphogenetic functions of BMPs depend on SMAD phosphorylation but phospho-SMAD (pSMAD)-independent BMP signaling read-outs have also been described [34],[51]–[53]. Robust SMAD phosphorylation was detected when recombinant BMP4 was added to cultured pontine explants in vitro and phosphorylation was prevented by co-application of the antagonist noggin (Figure S5A). Quantitative evaluation of SMAD phosphorylation in PGN in vivo using Western blot and immunohistochemistry revealed a dynamic regulation, with moderate levels at P0, strongly increased levels at P14, and persistent pSMAD immune-reactivity at P21 (Figure S5B–D). SMAD1,5,8 protein levels were not significantly altered during this developmental time period, suggesting that regulation of SMAD signaling occurs primarily at the level of SMAD phosphorylation (unpublished data). This demonstrates a functional BMP signaling pathway in developing pontine neurons in vitro and in vivo.
Given that BMP4 was dynamically expressed in Purkinje cells we asked whether Purkinje cell-derived BMP4 was required for SMAD activation in pontine neurons. We analyzed conditional BMP4fl/fl::Pcp2cre/cre knockout (BMP4 cKO) mice lacking BMP4 expression selectively in Purkinje cells. In the Pcp2cre knock-in line, cre-mediated recombination is detected during late embryonic stages and specifically in Purkinje cells ([54] and Figure S6). Ablation of BMP4 expression was verified by in situ hybridization (Figure S6C). Interestingly, SMAD activation in the pontine gray nucleus was not dramatically altered in BMP4 cKO mice (Figure S5D,E). While we cannot completely exclude that some of the pSMAD signal is due to incomplete ablation of the BMP4 expression in Purkinje cells, these results indicate that Purkinje cell-derived BMP4 might not be essential for pSMAD activation in pontine nuclei during postnatal development. Notably, Purkinje cells express significant amounts of BMP7 and other BMP growth factors during postnatal development which might be responsible for the persistent SMAD phosphorylation in the absence of BMP4 (unpublished data).
Importantly, signaling activities have been described for specific BMP growth factors that control cytoskeletal rearrangements through pSMAD-independent pathways [34],[51]–[53]. In commissural spinal neurons such BMP signals represent extrinsic cues for the initial polarization of axons [55],[56]. Therefore, we examined the possibility that target-derived BMP signaling might regulate axon development and axon-target interactions of pontine mossy fibers.
Previous work demonstrated that cerebellar explants cultured in vitro release a growth inhibiting activity for mossy fibers which is thought to resemble a target-derived stop signal for afferents [57]. Explants from the PGN exhibit robust radial axon outgrowth. However, when pontine explants are co-cultured with cerebellar tissue, axon growth on the side facing the cerebellar tissue is reduced, suggesting the presence of a growth inhibiting activity derived from cerebellar tissue (Figure 5A). In order to assess a possible role for BMPs in this process, we applied the soluble BMP antagonist noggin to the culture medium. Noggin addition blocked cerebellar growth retardation activity in this assay (Figure 5B,D). To directly examine whether BMP4 might exert such growth-inhibiting activity, we combined pontine explants with BMP4-expressing HEK293 cells in collagen gel co-cultures. Using this assay, we observed that BMP4 was sufficient to negatively regulate mossy fiber growth in vitro (Figure 5C). Importantly, this activity could be neutralized by addition of noggin to the collagen gel matrix, indicating that the growth regulation was indeed mediated by BMP signaling (Figure 5E). Therefore, BMP4 negatively regulates pontine mossy fiber growth in vitro.
Based on the dynamic regulation of BMP4 expression in Purkinje cells and the repulsive activity of BMP4 towards mossy fiber axons in vitro, we hypothesized that BMP4 might control either initial mossy fiber–Purkinje cell interactions, the detachment of mossy fiber–Purkinje cell contacts, or both. To explore these possibilities, we further examined the BMP4 cKO mice. Given the important patterning functions of BMP signaling in early cerebellar development [46],[47],[58] we first asked whether the overall anatomical organization of the cerebellar cortex or specification of cerebellar cell types was perturbed in the mutant mice. No significant changes were detected in the foliation pattern, cerebellar layering, specification of the major cell types, and expression of transcriptional markers and signaling molecules (Figure S7). In DiI labeled preparations, the number of labeled pontine mossy fiber axons or density of Purkinje cells observed in the cerebellar cortex of BMP4 cKO mice was not significantly different from control littermates or wild-type animals (Figure S7G and unpublished data). Finally, the development of climbing fibers and formation of vGlut2-positive climbing fiber synapses on the Purkinje cell dendrites was not noticeably altered in the BMP4 cKO mice (Figure S8).
Next, we examined whether loss of Purkinje cell-derived BMP4 resulted in defects in mossy fiber–Purkinje cell contact formation, synapse formation, and/or synapse elimination. In BMP4 cKO mice the fraction of pontine mossy fibers that penetrated into the Purkinje cell territory at postnatal day 0 was increased approximately 2-fold as compared to control (Figure 6A,B). Moreover, the number of Purkinje cells receiving mossy fiber contacts was increased 7-fold at P0 and remained significantly increased over the following 2 wk. In our wild-type analysis (Figure 2) we identified a peak in mossy fiber elimination from the Purkinje cell territory in the second postnatal week. Mossy fiber elimination was quantitatively compared using an elimination index for the fraction of mossy fibers removed from the PCL between P7 and P14 ([MFsPCL P7–MFsPCL P14] / MFsPCL P7). In the cKO animals elimination of mossy fibers still occurred but the elimination index for control and cKO mice was reduced to about 50% of that in control animals (Figure 6C). When normalized to the length of mossy fiber segments in the Purkinje cell territory, the density of contacts per 100 µm mossy fiber length was more than 3-fold increased at postnatal day 7 (Figure 6C). These observations highlight an essential function for BMP4 in the control of initial mossy fiber–Purkinje cell contact formation during the first postnatal week as well as the subsequent removal of mossy fiber processes from the Purkinje cell territory.
Considering the developmental regulation of the filling fraction observed in wild-type animals (Figure 3) we further examined mossy fiber–Purkinje cell contacts, synapses, and filling fractions at P7 using correlated light-electron microscopy. As in wild-type and control tissue, the majority of putative mossy fiber–Purkinje cell somatic contacts identified in BMP cKO mice indeed represented direct cellular appositions (38 direct contacts out of 40 potential contacts analyzed, Figure 6D,E). Some mutant contacts were characterized by unusual, irregular synapse-like profiles (Figures 6D, S4B). However, the filling fraction was substantially reduced in the cKO as only 11% of these contacts exhibited synaptic ultrastructure (Figure 6E). Based on the correlated light-EM analysis and the calculated filling fraction, the total density of mossy fiber–Purkinje cell synapses was not significantly changed, indicating that the excess contacts do not efficiently differentiate into synaptic structures. These experiments identify BMP4 as a retrograde signal that specifically controls mossy fiber-Purkinje cell contact formation and highlight that independent programs regulate contact versus synapse formation during postnatal development.
If loss of BMP4 from Purkinje cells results in exuberant pontine mossy fiber–Purkinje cell contacts during early postnatal development, do these aberrant interactions perturb the placement or specificity of synapses in the mature cerebellum? Using in utero electroporation, we marked pontine mossy fibers in control and BMP4 cKO animals and examined their projection pattern at P21 when cerebellar development is essentially complete (Figure 7). While in wild-type mice mossy fibers were restricted to the IGL and did not protrude into the molecular layer, we observed overshooting axons in the BMP4 cKO mice (Figure 7A). A subset of mossy fiber axons penetrated more than 20 µm beyond the Purkinje cell somata into the molecular layer, a phenotype never observed in control cerebella (Figure 7B). Within the molecular layer, most mossy fiber axons had a smooth appearance but some developed swellings comparable to simple mossy fiber rosettes. Overshooting mossy fiber axons have been observed previously in mouse mutants with perturbed granule cell migration [59]. However, we did not observe ectopic granule cells in the molecular layer of BMP4 cKO mice (Figure 7C). Instead, high-resolution analysis of the overshooting mossy fiber axons revealed that some established direct contacts with the dendritic tree of Purkinje cells. Other overshooting axons formed contacts with neurogranin-positive Golgi cells (Figure 7C). Notably, Golgi cells are one of the specific synaptic targets of ponto-cerebellar mossy fibers in the IGL. However, in wild-type mice mossy fiber synapses are excluded from the distal dendritic arbors in the IGL. Finally, we examined the position of pontine mossy fiber synapses in the IGL and observed a significant shift of mossy fiber rosettes towards the PCL (within 40 µm of the Purkinje cell somata, Figure 7). Therefore, loss of Purkinje cell-derived BMP4 results in persistent alterations in mossy fiber connectivity in the mature cerebellum.
Cerebellar circuits and the “crystalline” architecture of the cerebellar cortex are a prime example of the precision of neuronal connectivity. In this study, we identified cellular and molecular mechanisms orchestrating aspects of afferent-target specificity in cerebellar networks. First, we demonstrate that synaptic specificity of pontine mossy fibers emerges in a protracted, stepwise process that encompasses extensive contacts and synapse formation with Purkinje cells. Second, we identify BMP4 as a retrograde, Purkinje cell-derived signal that negatively regulates mossy fiber-Purkinje cell contacts and synaptic specificity.
BMPs are key regulators of patterning and cell fate decisions, but novel functions in neuronal wiring are emerging [49],[60]–[62]. In the vertebrate central nervous system BMPs (and the related TGFbeta growth factors) control initial axon orientation and axon regeneration [55],[56],[63]–[65]. Moreover, retrograde, target-derived BMP signaling has been examined in the peripheral nervous system [66]–[69]. At the Drosophila neuromuscular junction a muscle-derived BMP-analogue regulates synaptic growth and homeostatic signaling [27],[29],[30],[43],[70]. Whether BMP growth factors have similar retrograde signaling activities in the central nervous system and, specifically in axon-target interactions in vertebrates, has remained unclear. In our experiments, we explored retrograde BMP signaling in the mouse pontocerebellar system and uncovered a novel function during the development of synaptic target specificity. In this system, BMP4 acts as a negative signal that limits interactions of mossy fibers with Purkinje cells, a transient target cell. The dynamic regulation of BMP4 expression in Purkinje cells mirrors the assembly of mossy fiber-Purkinje cell contacts and synapses, with a transient peak at P7 where BMP4 expression is low. Thereafter, BMP4 is strongly up-regulated and mossy fiber-Purkinje cell contacts are eliminated.
The mossy fiber phenotypes in the BMP4 cKO mice highlight a critical function of Purkinje cell-derived BMP4 in mossy fiber–Purkinje cell interactions. In the cKO mice, there is a substantial increase in mossy fiber–Purkinje cell contacts at early postnatal stages (P0–P7). This supports an essential repulsive role for BMP4 in target recognition which limits the initial mossy fiber–Purkinje cell contacts and restricts the invading mossy fiber axons to their target territory in the IGL. The correlated light-electron microscopy analysis enabled us to dissociate changes in contact and synapse formation in the cerebellar system. Notably, while BMP4 cKO mice exhibit a 3-fold increase in mossy fiber–Purkinje cell contact density we did not detect a comparable increase in synapse density. Therefore, axon target contacts and synapse formation are controlled by different signaling systems.
The subsequent, removal of mossy fiber–Purkinje cell contacts and elimination of mossy fiber processes from the Purkinje cell territory was significantly delayed, and after completion of cerebellar development, we observed persistent overshooting mossy fiber projections in the Purkinje cell and molecular layers. Some overshooting axons retain interactions with Purkinje cells, while others form contacts on distal Golgi cell dendrites. Notably, Golgi cells are appropriate synaptic partners of mossy fibers, but in BMP4 cKO cerebella mossy fiber–Golgi cell interactions are observed ectopically in the molecular layer. These findings support an important role for Purkinje cell-derived BMP4 in eliminating mossy fiber projections from this area, in addition to its function in regulation of the early mossy fiber–Purkinje cell contacts.
Within the IGL, the placement of mossy fiber rosettes was shifted towards the Purkinje cell layer, further supporting a repulsive role for Purkinje cell-derived BMP4. However, the fact that most mossy fiber axons did not overshoot to the molecular layer indicates that there are additional signals that restrict mossy fibers to the IGL. BMP2 and 7 transcripts are up-regulated in Purkinje cells of the cKO mice (unpublished data) and may partially compensate for the loss of BMP4. Moreover, the specificity of mossy fiber connectivity is likely to emerge not only from negative, Purkinje cell-derived signals but from an interplay with positive signals derived from the appropriate target cells. Granule cells express FGF22, Wnt7a, and neuroligins which all have been demonstrated to have positive, synaptogenic activities towards mossy fiber afferents [16]–[18]. Therefore, presentation of these synaptogenic signals by mature granule cells which strongly increase in number at later postnatal stages (P7–P21) may compete with the constant number of Purkinje cells for mossy fiber contact. A prediction of this model is that direct mossy fiber-Purkinje cell synapses would persist in the absence of granule cells. This is, indeed, observed in agranular cerebella of mouse mutants or after irradiation where mossy fiber target selectivity can be examined in the absence of the appropriate synaptic targets [71]. Importantly, while most mossy fibers were appropriately restricted to the IGL, the positioning of mossy fiber rosettes within the IGL was shifted closer to the Purkinje cell layer (Figure 7D), consistent with the loss of a negative regulator of synaptic connectivity in Purkinje cells of BMP4 cKO mice.
The finding that the development of mossy fiber target specificity involves not only extensive contact but also synapse formation with Purkinje cells argues against a model of absolute recognition specificity for unique synaptic targets. This remodeling of transient target interactions is reminiscent of interactions in the thalamo-cortical projection and for Cajal Retzius cells in the hippocampus [72]–[74]. In both cases, afferents enter the target territory before their appropriate target cells have fully differentiated and form transient synapses on a third cell type (subplate neurons and Cajal Retzius cells, respectively). This situation in the hippocampus is comparable to the transient mossy fiber–Purkinje cell synapses described in our study that are elaborated during early postnatal development when only few granule cells have descended into the forming IGL. While the initial assembly of such transient contacts is comparable, the mechanism of contact removal is fundamentally different. Elimination of transient synapses received by subplate and Cajal Retzius neurons occurs via programmed cell death of the transient target cells. By contrast, removal of mossy fiber–Purkinje cell interactions occurs independently of Purkinje cell death and requires signals for contact destabilization.
The existence of widespread mossy fiber–Purkinje cell interactions during development poses the question of whether these synapses simply represent an imprecision in the initial trans-synaptic interactions or whether transient contacts serve a purpose in the development of functional cerebellar circuits. In the hippocampus, Cajal Retzius cells appear to be required for the laminar specificity of entorhinal axon projections [74]. Similarly, in the absence of subplate neurons, thalamocortical axons do not establish appropriate synaptic connectivity [75],[76]. Therefore, transient mossy fiber–Purkinje cell interactions might similarly contribute to the assembly of cerebellar circuits. The cerebellar cortex is subdivided into longitudinal bands identified by specific molecular codes of gene expression in Purkinje cells [77],[78]. This code develops during the first postnatal weeks, coincident with emergence of mature cellular and sub-cellular targeting specificity of both climbing and mossy fiber afferents. Recent tracing studies indicate that there is a precise somatotopic matching of pontine and climbing fibers [79],[80]. This raises the possibility that transient mossy fiber–Purkinje cell interactions might provide a mechanism to coordinate mossy fiber and climbing fiber development and, thereby, serve a functional role in the assembly of cerebellar circuits.
All animal experiments were reviewed and approved by the institutional animal care and use committee of Columbia University and the cantonal veterinary office Basel, respectively. Mice were of the NMRI (Figure 1) and C57BL/6 strains (all other experiments). PCP2cre knock-in mice were previously described [54]. The conditional BMP4 floxed allele (BMP4fl) was generously provided by Dr. Brigid Hogan [81]. Htr5b-GFP mice are BAC transgenic mice generated by the GENSAT consortium [82] and were obtained from the MMRRC repository. Thy1.2-GFP (GFP-O) mice were generated by Drs. Sanes and Feng [38] and were obtained from the Jackson Laboratory. R26-lox-stop-lox-YFP were described in [83].
Timed-pregnant mice (NMRI or C57BL6 background) were used at embryonic day 14.5 following the protocol described in [37]. After electroporation, the mice were brought to term, pups were sacrificed by transcardial perfusion with 4% paraformaldehyde in 100 mM Na-phosphate buffer (pH7.4), and tissue from successfully electroporated pups (P7, P14, P21) was processed for immunohistochemistry.
The following primary antibodies were used in this study: rabbit anti-Shank1a [84], goat anti-Car8 (Frontiers Institute), rabbit and mouse anti-Calbindin D-28K (Swant), rabbit anti-GFP [85], rabbit anti Pax6 (Covance), guinea pig anti-vGlut1 (Chemicon), mouse anti-vGlut2 (Chemicon), goat anti-Parvalbumin (Swant), rabbit anti-Smad1 (Millipore), rabbit anti-phospho-SMAD1/5/8 (Millipore), rabbit anti-SMAD1,5,8 (Imgenex), mouse anti-NeuN (Millipore), rabbit anti-neurogranin (Abcam), mouse anti-actin (Sigma), and mouse anti-VAMP2 (Synaptic Systems).
Most procedures followed standard protocols; see Text S1 for details.
High-resolution images of 30 to 40 µm z-stacks consisting of 0.45 µm thick optical sections were acquired using Zeiss LSM510, a Zeiss LSM5 Exciter, or a LIS-spinning disk confocal system. Direct apposition of cellular markers was identified by rotating the 3D reconstruction of the stacks using Imaris Software (Bitplane).
Quantitative assessment of SMAD1,5,8 activation was performed using pSMAD1,5,8 and NeuN immunolabelling on 50 µm thick sagittal section using Metamorph software. The percentage of pSMAD1,5,8 positive cells among the NeuN positive cells and the pSMAD1,5,8 fluorescence intensity per NeuN area was determined through intensity thresholding and integrated morphometry analysis using MetaMorph software.
DiI (1,1′-dioctadecyl-3,3,3′,3′-tetraindocarboyanine perchlorate, Molecular Probes) labeling was modified from a previously published procedure [86]; see Text S1 for details.
Quantitative assessment of mossy fiber invasion into Purkinje cell territory was performed on camera lucida drawings of DiI and calbindin double-labeled material (50 µm coronal “thick” sections, 100× objective) of P0, P7, P14, and P21 cerebellar hemispheres (crus1, crus2, and simplex lobules). All camera lucida drawings contained all of the labeled mossy fiber segments and Purkinje cell outlines drawn from a fixed area size of 175 µm horizontal×120 µm vertical×30 µm deep (thickness of one Purkinje cell soma), encompassing the upper IGL and PCL, and spanning a stretch containing on average 40 Purkinje cells at P0 (before PC alignment occurs), and 9 Purkinje cells at P7–P21. For the sake of consistency, and since Purkinje cell density and the angle of the mossy fiber segments penetrating the PCL differs at the base, versus apex, versus sides of the folia, areas for analysis were always drawn from the sides of the folia. The percentage of mossy fiber segments invading into the PCL out of all mossy fiber segments drawn per area was scored (>20 segments per area from >25 areas obtained from 5–9 animals per time point, >3 50 µm section per animal were analyzed). For the quantification of Purkinje cells receiving putative somatic contacts from mossy fibers, contacts were judged as varicosities in the mossy fiber axon at the site apparently immediately adjacent to Purkinje cell soma. For quantification of contact density per mossy fiber length (in Figure 7D) camera lucida drawings were scanned at 600 dpi, and the length of mossy fiber segments in the PCL was measured using line tool in NeuronJ [87]. The number of putative contacts on Purkinje cell somata per mossy fiber segment was scored visually, using the criteria described above.
The filling fraction was calculated as actual synapses divided by the number of contacts (EM-verified). The elimination index for mossy fibers projecting into the Purkinje cell layer (MFsPCL) was calculated using the data points for P7 and P14 (graph Figure 6B) as follows: [MFsPCL P7–MFsPCL P14] / MFsPCL P7.
|
10.1371/journal.ppat.1000605 | Human Papillomaviruses Activate the ATM DNA Damage Pathway for Viral Genome Amplification upon Differentiation | Human papillomaviruses (HPV) are the causative agents of cervical cancers. The infectious HPV life cycle is closely linked to the differentiation state of the host epithelia, with viral genome amplification, late gene expression and virion production restricted to suprabasal cells. The E6 and E7 proteins provide an environment conducive to DNA synthesis upon differentiation, but little is known concerning the mechanisms that regulate productive viral genome amplification. Using keratinocytes that stably maintain HPV-31 episomes, and chemical inhibitors, we demonstrate that viral proteins activate the ATM DNA damage response in differentiating cells, as indicated by phosphorylation of CHK2, BRCA1 and NBS1. This activation is necessary for viral genome amplification, as well as for formation of viral replication foci. In contrast, inhibition of ATM kinase activity in undifferentiated keratinocytes had no effect on the stable maintenance of viral genomes. Previous studies have shown that HPVs induce low levels of caspase 3/7 activation upon differentiation and that this is important for cleavage of the E1 replication protein and genome amplification. Our studies demonstrate that caspase cleavage is induced upon differentiation of HPV positive cells through the action of the DNA damage protein kinase CHK2, which may be activated as a result of E7 binding to the ATM kinase. These findings identify a major regulatory mechanism responsible for productive HPV replication in differentiating cells. Our results have potential implications for the development of anti-viral therapies to treat HPV infections.
| Over 100 types of human papillomavirus (HPV) have been identified, and approximately one-third of these infect epithelial cells of the genital mucosa. A subset of these HPV types are the causative agents of cervical and other anogenital cancers. The infectious life cycle of HPV is dependent on differentiation of the host epithelial cell, with viral genome amplification and virion production restricted to differentiated suprabasal cells. While normal keratinocytes exit the cell cycle upon differentiation, HPV positive suprabasal cells are able to re-enter S-phase to mediate productive replication. The mechanisms regulating the activation of differentiation-dependent viral replication are largely unknown. In this study, we demonstrate that HPV induces an ATM-dependent DNA damage response that is essential for viral genome amplification in differentiating cells. In addition, we have found that ATM signaling to its downstream target CHK2 is critical for providing an environment that is conducive to HPV productive replication. Our findings identify an important regulatory mechanism by which HPV controls replication during the productive phase of the life cycle and may identify new targets for the development of therapeutics to treat HPV-induced infections.
| Human papillomaviruses (HPV) are the etiological agents of most anogenital cancers and their productive life cycle is dependent upon epithelial differentiation [1],[2]. HPVs infect cells in the basal layer of stratified epithelia, but restrict the productive phase of the life cycle to highly differentiated suprabasal cells [3]. Viral genome amplification, late gene expression and virion production are induced in suprabasal cells that have re-entered S-phase. In undifferentiated basal cells, viral genomes are maintained as episomes at approximately 100 copies per cell and replicate in synchrony with cellular replication. In contrast, upon differentiation HPV genomes are replicated to thousands of copies per cell in a process referred to as amplification [4]. While normal epithelial cells exit the cell cycle upon differentiation, HPV-infected cells are able to over-ride normal checkpoint controls and remain active in the cell cycle, allowing for the synthesis of cellular proteins that are necessary for viral replication [5],[6]. HPV proteins activate low levels of caspases belonging to the intrinsic pathway in differentiating cells, and this is necessary for viral replication [7]. The mechanisms regulating productive replication of HPVs upon differentiation, however, remain largely unknown.
The fidelity of cellular replication is controlled by signaling pathways that block the propagation of damaged DNA [8],[9]. Central to these repair pathways are the ATM (ataxia-telangiectasia mutated), and ATR (ATM and Rad3-related) kinases, which belong to a structurally related family of serine-threonine kinases that share a PI-3 kinase-like domain, but only phosphorylate proteins [9]. ATM is a prime mediator of the cellular response to double strand breaks [10], while ATR controls the response to UV damage, as well as stalled DNA replication forks [11]. Both kinases act in part by controlling cell cycle checkpoints at G1, S and G2. Several viruses have been shown to interact with and/or affect components of the ATM DNA damage pathway [12]. Herpes simplex virus (HSV) induces an ATM-damage response as soon as pre-replication centers are formed, and this activation is essential for productive replication [13],[14]. In contrast, adenovirus must mislocalize and degrade DNA repair proteins to ensure viral replication [15]. Using recombinant adenoviruses, high-level expression of HPV-16 E7 in fibroblasts was shown to activate the ATM pathway [16], but it is unclear whether these effects are physiologically significant, or if they play any role in the viral life cycle.
ATM activates a number of downstream targets that are involved in cell cycle control, apoptotic responses and DNA repair [17]. These proteins can be divided into three pathways that lead to activation of cell cycle checkpoints: a p53/mdm2 pathway, a CHK2 branch, and a NBS1/BRCA1/SMC1 pathway. ATM directly activates p53 by phosphorylation at serine 15, as well as by phosphorylating Mdm2, the ubiquitin ligase that regulates p53 stability [18],[19],[20]. In the second pathway, ATM phosphorylates CHK2 leading to arrest in S- and G2-phases by inhibiting the action of Cdc25 phosphatases [21],[22]. Another branch of S-phase checkpoint control involves ATM targeting of NBS1, a member of the MRN double strand break repair complex [23]; BRCA1, the breast cancer susceptibility protein [24]; and SMC1, a cohesin binding protein [25],[26]. An additional downstream activity of ATM, as well as ATR, that is important for S-phase checkpoint control is phosphorylation of the tail of a histone variant H2AX (γ-H2AX), which leads to recruitment of DNA damage regulatory factors to distinct nuclear foci [27],[28]. Phosphorylation of these, as well as other targets allows for DNA damage to be assessed and for repair to take place.
Given the importance of this pathway in controlling replication, we investigated if ATM signaling was necessary for stable HPV replication in undifferentiated cells, as well as productive replication in differentiated cells. Our studies indicate that HPV proteins induce an ATM response in both undifferentiated and differentiated cells. Importantly, we found that ATM kinase activity is necessary for viral genome amplification in differentiating cells, but not for stable maintenance in undifferentiated cells. These studies implicate HPV activation of DNA damage signaling in controlling productive viral replication upon differentiation.
To determine if HPV induces a DNA damage response in infected cells we first examined the expression level, as well as activation status, of ATM by Western blot analysis (Figure 1C). For these studies, we examined undifferentiated, as well as differentiated human keratinocyte cell lines that maintain HPV-31 episomes at approximately 50 copies per cell, and compared effects to a matched set of normal human keratinocytes isolated from foreskin circumcisions (HFKs). We have shown previously that calcium-induced differentiation is sufficient to activate the productive phase of the HPV life cycle by 48 hours [7]. Upon differentiation, approximately 25% of HPV positive cells re-enter S phase and undergo viral genome amplification, resulting in high levels of episomal DNA [29] (Figure 1A). The induction of differentiation is indicated by the expression of cellular proteins, such as keratin 10 (K10) and involucrin (Figure 1B). As shown in Figure 1C, the total levels of ATM were similar between HPV-31 positive HFKs (HFK-31) and matched normal HFKs in undifferentiated cultures, as well as after 48 and 96 hours of differentiation in high calcium medium. In response to double-strand breaks or changes in chromatin, ATM is activated through the autophosphorylation of inactive, dimeric ATM on serine 1981, followed by dissociation into active monomers [30]. The phosphorylated form of ATM (Ser1981) was detected in undifferentiated HPV-31 positive cells and maintained at a similar level through 48 hours of differentiation, with a decrease occurring at 96 hours (Figure 1C). In contrast, only a low level of pATM Ser1981 was observed in normal HFKs. This pattern of ATM phosphorylation was observed in multiple independently derived polyclonal populations of HFK-31 cells, and matched normal HFKs, as well as in CIN612 cells, which is a clonal cell line derived from a HPV-31 positive biopsy (Bedell MA, 1991) (data not shown).
To determine if phosphorylation of ATM correlated with activation of downstream targets, we examined the phosphorylation status of three of its substrates: CHK2, NBS1 and BRCA1. CHK2 is an important transducer of the ATM signaling pathway, and its activation is initiated by ATM phosphorylation on threonine 68 (pCHK2) [21],[22]. While pCHK2 was detected in both undifferentiated and differentiated HPV-31 positive cells, only low levels were detected in normal HFKs (Figure 1C), which correlates with the low level of pATM detected in these cells. Activation of the ATM pathway can also result in the phosphorylation of NBS1 and BRCA1, both of which play important roles in DNA repair and the regulation of S and G2 checkpoints [24],[31],[32],[33]. In HFK-31 cells, BRCA1 was phosphorylated at all states of differentiation (Figure 1C). While we also observed pBRCA1 in undifferentiated normal HFKs, the levels rapidly declined along with total BRCA1 upon differentiation. Phosphorylation of BRCA1 in normal HFKs may be due to either low level activation of ATM, or through the action of ATR, which can also phosphorylate BRCA1 [24]. Importantly, phosphorylation of NBS1 at Ser343 was induced only upon differentiation of HPV positive cells, with low levels observed in undifferentiated cells (Figure 1C). In contrast, little phosphorylation of NBS1 was observed in either undifferentiated or differentiated normal HFKs (Figure 1C). These results indicate that phosphorylation of NBS1 in HPV positive cells is differentiation-specific, and correlates with the induction of productive viral replication (Figure 1A). Interestingly, we observed phosphorylation of the ATR substrate CHK1 on Ser317 in undifferentiated HPV positive cells (Figure 1C), suggesting that ATR is also active. However, the levels of phosphorylated CHK1, as well as total CHK1, decreased by 48 hours post-differentiation. Similar effects were observed in multiple independently derived HFK-31 cell lines, as well as in CIN612 cells (data not shown). Overall, these results indicate that HPV proteins induce a DNA damage response that is maintained throughout the viral life cycle and characterized by the activation of the ATM substrates CHK2, NBS1 and BRCA1.
We next wanted to confirm that ATM activity was responsible for CHK2 phosphorylation in HPV positive cells, as CHK2 can also be phosphorylated and activated by the ATR kinase [22],[34]. HFK-31 cells were treated with a small molecule inhibitor of ATM, KU-55933, that inhibits kinase activity without affecting the total levels of ATM [35] (Figure 1D). Treatment of HFK-31 cells with 5 or 10 uM of the ATM inhibitor resulted in significantly reduced phosphorylation of CHK2, as well as ATM itself in differentiated cells (Figure 1D). Treatment of HPV-31 cells with this inhibitor also resulted in decreased phosphorylation of NBS1 and BRCA1 (data not shown), while phosphorylation of CHK1 Ser317 was only minimally affected with 10 uM of the inhibitor (Figure 1D). Treatment of undifferentiated HPV-31 positive cells also resulted in inhibition of CHK2 phosphorylation without affecting CHK1 phosphorylation (data not shown). These results indicate that ATM activity is necessary for CHK2 phosphorylation in cells maintaining HPV genomes. In addition, our findings indicate that the low level of CHK1 phosphorylation observed in differentiating HPV positive cells is ATM-independent, and may suggest a role for ATR in the non-productive phase of the life cycle.
In response to DNA damage, ATM is recruited to distinct nuclear foci by the MRN complex, which consists of NBS1, MRE11 and RAD50 [36],[37]. Recruitment to double strand breaks allows for ATM-dependent phosphorylation of at least a subset of downstream targets [10],[38]. Adenoviruses abrogate the ATM response by relocalizing and degrading components of the MRN complex, which would normally promote the detrimental concatemerization of viral genomes [15],[39]. For HPV, the amounts of MRE11 and RAD50 were found to be similar in both undifferentiated and differentiated cells, as well as in HFKs, while the levels of NBS1 were consistently higher in HPV-31 cells compared to normal HFKs (Figure 2). In addition, components of the MRN complex were localized to nuclear foci in HPV positive cells (Figure S1). Overall, these results indicate that HPV does not induce degradation of these repair proteins in order to facilitate the viral life cycle, but instead maintains them at high levels throughout differentiation.
The detection of phosphorylated CHK2, NBS1 and BRCA1 in HPV positive cells suggested that ATM may be localized to nuclear foci, as is observed in cells undergoing a DNA damage response [8]. To investigate this possibility, we examined the localization of the phosphorylated form of ATM (Ser1981) in HPV-31 positive cells by confocal fluorescence microscopy. As shown in Figure 3A, pATM Ser1981 was found in distinct nuclear foci in undifferentiated HPV-31 positive cells. pATM Ser1981 colocalized with γ-H2AX, a modified histone associated with double-stranded breaks [40], and these foci were retained upon differentiation in a similar number of HPV-31 positive cells (39.7±6.6%). In contrast, normal HFKs exhibited diffuse staining of pATM, with only a few foci being detected in both undifferentiated and differentiated cells (3.2%±.8 and 3.1±.6%, respectively). Inhibition of ATM activity by treatment with KU-55933 abrogated formation of pATM-Ser1981 foci in HPV-31 positive cells (Figure 3A), and correlates with the decrease in pATM levels observed in the presence of the inhibitor by Western blot analysis (Figure 1D). Interestingly, the number of HPV-31 positive cells that exhibited γ-H2AX foci upon 48 hours of differentiation was reduced but not completely inhibited in the presence of KU-55933 (40.7% to 10.8%), indicating that other kinases, such as ATR [27], may contribute to this activity in HPV positive cells. Similar results were observed in multiple HFK-31 lines, as well as CIN612 cells. These results suggest that HPV induces the activation and accumulation of ATM and γ-H2AX at distinct sites in the nucleus.
We next examined the localization of phosphorylated CHK2 Thr68 in undifferentiated and differentiated cells. In undifferentiated HPV-31 positive cells, a number of foci containing both pCHK2 Thr68 and γ-H2AX were detected, although every cell did not stain positive for these foci (Figure 3B). Upon differentiation in high calcium medium for 48 hours, the number of foci containing both pCHK2 and γ-H2AX was retained (21.5±1.1%). In contrast, only smaller, less intense foci containing pCHK2 and γ-H2AX were detected in undifferentiated HFKs (3.3±0.5%), and the number did not increase upon differentiation. When HPV-31 positive cells were treated with the ATM inhibitor, the formation of the pCHK2 foci was substantially diminished (21.5% to 3.4%), further supporting a dependence on ATM activity for CHK2 activation (Figure 3B). Calcium-induced differentiation of HPV-31 positive cells and normal HFKs was confirmed by staining for K10 (Figure 3C).
It was next important to confirm that pCHK2 Thr68 localization to nuclear foci was not specific to calcium-induced differentiation. For these studies, HPV-31 positive keratinocytes, as well as normal keratinocytes, were induced to differentiate by growth in organotypic raft cultures, and immunohistochemistry was performed on cross sections of the rafts. As shown in Figure 3D, pCHK2 Thr68 was detected in the basal layer of HPV-31 positive rafts, as well as in a large number of suprabasal cells (67%), while only a few cells in the basal layer of normal HFK rafts had a comparable signal (5%). These results correlate with the CHK2 activation and nuclear localization observed upon calcium-induced differentiation. In addition to CHK2, we also observed similar staining patterns for pATM Ser1981, γ-H2AX and MRE11 in rafts of HPV-31 positive cells (Figure S2). Differentiation of HFK-31 cells, as well as normal HFKs, using the raft system was confirmed by staining for K10 (Figure S2). Overall, these results indicate that an ATM DNA damage response is activated in HPV-positive cells.
We next investigated if activation of the ATM pathway is necessary for stable replication of HPV genomes in undifferentiated cells, or viral genome amplification in differentiated cells. Since our studies identified activated CHK2 and BRCA1 in undifferentiated HPV-positive cells, we first examined the effect of inhibiting ATM kinase activity on episomal maintenance (Figure 4A). HPV-31 positive cells were grown in monolayer cultures and treated every two days with either DMSO as a vehicle control, or 5 uM KU-55933 to inhibit ATM kinase activity. The cells were passaged every five days and DNA was harvested at every other passage. Southern blot analysis was then performed to examine the status of the episomal viral DNA. As shown in Figure 4A, cells treated with DMSO maintained a similar number of episomes over time. We also found that cells treated with the ATM inhibitor exhibited only modest fluctuations in copy number as a function of extended passage. This experiment was performed multiple times with similar results. Since viral episomes were not rapidly lost upon passaging, these results indicate that ATM activity is not essential for the stable maintenance of episomes in undifferentiated cells.
To determine if ATM kinase activity is necessary for differentiation-dependent viral genome amplification, HPV-31 positive CIN612 cells were induced to differentiate in high calcium medium in the presence of DMSO or KU-55933. DNA was harvested from monolayer cells (0 hour), as well as 48 and 96 hours post-differentiation. As shown in Figure 4B, Southern blot analysis for HPV DNA demonstrated that cells treated with the ATM inhibitor exhibited significantly impaired viral genome amplification as compared to cells treated with DMSO alone. This experiment was repeated five times with identical results, demonstrating that ATM activity is necessary for productive viral replication. Since CHK2 is a major transducer of ATM signaling, we next wanted to determine if CHK2 activity is necessary for viral replication in differentiating cells. For this study, we used a specific inhibitor of CHK2 activity that effectively blocks phosphorylation of downstream targets, without affecting its own phosphorylation (Figure S4). Similarly to the ATM inhibitor, treatment of HPV-31 positive cells with the CHK2 inhibitor resulted in greatly diminished viral genome amplification upon differentiation (Figure 4C), indicating that ATM signaling to CHK2 is essential for this activity. Similar effects were also seen upon treatment of HFK-31 cell lines with the ATM and CHK2 inhibitors (data not shown), as well as when viral DNA was linearized by restriction digestion (Figure S3). Since epithelial differentiation is necessary for activation of viral genome amplification [41], we wanted to ensure that KU-55933 treatment did not act indirectly by blocking epithelial differentiation. For this analysis, we examined the expression levels of the differentiation-specific markers K10 and involucrin by Western blot analysis (Figure 4D). The subpopulation of differentiating cells that are amplifying HPV DNA are in S or G2 phase and do not express K10, while adjoining cells in G0/G1 synthesize high levels of K10 [29]. This is consistent with studies showing that upon keratinocyte differentiation, K10 is expressed in post-mitotic cells that are still metabolically active [42]. Treatment of HPV-31 positive cells with concentrations of KU-55933 up to 10 uM minimally affected K10 and involucrin expression, indicating that ATM inhibition does not alter differentiation (Figure 4D). To ensure that DMSO does not affect epithelial differentiation, we compared K10 expression between HPV-31 positive cells induced to differentiate in high calcium in the presence or absence of DMSO, and found no differences (Figure 4E).
Since we observed the formation of nuclear foci containing pATM Ser1981 in HPV-31 positive cells and found that ATM kinase activity is necessary for viral genome amplification, we investigated whether ATM activity is required for the formation of HPV replication foci in differentiating cells. For these studies, tyramide-enhanced fluorescence in situ hybridization (FISH) was used to detect viral genomes in cells that were treated with DMSO, or 10 uM of the ATM inhibitor. In monolayer cells, only single foci of viral genomes were detected in a small number of cells (9.6±2.3%) (Figure 5), which is consistent with previous observations [29]. In contrast, after differentiation for 48 hours in high calcium medium, the number of cells containing viral genome foci, as well as number and size of the foci, greatly increased (48±1.9%). Treatment with the ATM inhibitor prevented the formation of multiple foci per cell, resulting in a staining pattern similar to that of undifferentiated cells, providing further evidence that ATM activity is necessary for viral replication in differentiating cells.
We previously showed that HPVs induce low-levels of caspase 3/7 activation upon differentiation and that this is important for cleavage of the E1 replication protein and genome amplification [7]. The studies described above indicate that HPV requires CHK2 activity for productive replication (Figure 4C). In addition to its activity in DNA repair and cell cycle checkpoint function, CHK2 also plays a role in damage-induced apoptosis [16],[43],[44],[45], and thus could potentially contribute to the caspase activation observed in differentiating HPV-31 positive cells. To investigate this possibility, HPV-31 CIN612 cells were induced to differentiate in high calcium in the presence of DMSO or 5 uM of the CHK2 inhibitor. Cells extracts were harvested as a function of time and examined by Western blot analysis for cleavage of caspase-7. As shown in Figure 6, inhibition of CHK2 significantly impaired caspase-7 cleavage as compared to cells treated with DMSO alone, without affecting total levels of caspase-7. Similar results were found upon treatment of HFK-31 cells with the CHK2 inhibitor (data not shown). This finding, coupled with the observation that CHK2 is required for productive replication, indicates that HPVs utilize ATM signaling to promote caspase activation through CHK2, allowing for enhanced viral replication in differentiating cells.
To determine which viral proteins could be responsible for ATM activation, we first examined a possible role of E7 in this process, since it has been shown to de-regulate cell cycle control upon differentiation [46]. Our studies indicate that keratinocytes stably expressing HPV-31 E7 from the pLXSN retroviral vector exhibit phosphorylation of CHK2 on Thr68 in undifferentiated and differentiated keratinocytes, with little activation observed in cells expressing the pLXSN vector alone (Figure 7A). Since E7 mediates the majority of its functions through protein-protein interactions [46], we next investigated if E7 interacts with ATM. For these studies, we utilized the osteosarcoma cell line U2OS, which contains both wild-type p53 and Rb. U2OS cells were transfected with expression vectors for HA-tagged HPV-31 E7, HA-E7 lacking the LXCXE Rb binding motif (ΔLHCYE), or with an HA-E7 HDAC binding mutant (L67R). Following transfection, proteins associated with endogenous ATM were isolated by immunoprecipitation using an antibody to ATM, followed by Western blot analysis using an antibody to HA to detect precipitated E7 proteins. As shown in Figure 7B, wild-type E7 and the HDAC binding mutant were both able to co-immunoprecipitate with ATM, although the HDAC binding mutant was less efficient in doing so. In contrast, the interaction between E7 and ATM was abrogated in cells transfected with the LXCXE Rb-binding mutant, which is also defective for productive HPV replication [47]. Several proteins in addition to Rb have been shown to interact with E7 at this site [46], and it is unclear whether the binding of E7 to ATM is direct or is mediated through another protein or complex of proteins. We also performed immunoprecipitations using an antibody to the HA tag, followed by Western blot analysis for pATM Ser1981, and found that E7 is able to bind the phosphorylated form of ATM (Figure 7C). Again, the HDAC binding mutant was still able to interact with pATM, but to a lesser degree than wild-type E7. We consistently observed that the L67R mutant was expressed at lower levels than wild-type HA-E7, as well as the Rb-binding mutant, and likely accounts for the decreased binding of this mutant to ATM. To determine if ATM phosphorylation is necessary for E7 to interact with ATM, U20S cells transfected with wild-type HA tagged E7 were treated with 10 uM of the ATM inhibitor KU-55933, which inhibits the phosphorylation of ATM (Figure 1D). Endogenous ATM was precipitated, followed by Western blot analysis using an antibody to HA to examine the presence of E7 in the precipitated complexes. As shown in Figure 7D, E7 co-precipitated with ATM in cells treated with DMSO alone, but not in cells treated with the ATM inhibitor. Since total levels of ATM were not affected by treatment with the ATM inhibitor (Figure 7D), these results suggest that E7 interacts primarily with the phosphorylated form of ATM. Overall, these results indicate that E7 activates CHK2, possibly through its association with ATM, which may in turn be important for the productive phase of the viral life cycle.
In this study, we show that human papillomaviruses activate the ATM DNA damage response and that this is necessary for productive viral replication upon differentiation. These findings identify a primary regulatory mechanism responsible for HPV genome amplification. Our studies indicate that papillomaviruses induce phosphorylation of ATM, as well as its substrates CHK2 and BRCA1, in undifferentiated cells, but this activation has minimal effect on the long-term maintenance of HPV episomes. In differentiating cells, the phosphorylation of NBS1, well as CHK2 and BRCA1, was observed and inhibition of the ATM pathway completely blocked amplification of viral genomes. In addition, we found CHK2 activity to be required for HPV-mediated caspase activation, as well as viral genome amplification. It is possible that activation of the ATM response in differentiating cells induces an S or G2/M arrest that provides an environment conducive to productive viral replication. HPV genomes replicate bi-directionally via theta structures in basal cells, but may switch to replication by a rolling circle mechanism during amplification [48]. This switch in replication modes may also require activation of the ATM pathway specifically in differentiating cells.
HSV-1 and SV40 have been shown to recruit members of the ATM DNA damage pathway to specific sites of replication in the nucleus [49],[50],[51]. In differentiating cells undergoing productive replication of HPV genomes, we observed the formation of nuclear foci containing pATM, pCHK2 and γ-H2AX, as well as the MRN components, MRE11 and RAD50. The addition of ATM inhibitors prevented formation of these foci and blocked viral genome amplification, further implicating these DNA repair proteins as important regulators of HPV productive replication. We did not observe a complete loss in the formation of γ-H2AX foci upon treatment with the ATM inhibitor, however this may be due to the actions of other checkpoint kinases, such as ATR and DNA-PK [27]. Consistent with this idea, we observed ATM-independent phosphorylation of the ATR target CHK1 on Ser317 in undifferentiated cells, as well as in differentiated cells, although at decreased levels.
Our studies further demonstrate that the addition of a specific inhibitor to CHK2 blocks viral genome amplification. Upon activation, CHK2 phosphorylates the Cdc25a and Cdc25c phosphatases to initiate cell cycle arrest in G1/S or G2/M through their degradation or cytoplasmic sequestration, respectively [52]. Cdc25c is necessary for the activation of Cdc2/Cdk1, as well as entry into mitosis, and its inactivation plays a central role in inducing G2/M arrest [53]. Preliminary results indicate that phosphorylation of Cdc25c on Ser216 increases in HPV positive differentiating cells (Figure S4), as does inhibitory phosphorylation of Cdc2/Cdk1 on Tyr15 (Moody and Laimins, unpublished data), and this occurs in a CHK2-dependent manner. These results are consistent with ATM-dependent activation of CHK2 leading to a G2/M arrest and viral genome amplification. Recent studies suggest that HPV infected cells undergoing productive replication are arrested at G2/M rather than in S-phase and is consistent with our findings [54].
The MRN complex appears to be important for HPV amplification, as the phosphorylated forms of CHK2, NBS1 and BRCA1 were detected in differentiated HPV positive cells. For some viruses, the DNA damage response represents an obstacle that must be overcome for efficient replication. For example, the adenovirus E1b55K/E4orf6 proteins induce degradation of the MRN complex, blocking NBS1 phosphorylation and preventing concatemerization of viral genomes [15],[39]. In contrast, differentiating HPV positive cells exhibited high levels of NBS1, MRE11 and RAD50, which were maintained throughout differentiation. In addition, we observed that NBS1 phosphorylation occurs concomitantly with viral genome amplification, implicating NBS1 as a potential regulator of replication. The Rb associated transcription factors E2F1 and E2F2 have been shown to interact with MRE11 and NBS1 at both viral and cellular origins of replication [55]. Upon differentiation, E7 induces increased expression of E2F2, as well as its re-localization to nuclear foci, which is necessary for viral genome amplification [56]. Recent studies indicate that E2F2 binds to the region around the HPV-31 replication origin and increased binding was observed upon differentiation [57]. It is possible that E2F2 directs MRN components to HPV origins to ensure integrity of replication forks and promote replication in differentiating cells.
Caspase activation has been shown to be an important and novel means by which HPV proteins regulate amplification [7]. Low level caspase activation by E6 and E7 upon differentiation induces cleavage of the E1 protein, which is required for efficient viral genome amplification. Cleavage of E1 results in enhanced binding of E1 to the origin and the ability to replicate in an E2-independent manner (Moody, Archambault et al. unpublished data). In the present study, we have found that CHK2 kinase activity is necessary for caspase activation, as well as viral genome amplification in differentiating cells. These results provide a possible link between caspase cleavage and the activation of CHK2 through the binding of ATM to E7. In preliminary studies, we have found that E6 can also activate CHK2 (Moody and Laimins, unpublished data), although it is unclear whether E6 utilizes the DNA damage response to induce caspase activation in differentiating cells. Several recent studies have demonstrated a role for CHK2 in DNA damage-induced apoptosis [16],[43],[44],[45]. In response to DNA damage, CHK2 phosphorylates E2F1, resulting in its stabilization, transcriptional activation, and induction of p53-dependent or -independent apoptosis [16],[43]. E2F1 is expressed at high levels in differentiating HPV positive cells [56], and it is possible that CHK2 mediates caspase activation through E2F1. Interestingly, our studies indicate that HPV induces caspase activation only in differentiating cells, while activation of CHK2 is observed in both undifferentiated and differentiated cells. This suggests that while CHK2 activation is necessary for caspase activation, it alone is not sufficient and may require differentiation-specific factors as well as other members of the ATM pathway for this function. Several viruses have been shown to utilize DNA damage response for productive replication, and it is possible that these viruses also utilize low-level caspase activation as part of their life cycle.
Our studies indicate that in stable cell lines E7 activates CHK2, and that it forms a complex with the phosphorylated form of ATM. Deletion of the LXCXE Rb binding domain in E7 abrogated ATM binding, however binding could occur directly or indirectly through another protein. Preliminary results using Rb-deficient Saos-2 cells indicate that E7 binding to ATM does not require Rb (Moody and Laimins, unpublished data). Previous studies have shown that deletion of the LXCXE binding domain in E7 blocks HPV genome amplification [47], and in our studies this motif is important for ATM binding. This is consistent with the idea that E7's interaction with ATM may be necessary for productive replication.
In addition to E7, the HPV replication protein E1 may also contribute to the ATM response. Upon differentiation, the expression of E1 increases, contributing to enhanced viral replication [58]. Kadaja et al demonstrated that heterologous high-level expression of E1 initiates replication from integrated HPV origins multiple times in a single S-phase [59]. This suggests that E1 may disrupt normal licensing control in differentiating cells, allowing for re-replication of HPV DNA and activation of an ATM response. It is possible that multiple HPV proteins act cooperatively to activate the full ATM response.
In summary, our studies demonstrate that HPV proteins activate the ATM DNA damage response and that this is necessary for amplification of viral genomes upon differentiation. The formation of HPV replication foci in differentiating cells is dependent upon ATM activity and suggests that DNA repair proteins may participate directly in viral replication. Importantly, we have established a link between caspase activation and the DNA damage response. Caspase 3/7 consensus cleavage motifs are found at conserved locations in E1 proteins of almost all papillomavirus types, and we suspect caspase activation may be necessary for their productive replication. We believe that activation of the ATM pathway will prove to be a common mechanism utilized by HPVs to promote viral replication in differentiating cells. These observations suggest that the ATM pathway may be an effective therapeutic target to block the spread of HPV infections.
Human foreskin keratinocytes (HFKs) were derived from neonatal human foreskin epithelia and maintained in E medium containing mouse epidermal growth factor (EGF) and mitomycin-treated J2 fibroblasts as previously described [60]. Human osteosarcoma cells (U2OS) were maintained in Dulbecco's modified Eagle's medium (DMEM) containing 10% bovine serum. CIN612 is a clonal cell line that stably maintains HPV-31 episomes. CIN612 cells were maintained in E medium with EGF and J2 fibroblast feeders. Before harvesting DNA or protein, fibroblast feeders were removed by treatment with phosphate-buffered saline (PBS) containing 0.1% of .5 M EDTA for two minutes, followed by two washes in PBS. Creation of HFK cell lines containing retroviral constructs has been previously described [47].
The pBR322min-HPV31 plasmid has been described [61]. The HA-tagged E7 proteins were previously described [47] and are as follows: HA-E7 ΔLHCYE contains an in-frame deletion of the Rb binding domain and HA-E7 L67R contains a point mutation in the HDAC binding site, converting a leucine to an arginine. KU-55933 and the CHK2 inhibitor II were obtained from Calbiochem.
Transfection of HFKs and selection for cells stably maintaining HPV-31 genomes were performed as described previously [62]. Briefly, HPV-31 genomes were released from the pBR322 plasmid by digestion with HindIII. Viral genomes were then unimolecularly ligated with T4 DNA ligase (New England Biolabs) and precipitated with isopropyl alcohol. HFKs were co-transfected with 1 ug of religated genomes and 1 ug pSV2-Neo using FuGene6 according to the manufacturer's protocol (Roche). Selection was carried out for eight days in the presence of G418 (Sigma). After selection was complete, pooled populations were expanded for further analysis.
Differentiation in high calcium was performed as described previously [7]. Briefly, upon reaching 90% confluency, HPV-positive cells and normal HFKs were cultured in keratinocyte basal medium (KBM) with supplements (Invitrogen) for 24 hours. Cells were then switched to KBM (without supplements) containing 1.5 mM calcium chloride, and where indicated were cultured with either DMSO, 5 or 10 uM KU-55933, or 5 uM of the CHK2 inhibitor. At 48 and 96 hours, DNA was harvested from one half of the cells, and protein was harvested from the other half. Viral genome amplification was then measured by Southern blot analysis to examine the episomal (supercoiled) form of DNA to ensure that the productive phase of the viral life cycle was activated. Western blot analysis was performed to analyze the expression of differentiation-specific proteins.
Whole cell extracts were harvested in RIPA lysis buffer and quantified using the Bio-Rad protein assay. Western blot analysis was performed as described [47]. Equal amounts of protein were electrophoresed on SDS-polyacrylamide gels and subsequently transferred to polyvinylidene difluoride membrane (Immobilon-P; Millipore). Primary antibodies were as follows: phospho-ATM Ser1981 was purchased from R&D Systems. Phospho-CHK2 Thr68, CHK2, BRCA1, K10 and Involucrin were purchased from Santa Cruz. ATM was purchased from Calbiochem. Phospho-CHK1 Ser317, CHK1, phospho-CHK2 Thr68, phospho-BRCA1 Ser1524, Cleaved Caspase-7, and Caspase-7 were purchased from Cell Signaling. Phospho-NBS1 Ser343, MRE11 and RAD50 were purchased from Genetex. NBS1 was purchased from Novus Biologicals. Secondary antibodies included horseradish peroxidase-linked anti-rabbit (Cell Signaling Technologies) and horseradish peroxidase-linked anti-mouse (Santa Cruz). DNA isolation and Southern blot analysis were performed as described [63].
U20S cells were transfected at 30% confluency with one microgram of wild-type or mutant HA-tagged HPV-31 E7 proteins using FuGene6, according to the manufacturer's instructions (Roche). After 48 hours, lysates were harvested as previously described [47]. Immunoprecipitations were performed using one milligram of protein lysate. The samples were pre-cleared for 1 hour with 40 µl of protein G agarose (Roche) at 4°C, then incubated overnight with either a mouse anti-HA antibody (Santa Cruz) or rabbit anti-ATM antibody (Calbiochem). Protein complexes were then precipitated using 50 µl of protein G agarose for 4 hours at 4°C. Immunoprecipitated complexes were then washed three times with lysis buffer, and subsequently analyzed by Western blot analysis using either mouse anti-HA, mouse anti-pATM Ser1981 (Rockland), or rabbit anti-ATM antibodies.
HPV positive cells and normal HFKs were grown on coverslips and induced to differentiate in high calcium in the presence of DMSO or 10 uM KU-55933. At time 0 (undifferentiated) and 48 hr post-calcium induced differentiation, the cells were washed three times in cold phosphate buffered saline (PBS), fixed in 4% paraformaldehyde in PBS for 15 minutes, then permeabilized in 1%Triton X-100-PBS for 10 minutes. Cells were blocked with PBS containing 10% bovine serum albumin (BSA) for one hour at room temperature. Primary antibodies were diluted in PBS containing 10% BSA and incubated on coverslips overnight at 4°C. The samples were then washed in PBS and stained with fluoroscein isothiocyanate (FITC)-conjugated anti-rabbit antibody (1∶50 dilution) (Zymed) or AlexaFluor 568 anti-mouse antibody (Invitrogen) (1∶400 dilution) for one hour at room temperature. Primary antibody dilutions for mouse anti-phospho-ATM Ser1981 (Rockland), anti phospho-H2AX Ser139 (γ-H2AX) (Upstate) were 1∶400 and 1∶500, respectively. Rabbit anti-phospho-CHK2 Thr68 (Cell Signaling) and anti-phospho-H2AX Ser139 (γ-H2AX) (Cell Signaling) were diluted 1∶50. Mouse anti-MRE11 and mouse anti-Rad50 were diluted 1∶200. Mouse anti-keratin 10 (K10) (Santa Cruz) was diluted 1∶100. Sections from normal keratinocyte raft cultures or HPV-31 transfected keratinocyte raft cultures were examined by immunofluorescence as described previously [63]. Cross sections of rafts were stained using a 1∶50 dilution of anti-phospho-CHK2 Thr68 (Cell Signaling), 1∶200 of anti-pATM Ser1981, or MRE11, 1∶500 of γ-H2AX Ser139 (Millipore), and 1∶100 of anti-K10 (Santa Cruz). Cellular DNA was counterstained with DAPI, and the coverslips or slides were mounted in Vectashield (Vector Laboratories). Confocal images were acquired by a UV LSM 510 confocal laser-scanning microscope (Zeiss). To quantify number of cells containing DNA repair foci, at least 100 cells were counted for three independent experiments. The average number of cells containing foci is indicated, along with the standard deviation.
HPV-31 genomic DNA probes for FISH were prepared by nick translation of plasmid genomic DNAs using the BioNick labeling system according to the manufacturer's instructions (Invitrogen). Viral DNA was detected by tyramide fluorescent in situ hybridization as previously described [64]. Briefly, 1×106 undifferentiated HFK-31 cells, or HFK-31 cells differentiated in high calcium for 48 hr were spread on Superfrost Plus slides (Fisher) and allowed to air-dry. Cells were fixed with 4% paraformaldehyde at room temperature followed by permeabilization in 1× PBS, 0.5% Triton X-100 for 10 min. The slides were treated with 100 ug/ml RNase A in 2× SSC for 1 hour at 37°C. Subsequently, the slides were washed three times with 2× SSC, then dehydrated for 2 min each in 70% EtOH, 85% EtOH and 100% EtOH at room temperature. Slides were then denatured in 70% formamide-2× SSC at 74°C for 2 minutes, followed by dehydration for 2 min each in 70% EtOH (−20°C), 85% EtOH and 100% EtOH at room temperature. The probe was denatured at 74°C for 10 minutes, and then 10 ul of probe was hybridized overnight to the denatured slide at 37°C. After overnight incubation, the slides were washed multiple times, and tyramide-enhanced fluorescence was carried out according to the manufacturer's instructions (Perkinelmer). The cellular DNA was counterstained with DAPI, and the slides were mounted in Vectashield. Images were collected using a UV LSM 510 confocal laser-scanning microscope (Zeiss).
|
10.1371/journal.pcbi.1003291 | On the Role of Aggregation Prone Regions in Protein Evolution, Stability, and Enzymatic Catalysis: Insights from Diverse Analyses | The various roles that aggregation prone regions (APRs) are capable of playing in proteins are investigated here via comprehensive analyses of multiple non-redundant datasets containing randomly generated amino acid sequences, monomeric proteins, intrinsically disordered proteins (IDPs) and catalytic residues. Results from this study indicate that the aggregation propensities of monomeric protein sequences have been minimized compared to random sequences with uniform and natural amino acid compositions, as observed by a lower average aggregation propensity and fewer APRs that are shorter in length and more often punctuated by gate-keeper residues. However, evidence for evolutionary selective pressure to disrupt these sequence regions among homologous proteins is inconsistent. APRs are less conserved than average sequence identity among closely related homologues (≥80% sequence identity with a parent) but APRs are more conserved than average sequence identity among homologues that have at least 50% sequence identity with a parent. Structural analyses of APRs indicate that APRs are three times more likely to contain ordered versus disordered residues and that APRs frequently contribute more towards stabilizing proteins than equal length segments from the same protein. Catalytic residues and APRs were also found to be in structural contact significantly more often than expected by random chance. Our findings suggest that proteins have evolved by optimizing their risk of aggregation for cellular environments by both minimizing aggregation prone regions and by conserving those that are important for folding and function. In many cases, these sequence optimizations are insufficient to develop recombinant proteins into commercial products. Rational design strategies aimed at improving protein solubility for biotechnological purposes should carefully evaluate the contributions made by candidate APRs, targeted for disruption, towards protein structure and activity.
| Biotechnology requires the large-scale expression, yield, and storage of recombinant proteins. Each step in protein production has the potential to cause aggregation as proteins, not evolved to exist outside the cell, endure the various steps involved in commercial manufacturing processes. Mechanistic studies into protein aggregation have revealed that certain sequence regions contribute more to the aggregation propensity of a protein than other sequence regions do. Efforts to disrupt these regions have thus far indicated that rational sequence engineering is a useful technique to reduce the aggregation of biotechnologically relevant proteins. To improve our ability to rationally engineer proteins with enhanced expression, solubility, and shelf-life we conducted extensive analyses of aggregation prone regions (APRs) within protein sequences to characterize the various roles these regions play in proteins. Findings from this work indicate that protein sequences have evolved by minimizing their aggregation propensities. However, we also found that many APRs are conserved in protein families and are essential to maintain protein stability and function. Therefore, the contributions that APRs, targeted for disruption, make towards protein stability and function should be carefully evaluated when improving protein solubility via rational design.
| Irreversible β-strand driven protein aggregation and amyloidogenesis is a tremendous burden to biological organisms. Protein loss-of function due to aggregation causes stress to the cell and metabolic energy is lost on the expression, synthesis, and degradation of proteins which aggregate. To overcome these challenges and build cellular machineries that can sustain metabolic flux, higher organisms have developed sophisticated protein quality control mechanisms, including molecular chaperones, post-translational modifications, and degradation/clearance pathways to prevent aggregation from disrupting homeostasis [1]–[3]. When quality control mechanisms are impaired, due to aging or otherwise, protein aggregation can lead to ‘conformational diseases’ in humans and animals [1], [3]–[5].
Despite its deleterious effects, protein aggregation remains unavoidable due to the inherent physico-chemical properties of protein sequences and the formation of non-native conformations due to sequence mutation or unfolding events in response to environmental stress. However, studies of amyloidogenic proteins have revealed that different protein sequences vary in their propensity to aggregate, which can be attributed to the presence of aggregation-nucleating short sequence stretches, capable of forming the cross-β steric zipper motif, called aggregation prone regions (APRs) [6]–[10]. Analyses of APRs indicate common sequence properties including a high preference for β-branched hydrophobic residues, strong β-sheet propensity, low net charge, and in the case of fibril forming patterns, position-specific charged residues [11], [12]. Knowledge of these properties has enabled the development of phenomenological and first-principle based methods to predict APRs in any protein sequence [13]–[20].
The availability of computational APR prediction tools has facilitated large-scale investigations into the aggregation propensities of protein sequences [21]–[27]. Analyzing intrinsically disordered protein (IDP) sequences using APR prediction tools has revealed that the number of APRs found in IDPs is three times less than those found in sequences for ordered proteins [21]. Given the tendency for APRs to exist in ordered sequence regions, it was proposed that APRs may have a role in promoting structural order in native folds. More recent studies have extended the concept that APRs can play a role in promoting structural order based on the prevalence of APRs in protein-protein interactions sites [22], [28], including antibody-antigen interfaces [23]. In fact, the trend for APRs to exist in protein-protein interaction sites has led some research groups to repurpose their APR prediction methods into tools for identifying potential protein-ligand [29] and protein-protein interaction sites [28]. On the other hand, mounting evidence from analyses of large sequence datasets strongly suggests that nature is actively minimizing the occurrence/impact of aggregation prone regions in protein sequences. For example, APRs are frequently punctuated by charged or proline residues (labeled gate-keepers) [24], proteins with higher aggregation propensities have shorter half-lives in the cell [25], and mRNA expression levels in E. coli are lower for proteins with higher aggregation propensities [26]. A study linking protein evolution and aggregation discovered an overall decreasing trend for aggregation propensity among organisms with increasing complexity and longevity [30]. This implies that organisms have evolved by minimizing the aggregation propensity of their proteomes. Subsequently, it was determined that the aggregation propensity differences between prokaryotic and eukaryotic proteomes could be explained by differences in their number of IDPs, which have fewer APRs than ordered proteins do [27]. This contradictory finding, suggests that organisms may not have minimized the aggregation propensity of their proteomes during the course of evolution. In light of the above reports, there is a need to further investigate the various roles that aggregation prone regions have in protein structure and the concept that nature is actively minimizing the aggregation propensity of protein sequences.
This report presents findings from comprehensive analyses on multiple non-redundant datasets of protein sequences and structures (see Methods). Datasets were analyzed for aggregation propensity differences among monomeric proteins, IDPs, and randomized sequences with uniform and natural amino acid compositions. To assess if evolutionary selective pressure has minimized the aggregation propensity of protein sequences through APR disruption, average sequence identity among homologous proteins was compared to percent APR conservation among the same sequences. IDPs and monomeric proteins were used to evaluate the role that APRs have in promoting structural order and stabilizing interactions. Proximity of APRs to catalytic sites in enzyme structures was also investigated. Our findings suggest that proteins have evolved by optimizing their risk of aggregation for cellular environments through the overall minimization of aggregation prone regions and the conservation of those important for folding and function. In addition to promoting aggregation under conditions that destabilize proteins, APRs also stabilize protein structure and resist disorder, particularly, in structural areas that are important for protein function. Therefore, strategies employing site directed mutagenesis to improve protein solubility should carefully evaluate the contributions made by candidate APRs, targeted for disruption, towards protein structure and activity. The major findings from this work are summarized in Table 1.
To assess the various roles that aggregation prone regions have in protein structure and the concept that evolution is actively minimizing the aggregation propensities of protein sequences, several non-redundant sequence datasets were generated (see Methods) that include: a library of experimentally proven amyloid-like fibril forming peptide sequences [31], [32] (Amylsegs); sequences and structures of 495 small (sequence length, 52–200 residues; average, 152±34) monomeric (both in crystal asymmetric units and in biological units) and non-homologous (sequence identities ≤30% in all against all alignments) proteins with high resolution crystal structures (R≤2.0 Å), (F495); 536 non-homologous IDP sequences obtained from the DisProt database [33] (IDP536); and 961 catalytic residues in 314 non-homologous protein chains (299 proteins) derived from a dataset of functional residues compiled by Xin et al. [34] (Cata). These datasets were supplemented with two random sequence datasets (R10000 and N10000) each with 10,000 amino acid sequences, 100 residues long. A uniform distribution of amino acids (5% for each amino acid) was used to generate random sequences for R10000. Random sequences in N10000 were generated from the amino acid distribution of naturally occurring protein sequences found in F495. Each protein sequence in F495 was also scrambled one hundred times to obtain a dataset, SF49500, which contains 49,500 sequences that have the same amino acid distribution, but not the same patterning as sequences in F495 (see Methods). The F495 dataset was also divided into two datasets, F1 and F2, of roughly equal number of sequences but significantly different amino acid compositions. Both F1 and F2 retain their natural protein sequence patterning but have different amino acid compositions. The results described in this report are organized into five sections. In the first section, evidence is presented that indicates the aggregation propensities of protein sequences have been minimized in comparison to random sequences with uniform or natural amino acid compositions. The second section describes an examination into whether evolution disrupts or conserves APRs in the sequences of homologous proteins. The third section compares the incidence of APRs in ordered versus disordered residues in proteins. This is followed by an evaluation of the contributions that all predicted APRs make towards stabilizing the proteins in which they exist compared to other protein segments of equal length from the same protein. The last section reports on the spatial proximity of catalytic residues and APRs to assess if APRs have a role in maintaining protein function.
Tables 2–4 present data on the various statistical measures of aggregation that were used in this study. Aggregation propensities were computed by normalizing total TANGO [11] and WALTZ [12] aggregation scores for protein sequences by their lengths. The average aggregation propensities for sequences of our various datasets are summarized in Table 2. For each dataset, the average length of protein sequences which contain at least one APR, total number of predicted APRs, average APR lengths and average proportion of APR residues in protein sequences are summarized in Table 3. The incidence of gate-keeper residues in APR flanking regions for sequences in R10000, N10000, SF49500, F495 and IDP536 datasets is presented in Table 4. Figure 1 shows box and whisker plots of the aggregation propensities for all datasets. These plots indicate an overlap among the aggregation propensity distributions for the various datasets. Overlaps are expected because the amino acid sequences in all datasets are composed of the same twenty types of naturally occurring amino acids found in proteins. Notwithstanding the overlaps, it can be seen that there are important differences among datasets in the inter-quartile ranges for both TANGO and WALTZ predicted aggregation propensities. To determine the statistical significance of aggregation propensity distribution differences among datasets, two sample t-tests were performed (See Methods) and the results are summarized in Table S1 of the Supplementary Material. In the following discussion, aggregation property averages (e.g. propensity or APR length, etc.) are compared among the various datasets. Standard deviations are provided along with averages.
Amino acid sequences in N10000 and SF49500 have a lower TANGO average aggregation propensity (i.e. average normalized aggregation score) than those in R10000 (Table 2). The average TANGO aggregation propensity for N10000 is 5.42±5.18, which is ∼16% lower than the corresponding value for R10000, 6.48±5.89. The average TANGO aggregation propensity for SF49500 is 5.20±4.33, almost 20% lower. Consistent with these averages, boxes representing inter-quartile ranges of TANGO aggregation propensities for N10000 and SF49500 are smaller (Figure 1) and the distributions of TANGO predicted aggregation propensities for N10000 and SF49500 are significantly different from those in R10000 at the 95% confidence level (Table S1). Interestingly, the percentage of sequences with at least one TANGO predicted APR in SF49500 is 71.4%, which is higher than that in N10000 (62.6%) or R10000 (68.9%). The number of APRs per sequence is also higher for SF49500 (59878 APRs/49500 sequences) = 1.21) than that for R10000 (10474/10000≈1.05) or N10000 (8965/10000≈0.90), (Table 3). However, the average length of APR containing sequences in SF49500 is 149±35, which is greater than that for R10000 and N10000 (length 100), and the average proportion of APR residues (see Methods, Equation 1) in SF49500 (10.3±5.4%) and N10000 (12.2±6.3%) is lower than those in R10000 (13.4±7.2%), (Table 3). These values clarify why the percentage of sequences with at least one TANGO predicted APR is greater in SF49500 than in R10000 but the average aggregation propensity is lower for SF49500 than in R10000.
The smaller average TANGO aggregation propensity of N10000 and SF49500 compared to R10000 is a consequence of their differences in amino acid composition. The amino acid distributions in N10000 and SF495000 are identical and significantly different from the amino acid distribution in R10000. The χ2 value for the amino acid distributions in R10000 versus N10000 (and SF49500) is 160.84. For 19 parameter distributions, χ2 values >43.82 reject the null hypothesis at the 99.9% confidence level (p-value<0.001) that the two amino acid distributions are the same [35]. N10000 and SF49500 contain both fewer aggregation-promoting residues and more gate-keeper residues than R10000. The total proportion of aggregation-promoting β-branched nonpolar (Ile and Val), aromatic (Phe, Tyr and Trp) and polar (Asn and Gln) residues is smaller in N10000 and SF49500 (29%) than in R10000 (35%) while the total proportion of gate-keeper charged and proline residues (Asp, Glu, Lys, Arg and Pro) is greater in N10000 and SF49500 (29.3%) than in R10000 (25%).
For WALTZ predicted APRs, the average aggregation propensities for N10000 and SF49500 are again smaller (2.40±2.75 and 2.34±2.48, respectively) than that for R10000 (3.39±3.33), (Table 2). WALTZ predicted APRs also constitute 7.2±3.8% of residues in SF49500 (8.5±3.9% in N10000) as compared to 9.4±4.7% in R10000, (Table 3). Note, only sequences that contain at least one APR were used in these calculations. WALTZ aggregation propensity box and whisker plot shows that the inter-quartile ranges for N10000 and SF49500 are smaller than the inter-quartile range for R10000 (Figure 1). The distribution of WALTZ aggregation propensities in R10000 is also significantly different from those in N10000 and SF49500 at 95% confidence level (Table S1). Other observations for N10000 and SF49500 versus R10000 show similar trends as in TANGO predictions (Tables 2 and 3). In summary, the above statistical measures have highlighted the importance of amino acid composition in explaining the average TANGO and WALTZ aggregation propensity differences between N10000 and SF49500 versus R10000.
Sequences of monomeric proteins in F495 have a lower average TANGO aggregation propensity than randomized sequences with natural amino acid composition (N10000 and SF49500), (Table 2). The amino acid sequences in N10000 and SF49500 lack the sequence patterning features of F495 due to randomization and scrambling (see Methods). The average TANGO aggregation propensity for F495 (3.41±2.98) is 34% lower than that for SF49500 (5.20±4.33) and 37% lower than that for N10000 (5.42±5.18)), (Table 2). Box and whisker plots of TANGO aggregation propensities indicate that the inter-quartile range for F495 is smaller and the third quartile (Q3) is shifted to lower values compared to SF49500 and N10000 (Figure 1). The distribution of TANGO aggregation propensities in F495 is also significantly different, at the 95% confidence level, from those in N10000 and SF49500 (Table S1). The aggregation propensity differences between F495 versus N10000 and SF49500 result from a similar, but broader, set of APR differences as those between N10000, SF49500 and R10000. Specifically, the percentage of sequences with at least one TANGO predicted APR decreases from 71.4% in SF49500 (62.6% in N10000) to 60.4% in F495, even though, the average length of APR containing sequences in F495 (152±34 residues) is similar in SF49500 (149±35), and longer than that in N10000 (100 residues), (Table 3). The average proportion of APR residues also decreases from 10.3±5.4% in SF49500 (12.2±6.3% in N10000) to 7.8±4.1% in F495 and the average APR length in F495 decreases by one residue from 8.48±2.85 in SF49500 (8.49±2.88 in N10000) to 7.47±1.99 in F495 (Table 3). The above analysis of TANGO predicted APRs reveals the importance of sequence patterning in explaining the average aggregation propensity differences between N10000 and SF49500 versus F495. Similar observations have been made by Chiti and coworkers via experiments on a 29 residue peptide from horse heart apomyoglobin. Sequence scrambled variants of the peptide showed substantial increases in aggregation propensity compared to the wild-type sequence due to clustering of amyloidogenic residues [36].
Aggregation propensity differences between F495 versus N10000 and SF49500 for WALTZ predicted APRs are dissimilar to those of TANGO predicted APRs (Tables 2 and 3, and Figure 1). The inter-quartile range for WALTZ predicted aggregation propensities in F495 is similar to the inter-quartile ranges for SF49500 and N10000 (Figure 1) and t-tests show that the distribution of WALTZ predicted aggregation propensities in F495 is statistically similar to SF49500 and N10000 distributions. The average WALTZ aggregation propensity for sequences in F495 (2.52±2.70) is comparable to that for sequences in N10000 (2.40±2.75) and SF49500 (2.34±2.48). The average APR lengths for WALTZ predicted APRs in SF49500 (6.13±0.61), N10000 (6.12±0.63) and F495 (6.13±0.68) are also the same (Table 3). The percentage of sequences that contain at least one WALTZ predicted APR in SF49500 and F495 are again similar at 63.6% and 63.2% respectively. The average proportion of APR residues in sequences of SF49500 (7.2±3.8%), N10000 (8.5±3.9%) and F495 (7.7±3.9%) are also similar (Table 3). From the observations above, scrambling sequences (SF49500) or producing random sequences with natural amino acid compositions (N10000) did not increase the number of WALTZ predicted APRs as it did for TANGO. WALTZ uses position-specific matrices that enable the program to predict APRs with fibril forming charged residues [12]. Consequently, sequence randomization may not have produced sequence patterns that WALTZ position-specific matrices have defined as aggregation prone.
After dividing F495 into different SCOP subclasses, the average aggregation propensity and other statistical measures of aggregation, such as proportion of sequences with at least one APR, number of APRs per sequence and average proportion of APR residues, remain similar to F495 for all subclasses (Tables 2 and 3), except in the α/β subclass. The α/β subclass contains only 43 (∼9%) proteins from F495. The average TANGO aggregation propensity of the α/β subclass (4.95±2.72) is higher than that for F495 (3.41±2.98), but the corresponding values for WALTZ aggregation propensities are similar (2.37±2.12 for α/β and 2.52±2.70 for all F495), (Table 2). A greater percentage of sequences in the α/β subclass contain one or more TANGO/WALTZ predicted APRs (TANGO, 79.1% for α/β subclass versus 60.4% for F495; WALTZ, 72.1% for α/β subclass versus 63.2% for F495) and the number of APRs per sequence is also higher for the α/β subclass (TANGO, 1.53 for α/β subclass versus 0.91 for F495; WALTZ, 1.35 for α/β subclass versus 1.09 for F495), (Tables 2 and 3). However, the average APR lengths in the α/β subclass are similar to that for all F495 protein sequences. Excluding the α/β subclass, the results in Tables 2 and 3 for the various SCOP classes suggest that differences in topology and secondary structure content alone do not produce average aggregation propensity changes that are similar to those observed after modifying sequence composition and patterning.
Average TANGO aggregation propensities are similar between IDP536 and F495 (3.85±4.05 for IDP536 and 3.41±2.98 for F495, Table 2). Box and whisker plots for IDP536 and F495 also show that these datasets have similar predicted aggregation propensities (Figure 1). The χ2-test on amino acid compositions of F495 and IDP536 yields a value of 20.4, thereby, accepting the null hypothesis that they have the same amino composition. The number of TANGO predicted APRs per sequence (2.46, Table 3) and the average APR length (9.13±3.90) is considerably larger in IDP536 compared to F495 (0.91 and 7.47±1.99, respectively). However, TANGO predicted APR differences between IDP536 and F495 are offset by longer sequence lengths in IDP536 (763±562), which results in a similar average proportion of APR residues (7.8±4.1% for F495 and 6.9±5.4% for IDP536, Table 3) and similar average TANGO aggregation propensities (Table 2). Longer APR lengths may result from lower sequence complexity among IDPs [33], [37]. For WALTZ, the average aggregation propensity (1.74±1.60) and average proportion of APR residues (4.7±2.7%) for IDP536 is lower than that for F495 (2.52±2.70 and 7.7±3.9%, respectively), but the average APR lengths (6.09±0.54 for IDP536 and 6.25±1.15 for F495) remain similar (Table 3). To assess if IDP536 and F495 actually have similar or different average aggregation propensities, the total proportion of ordered and disordered residues is needed for both datasets. This data is available for IDP536, but it is not for F495. Therefore, comparisons among ordered and disordered regions of protein sequences, instead of full length sequences from folded and intrinsically disordered proteins, are needed to better understand the relationship between aggregation and structural disorder (see section entitled APRs have fewer disordered than ordered residues).
An analysis of APR flanking residues among sequence datasets indicates that APRs are, on average, flanked more often by gate-keeper residues (Asp, Glu, Lys, Arg, or Pro) in F495, N10000, SF49500 and IDP536 datasets than in R10000 (Table 4). The average number of gate-keeper residues flanking TANGO predicted APRs in F495 is 2.6±1.2. Comparable values are observed for N10000 (2.4±1.1), SF495000 (2.5±1.1) and IDP536 (2.4±1.1). However for R10000, the corresponding value is lower at 2.2±1.1. For WALTZ predicted APRs, the average number of flanking gate-keeper residues in F495 is 2.0±1.1 which is comparable to averages for N10000 (1.9±1.1), SF49500 (1.9±1.1), and IDP536 (2.0±1.2). Gate-keeper residues flank WALTZ predicted APRs in R10000 less often at 1.7±1.1 (Table 4). Comparing the incidence of flanking gate-keeper residues in TANGO and WALTZ predicted APRs, gate-keeper residues flank WALTZ predicted APRs with lower frequencies (∼20–25% lower) than TANGO predicted APRs (Table 4). Charged gate-keeper residues oppose aggregation by keeping APRs solvated when they become exposed due to local flexibility or protein destabilization. However, WALTZ predicted APRs are more likely to contain charged or polar residues than TANGO predicted APRs due to the parameterization of WALTZ. As a result, the need to gate-keep WALTZ predicted APRs under normal conditions may be lower. Gate-keeper residues for all SCOP classes show similar trends as for F495, except for the α/β class which has slightly more gate-keeper residues (2.8±1.1 for TANGO predicted APRs and 2.2±1.1 for WALTZ predicted APRs, Table 4). Taken together, these observations suggest that the aggregation propensity of protein sequences has been minimized by increasing the number of gate-keeper residues in APR flanking regions, which is consistent with earlier studies [24], [38]. Gate-keeper residues occur with greater frequency in the flanking regions of APRs in F495, SF49500 and N10000 than in R10000 due to amino acid composition differences. However, N10000, SF49500 and F495 have identical amino acid compositions. Differences in the number of flanking gate-keepers residues between N10000/SF49500 versus F495 suggests that number of gate-keeper residues flanking APRs is also determined by sequence patterning as well as amino acid composition.
Selective pressure to reduce the burden of protein aggregation in biological organisms has minimized the aggregation propensities of protein sequences by directing changes to amino acid composition and patterning during protein evolution. Results presented in this section indicate that changes in sequence patterning via scrambling and randomization of protein sequences increases the TANGO predicted aggregation propensity more than do changes in amino acid composition. This is reflected in a greater difference in average aggregation propensity between F495/IDP536 and N10000/SF49500 than between SF49500/N10000 and R10000 (Tables 2 and 3) for TANGO, but not for WALTZ predicted APRs. To support the finding that changes to patterning impacts sequence aggregation propensity more than changes to amino acid composition, the F495 dataset was divided into two datasets (F1 and F2) that retain their natural protein sequence patterning but have significantly different amino acid compositions (see Methods). The average TANGO aggregation propensities for F1 (3.66±3.16) and F2 (3.21±2.77), and average WALTZ aggregation propensities for F1 (2.84±2.67) and F2 (2.19±2.51), are both similar to their respective values for F495 (TANGO 3.41±2.98; WALTZ 2.52±2.70), (Table 2). Furthermore, in the case of TANGO predicted APRs, the difference in average aggregation propensity between F1 and F2 is considerably less than it is between F495 and N10000/SF49500. Thus, changes to sequence patterning (F495 versus N10000) produce greater differences in average aggregation propensity than changes to amino composition do (F1 versus F2), at least for TANGO predicted APRs. Overall, the use of three different randomly generated sequence datasets, R10000, SF49500 and N10000, and three natural protein sequence datasets, F1, F2, F495, has allowed us separate the impact evolutionary selective pressure has had on amino acid composition and sequence patterning of protein sequences.
A lack of APR conservation among homologous sequences is expected if protein evolution disfavors the tendency for proteins to aggregate [27], [30]. Here, the conservation of APRs among homologues of 9 proteins, selected from F495, was studied at two sequence identity cut-off values (see Methods). These 9 proteins contain ≥3 TANGO predicted APRs and ≥3 WALTZ predicted APRs or ≥3 Amylsegs, indicating these proteins have a high propensity to aggregate under suitable conditions. Table 5 shows APR and Amylseg conservation (Equation 11) among homologues of these 9 proteins at 80% and 50% sequence identity. Note these analyses did not include remote homologues (<30% sequence identity). The term ‘distant homologues’ used below refers to homologues that have at least 50% sequence identity with a parent sequence and the term ‘close homologues’ refers to homologues that have at least 80% sequence identity with a parent sequence. All homologues have the same fold as the parent protein. For the majority of these proteins, APR conservation levels are slightly lower than average sequence identity among close homologues. This observation is in agreement with that of Gromiha and coworkers [32] on protein sequence families containing highly homologous thermophilic and mesophilic proteins. However, when the sequence identity cut-off among homologues was dropped from 80% to 50%, the corresponding decrease in percent APR conservation is smaller. In other words, APRs are more conserved than average sequence identity among distant homologues that have at least 50% sequence identity to a parent sequence. Furthermore, amyloid-like fibril forming peptide segments (Amylsegs) in PDB entries 1KCQ (human gelsolin domain 2), 2D4F (human β-microglobulin), and 2VB1 (hen egg-white lysozyme), are highly conserved among their homologues at both sequence identities (50% and 80%). Overall, the data in Table 5 indicates that among close homologues, APRs are slightly less conserved than average sequence identity. A lack of APR conservation among closely related homologues suggests that evolution is continuing to disrupt aggregation prone sequence regions. On the other hand, among distantly related homologues, APRs are more conserved than average sequence identity, indicating that some of these regions may play a role in maintaining protein folds and activity.
Sequences from IDPs contain both ordered and disordered regions, which facilitates an investigation into the structural ordering of residues located in predicted APRs. Of the 232,321 residues in IDP536 sequences, 60,638 (26.1%) fall into regions annotated as intrinsically disordered by DisProt [33], the source database for IDP536. Amino acid residues not annotated as disordered regions were assumed to be ordered in this analysis. Therefore of the 232,321 residues in IDP536, 171,683 (73.9%) were assumed to be ordered. After predicting APRs in IDP536 using TANGO, it was found that 12,066 (5.2%) of the 232,321 IDP536 residues fall within one of 1,316 predicted APRs. Of the 12,066 APR residues, 1,327 (11% of the 12,066 APR residues) residues are disordered leaving the remaining 10,739 (89% of 12,066) residues as ordered. Using this data, a contingency table was prepared to categorize IDP536 residues as ordered and not part of an APR, or ordered and part of an APR (likewise for disordered residues, see Table 6). Performing a χ2 analysis on the contingency table rejects the null hypothesis (p<0.001) that the structural classification of an amino acid residue in IDP536, as ordered or disordered, is independent from the likelihood that the residue is part of an APR. Furthermore, the contingency table produces an odds ratio of 2.98∶1 indicating that an ordered residue is three times more likely to be part of an APR than a disordered one (see the tutorial by McHugh for more information on contingency tables and odds ratios [39]). WALTZ predicted APRs have comparable results (odds ratio 2.02∶1, see Table 6). Thus, these observations are independent of the APR prediction tool used. Similar to the results reported above, Schymkowitz and coworkers [21] have reported that ordered proteins have three times the number of APRs that IDPs do. Here, we report that APRs are three times more likely to contain ordered versus disordered residues which provides direct evidence that APRs are located in order forming regions. At a fundamental level, APRs are promoting local structural order in all conditions, irrespective of whether conditions stabilize or destabilize proteins.
The observation that APRs are more likely to contain ordered versus disordered residues suggests the need for an examination of APRs within globular protein structure. To perform the analysis, protein sequences from the F495 dataset, for which atomic coordinates are available, were used to investigate the secondary structure, solvent exposure, and relative solvent isolatedness of predicted APRs. Within F495 sequences regions with atomic coordinates, TANGO predicted 409 APRs, WALTZ predicted 516 APRs, and 19 Amylsegs of length ≥6 were found. Residues in these APRs occupy all three secondary conformational states (helix, coil, and strand) as measured by binning STRIDE [40] output (see Table S2 in Supplementary material). Overall, Amylsegs, TANGO, and WALTZ predicted APRs exist primarily in β-strands. This observation differs from the work of Doig and coworkers, who reported that amyloidogenic regions were often located in α-helical secondary structure [41]. Figure 2 compares the percent solvent exposure of TANGO and WALTZ predicted APRs and Amylsegs in F495 structures (Equations 2, 3 and 5). The average percent solvent exposure for TANGO predicted APRs is 16.9±12.2% (range 0–73%) and WALTZ predicted APRs is 23.3±14.9% (range 0–66%). On average, Amylsegs in F495 are more solvent exposed than TANGO or WALTZ predicted APRs (average percent solvent exposure 32.1±18.5% and range 1–62%). However, Amylsegs are longer than TANGO and WALTZ predicted APRs, with an average length of 11.5±10.4 residues. In comparison, the average APR lengths are 7.5±2.0 for TANGO predicted and 6.1±0.7 for WALTZ predicted APRs. Because proteins in F495 are small globular monomers with larger surface area to volume ratios, longer sequence segments are more likely to contain solvent exposed portions than smaller segments do. The low average percent solvent exposure of TANGO and WALTZ predicted APRs implies that these sequence regions are buried in their protein structures. Similar observations have been previously reported [22], [41].
Predicted APRs are expected to be buried in protein cores since sequence hydrophobicity is a differentiating physico-chemical property for APR prediction programs. Therefore, it is of interest to know whether predicted APRs are buried in protein cores beyond what is expected from the hydrophobicity of their constituent residues. To investigate this question, a measure called Burial Preference (BurPref) was devised (see Methods). The BurPref of an APR is the ratio of the observed APR solvent exposure to the expected APR solvent exposure calculated from the average surface areas of its constituent amino acid residues in F495 (Equations 2–4 and 6). If the BurPref of an APR is less than one, then the APR is more buried than expected and vice versa. The average BurPref for TANGO predicted APRs in F495 is 0.82±0.56 (range 0–4.15) and for WALTZ predicted APRs is 0.89±0.55 (range 0–3.69). This indicates that TANGO and WALTZ predicted APRs are, on average, more buried in the protein cores of F495 than expected. On the other hand, the average BurPref for the 19 Amylsegs is 0.97±0.47 (range 0.05–1.65), which indicates that these segments are not preferentially buried. BurPref values for TANGO and WALTZ predicted APRs suggests that sequence hydrophobicity alone cannot explain their degree of burial. Other factors are also important such as the role an APR has in providing protein stability or function. Overall, BurPref values for TANGO/WALTZ predicted APRs support the view [41] that the risk associated with aggregation prone sequence regions is minimized by their burial within protein structures. As a consequence, APR initiated aggregation may be limited to conditions in which the native structure is destabilized or in which APRs become exposed during co-translation or due to local protein flexibility.
Buried APRs can form multiple interactions within their protein structure and contribute towards stabilizing native folds. To quantify the contribution an APR makes towards the stability of a protein, solvent isolatedness (Iso) [42] was computed. Solvent isolatedness of a protein segment (such as an APR) measures the fraction (0–1) of its surface buried by the rest of the protein structure (Equations 3, 7, and 8). A large solvent isolatedness value for a protein segment indicates that most of its surface is interacting with the rest of protein. On the other hand, a small solvent isolatedness value for a protein segment indicates that most of its surface is solvent exposed and thus contributes few stabilizing protein interactions. The average solvent isolatedness (Iso) values for TANGO predicted APRs (0.83±0.12; range 0.27–1.0), WALTZ predicted APRs (0.77±0.15; range 0.34–1.0), and Amylsegs (0.68±0.19; range 0.38–0.99), all indicate that these sequence regions have multiple interactions within their parent protein structures and are thus mostly solvent protected.
To assess if APR segments are more solvent isolated than expected, relative solvent isolatedness (RIso) was computed (Equations 7, 8 and 10). This quantity provides a measure of how much more an APR contributes to the stability of a protein than expected, based on the average solvent isolatedness of equal length segments from the same protein [42]. The relative solvent isolatedness values for TANGO and WALTZ predicted APRs and Amylsegs are greater than one which indicates that they are more solvent isolated than expected (TANGO average RIso 1.33±0.18, range 0.47–1.73; WALTZ average RIso 1.20±0.22, range 0.56–1.67; Amylsegs average 1.19±0.29, range 0.66–1.64), (Figure 3(a)). Therefore, TANGO and WALTZ predicted APRs, as well as Amylsegs, contribute towards protein stability more than expected. To assess the significance of APR solvent isolatedness values, Z-scores of solvent isolatedness were computed for predicted APRs and Amylsegs (Equations 7, 8 and 9). Although average Z-scores for WALTZ predicted APRs (0.88±0.95; range −1.9–2.8) and Amylsegs (0.78±1.14; range −1.2–2.4) are below one, the average Z-score for TANGO predicted APRs is 1.44±0.76 (range −1.9–3.3). Therefore, the contributions made by TANGO predicted APRs towards stabilizing the protein, in which they exist, are more than one sigma greater than the average contribution made by equal length segments (Figure 3(b)). Others have previously proposed that APRs make stabilizing interactions in native folds [25], [43]. However, this is the first report that TANGO/WALTZ predicted APRs and Amylsegs make more stabilizing interactions in native folds than, on average, those made by equal-length segments from the same protein.
Evidence that APRs stabilize native protein folds calls for an examination into whether functional sites are located within APRs or are structurally proximal to them. In this report, co-localization of catalytic sites and TANGO or WALTZ predicted APRs in crystal structures of enzymes is investigated using the Cata dataset which contains 961 catalytic residues from 314 non-homologous protein chains (299 enzymes). Catalytic residues form a subset of active site residues in enzymes. Thus, most enzymes contain few catalytic residues and the likelihood that these residues are located either within APRs, or near APRs, by random chance is low. Incidences of catalytic residues within TANGO/WALTZ predicted APRs and their flanking regions were estimated using equation 12 (see Methods). These estimated numbers were compared with the observed incidence of the catalytic residues within TANGO/WALTZ predicted APRs and their flanking regions. Ninety nine of the 961 catalytic residues (10.3%) were estimated to fall within TANGO predicted APRs and flanking regions and 57 of them (5.9%) were observed within these regions. Similarly, 103 out of the 961 catalytic residues (10.7%) were estimated and 69 (7.2%) were observed to fall within WALTZ predicted APRs and flanking regions. Therefore, the observed incidence of the catalytic residues within APRs and their flanking regions is lower than the estimates and it can be concluded that the catalytic residues are not usually found within these regions. This is expected given that most catalytic residues are charged or polar. On the other hand, catalytic residues do tend to be in structural contact with APRs (structural contacts are inferred when a catalytic residue heavy atom is within 4.5 Å from an APR residue heavy atom, see Methods). Of the 961 catalytic residues, 373 (38.8%) are in structural contact with at least one neighboring TANGO predicted APR residue. Similarly, 310 (32.3%) catalytic residues are in contact with at least one neighboring residue within WALTZ predicted APRs. Although, it has previously been observed that APRs and amyloidogenic regions can be located in protein-protein interfaces [22], [28], including antigen-antibody interfaces [23], this is the first report of structural proximity between APRs and catalytic residues to our knowledge.
To assess the significance of the observed structural proximity between APRs and catalytic residues, statistical simulations were performed by generating one million catalytic decoy lists. Each list contained the residue coordinates of 961 randomly chosen decoy catalytic residues from the atomic coordinates of protein chains in the Cata dataset. Randomly chosen decoy catalytic residues were selected for each true catalytic residue in Cata and were limited to any residue within the same protein structure as the true catalytic residue. Thus, if there are X number of true catalytic residues from protein Y in the Cata dataset, all decoy lists contained X number of decoy catalytic residues from the same protein Y. Using decoy catalytic lists enabled us to calculate an expected number of residues in structural contact with predicted APRs. This expected number was compared to the observed number of true catalytic residues in structural contact with predicted APRs. Figure 4(a) shows the distribution of the number of decoy catalytic residues which are in structural contact with TANGO predicted APR residues for all one million lists. The average number of decoy catalytic residues in contact with TANGO predicted APRs is 254±13. This yields a Z-score of 9.3 for the 373 of 961 true catalytic residues in the Cata dataset that are in structural contact with TANGO predicted APRs and suggests the number of true catalytic residues in contact with TANGO predicted APRs is highly significant. These calculations were repeated by varying the distance cut-off for inferring structural contacts (3.5 Å and 6.0 Å). The observed number of true catalytic residues in contact with TANGO predicted APRs is 278 and 461 at 3.5 Å and 6.0 Å, respectively. Statistical simulations to compute the expected number of randomly chosen decoy catalytic residues in contact with TANGO predicted APRs yield Z-scores of 10.1 at the 3.5 Å cut-off distance and 11.7 at the 6.0 Å cut-off distance for the number of true catalytic residues in contact with TANGO predicted APRs. Catalytic residues in the Cata dataset are also in close structural proximity to WALTZ predicted APRs, significantly more often than expected. There are 224, 310 and 405 catalytic residues in contact with WALTZ predicted APRs at cut-off distances of 3.5 Å (Z-score, 5.8), 4.5 Å (Z-score, 5.7) and 6.0 Å (Z-score, 8.5) respectively. To further probe our observation of catalytic residues in contact with APRs, an additional condition on the solvent exposure of randomly selected residues as decoy catalytic residues was imposed. Decoy catalytic residues were required to have a solvent exposure that was similar to the solvent exposure of their corresponding true catalytic residue (ASA value of each decoy catalytic residue must be within ±10% of the ASA value for the corresponding true catalytic residue). At the structural contact cut-off distance of 4.5 Å, Z-scores for catalytic residues in contact with TANGO and WALTZ predicted APRs are 5.9 and 3.1 respectively when ±10% ASA condition was imposed. As such, Z-scores decreased but remain statistically significant. Z-scores decreased when the ASA condition was imposed because the probability that a decoy catalytic residue is in contact with an APR increases when decoy catalytic residues are limited to the same solvent exposure as their true catalytic residues. Computing expected values for the number of residues in contact with APRs, using both ASA limitations and multiple distance cut-offs, has supported our finding that catalytic residues are in close structural contact with APRs significantly more often than expected by random chance.
To visualize the co-localization of APRs and catalytic residues in enzyme structures, an example from the Cata dataset is shown in Figure 4(b)). Cholesterol Oxidase from B. Sterolicum (PDB entry: 1I19, UniProt entry: Q7SID) is a 561 residue monomeric enzyme with covalently bound FAD that catalyzes oxidation and isomerization of steroids [44]. This enzyme contains three catalytic residues, namely, Glu 311, Glu 475 and Arg 477, and two TANGO identified APRs, 229-LTAVVW-234 and 511-VAIWLNVL-518. Two catalytic residues, Glu 475 and Arg 477, make several close contacts with the second APR (Figure 4(b)), which lies in the substrate binding domain [44]. Considering the large structural size of Cholesterol Oxidase, and the fact that is has only three catalytic residues and two TANGO predicted APRs, which are short in length, it is surprising to find its catalytic residues in structural contact with its APRs.
Are there examples of catalytic residues making structural contact with experimentally validated amyloid-fibril forming peptide segments, Amylsegs? To answer this question, the AmylSegs dataset was searched for peptide segments from enzymes. Amylsegs contain amyloid-fibril forming peptides derived from nine different enzymes. Six of these nine enzymes have been annotated in UniProtKB for catalytic residues and have at least one crystal structure deposited in the PDB (Table 7). Of these six, three enzymes (pancreatic ribonuclease A, hen egg white lysozyme and human lysozyme) have a catalytic residue that is in contact with at least one Amylseg residue. The criterion used here to identify a structural contact is the same as in analyses on the Cata dataset (distance cut-off of 4.5 Å for a pair of heavy atoms). Figure 5 shows an example of catalytic residues from human lysozyme, Glu 35 and Asp 53, which are in contact with its Amylseg regions. There are three peptides in the Amylsegs dataset from human lysozyme that have been experimentally shown to form amyloid fibrils. These are 5-RCELARTLKR-14, 25-LANWMCLAKW-34 and 56-IFQINS-61 [45]. Catalytic residue, Glu 35, lies immediately after the second peptide and makes contact with residues 30-CLAKW-34 as well as contact with residues 56-IFQ-58 from the third peptide. Catalytic residue Asp 53 also contacts the third peptide. Indeed, catalytic residues are making structural contact with experimentally validated amyloid-fibril forming peptide segments.
Protein functional sites require optimal combinations of flexibility and stability to fulfill their biological purposes. Therefore, it is logical for catalytic residues, which require consistent and specific orientations, to be in contact with regions that promote local structural order and form stabilizing interactions. While this report has focused on catalytic residues, an interesting example of an Amylseg, 182-SFNNGDCFILD-192, containing a Ca2+ binding residue at D187 in human gelsolin has also been identified. The D187N mutation, which disrupts the metal binding site, leads to protein instability and amyloidosis in patients with a disease called familial amyloidosis-Finnish type [46]. Co-localization of metal catalyzed oxidation (MCO) sites and APRs have also been observed in the structures of therapeutic proteins where metal-ion leachates contribute towards drug product degradation [47].
The results presented above show that protein sequences have evolved by optimizing their risk of aggregation for cellular environments by both minimizing aggregation prone regions and conserving those that are important for folding and function. Outside the cell, protein aggregation is commonly encountered in the laboratory and is a major hurdle in successful development of protein based biotechnology products, such as biotherapeutics. For these applications, the aggregation propensity may need to be further reduced in order to enhance protein yields from cell cultures and to improve protein solubility, especially at high concentrations. Disruption of APRs via site directed mutagenesis is an attractive protein engineering strategy to improve protein solubility [13], [14], [23], [48]–[50]. Alternatively, ‘supercharging’ functional proteins with very high electrostatic surface charge has also been shown to improve protein solubility beyond levels normally observed for natural proteins [51]–[53]. Because APRs can also form part of HLA-DR binding T-cell immune epitopes [54], [55], disrupting APRs potentially leads to a lower risk of immunogenic reactions in patients receiving biotherapeutic drugs. Insights gained from this report caution that the contributions made by candidate APRs, targeted for disruption, towards protein stability and function should be considered when identifying sites that are suitable for rational mutagenesis. Disruption of APRs, without knowledge of their contributions, can lead to undesirable consequences, such as protein destabilization and/or loss-of-function. This work also complements our efforts to distinguish between ‘active’ and ‘inactive’ APRs in proteins [56], [57].
A broader implication of this research is that a general strategy for identifying mutation sites for improving solubility of a candidate protein can be proposed. This strategy is presented in Figure 6 and the major steps are described below. If a protein of biotechnological interest aggregates at higher than a desired level, the following information is needed to employ the strategy: protein sequence, three dimensional structure, homologues, potential cross-β motif forming APRs and functional sites. APR prediction programs often identify several potential APRs in the sequence of a protein. For each APR, its contribution towards protein stability should be evaluated. This can be done by computing solvent isolatedness for the APR and equal length segments from the same protein (Equation 8). If the APR has a high solvent isolatedness (low solvent exposure, buried in protein core) then it should not be a target for disruption. If the APR has a low solvent isolatedness (high solvent exposure, located at or near protein surface) then it is expected to make a smaller contribution to protein stability and can be marked for disruption depending upon the outcomes of the following tests. The protein structural region around the APR should be examined for contacting functional residues and for sequence conservation among homologues. If the APR contains residues that are in structural contact with functional residues and the APR is more conserved than average sequence identity among homologues, then it should not be targeted for disruption. If the APR is not in contact with functional residues and is less conserved than the average sequence identity among its homologues, then it is a priority target for disruption. If the APR is not structurally proximal to functional residues, but is more conserved than average sequence identity among its homologues, it can still be targeted for disruption after verifying that the conservation is not due other structure-function purposes such as allostery. If the APR is structurally proximal to functional residues, but is less conserved than average sequence identity among the homologues, it can still be targeted for disruption if molecular modeling can offer clues into potential sites and mutations that can be safely substituted. In this case, care should be taken to avoid disturbing the conformations of functional and contacting residues. This can be done by choosing a residue within the APR which is not in direct contact with functional residues, but whose mutation disrupts the APR.
TANGO [11] and WALTZ [12] were used to predict APRs in all sequence datasets. Both programs have been extensively validated using independent testing sets and found to be highly accurate. The following options were used as input parameters for both programs: Temperature, 298K; pH, 7.0; Ionic Strength, 150 mM; Concentration 1 mM; TANGO/WALTZ aggregation, ≥10%; Minimum window size, 6; Flanking residues, 3.
The outputs from TANGO and WALTZ yield data on the total sequence aggregation score and sequence length along with position, length, sequence and flanking residues for each predicted APR. The total aggregation score for each sequence was normalized by the length of the sequence to obtain its aggregation propensity. The proportion of APR residues (APRprop(%)) in a sequence was also computed as follows:(1)APRlen(i, j) is the length of the ith APR in the jth sequence, nAPR(j) is the number of APRs in the jth sequence, seqlen(j) is the number of residues in the jth sequence. Averages and standard deviations were computed for both aggregation propensities and proportional APR residues. These results are reported in Tables 2 and 3.
Flanking residue positions that precede (PB−1, PB−2, PB−3) and succeed (PE+1, PE+2, PE+3) each APR were searched for the presence of gate-keeper residues (Asp, Glu, Lys, Arg and Pro) [24], [25] and their frequencies were computed at each of these positions. These results are presented in Table 4.
Atomic coordinates from F495 proteins were submitted to STRIDE [40] to obtain secondary structure information and solvent accessible surface areas (ASA) for all residues within their three dimensional context. Average ASA values for each of the twenty amino acids in the F495 dataset were obtained by summing the ASAs of amino acid, k, in all proteins and dividing the sum by the number of k amino acids in F495, where k runs from 1 to 20.(2)AvASAResk is the average ASA for amino acid of type k. ASAReskj is the ASA of each amino acid residue of type k in protein j and NResk is the number of amino acid residues of type k in F495 dataset. ASA values for individual residues of APRs were summed to obtain observed solvent accessible surface areas (SASAobs) for each predicted APR.(3)APRij is the ith APR in the jth protein from F495. SASAobsAPRij is the observed SASA of APRij. ASARes(k)APRij is the ASA of an individual residue, k, in APRij. The summation runs over the length of APRij, APRlen(i,j). Expected ASA values for each APR (SASAexp) were computed by summing the average ASA values from F495 for each of the constituent amino acid residues.(4)SASAexpAPRij is the expected SASA of APRij. AvASARes(k)APRij is the average ASA of residue type k in APRij, computed using Equation 2. The total surface area (TotSA) for an APR outside of its three-dimensional context was computed by submitting atomic coordinates, only from APR segments, to STRIDE. Note that in both calculations, APRs have identical conformations. TotSA and SASA values for each APR were used to compute percent solvent accessibility (SolvAcc) of the APR, burial preference (BurPref) for an APR, and solvent isolatedness (Iso) of an APR from solvent [42] as follows:(5)(6)(7)(8)SolvAccAPRij is the percent solvent accessibility of APRij. SASAobsAPRij is the solvent accessible surface area of APRij within its three-dimensional context. TotSAAPRij is the total surface area of the APR outside of the three-dimensional context of protein j. SASAexpAPRij is the expected solvent accessible surface area of APRij computed from average ASA values for its constituent amino acids in F495. BurPrefAPRij is a ratio of the observed to expected solvent accessible surface area for APRij. It indicates the preferential burial of APRij in protein cores. If BurPrefAPRij is below one, APRij is more buried than expected from the average burial of its constituent residues. If BurPrefAPRij is above one, APRij is more solvent exposed than expected from the average solvent exposure of its constituent residues. ProtBurSAAPRij is the surface area of APRij that is buried by the rest of the protein j. IsoAPRij values can be interpreted as the contribution APRij makes towards the stability of protein j. To evaluate the significance of this contribution, IsoAPRij values were also computed for all segments the equal length as APR, i in protein, j. Each segment was obtained by sliding a window the equal length as APR, i over the structure of protein, j one residue at a time. Average (<Isoij>) and standard deviation (σIsoij) values for segments within protein j were used to compute Z-scores (Z-score (IsoAPRij)) and relative values (RIsoAPRij) for solvent isolatedness of APRij using the following equations:(9)(10)
Nine selected monomeric proteins from F495 were used for sequence conservation analyses. These 9 proteins contain ≥3 TANGO predicted APRs and ≥3 Waltz predicted APRs or ≥3 Amylsegs, indicating these proteins have a high propensity to aggregate under suitable conditions. PDB entries for these proteins are 1FUK (C-terminal domain of yeast initiation factor 4A), 1JEO (Hypothetical protein MJ1247 from Methanococcus jannaschii), 1KCQ (Human Gelsolin Domain 2), 1OW1 (SPOC domain of human transcriptional factor SHARP), 1SK7 (Hypothetical protein pa-HO from Pseudomonas aeruginosa), 1Z77 (Transcriptional regulator (tetR family) from Thermotoga maritima), 2D4F (Human β-microglobulin), 2VB1 (Hen Egg White Lysozyme) and 3NR5 (Human RNA polymerase III transcription repressor Maf1). Note that Hen egg-white lysozyme, human β-microglobulin and human gelsolin are well studied amyloidogenic proteins [61]–[63].
Sequence conservation analyses for the above mentioned proteins were performed at two arbitrarily chosen levels of sequence identity, 80% and 50%. The procedure for selecting homologues at the 80% level is described below. Sequences of the nine proteins (query sequences) were searched for homologues in the UniProtKB database (www.uniprot.org) [64], [65] using blastp and all default options. For each query, hit sequences with ≥80% sequence identity were selected, provided that homologous regions of hit sequences covered the query sequence completely. Since hits to query sequences were sometimes longer than the length of the query sequence, only portions of hit sequences that aligned with the query sequence were taken for conservation analysis. All the retrieved homologous sequences were re-aligned using ClustalW [59]. For each query sequence, the ClustalW input file included the sequence from the PDB file as the first sequence. Any sequence with a ClustalW alignment score of <80 to the first sequence was deleted from the alignment. Sequences with alignment scores of 100 to the first sequence were also removed. All the above steps were repeated to obtain homologous sequences with ≥50% sequence identity to selected PDB files.
An APR was labeled as ‘conserved’ between two homologous sequences, if the APR has the same sequence in both homologues. Percent APR conservation in a multiple sequence alignment was computed using the following formula:(11)PAPRconserved is the proportion of conserved APRs in an alignment, nAPRtotal is the total number of APRs for all the sequences in the alignment and nAPRuniq is the number of unique APRs (non-identical sequence) over all sequences in the alignment. These calculations were performed for both TANGO and WALTZ predicted APRs. Analogous calculations were also performed for all peptide sequences from Amylsegs that were detected in the 9 proteins and their homologues.
Number of the catalytic residues from the Cata dataset that fall within the TANGO/WALTZ predicted APRs and their flanking regions were estimated using the following equation:(12)NCata-APRs is the estimated number of catalytic residues that fall within TANGO/WALTZ predicted APRs and their flanking regions. Three residues preceding and three residues succeeding an APR are considered as its flanking regions. NCata is the number of catalytic residues in the Cata dataset. This number is 961. NAPRs is the number of residues in APRs and their flanking regions predicted using TANGO and WALTZ in the sequences of the 299 enzymes (314 Chains) in the Cata dataset. For 823 TANGO predicted APRs and their flanking regions, NAPRs is 11,414 (6498 residues in TANGO APRs plus 4916 residues in the flanking regions). NAPRs is 11,856 for 982 WALTZ predicted APRs and their flanking regions (5988 residues in the APRs and 5868 residues in the flanking regions). NTot is 110,334, the total number of residues in the sequences of the 299 enzymes.
TANGO and WALTZ predicted APRs were also mapped onto protein structures from the Cata dataset to search for structural contacts made by catalytic residues to residues in predicted APRs. A catalytic residue is considered to be in structural contact with an APR, if at least one of its heavy atoms is within 4.5 Å from a heavy atom in any residue that falls within an APR. The choice of a 4.5 Å cut-off is arbitrary but was used here because it is common in the literature [66]–[68]. Catalytic residues in structural contact with APRs residues were counted for both TANGO and WALTZ predicted APRs.
To assess the significance of the observed structural proximity between APRs and catalytic residues, statistical simulations were performed by generating one million decoy catalytic lists. Each list contained the residue coordinates of 961 randomly chosen decoy catalytic residues from the atomic coordinates of protein chains in the Cata dataset. Randomly chosen decoy catalytic residues were selected for each true catalytic residue in Cata and were limited to any residue within the same protein structure as the true catalytic residue. For each of the 1,000,000 randomly generated lists, the number of residues making structural contact with APR residues (APR contacting) was computed again in the same way as for the Cata dataset.
The number of APR contacting residues was counted for each random list to generate a distribution of expected APR contacting residues. This distribution was also used to compute Z-scores for the incidence of APR contacting residues in the Cata dataset in the same way as the Z-score for solvent isolatedness was calculated (Eq. 9). The analogous calculations were also performed using the contact distance cut-off values of 3.5 Å and 6.0 Å.
To further probe our observation of catalytic residues in contact with APRs, a restriction on the solvent exposure of randomly selected residues as catalytic decoys was imposed. Decoy catalytic residues were required to have a solvent exposure that was similar to their corresponding true catalytic residue (ASA value of each decoy must be within ±10% of ASA of true catalytic residue). For each of the 1,000,000 randomly generated lists, the number of residues making structural contact with APR residues (APR contacting) was computed again in the same way as for the Cata dataset.
|
10.1371/journal.ppat.1000836 | T Cell-Dependence of Lassa Fever Pathogenesis | Lassa virus (LASV), the causative agent of Lassa fever (LF), is endemic in West Africa, accounting for substantial morbidity and mortality. In spite of ongoing research efforts, LF pathogenesis and mechanisms of LASV immune control remain poorly understood. While normal laboratory mice are resistant to LASV, we report that mice expressing humanized instead of murine MHC class I (MHC-I) failed to control LASV infection and develop severe LF. Infection of MHC-I knockout mice confirmed a key role for MHC-I-restricted T cell responses in controlling LASV. Intriguingly we found that T cell depletion in LASV-infected HHD mice prevented disease, irrespective of high-level viremia. Widespread activation of monocyte/macrophage lineage cells, manifest through inducible NO synthase expression, and elevated IL-12p40 serum levels indicated a systemic inflammatory condition. The absence of extensive monocyte/macrophage activation in T cell-depleted mice suggested that T cell responses contribute to deleterious innate inflammatory reactions and LF pathogenesis. Our observations in mice indicate a dual role for T cells, not only protecting from LASV, but also enhancing LF pathogenesis. The possibility of T cell-driven enhancement and immunopathogenesis should be given consideration in future LF vaccine development.
| Lassa virus (LASV) is the causative agent of Lassa fever (LF), accounting for substantial morbidity and mortality in West Africa. Yet the mechanisms leading to disease remain poorly understood. Here we propose a concept whereby the body's immune defense either defeats LASV rapidly or, if unsuccessful, becomes an essential facilitator of disease. This latter paradoxical postulate stems from observations in genetically engineered (HHD) mice, which we found to be susceptible to LF. HHD mice differ from resistant wild type mice in that they have a humanized repertoire of T cells, a main component of the mammalian immune system. Counterintuitively, we could protect HHD mice against LF by experimentally removing their T cells. We further found that LF correlated with widespread activation of macrophages, which again depended on T cells. Similar to T cells, macrophages are important players in our body's defense system, but their inflammatory products are also candidate mediators of LF. Taken together, these findings suggest that LF may represent an inappropriate host response to infection. Specifically, our study demonstrates a two-faced role of T cell responses against LASV. Such detrimental aspects of immune defense need to be given consideration in future LF vaccine development, to avoid enhancement of disease in vaccinated individuals.
| Lassa virus (LASV) is the causative agent of Lassa fever (LF) [1]. It accounts for an estimated number of 300′000 infections and several thousand deaths in endemic areas each year [2], while imported cases have been reported from around the globe [3]. The virus is listed category A by the Center for Disease Control and Prevention [4]. So far, LASV vaccines have remained unavailable for clinical use, and Ribavirin, the only available therapy, has shown limited efficacy [5]. The development of effective vaccination strategies would therefore benefit from further insight into the mechanisms of successful LASV immune control, as well as into the processes underlying LF development and pathogenesis.
It is generally agreed upon that the level of tissue damage observed at autopsy cannot by itself account for the severe nature of LF. Therefore, as with other viral hemorrhagic fevers [6],[7], a contribution of the host response to LF pathogenesis has long been a matter of debate. For instance, the manifestation of Dengue Hemorrhagic Fever (DHF) has long been accredited to pre-existing immunity [8],[9]. Apart from serotype cross-reactive antibodies [8],[9], memory T cells were recently identified as important players in the disease process [10], and susceptibility as well as resistance to DHF have been linked to particular MHC alleles [11],[12]. In addition, infected monocytes and macrophages play an important role in DHF by secreting inflammatory cytokines [13],[14].
Such contributions of the immune response to disease severity can represent a major hurdle in vaccine development [15]. For instance, formalin-inactivated vaccines to respiratory syncytial virus (RSV) and measles virus resulted in enhanced morbidity and mortality in response to natural infection [16],[17]. Animal models for RSV have since provided evidence that T cell subsets play an important role in disease enhancement [16],[18]. Interestingly, innate immune cells including eosinophils and polymorphonuclear granulocytes dominate the histological picture upon T cell-driven enhancement of RSV [18]. Similarly, inflammatory macrophage responses were found to be a common feature of viral hemorrhagic fevers [6]. In accordance with the “cytokine storm” hypothesis, macrophage-derived inflammatory cytokines [19],[20],[21],[22],[23] and nitric oxide (NO) [24],[25] are candidate mediators of capillary leakage and shock [26], and elevated levels of such mediators correlate with increased disease severity and worsened clinical outcome.
Still, LASV lacks a clear pathognomonic signature, and clinical manifestations of LF are largely unspecific, making it difficult to diagnose the infection accurately via clinical criteria alone [27]. In contrast to other hemorrhagic fevers, coagulation abnormalities and bleeding are largely absent in LF [27],[28], leading some to argue on pathological grounds that Lassa fever ought not be considered a hemorrhagic fever at all [6],[29]. More characteristic of severe LF cases are the vascular leakage with edema and effusions in the pleural and pericardial cavities [30],[31],[32]. At necropsy, liver and lung count among the organs most commonly affected during LF [29],[30],[31],[32],[33].
One of the few well-documented characteristics of primary LASV-directed immune response is that neutralizing antibody responses develop only weeks or months after the virus has been eliminated [28]. Also studies of vaccination-induced LASV immunity point toward a cell-mediated mechanism at the frontline of antiviral defense [34],[35]. Still, this notion remains to be addressed and verified directly, and the responsible T cell subtypes to be characterized. Further, a potential disease-enhancing effect of T cell responses in LF has not yet been given sufficient consideration.
Although normal laboratory mouse strains develop acute disease of the central nervous system when infected with LASV intracerebrally [1],[28],[36], they remain resistant to the systemic disease so characteristic of human LF, irrespective of the route used to infect them. Research on LF has therefore been limited to the use of guinea pigs and non-human primates [28], complicating mechanistic studies on immunity and pathogenesis. Here we report on a series of experiments triggered by accidental observations of serious disease in LASV-infected humanized mice (HHD mice [37], C57BL/6 background). HHD mice are genetically engineered to express a human/mouse-chimeric HLA-A2.1 molecule instead of the murine MHC class I gene products and are widely used to identify human HLA-A2.1-restricted peptide epitopes. Stimulated by these unexpected results, we were able to identify T cell-dependence of LASV control, but also of LF pathogenesis. These findings, combined with the propensity of LASV to target monocyte/macrophage lineage cells in vivo, followed by T cell-dependent activation of this cell population, provide a novel concept for virus-host relationship and pathogenesis of LASV. We anticipate that such understanding may aid rational refinement of both vaccine-mediated prevention and treatment of LASV infection.
We first compared viral replication in HHD mice and wild type C57BL/6 controls. C57BL/6 mice cleared LASV within about seven days after infection whereas HHD mice remained viremic for substantially longer periods of time (Fig. 1A and data not shown). A detailed analysis of the initial phase of infection documented a virtually immediate uptake of the inoculum into tissues (no virus in blood 2.5 hours after inoculation), followed by identical levels of viremia in wild type and HHD mice up to around day 4 (Fig. 1A). This demonstrates that viremia reflected viral replication in tissues rather than residual inoculum, and that the early phase of virus replication was identical in HHD and C57BL/6 mice. Clear differences in virus control became evident no earlier than seven days after infection (Fig. 1A). These differences in kinetics were compatible with differential adaptive immune control in HHD and C57BL/6 mice. Given that MHC class I (MHC-I) represents the only genetic difference between HHD and C57BL/6 mice, these findings suggested that H-2Db/H-2Kb-restricted T cell responses in C57BL/6 mice played an important role in virus control. Hence, we extended our study to further analyze the contribution of MHC-I- and MHC-II-restricted T cell responses to LASV control (Fig. 1B). MHC-II-deficient mice (lacking CD4+ T cells) efficiently resolved the infection whereas MHC-I-deficient animals (MHC-I-/-; targeted mutation of the β2-microglobulin gene; devoid of CD8+ T cells) developed persistent high-level viremia. This corroborated the key role of MHC-I-restricted T cell responses in LASV control and indicated further that MHC-II-restricted responses were of lesser importance. A time course analysis of viral titers in kidney, lung, liver and spleen of LASV-infected HHD, C57BL/6 and MHC-I-/- mice confirmed that viral replication was comparable on day 2 and day 4. By day 8, however, LASV in the organs of C57BL/6 mice approached the detection limit whereas comparably high titers of virus persisted in tissues of HHD and MHC-I-/- mice. These data provided additional independent support for the above conclusions on productive replication of LASV in mice and the key role of MHC-I-restricted T cells in its control.
Between 7 and 12 days after LASV infection, HHD mice developed ruffled fur and reduced spontaneous activity. Some of them rapidly and unexpectedly deteriorated and progressed to a state of agony and death: Five out of twenty three mice (∼22%) infected in a total of five experiments succumbed to disease. In contrast but in accordance with the literature, all thirteen wild type C57BL/6 mice, serving as controls in three of these experiments, survived without clinical evidence of disease. Elevated serum aspartate aminotransferase activity (AST) represents the primary parameter for monitoring LF, and AST combined with viremia represent the best predictors for clinical outcome in humans [38]. In keeping with this manifestation of LF, serum AST activity remained mostly within normal ranges in wild type C57BL/6 controls but was significantly elevated in the serum of HHD mice, with a peak around day eight to twelve (Fig. 2A). To further investigate how T cell responses restricted to H-2Kb/H-2Db and to HLA-A2.1 influenced LASV control and disease, we crossed HHD mice to C57BL/6 mice. C57BL/6 x HHD F1 mice express H-2Kb, H-2Db and HLA-A2.1 molecules owing to hemizygosity at all relevant genetic loci. C57BL/6 x HHD F1 mice controlled LASV infection as efficiently as did C57BL/6 wild type mice, and their AST levels remained within normal ranges (Fig. 2B, C). This showed that H-2Kb/H-2Db-mediated virus control prevented disease even in the presence of the HLA-A2.1 molecule. At first it suggested also that persistent and high virus load was directly responsible for pathogenesis. Contrary to this notion, however, the experiments in MHC-I-deficient mice had not resulted in obvious disease despite persistent high-level viremia. This raised the possibility that primary T cell responses may contribute to LF in an immunopathological fashion, similar to the role of memory T cell responses in DHF [10],[11],[39]. To address this possibility, we infected HHD mice with LASV, and prior to infection depleted either CD8+ T cells or CD4+ T cells or both using monoclonal antibodies. MHC-I-/- mice are devoid of a CD8+ T cell compartment and were also included in the experiment. Unlike in untreated HHD mice and irrespective of comparably high levels of viremia in all groups (Fig. 2E), serum AST levels of CD8/CD4-double-depleted HHD mice remained in normal ranges (Fig. 2D). Also, depletion of only CD4+ or CD8+ T cells or genetic deficiency for MHC-I (affecting the CD8+ but not the CD4+ T cell compartment) afforded at least partial protection. In agreement with the results shown in Figs. 1B and 1C, these data suggested that T cells of HHD mice were unable to significantly influence viremia. Nevertheless CD8+ and CD4+ T cells played apparently an essential role in the pathogenesis of LF in HHD mice.
To further characterize the role of T cells in the HHD mouse model for LF, we analyzed whether such animals could be immunized against LASV. For this, we used Mopeia virus (MV), an apathogenic close relative of LASV. MV infection is known to elicit heterologous immunity against LASV in monkeys (analogous to vaccinia virus protecting against smallpox), and MV and recombinants thereof have therefore been postulated as LASV vaccines [40],[41]. MV infection of HHD mice did not result in detectable viremia (Fig. 3A) nor was AST elevation recorded at any time point (Fig. 3B). This pattern of susceptibility of HHD mice to LASV but not MV reflected the one reported in non-human primates [40]. Next we tested whether MV immunization could induce HLA-A2.1-restricted immunity against LF. A recent study has characterized HLA-A2.1-restricted T cell epitopes in the glycoprotein (GP) of LASV [42]. Here we found that MV infection of HHD mice elicited high frequencies of CD8+ T cells specific for the GP42-50 epitope of LASV and a somewhat weaker but clearly detectable response against the GP60-68 epitope (Fig. 3C). CD8+ T cells specific for a third known epitope in LASV-GP (GP441-449) were not induced to a detectable extent, owing to only partial sequence homology of MV and LASV. When subsequently challenged with LASV, MV immunization prevented viremia and serum AST elevation in HHD mice (Fig. 3D, E), and by day 14 after challenge the spleen, liver, lung and kidney of MV-immunized mice were free of detectable LASV (data not shown). MV and LASV are serologically distinct i.e. neutralizing antibodies elicited against one virus do not crossreact nor crossprotect against the other [43]. This suggested that T cell immunity protected MV-immunized HHD mice against subsequent LASV challenge (Fig. 3D, E), albeit primary T cell responses facilitated apparently the disease process in unvaccinated animals (Fig. 2D).
Next we set out to characterize tissue alterations in LASV-infected HHD mice and to study their dependence on T cells. In all LASV-infected HHD mice analyzed, the lung showed severe pneumonitis with interlobular septal thickening and collapse of the alveolar lumen (Fig. 4A, B). In addition, macroscopic analysis at necropsy or in terminally diseased animals often revealed a substantial pleural effusion (up to about 0.5 ml in each hemithorax, not shown). Both observations matched those in human LF [30],[31]. In contrast to HHD mice, the lungs of C57BL/6 sacrificed at the same time point were only mildly affected or appeared normal (Fig. 4A, B). CD8/CD4-depletion prevented these alterations in HHD mice, and also MHC-I-/- mice exhibited considerably milder signs of peumonitis. Interestingly, HHD lungs contained dense infiltrations of rounded Iba-I+ monocyte/macrophage lineage cells (Fig. 4C), a finding that was less prominent or absent in C57BL/6 mice, T cell-depleted HHD mice or MHC-I-/- mice. Accumulation of T cells was also noted in HHD lungs (Fig. 4D) albeit to a lesser extent than for monocytes/macrophages (compare Fig. 4C). Moreover, similar infiltrations were also found in resistant C57BL/6 wild type mice and thus did not correlate with disease.
In the liver, nodules of mononuclear cells were found around the portal fields of all four groups of mice (Fig. 4E). Striking differences were, however, noted in the distribution, shape and orientation of hepatic monocyte/macrophage populations (including Kupffer cells, Fig. 4F). Like in uninfected mice, hepatic monocytes/macrophages of C57BL/6 and of T cell-depleted HHD mice formed predominantly a flat layer along liver sinusoids, oriented towards the central vein in a stellar pattern. In contrast, the architecture of this cell layer was disrupted in HHD mice (unless depleted of T cells) with the remaining cells enlarged, rounded up, disorganized and often accumulated in clusters, indicative of cellular activation and reminiscent of the vigorous hepatic macrophage response reported from human LF [29],[30],[31],[32],[33]. MHC-I-/- mice displayed an intermediate picture with only moderate monocyte/macrophage activation. Conversely, T cells were scattered at similarly moderate density throughout the liver parenchyma and in periportal inflammatory nodules of HHD and C57BL/6 mice (Fig. 4G and data not shown). The number of hepatic T cells did therefore not correlate with disease, similar to the findings in the lung.
Generalized immunosuppression is widely assumed to accentuate viral hemorrhagic fever [6]. A recent monkey study has tentatively attributed LASV immunosuppression to disorganization of the microarchitecture in secondary lymphoid organs [44]. Here we found that LASV infection of HHD mice resulted in disruption of the splenic white and red pulp compartments, whereas the spleens of C57BL/6, CD8/CD4-depleted HHD mice and MHC-I-/- mice were less affected (Fig. 4H). In correlation with these alterations, the marginal zone macrophage layer was lost in HHD mice but not in the other groups, and monocytes/macrophages were homogenously distributed throughout the splenic tissue of HHD mice (Fig. 4I). T cells were largely absent in CD8/CD4-depleted mice, as expected, and were also somewhat scarce in HHD mice (Fig. 4J), possibly indicating LASV-induced T cell depletion as reported from non-human primates [44].
The above morphological alterations had suggested T cell-dependent monocyte/macrophage activation in LASV-infected HHD mice. Classical activation of monocytes/macrophages e.g. by the T cell cytokine interferon gamma and cell-to-cell contact [45],[46],[47],[48] triggers the secretion of NO and inflammatory cytokines, and expression of the former is mediated by inducible NO synthase (iNOS) [45],[46],[47],[48],[49]. iNOS expression can therefore serve as a histological marker for inflammatory differentiation of monocytes/macrophages [50]. On day 8 after LASV infection we detected numerous iNOS-expressing monocyte/macrophage clusters in the liver parenchyma of HHD mice (Fig. 5A). Conversely, iNOS expression was not found in the liver of C57BL/6 mice, CD8/CD4-depleted HHD mice or MHC-I-/- mice infected with LASV. Furthermore, neither HHD nor C57BL/6 mice displayed hepatic iNOS expression or morphological evidence of monocyte/macrophage activation when assessed on day 2 and day 4 after infection (Fig. S1, analogous data in lung and spleen not shown), i.e. prior to the onset of the adaptive immune response.
Synthesis of the inflammatory cytokine subunit IL-12p40 is restricted to macrophages, monocytes and dendritic cells, and its production is greatly enhanced by T cell stimulation [48],[51]. Within eight days after LASV infection, susceptible HHD mice displayed vastly elevated serum IL12-p40 levels (Fig. 5B). Resistant C57BL/6 mice showed comparably moderate IL-12p40 elevation. In agreement with the above results on iNOS, IL-12p40 secretion was strongly reduced by CD4/CD8-depletion in HHD mice, and MHC-I-/- mice exhibited an intermediate IL-12p40 response. Together, these findings suggested that T cells triggered an inflammatory differentiation of monocytes/macrophages with subsequent production of NO, IL-12p40 and likely other inflammatory mediators (see Discussion section). This process occurred, however, solely under conditions of unchecked LASV replication i.e. in HHD mice but not in C57BL/6 mice where the infection and thus the antigen were rapidly cleared.
We had noted high virus loads in lung, liver and spleen of HHD mice (Fig. 1C) where the above pathological changes were found. The cellular distribution of LASV in vivo remains unknown, albeit the virus has been shown to replicate productively in cultured primate macrophages and dendritic cells [21],[52],[53]. To better understand how viral replication might be associated with disease we assessed virus distribution in tissues by immunohistochemistry on day eight after infection (Fig. 5C). LASV nucleoprotein (NP) was readily detected in a distinct population of cells in lung, liver and spleen. Morphological criteria suggested that these cells were predominantly monocytes/macrophages. We therefore performed immunofluorescence double-stains for LASV nucleoprotein (LASV-NP) and the monocyte/macrophage marker Iba-I (Fig. 5D). By this method, 82±6.8% of LASV-NP+ cells in liver, 65±8.5% in the lung and 79±11% in the spleen (mean±SD of six mice) could be identified as monocytes/macrophages, suggesting that they served as a major target of LASV.
The data in mice presented here suggest a dual role for T cells during LASV infection: T cells appear essential for rapid clearance of the virus, but if failing to do so they may play a key role in the ensuing disease process, too. Involvement of T cells in the pathogenesis of viral hemorrhagic fever has precedence in Dengue virus infection [10],[11],[39]. Unlike dengue hemorrhagic fever (DHF), where “original antigenic sin” of memory T cells appears to be involved, our findings suggest that T cell responses can have disease-enhancing effects during primary LASV infection. Importantly, we do not exclude roles for memory T cells in addition, but such aspects remain to be investigated (see below).
Our experiments modeled three prototypic scenarios of LASV-host balance, as defined by the parameters “T cells” (relative efficacy of antiviral T cell responses) and “Virus” (persistent virus load), thus delineating the extremes of a spectrum and resulting in the following outcomes:
Several mechanisms can be envisaged by which T cell responses enhance LF [10],[11],[39] but additional studies will be needed to address them directly. The histological picture in HHD mice supports the current view [7] that direct T cell-mediated cytolysis [54] unlikely is the main mechanism responsible for the tissue damage in LF, such as hepatocyte death with subsequent AST release. Based on the available evidence we postulate that during persisting viremia, T cells continuously encounter LASV epitopes on infected monocytes/macrophages in MHC-I and MHC-II context (Fig. 5E). Additional interaction via co-stimulatory molecules, or stimulation via T cell cytokines may trigger infected monocytes/macrophages to differentiate and subsequently secrete inflammatory mediators of their own [45],[48],[49],[51]. Overstimulation of macrophages can result in severe hepatic and pulmonary damage besides mediating a shock syndrome [26],[55],[56], and such overstimulation provides a plausible mechanism for indirect T cell involvement in LF pathogenesis. LASV and related viruses are known to replicate in cultured macrophages without causing cellular activation or production of inflammatory cytokines [21],[52],[53]. Hence, classical T cell-driven monocyte/macrophage activation by IFNγ and direct cell-to-cell contact [45],[46],[47],[48] may augment inflammatory differentiation and cytokine release from LASV-infected monocytes/macrophages in vivo [19],[21],[22], similar to the ability of LPS to induce the activation of LASV-infected macrophages in culture [21],[52],[53]. T cell stimulation may thus facilitate a systemic inflammatory condition [57] as a potential pathogenetic correlate of the diverse and non-specific clinical manifestations of LF.
This view of the role of T cells in LF correlates well with our observations in scenario I, where one would predict that efficient virus elimination by C57BL/6 and immune HHD mice results in only short and transient antigen presentation on a limited number of infected monocytes/macrophages, and therewith lack of disease, as was indeed found. Similarly, the concept explains our findings in scenario III, where the absence of T cells to properly activate infected monocytes/macrophages in CD8/CD4-depleted HHD mice results in mild or absent disease.
In addition, at least two – seemingly paradoxical – observations in LASV-infected non-human primates would support our mechanistic postulate. The first observation is that high doses of LASV tend to be less lethal than low ones [28]. This phenomenon may be explained through the mechanisms of T cell “exhaustion” or “deletion” under high virus loads [58]. As such, a high initial virus inoculum may weaken the T cell response, thus attenuating disease through shifting conditions from scenario II towards scenario III.
The second paradoxical observation stems from a vaccination study, in which a recombinant vaccinia virus expressing the NP protein of LASV was used in monkeys. The vaccine turned out to protect only a minority of the animals and, intriguingly, those LASV-challenged monkeys not protected by the vaccine displayed a more acute form of disease than control monkeys that had not been vaccinated at all [34]. A likely explanation may be that, although the vaccination may not have protected all animals, the still accelerated (memory) T cell response of non-protected animals also accelerated their disease process. Support for such a scenario comes from infection of mice with another arenavirus, lymphocytic choriomeningitis virus (LCMV). LCMV-induced immunopathological disease of the central nervous system is T cell-dependent and, similar to the observation with LASV in monkeys, can be enhanced by prior vaccination with recombinant vaccinia viruses expressing LCMV antigens [59].
Last but not least, our study introduces a mouse model for LF, the lack of which has long hampered progress in this field of research. Only the general versatility of the mouse as a research model, including the availability of gene-targeted strains, makes mechanistic studies as presented here possible. Despite certain shortcomings as listed below, we think that the humanized mouse model could prove useful in further studies on LF pathogenesis, especially as the model lends itself well to the assessment of CD8+ T cell-based vaccines (compare Fig. 3). Although the ∼20% lethality we found in HHD mice using the Ba366 strain of LASV may contrast with the uniform lethality observed in LASV strain Josiah-infected strain 13 guinea pigs or monkeys [28], our lethality rates match those reported for Josiah-infected outbred Hartley guinea pigs [60], as well as inbred strain 13 and strain 2 guinea pigs inoculated with other LASV isolates [61]. Nevertheless, we cannot exclude the possibility that the HHD model fails to recreate certain aspects of human LF.
For the mechanistic analyses presented here, we used relatively high intravenous LASV doses of 106 PFU. The need for such doses to cause severe disease likely reflects the imperfect adaptation of LASV to mice (compare also the Methods section). However, preliminary experiments indicate that viremia and AST elevation can already be observed at lower doses, albeit with higher variability, and that LASV can replicate in HHD mice after subcutaneous administration (data not shown). We would therefore argue that viremia and AST elevation in HHD mice may represent useful surrogates [38] to assess vaccine efficacy prior to an eventual confirmation in non-human primates.
T cell densities in the spleen of LASV-infected HHD and C57BL/6 mice were comparable (day 8: 3339±925 CD3+ cells/mm2 in HHD mice, 3826±2057 CD3+ cells/mm2 in C57BL/6 mice; mean±SD, n = 6; p = 0.61), arguing against preferential T cell depletion [44] as a potential reason for defective virus control in HHD mice. To characterize disease enhancing (HHD) vs. protective (C57BL/6) primary T cell responses against LASV, future studies will need to compare the magnitude, kinetics and effector-/cytokine-profile [62] of LASV epitope-specific CD4+ and CD8+ T cell responses. Similarly, serum IL-12p40 and iNOS expression have served as surrogates of the inflammatory monocyte/macrophage response in this study, but future work will have to determine its breadth in terms of cytokines, chemokines and inflammatory mediators such as NO, leukotrienes and prostaglandins, their production in tissues and systemic dissemination in blood. The experimental depletion and/or inhibition of monocyte/macrophage populations and of inflammatory mediators may provide additional insights into the cellular and molecular players in LF, indicating to which extent the “cytokine storm” hypothesis [57] can explain LF pathogenesis. With the availability of a mouse model for LASV such studies become possible, although the necessity for BSL-4 laboratory containment can represent a major practical hurdle.
Taken together, our results in mice suggest a two faced role of T cells in LASV infection, both in virus control and also in enhancing LF pathogenesis. This extends our understanding of LASV-host interactions and raises the possibility that heterogeneity in MHC-I and in overall T cell immunocompetence represents one explanation for the wide spectrum of clinical outcomes in a human population exposed to LASV [2]. Perhaps even more important, we think that beside beneficial also detrimental aspects of T cell-responses and -immunity [34],[59] should be given thorough consideration in future strategies for LF vaccine design.
C57BL/6, β2-microglobulin-deficient mice (MHC-I-/-) [63], MHC-II-/- [64] and HHD [37] mice were bred at the Institute for Laboratory Animal Sciences, University of Zurich, Switzerland. Experiments with Lassa virus were performed in the BSL-4 unit of the Bernhard Nocht Institute, Hamburg, Germany. Experiments with Mopeia virus were performed at the University of Geneva and at the University Hospital of Zurich, Switzerland. Permission for animal experiments was obtained from the authorities of the Freie und Hansestadt Hamburg, and from the Cantonal authorities of Geneva and Zurich, Switzerland, respectively. All experiments were performed in accordance with the Swiss and German law for animal protection, respectively.
This model is based on the Ba366 [65] strain of LASV. Pilot experiments with a range of isolates representing the different endemic areas (Josiah, Sierra Leone; Lib90, Liberia; Ba366, Guinea; AV, Ivory coast/Burkina Faso; CSF, Nigeria) had indicated that Ba366 was the virus that most efficiently replicated in HHD mice. LASV and Mopeia virus (AN21366), were grown on BHK21 and Vero cells, respectively, and were administered to mice at a dose of 106 PFU i.v. unless stated differently. Virus stocks and viral infectivity in blood samples were determined in immunofocus assays as described [66]. CD8+ and/or CD4+ T cell populations were depleted by i.p. administration of monoclonal antibodies YTS169 (anti-CD8) and YTS191 (anti-CD4) on day -3 and day -1 of LASV infection as previously described [67]. The efficiency of depletion was verified by flow cytometry and was >99%.
Serum AST and ALT activities were determined by using commercially available colorimetric assay kits (Reflotron, Roche Diagnostics, Germany). Mouse tissues were fixed in 4% formalin and were embedded in paraffin. Sections were stained with hematoxilin/eosin (H/E) or processed for immunohistochemistry as follows: Upon inactivation of endogenous peroxidases (PBS/3% hydrogen peroxide) and blocking (PBS/10% FCS) sections were incubated with the primary antibodies rat anti-human CD3 (crossreactive with murine CD3 on mouse T cells; Serotec), rabbit anti-Iba-1 (monocytes/macrophages, Wako Pure Chemical Industries) or rat anti-Lassa nucleoprotein (see below). Bound primary antibodies were detected with biotinylated rat-specific (DakoCytomation) or rabbit-specific (Amersham) secondary antibody, followed by incubation with extraavidin peroxidase (Sigma Aldrich), and bound peroxidase was visualized by 3,3′-diaminobenzidine as chromogen (Sigma Aldrich). Haemalaun was used for counterstaining of nuclei. For fluorescence double labeling, primary antibodies were visualized using species specific Cy3- or Cy2-conjugated secondary antibodies (all from Jackson ImmunoResearch Laboratories Inc.) with DAPI (Sigma-Aldrich) nuclear staining. To determine the percentage of monocytes/macrophages (Iba-1-positive cells) amongst LASV-infected cells, a total of 41 (liver) and 27 (lung) randomly captured 40x visual fields were analyzed.
Histological spleen sections stained with anti-CD3 antibody (T cells) and counterstained with Haemalaun (nuclei, see above) were scanned using the Dotslide System (Olympus GmBH) at a 200-fold magnification. For analysis, the images were automatically processed in a custom-programmed script of Cognition Network Language based on the Definiens Cognition Network Technology platform (Definiens Developer XD software). The Cognition Network Language is an object-based procedural computer language, designed for automated analysis of complex, context-dependent image analysis. In brief, the programmed script first discriminates spleen tissue and tissue-free surroundings by spectral difference detection. The surface of the resulting region of interest (spleen tissue, ROI) is calculated. Subsequently “CD3 positive cells” within the ROI are detected based on brown anti-CD3 staining and are counted, to calculate the number of CD3+ cells per mm2 of tissue.
Serum IL-12p40 was determined using a sandwich ELISA kit (eBioscience) according to the manufacturer's instructions.
Recombinant NP of LASV strain BA366 was expressed in E. coli using the pET28 expression vector system (Novagen). Supernatants of pET28 constructs were purified using the Talon Metal Affinity Resin (Clontech) in a batch procedure. Urea (8 M) lysates were brought to nondenaturing conditions by increasingly substituting the buffer for sonication buffer during the resin-batch procedure. Proteins were eluted with 250 mM imidazole in sonication buffer on a gravity column (Bio-Rad). Rat antisera were raised against purified recombinant NP by s.c. immunization with recombinant NP emulsified in complete Freund's adjuvant containing 1 mg of Mycobacterium tuberculosis (H37RA; Difco Laboratories, Detroit, MI). Four weeks after the first immunization, animals were boosted with recombinant NP emulsified in incomplete Freund's adjuvant (Difco). Terminal bleedings were performed 4 weeks after the boost. The specificity of the anti-LASV-NP antiserum was verified by immunofluorescence tests on LASV-infected cells as well as on LASV-infected or non-infected tissues. Pre-immune serum from the rats used for immunization was included as a control in both settings.
Epitope-specific CD8+ T cells were enumerated by an intracellular cytokine assay for IFNγ as previously described [68]. In brief, 106 splenocytes were incubated in 200 µl of IMDM supplemented with 10% FCS and penicillin/streptomycin for 5 h at 37°C at a 10−6 M concentration of the LASV-GP-derived peptide epitope GP42-50 (GLVGLVTFL), GP60-68 (SLYKGVYEL), GP441-449 (YLISIFLHL) or with medium alone as a negative control. To enhance intracellular accumulation of IFN-γ, brefeldin A was added at a final concentration of 5 µg/ml for the last 3.5 hours of culture. Subsequently, the cells were washed with FACS buffer (PBS supplemented with 2% FCS, 0.01% NaN3 and 20 mM EDTA) and surface staining was performed with anti-CD8β-PE and anti-B220-PerCP antibody conjugates (both from BD Biosciences) for 30 min at 4°C. After washing twice with FACS buffer, the cells were fixed with 100 µl of 4% paraformaldehyde in PBS for 5 min at 4°C. Two milliliters of permeabilization buffer (FACS buffer supplemented with 0.1% w/v saponin, Sigma) were added and the cells were incubated for 10 min at 4°C. Subsequently, they were spun down and stained intracellularly with anti-mouse-IFNγ-APC (BD Biosciences) in permeabilisation buffer for 60 min at 4°C. After two washes with permeabilization buffer, the cells were resuspended in FACS buffer and were analyzed on a FacsCalibur (Becton Dickinson). FACS plots were gated on B220− lymphocytes.
Between group differences were analyzed by 1-way ANOVA and 2-way ANOVA for individual or multiple values of different groups, respectively, followed by LSD post tests. SPSS vs. 13 was used for analysis. P values <0.05 were considered statistically significant (indicated as * in figures). P<0.01 was considered highly significant (indicated as ** in figures). P>0.05 was considered as not significantly different (“n.s.”).
|
10.1371/journal.pcbi.1002087 | Stimulus-Dependent State Transition between Synchronized Oscillation and Randomly Repetitive Burst in a Model Cerebellar Granular Layer | Information processing of the cerebellar granular layer composed of granule and Golgi cells is regarded as an important first step toward the cerebellar computation. Our previous theoretical studies have shown that granule cells can exhibit random alternation between burst and silent modes, which provides a basis of population representation of the passage-of-time (POT) from the onset of external input stimuli. On the other hand, another computational study has reported that granule cells can exhibit synchronized oscillation of activity, as consistent with observed oscillation in local field potential recorded from the granular layer while animals keep still. Here we have a question of whether an identical network model can explain these distinct dynamics. In the present study, we carried out computer simulations based on a spiking network model of the granular layer varying two parameters: the strength of a current injected to granule cells and the concentration of Mg2+ which controls the conductance of NMDA channels assumed on the Golgi cell dendrites. The simulations showed that cells in the granular layer can switch activity states between synchronized oscillation and random burst-silent alternation depending on the two parameters. For higher Mg2+ concentration and a weaker injected current, granule and Golgi cells elicited spikes synchronously (synchronized oscillation state). In contrast, for lower Mg2+ concentration and a stronger injected current, those cells showed the random burst-silent alternation (POT-representing state). It is suggested that NMDA channels on the Golgi cell dendrites play an important role for determining how the granular layer works in response to external input.
| Intensive studies of Pavlovian delay eyelid conditioning suggest that the cerebellum can memorize a passage-of-time (POT) from the onset of an external stimulus. To account for possible mechanisms of such POT representation, some network models have been proposed to show that granule cells (grcs) in the cerebellar granular layer can exhibit random alternation of burst and silent modes under feedback inhibition from Golgi cells, resulting in non-recurrent generation of active granule cells populations. On the other hand, the oscillation of local field potential (LFP) has been observed in the cerebellar granular layer when animals stay at rest. Some network models have shown that grcs can elicit synchronous spikes in an oscillatory manner. These qualitatively different neural dynamics of the granular layer raises a question of how they can be accounted for by an identical network in the granular layer. Here we report that grc activities of a biologically plausible spiking network model undergo the state transition between synchronized oscillation and random burst-silent alternation, depending on the activation of NMDA channels on the Golgi cell dendrites and the strength of a current injected to grcs.
| The cerebellar granular layer is one of the stations receiving external stimuli for information processing of the cerebellar cortex. The granular layer is thought to transform spatial patterns of mossy fibers (MFs) input signals into a population of active granule cells (grcs) [1], [2]. Recently, Yamazaki and Tanaka [3] have proposed that the granular layer transforms spatiotemporal patterns of MF input signals into a sparse population of active grcs in the presence of inhibitory Golgi cells (Gocs), and suggested that the passage of time (POT) from the onset of MF signals is represented by the granular layer network.
On the other hand, when an animal stays at rest without any external stimuli, oscillatory local field potential (LFP) is observed in the granular layer at 7–8 Hz in rats [4] and at 13–14 Hz in monkeys [5], [6]. Although the origin of this oscillation remains unknown, Maex and De Schutter [7] proposed that grcs and Gocs become active alternatingly and repeatedly by the recurrent connections, and their average oscillatory firing may be observed as the oscillation of LFP.
There are two distinct dynamics in the cerebellar granular layer: one for the POT representation and the other for the synchronized oscillation. Assuming that the two dynamics emerge from the same neural circuit, one may expect that the strength of external input controls the transition between the two dynamics. However, models accounting for the POT representation do not exhibit the synchronized oscillatory firing of grcs and Gocs but generate spikes randomly for weak external input [3], whereas models accounting for the synchronized oscillation persistently show the oscillatory state even for strong external input [7]. Therefore, it is not trivial to explain the possibility that the two dynamics take place in the same model of the cerebellar granular layer. In this study, we demonstrate that a spiking network model of the cerebellar granular layer can generate synchronized oscillation for weak external input and active grc populations representing the POT for strong external input.
We build our model on the basis of the GENESIS script of the granular layer model for synchronized oscillation, which was written by Maex and De Schutter [7]. Briefly, we extend the original single-compartment Goc model to a multi-compartment model composed of a soma and a dendrite, on which N-methyl-D-aspartate (NMDA) channels are distributed (Fig. 1A). We also extend their one-dimensional network structure to a two-dimensional one and set random sparse connections between the model Gocs to the grcs (Figs. 1B and C). We modified the values of the time constant and conductance of GABAA channels on a grc soma, and that of the conductance of AMPA channels on a Goc dendrite (see below and Table S1) so that simulation reproduces the synchronized oscillation reported by Maex and De Schutter [7].
We build a model granular layer of the cerebellum using the network structure employed in our previous study [3]. That is, 32×32 model Gocs are arranged in two-dimensional grids (Figs. 1B and C), which is in contrast with the one-dimensional arrangements of model Gocs and grcs in the study of Maex and De Schutter [7]. Our model Gocs are evenly positioned at 35 µm intervals within a square sheet of 1,085×1,085 µm2. It was estimated that there are 1,000 times more grcs than Gocs [8], [9]. Numerous grcs are connected with a glomerulus [10]. However, simulation with more than 1 million model neurons is beyond the power of our computers. In Yamazaki and Tanaka [3], 100 nearby grcs that were assumed to be connected with a glomerulus exhibited similar firing patterns despite each of the grcs independently received noisy signals through MFs. Such redundant activity patterns of grcs suggest that a large number of grcs behave as a single cluster when they receive inputs from the nearest Goc through a single glomerulus. In the present model, for the sake of the economy of computer power, we assume that a single model grc represents a grc cluster composed of about 1,000 neurons.
We also assume that a model Goc randomly receives 10% of 9×32 PF inputs from model grcs with their dendritic arborization, whose diameter is set at 315 µm. The mean and standard deviation of the actual number of PF inputs to a model Goc were 26.80 and 6.50, respectively. The model Goc, in turn, sends inhibitory inputs to model grcs located within the extent of axonal arborization (Fig. 1B), which is set at 315μm. The number of model Gocs located inside a circle of the diameter of 315μm amounts to 69 by actual counting, which can be roughly estimated by π(315/2)2/(35-1)2+1 = 68.4. In the present model we assume that each grc randomly receives 10% of 69 connections from Gocs. The mean and standard deviation of the actual number of connections were 6.11 and 2.57, respectively. For simplicity, we omit MF inputs to model Gocs.
We use the same model grcs as those adopted by Maex and De Schutter [7] except for the synaptic channels. Brickley et al. [11] reported that inhibitory postsynaptic potentials (IPSPs) can be fitted well with the sum of three exponentials, and that the largest component, which has the decay time constant of 100 ms, contributes to 20% of the IPSPs. Considering this finding, we simulate IPSPs using a double-exponential function with rise and decay time constants of 5 and 100 ms, respectively [11]. Further details are shown in Table S1.
In Maex and De Schutter [7], a model Goc was also a single-compartment Hodgkin-Huxley unit with realistic ion channels. Their model Gocs received excitatory inputs from model grcs through α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) channels with rise and decay time constants of 0.03 and 0.5 ms, respectively [12].
It has been found that Gocs receive excitatory input signals from PFs through not only AMPA channels but also NMDA channels [13]. However, even when we add NMDA channels on the somas of model Gocs, the NMDA channels do not open effectively to evoke sustained depolarization because the after-hyperpolarization (AHP) [14], [15] following the generation of each action potential rapidly decreases the somatic membrane potential. On the other hand, it is known that the dendritic potential tends not to be affected by AHP at cell somas [14]. Moreover, it has been shown that direct dendritic excitation produces sustained and burst responses although somatic excitation does not [14]. If NMDA channels really induce a prolonged activation of Gocs, it implies that NMDA channels are located on the Goc dendrites. In this study, a model Goc is represented as a soma and a dendrite whose length is 300 µm (cf. [16]) (Fig. 1A). The dendrites of model Gocs are assumed to possess AMPA and NMDA channels. We simulate NMDA receptor (NMDAR)-mediated excitatory postsynaptic potentials (EPSPs) with a double-exponential function with rise and decay time constants of 5 and 100 ms, respectively, in accordance with Misra et al. [17]. Further details are shown in Table S1.
It is known that Gocs are morphologically heterogeneous and their action potentials are variable [18]. Although it is likely that random connectivity in the network would have smeared individual heterogeneities, a reversal potential for the leak current at model Gocs is distributed uniformly between −60 and -50 mV according to Maex and De Schutter [7].
We model MF input signals as a current injected directly to the soma of each model grc, instead of feeding spike trains. Freeman and Muckler [19] have reported that the spontaneous firing rate of MFs is as low as 5 Hz, whereas the firing rate increases up to 30 Hz when an animal is stimulated with a tone. We assume that the influx of a current to grcs increases with the frequency of firing conveyed through MFs. In the simulation, we inject a current IMF of 10.7 pA to all model grcs for 2 s. For the succeeding 2 s, we inject a current IMF of 22.7 pA to evoke activity in grcs. Subsequently, we inject a current IMF of 10.7 pA for 2 s again to resume the baseline activity of grcs. In total, we simulate the dynamics of the granular layer network for 6 s. Note that we omit noise in the injected current.
We build the present model and perform numerical simulations using the GENESIS simulator [20]. Differential equations are solved using the Crank-Nicolson method with a fixed time step of 20 µs. When we carry out simulation of eyelid conditioning, we do not need to calculate activities of the granular layer at the same time. We have only to use activity patterns of grcs, which have been stored in a file recorded during the simulation of activities of the granular layer using GENESIS. Because we assume a simple PC model, the simulation does not require the usage of GENESIS that is an expert for simulation of realistic neuron models. Therefore, the simulation program of the model Purkinje cell (PC) for eyelid conditioning, which is just solving differential equations using the first order Euler method with a fixed time step of 20μs, was written in C and C++ programming languages.
Let be the spike activity of model neuron (grc or Goc) i (1≤i≤N, N = 32×32 = 1024) at time t:(1)The fluctuation in the number of neurons that elicit spikes at time t, , is given by(2)The normalized autocorrelation function of the numbers of active neurons at times t and t + τ is defined by(3)where τ is a time lag and T is the duration of weak or strong external input. Calculating the normalized Fourier cosine transform of this normalized autocorrelation function, we define the oscillation index OI, which measures the degree of synchronized oscillation at neurons as follows:(4)where is the frequency giving a maximum power defined by the Fourier cosine transform of the normalized autocorrelation function. When OI takes a value of 1, a population of active neurons appears periodically. When OI takes a value of 0, NAC(τ) is a constant function of τ, and populations of active neurons appear uniformly in time.
is defined as the AMPA receptor (APMAR)-mediated EPSPs on a PC from grc i, as follows:(5)where is the decay time constant of AMPAR-mediated EPSPs on a PC, which is set at 30 ms. As defined in the previous study [3], we use the similarity index of the function of τ, which is given by(6)The right-hand side of this equation represents the extent to which two populations of active neurons at different times with interval τ are correlated. If takes a value of 1, the populations at times t and t + τ are identical, and when takes a value of 0, different populations are active at times t and t + τ. The monotonic decrease in with increasing |τ| indicates that the population of active neurons changes with time without recurrence [21]. In addition, we also define another index that measures the ability of POT representation. The ideal POT representation is achieved by monotonically decreasing with respect to |τ| with a large difference between and the minimum of [3]. However, even in the cases of oscillatory activities overlaid on the random repetition of burst and silent activities, if the temporal coherence of the oscillation decreases with increasing τ, the POT representation is possible to some extent. To extract a residual POT-representing component of the active neurons' population, we first fit with a Gaussian function and calculated the height between and the minimum of . The height is then divided by 0.376, which is the largest value of the height among all similarity indices examined by a brute-force search varying Mg2+ concentration ([Mg2+]) and input current strength. We call this normalized height the POT-representation index (PI).
To examine the reproducibility of active neurons' population among different trials of the same external input, we use the previously introduced reproducibility index [3], [21], given by(7)Here, the superscripts (1) and (2) indicate two different trials. represents the similarity between the populations of active neurons at time t measured from the onset of injection of a current to grcs in two different trials. When takes a value of 1, the populations of active neurons in the two trials are identical, and when it takes a value of 0, they are completely different [3].
In order to confirm that the POT-representing state generated by our model functionally serves for inter-stimulus interval representation in Pavlovian delay eyelid conditioning [22]–[24], we conduct simulations assuming that strong current injection to grcs corresponds to a neural signal of a conditioned stimulus (CS) conveyed through MFs and a neural signal from the inferior olive through climbing fibers corresponds to an unconditioned stimulus (US) (Fig. 1C). The sustained CS is fed to grcs, whereas the US is sent to PCs with a certain delay. We assume that a large current injection to all grcs corresponds to a CS presentation, and a small current injection corresponds to the input of spontaneous MF activity. We employ a Hodgkin-Huxley unit as a model PC assuming that the PC receives excitatory inputs from all Parallel fibers (PFs) (Fig. 1C). PF input signals are modeled as AMPAR-mediated EPSPs. The decay time constant is set at 30 ms [25], and the peak conductance is set so that the model PC elicits spikes at a maximum rate of 100 spikes/s in response to a CS [26]. We assume that the US is fed either 0.5, 0.75 or 1.0 s after the CS onset and that simulated long-term depression (LTD) is induced when PFs and a climbing fiber are coactivated within a brief time window [16]. We set initial synaptic weights of PFi in the model PC, wi(0), to 1. When grc i fires 0.05–0.1 s before the onset of a US, wi is set to 0; otherwise, wi is not changed. For simplicity, we assume that US signals contribute to the induction of LTD at PFs, and we do not take into account their contributions to the dynamics of the PC activity in simulations.
Figures 2B and C represent the temporal patterns of spikes generated by 200 representative model Gocs and grcs, respectively, in response to the injection of a current to the grcs (Fig. 2A). For the first 2 s, in which a small current was fed to the grcs, simulating a response to spontaneous MF activity, model grcs and Gocs elicited spikes rhythmically and synchronously with a frequency of 9 Hz. The emergence of the synchronized oscillation in the grcs and Gocs is consistent with the findings of Maex and De Schutter [7]. For the successive 2 s, in which a large current was injected to the grcs, the model grcs exhibited random alternation between burst and silent modes, consistent with the findings of Yamazaki and Tanaka [3]. After the current amplitude was decreased to that for the first 2 s, the network quickly returned to the oscillatory state at 9 Hz. These simulation results show that the model cerebellar granular layer can exhibit two qualitatively distinct dynamical states depending on the strength of the MF input signal. In the following subsections, we examine the properties of each dynamical state in detail.
To confirm the generation of synchronized oscillation at model grcs in response to the injection of a small current (0–2 s), we calculated a normalized autocorrelation function NAC(τ) of the activities of the grcs using Equation 3 and the value of the oscillation index OI using Equation 4. We found a clear oscillation at 9 Hz, as shown in Fig. 2D. We also obtained OI = 0.836, suggesting robust synchronization of the activities of the model grcs. This oscillation frequency was in the range of frequencies of oscillatory LFP observed in the cerebellar cortex by Hartmann and Bower [4] and Pellerin and Lamarre [5] (7–8 Hz in rats and 13–14 Hz in monkeys). This frequency was lower than that shown by Maex and De Schutter [7] due to a larger decay constant of γ-aminobutyric acid-A receptor (GABAAR)-mediated IPSPs on model grcs (0.31 ms (rise) and 8.8 ms (decay) for Maex and De Schutter [7]; 5.0 ms (rise) and 100 ms (decay) for the present model). These results suggest that the granular layer of the biological cerebellum is in a state of synchronized oscillation under spontaneous MF signal input. The similarity between spike activity patterns at different times decreased with the time interval, but patterns showed weak oscillation associated with synchronized oscillation (Fig. 2F). The NAC(τ) and similarity index for model Gocs were similar to those for model grcs (Fig. S1).
The emergence of synchronized oscillation may be interpreted in terms of the following circuit mechanism between grcs and Gocs. The relatively low average firing rate of grcs induced by a small input current evokes the weak depolarization of Gocs, which rarely opens voltage-gated NMDA channels on the Goc dendrites. Thereby, the Gocs elicit spikes only sparsely followed by a refractory period. The low-frequency spike activity of the Gocs inhibits grcs, and the grcs become inactive. After the recovery from the inhibition from the Gocs, the grcs become active due to the sustained external current input and elicit low frequency spikes again. The repetition of such activation and deactivation processes results in the synchronized oscillation of the model grcs and Gocs. This mechanism for the emergence of the oscillatory state is common to that reported by Maex and De Schutter [7].
The average firing rate of individual grcs was less than half (≈4 spikes/s) of the oscillation frequency (9 Hz) in our simulations. This indicates that not all grcs were activated at every oscillation cycle, which is attributed to the random and sparse connectivity between the Gocs and grcs. When the inhibition of the grcs by the Gocs was removed, the grc firing rate increased to 44 spikes/s. This result suggests that the synchronized oscillation was generated by the network dynamics, rather than the intrinsic mechanisms of individual grcs under the constant current injection.
Figures 2B and C also represent spike patterns of 200 model Gocs and grcs, respectively, in response to the injection of a large current to the grcs in the interval from 2 to 4 s. Model grcs exhibited random alternation between burst and silent modes. The normalized autocorrelation NAC(τ) is markedly reduced except at τ = 0 ms (Fig. 2E). Different grcs showed different patterns of spike trains. To determine whether the same active grc populations appear more than once, we calculated the similarity index using Equation 6 and plotted it in Fig. 2G. The value of was 1 at τ = 0 ms because of the trivial identity. monotonically decreased as |τ| increased with the POT-representation index PI = 0.923. This indicates that the population of active grcs changed gradually with time from the onset of a large current injection without the recurrence of active grc populations, as reported by Yamazaki and Tanaka [3]. The NAC(τ) and for model Gocs were similar to those for model grcs (Fig. S1). To determine the effect of random connections between model grcs and Gocs, we performed simulations using a modified model in which grcs and Gocs were all connected with each other so that the network did not have any randomness. As a result, activities of these cells exhibited only coherent oscillations (data not shown). This suggests that the random connectivity is a major cause of generating the random alternation between burst and silent modes in grcs and Gocs.
The random alternation between burst and silent modes of grcs can be accounted for by the following mechanism. The strong activation of grcs by the large current injection vigorously depolarizes randomly connected Gocs, resulting in the activation of voltage-gated NMDA channels on the dendrites of the Gocs. Because of the long decay time constant of NMDAR-mediated EPSPs, the Gocs send sustained inhibitory signals to nearby grcs, so that the grcs become inactive. Then a Goc that receives inputs from these grcs decreases its activity, which reactivates the grcs. However, due to the random connections, the timings of the reactivation and deactivation of grcs are different for different grcs, resulting in the random alternation between burst and silent modes of grcs.
To confirm that the present model achieves reproducible POT representation for different trials of the strong current injection, we carried out the following simulations. We injected a large current to grcs twice during 2–4 and 6–8 s, whereas we injected a small current at other times. We evaluated the reproducibility index between active grc populations in response to the first and second injections of a large current using Equation 7 (Fig. 3). immediately after the onset of the current injection (t<50 ms) was 1, indicating a perfect reproducibility. Afterward, it tended to decrease with the duration of the current injection. The perfect reproducibility observed around the onset of the current injection originates in the simultaneous activation of all grcs at the onset followed by strong inhibition from Gocs. This resetting mechanism enables the network to lose its history of activities. Consequently, the present model can represent the POT robustly and reproducibly for different trials of a large current injection.
To confirm the importance of NMDA channels assumed on the Goc dendrites for the network dynamics, we blocked the channels by increasing the concentration of Mg2+ (1.20 mM → 13.0 mM) and carried out a simulation.
Figure 4 shows spike trains elicited by 200 Gocs (Fig. 4B) and 200 grcs (Fig. 4C) in response to a current injection (Fig. 4A) under the blockade of the NMDA channels on the model Gocs. The synchronized oscillation of grcs and Gocs was unaffected during the injection of a small current to the grcs for 0–2 and 4–6 s. This was observed by the appearance of oscillatory behavior of the normalized autocorrelation function NAC(τ) (Fig. 4D). The similarity index resembled that in the default case (Fig. 4F). As mentioned previously, when the injected current was small, the NMDA channels on the Goc dendrites rarely opened, irrespective of whether NMDA channels were blocked. Therefore, the oscillatory behavior of grcs and Gocs during small current injections was preserved under the blockade of NMDA channels. On the other hand, in response to a large current injection to the grcs for 2–4 s, the grcs elicited spikes at a higher firing rate than in the default case (13 spikes/s → 32 spikes/s) and the spike patterns were random and uniform, as shown in Fig. 4C. This observation was justified by the fact that NAC(τ) showed oscillation whose amplitude was sufficiently small compared with its value at τ = 0 (Fig. 4E). The grcs did not show random burst-silent alternation. This was caused by the weaker inhibition from Gocs due to the blockade of NMDA channels. As shown in Fig. 4G, became much higher than in the default case. The POT-representation index PI was as small as 0.009, compared with the value of 0.923 in the default case. This indicates that the temporal sequence of active grc populations does not represent the POT even during a large current injection. We confirmed a similar tendency for the activity patterns of model Gocs (Fig. S1).
Taken together, the activation of NMDARs on the Goc dendrites is not important for the generation of synchronized oscillation of grc and Goc activities. In contrast, the activation of the NMDARs is indispensable for the generation of random burst-silent alternation, which enables the network to represent the POT.
Next, to examine whether the oscillatory states are affected by the reduced Mg2+ concentration, we performed a simulation under a nearly [Mg2+]-free condition (0.013 mM).
Figures 5B and C show spike trains of 200 Gocs and grcs, respectively, in response to the current injection to the grcs (Fig. 5A). The spike trains of the grcs exhibited random burst-silent alternation in response to a large current injection for 2–4 s. The observation of randomness in the spike activity patterns was justified by the fact that the normalized autocorrelation function NAC(τ) took almost zero except at τ = 0 (Fig. 5E). Although the firing rate of the grcs was reduced by the enhanced inhibition from the Gocs, the similarity index decreased monotonically with τ (Fig. 5G), as observed in the case where the POT is represented. On the other hand, for 0–2 and 4–6 s, during which a small current was injected, the grcs and Gocs did not undergo synchronized oscillation (Fig. 5D). The grcs generated spikes sparsely in a stochastic manner (the oscillation index OI = 0.207), whereas the Gocs elicited spikes at almost the same frequency as that in the default setting. monotonically decreased, reflecting the absence of synchronized oscillation (Fig. 5F). These results indicate that the persistent reduction of Mg2+ concentration disrupts the synchronized oscillation. We confirmed a similar tendency for the activity patterns of model Gocs (Fig. S1).
Here we examine how the synchronized oscillation and POT-representing states emerge depending on the Mg2+ concentration.
Figure 6A shows the changes of the oscillation index OI and POT-representation index PI with Mg2+ concentration at IMF = 22.7 pA. For 0.013 mM<[Mg2+]<0.260 mM, PI was almost constant at around 0.510. As Mg2+ concentration increased beyond 0.260 mM, PI decreased and reached 0.146 at [Mg2+] = 0.469 mM. Then, PI started to increase and reached a maximum at [Mg2+] = 1.40 mM. As Mg2+ concentration further increased, PI decreased again and vanished for [Mg2+]>6.52 mM. The POT-representing states emerged only in the interval of 0.782 mM<[Mg2+]<2.45 mM. OI was high in two separate domains for 0.417 mM<[Mg2+]<0.782 mM, and 5.21 mM<[Mg2+]<9.13 mM. Hence, the domain where the POT-representing state appeared was sandwiched by the two high-OI domains.
Figure 6B shows the firing rate averaged over all grcs and the oscillation frequency. The average oscillation frequencies in the two high-OI domains were 7–8 Hz and 20–21 Hz. In contrast, the firing rate monotonically increased from 5 to 32 spikes/s as Mg2+ concentration increased.
Next, we analyzed the changes of activity patterns as we varied continuously the strength of the injected current to grcs, IMF. Figure 7A shows the changes of the oscillation index OI and POT-representation index PI with the strength of the injected current at [Mg2+] = 1.2 mM. For IMF<6.5 pA, it was too small for the grcs to be activated, and hence PI could not be defined. For 6.5 pA<IMF<16.5 pA, there was a local maximum of PI (PI = 0.269) at IMF = 8.0 pA. PI exhibited a maximum at around IMF = 26.5 pA. For IMF>26.5 pA, PI gradually decreased as IMF increased. POT-representing states were well defined for 19 pA<IMF<33 pA. On the other hand, synchronized oscillation states were well defined in the interval of 9.5 pA<IMF<18 pA, whereas OI remained relatively small outside this interval.
Figure 7B shows the average firing rate and the oscillation frequency of grcs against the strength of the injected current. The oscillation frequency is shown only in the range of the injected current in which oscillation states were well defined. The oscillation frequency was dissociated from the average firing rate. As IMF increased up to IMF = 60 pA, the firing rate monotonically increased from 0 to 31 spikes/s.
Here, we explore all possible dynamical states in the present network by varying [Mg2+] and IMF exhaustively. Figure 8 shows the values of oscillation index OI (green) and POT-representation index PI (red) of grc spike patterns (see also Figs. S2A and B). In Figure 9, we show spike trains, similarity indices and normalized autocorrelation functions of the network dynamics at eight representative points marked by asterisks in the state map shown in Fig. 8. Point 1 is characterized by an extremely high firing rate of grcs without oscillations. Point 1 corresponds to the case of a strong blockade of NMDA channels on the Goc dendrites. Because of the weaker inhibition of grcs by Gocs than the excitation of grcs by the injected current, both PI and OI had small values (PI = 0.075 and OI = 0.133), while the average firing rate was high (38.4 spikes/s). Point 2 exhibits synchronized oscillation induced by a small injected current. PI, OI and the oscillation frequency at point 2 were 0.052, 0.991 and 13 Hz, respectively. Points 1 and 2 are irrelevant to the POT representation. Points 3 and 6 are relevant to the POT representation. In particular, point 3 gives a typical POT-representing state, at which PI and OI were 0.943 and 0.159, respectively. Also at point 6, the grcs showed a POT-representing state due to the sustained opening of the NMDA channels at a low Mg2+ concentration. However, PI was lower than that at point 3 (PI = 0.580 and OI = 0.243). Point 4 was located at the transition from a synchronized oscillation state to a POT-representing state, because synchronized oscillation appeared transiently at the beginning and changed to a POT-representing state. As IMF decreased, this transient synchronized oscillation became persistent as observed at point 5 (oscillation frequency = 7 Hz; OI = 0.969), which was caused by the repeated bumping of strong excitation of Gocs mediated by NMDA channels. The lower firing rate of the grcs at point 7 was caused by strong inhibition from the Gocs, which exceeded the excitation by a large current injected to the grcs. Finally, point 8 corresponds to an almost inactive state of the grcs exhibiting neither a synchronized oscillation state nor a POT-representing state, because the injected current is too small and Gocs activated by persistently open NMDA channels inhibit the grcs strongly.
We classified the eight points into four different states: a synchronized oscillation state (State I), POT-representing state (State II), a metastable state (State III) and a uniform firing state (State IV). State I to IV are represented by the dynamical states at points 2 and 5, points 3 and 6, point 4 and points 1, 7 and 8, respectively. State II is observed when the average firing rate of the grcs is approximately 1–20 Hz (Fig. 8B). When Mg2+ concentration was 1.0–1.5 mM and IMF was 22–33 pA, PI was higher and the average firing rate of grcs was almost constant. This shows that excitation and inhibition to grcs are balanced in this parameter range.
Finally, we conducted additional simulations to confirm that the POT-representing states identified in this study can work to encode inter-stimulus intervals (ISIs) and a Purkinje cell (PC) can read out the intervals to elicit conditioned responses (CRs). We embedded a model PC in the granular layer network so that the PC received excitatory inputs from all grcs.
Figure 10 shows the membrane potential of the model PC before and after the simulated conditioning. Before conditioning, the PC continuously elicited spikes at a high frequency during the CS presentation. In the simulated conditioning, the US signal was assumed to be given 0.75 s after the CS onset. The synaptic weights of the active PFs connected to the PC were set to zero 0.05–0.1 s before the US onset, whereas the synaptic weights of the other PFs were unchanged. After the conditioning, the model PC stopped firing about 200 ms earlier than the US onset and restarted at about 0.91 s. This result indicates that the model PC is able to learn the ISI between the CS and US onsets, which reproduces the experimental result of Jirenhed et al. [27]. We also confirmed that the model PC was able to learn other ISIs by setting the ISI to 0.5 or 1 s (Fig. 10). Moreover, we confirmed that other LTD time windows reported by Chen and Thompson [28] and Wang et al. [29] gave the same result (data not shown). When the CS signal was not presented, the model PC elicited burst spikes rhythmically at 9 Hz.
In the present study, we demonstrated that an identical computational model of the cerebellar granular layer can generate both synchronized oscillation states [7] and POT-representing states [3], [30]–[32] depending on the strength of a current injected to grcs (Fig. 11).When the network was driven by a small injected current which was assumed to represent spontaneous MF spike input, the model grcs and Gocs underwent synchronized oscillation at 9 Hz. This synchronized oscillation may be involved in the oscillation of LFP found in animals staying at rest [4], [5]. On the other hand, when a large current was injected to grcs, which was assumed to a CS signal fed to the network through MFs, the model grcs exhibited random burst-silent alternation of grc activities, which enables POT representation by active grc populations [3].
We have shown that a POT can be represented by populations of active grcs, each of which exhibits random alternation of burst and silent modes. This raises a question of what the random burst-silent alternation originates in. Maex and De Schutter [7], in which synaptic weights of any connections in their network were random, did not show the random burst-silent alternation of grc activities. On the contrary, Yamazaki and Tanaka [3], in which the synaptic weights were constant though the connection pattern was random, demonstrated that the random alternation emerged at grcs. Simulations using a modified model in which grcs and Gocs were all connected with each other without connection randomness resulted in only coherent synchronized oscillations of grc and Goc activities. These observations indicate that randomness in synaptic connections between grcs and Gocs is more important than random fluctuation in synaptic weights of these connections for the generation of the random alternation. For the randomness in the external input, Medina et al. [31] and Yamazaki and Tanaka [3] assumed that randomly generated spikes were fed to model grcs through MFs in order to demonstrate that a POT can be represented by populations of active grcs. However, the present model with a constant current injection to grcs showed that the random burst-silent alternation does not require noisy input to grcs. Taken together, it is concluded that randomness in the connectivity between grcs and Gocs is a major cause of generating the random burst-silent alternation in grcs and Gocs.
Buonomano and Mauk [32] have shown that the precision of a timed CR elicitation became worse with increase in weights of MF synapses on Gocs, whereas activity patterns of grcs became more resistant to noise in MF input signals. In addition, Maex and De Schutter [7] have observed that the synchronization of grc and Goc activities disappeared at stronger MF synapses on Gocs. These studies imply that the POT-representing states and synchronized oscillation states may disappear when we take into account direct MF inputs to Gocs, which are omitted in the present model. To examine the effects of the MF inputs to Gocs on the dynamical states, we performed additional simulations changing the ratio in strength of the current injected to Gocs to that injected to grcs, which would reflect the weight of MF synapses on Gocs relative to that on grcs, for two cases of a small and large currents injected to grcs. We found that OI was predominant over PI for any ratio of injected currents between 0 and 2 when the small current (10.7 pA) was injected to grcs, although both indices took small values when the ratio was large (Fig. S3A). For the large current injection (22.7 pA), PI was larger than OI when the ratio of injected currents was smaller than 0.3, whereas OI was larger than PI when the ratio was between 0.3 and 1.3 (Fig. S3B). The additional simulations by assuming a current injection to Gocs as well as grcs showed that there was a finite range of the strength of a current injected to Gocs, in which the POT-representing states or synchronized oscillation states appeared vigorously (Figs. S3A and B). These results are consistent with Buonomano and Mauk [32] for POT-representing states and Maex and De Schutter [7] for synchronized oscillation states. Therefore, the dynamical features of the present model are validated for weak MF inputs to Gocs.
Recently, Dugué et al. [33] have reported the existence of gap junctions between nearby Gocs and suggested that the oscillatory synchronization of Gocs can be induced from the electrical coupling. They also carried out computer simulation assuming the bath application of kainate to model Gocs and demonstrated that the Gocs mutually connected by gap junctions elicited spikes synchronously at 12 Hz. Their network model was composed of model Gocs alone. It is unclear whether the Goc synchronization is preserved when connections with grcs that receive spontaneous spike input through MFs are taken into account.
The importance of the voltage dependence of the NMDA channels was confirmed by simulations under various concentrations of Mg2+, which controls the opening of NMDA channels. We observed that the POT-representing states were disrupted by the blockade of NMDA channels increasing Mg2+ concentration, whereas the synchronized oscillation states were destabilized by the persistent opening of NMDA channels at a low Mg2+ concentration. Therefore, the NMDA channels are likely to be involved in the emergence of both the POT-representing states and the synchronized oscillation states. When the network was driven by the small current injection to grcs, simulating spontaneous MF spikes fed to grcs, the activities of the grcs were relatively low. Consequently, the excitatory input signals to Gocs were weak and Goc activities were also relatively low. This made the NMDA channels on the Gocs to open less effectively due to their voltage dependence. Thus, the depolarization of Gocs was largely driven by an AMPAR-mediated synaptic current, which resulted in the network behavior similar to that of Maex and De Schutter [7]'s model. In contrast, when a large current was injected to grcs, simulating strong MF signal input, the NMDA channels on the Goc dendrites opened persistently because of vigorous excitatory input signals from the grcs. This led to the random burst-silent alternation that can represent a POT, consistent with Yamazaki and Tanaka [3]'s model. Although the NMDA channels continually opened due to low Mg2+ concentration, synchronized oscillation was also observed (Figs. 6, 8 and 9). Synchronization in grc and Goc activities may appear at delicate balance among randomness in connectivity, the strength of a current injected to grcs, and the voltage dependence of NMDA channels. This problem should be one of the targets of future studies.
When we assumed the existence of NMDA channels only on the Goc somas, the random burst-silent alternation was not found at grcs even in response to the large current injection (data not shown). Specifically, the grcs did not exhibit persistent silent periods. This was because AHP following spike generation at Gocs quickly closed once-opened NMDA channels on the somas and the Gocs did not elicit burst spikes to produce the persistent silent periods of grcs. To open the NMDA channels for a longer time, NMDA channels should be located on the Goc dendrites somewhat electrically isolated from the somas, as assumed in the present model.
Gocs are regarded as playing an essential role in temporal information processing by the cerebellar cortex [34]. The present model specifically emphasizes the importance of voltage-dependent NMDA channels on the Goc dendrites. This leads to a prediction that blocking NMDA channels only on Gocs impairs delay eyelid conditioning due to the disruption of POT representation. This prediction may be examined experimentally. NMDA channels on grcs consist of NR1, NR2A and NR2C subunits [35], [36], whereas NMDA channels on Gocs consist of NR1, NR2B and NR2D subunits [36]. It is quite likely that NR2B subunits, which are exclusively expressed on Gocs, are responsible for the NMDA current of a long time constant [17], [37] that is needed for the temporal integration of grc activities. Therefore, one may be able to observe the impairment of delay eyelid conditioning by the bath application of selective NR2B antagonists [17], [37].
|
10.1371/journal.pntd.0006915 | Molecular analysis of clinical Burkholderia pseudomallei isolates from southwestern coastal region of India, using multi-locus sequence typing | The Gram-negative soil dwelling bacterium Burkholderia pseudomallei is the etiological agent of melioidosis. The disease is endemic in most parts of Southeast Asia and northern Australia. Over last few years, there has been an increase in number of melioidosis cases from India; however the disease epidemiology is less clearly understood. Multi-locus sequence typing (MLST) is a powerful genotypic method used to characterize the genetic diversity of B. Pseudomallei both within and across the geographic regions.
In this study, MLST analysis was performed on 64 B. pseudomallei clinical isolates. These isolates were obtained between 2008–2014 from southwestern coastal region of India. Broad population patterns of Indian B. pseudomallei isolates in context with isolates of Southeast Asia or global collection was determined using in silico phylogenetic tools.
A total of 32 Sequence types (STs) were reported among these isolates of which 17 STs (53%) were found to be novel. ST1368 was found as group founder and the most predominant genotype (n = 11, 17%). Most of the B. pseudomallei isolates reported in this study (or other Indian isolates available in MLST database) clustered in one major group suggesting clonality in Indian isolates; however, there were a few outliers. When analyzed by measure of genetic differentiation (FST) and other phylogenetic tools (e.g. PHYLOViZ), Indian STs were found closer to Southeast Asian isolates than Australian isolates. The phylogenetic analysis further revealed that within Asian clade, Indian isolates grouped more closely with isolates from Sri Lanka, Vietnam, Bangladesh and Thailand.
Overall, the results of this study suggest that the Indian B. pseudomallei isolates are closely related with lesser heterogeneity among them and cluster in one major group suggesting clonality of the isolates. However, it appears that there are a few outliers which are distantly related to the majority of Indian STs. Phylogenetic analysis suggest that Indian isolates are closely related to isolates from Southeast Asia, particularly from South Asia.
| Burkholderia pseudomallei, a gram negative bacterium, is the causative agent of melioidosis. B. pseudomallei is a soil saprophyte and causes infections in humans by percutaneous inoculation, inhalation or ingestion. Melioidosis is a life threatening disease, which requires prolonged antibiotic treatment and is classically characterized by pneumonia, septicemia and multiple abscesses. Melioidosis is widely prevalent in Southeast Asia and northern Australia. Of late it has been reported from tropical, subtropical and temperate regions. The predicted annual global burden of melioidosis is 165,000 cases. B. pseudomallei has been classified as a Category B threat agent by US Center for Disease Control. Melioidosis is an emerging disease in India that affects many regions. Over the past few years, there has been an increase in number of melioidosis cases, mainly from southwestern costal part of India. This study provides new insights into molecular epidemiology of melioidosis in India. By use of multi locus sequence typing (MLST), we show that Indian isolates are closely related and cluster in one major group suggesting clonality of the isolates. We further show that Indian isolates are more closely related to isolates from Asian countries particularly from South Asia.
| Melioidosis, caused by soil saprophyte Gram-negative bacterium Burkholderia pseudomallei, is classically characterized by pneumonia, septicemia and multiple abscesses. There are various predisposing factors for melioidosis with diabetes mellitus as one of the most important factors. Until recently, the disease was considered endemic only to Southeast Asia and northern Australia, however now it has been reported from tropical, subtropical and temperate regions [1,2]. Reports of melioidosis from India had been few and sporadic [3,4], however, over the past few years, there has been an increase in number of melioidosis cases. It has been reported from various states of India including Karnataka, Kerala, Maharashtra, Tamil Nadu and Puducherry [5]. The disease is quite prevalent in southwestern costal part of the country and is strongly associated with rainfall and diabetes mellitus [5,6]. A mathematical modeling study of 2015 has predicted global annual burden of melioidosis to be 165,000 cases with 89,000 deaths with Indian subcontinent to have highest burden of the disease [7].
Regional variations in melioidosis signs and symptoms have been reported. It has been seen that while prostatic abscess and encephalomyelitis are common in Australians whereas parotid abscess and hepatosplenic suppuration are most frequently seen in patients from Thailand [8]. The reasons for this diversity remain unclear but it could be due to host, bacterial and environmental factors [9]. Study of epidemiology by molecular methods provides insight about bacterial diversity and distribution. Various molecular methods e.g. pulse field electrophoresis, ribotyping have been used for phylogenetic reconstruction with different levels of success, however, multi-locus sequence typing (MLST) is a proven tool for the molecular typing of B. pseudomallei. MLST not only helps to explore the sequence types (STs) in a population or helps tracing an outbreak [10] but also helps to understand the microbial evolution. There is large MLST database for B. pseudomallei (http://pubmlst.org/bpseudomallei/).
In India, melioidosis is an emerging endemic infection and potentially fatal as early diagnosis is missed due to its varied manifestations such as localized or disseminated infection. Here, it has largely been reported from the coastal regions and it is generally believed that the disease is underreported or misdiagnosed. Using MLST, we had previously reported in a pilot study that Indian isolates were genetically diverse from the Australian or Southeast Asian isolates [11]. In this work, we aimed to study the genetic diversity among larger number of B. pseudomallei isolates using MLST. Further, using sequence data we attempted to find broad population patterns of Indian isolates with global collection of 5541 isolates and construct phylogeny of B. pseudomallei Indian isolates and their close relatives from Southeast Asia.
This study was approved by Institutional Ethical Committee of Kasturba Hospital, Manipal under protocol number IEC 141/2011.
Sixty-four (n = 64) clinical isolates of B. pseudomallei isolated from Karnataka and adjacent states (Kerala, Goa and Puducherry) of southern India were included in this study. These isolates were collected from melioidosis patients at Kasturba Medical College, Manipal, Karnataka during 2008–2014. All the isolates were identified using API20NE and later confirmed using TTS1-PCR assay [12] and latex agglutination test from the colonies. The isolates were also characterized for genetic markers linked to geographic origin Yersinia-like fimbriae (YLF) and B. thailendensis-like flagellum and chemotaxis (BTFC) gene [13].
For preparation of DNA, bacteria were grown overnight in Luria- Bertani (LB) broth at 37°C in high containment facility, a biosafety level 3 facility at Defence Research & Development Establishment (DRDE), Gwalior. Genomic DNA was isolated from culture using DNeasy blood and tissue genomic DNA kit (Qiagen Gmbh, Hilden), according to the manufacturer’s instructions. The genomic DNAs were stored at -20°C till further used.
MLST was carried out according to the method of Godoy et al [14]. Primer sequences targeting the conserved regions of seven housekeeping genes (ace- gltB- gmhD- lepA- lipA- narK- ndh) of B. pseudomallei were used as shown in MLST site (http://pubmlst.org/bpseudomallei/). PCR amplification, sequence analysis and determination of ST for each isolate were carried out by our earlier reported procedure [11]. The purified PCR amplicons were double pass sequenced using commercial SANGER sequencing services (M/s Genotypic Technology Pvt. Ltd., Bengaluru).
The relatedness of MLST profiles of isolates of this study or Indian isolates in B. pseudomallei MLST database was performed using eBURST software [15] (launched at http://eburst.mlst.net/) with single locus variant (SLV) selected. eBURST or global optimal based upon related sequence type (goeBURST) allows for an unrooted tree-based representation of the relationship of the analyzed isolates. The diversification of the "founding" genotype is reflected in the appearance of STs differing only in one housekeeping gene sequence from the founder genotype–single locus variants (SLVs). Further diversification of those SLVs results in the appearance of variations of the original genotype with more than one difference in the allelic profile: double locus variants (DLVs), triple locus variants (TLVs) and so on. The final eBURST tree provides a hypothetical pattern of descent for the strains analyzed. Basic quantities such as number of alleles, number of variable sites per allele, number and frequency of single nucleotide polymorphism (SNPs) in each locus were determined. Nucleotide sequence diversity (π) was calculated using DNAsp V5.1 software [16]. The association of individual STs with the type of infection among the study population was carried out using Fisher exact test (Graph Pad Prism software, La Jolla, USA).
Measures of genetic differentiation (FST) between the concatenated sequences of Indian origin and from Asia or Australia was estimated using DNAsp V5.1 [16]. Relationship of Indian STs with global collection of STs was studied using goeBURST [17] implemented in PHYLOViZ programme [18] available at MLST site (http://pubmlst.org/bpseudomallei). PHYLOViZ is a flexible and expandable plugin based tool that is able to handle large datasets and builds on goeBURST implementation. In order to further determine the relationship of Indian STs with STs from Asia and Australia, topology and grouping of all Indian STs were displayed on constructed boot strapped phylogenetic tree using Unweighted Pair Group Method with Arithmetic average (UPGMA) method in molecular evolutionary genetic analysis version -6 (MEGA 6) software [19]. Indian STs including STs of this study were analyzed with selected 57 STs from other Asian countries e.g. Sri Lanka, Vietnam, Bangladesh, Cambodia, Malaysia, Thailand, China and Laos and, 6 most predominant STs from Australia. The 57 STs were either SLV or DLV of the Indian STs. Concatenated sequences of all STs used for this study were downloaded from MLST site (http://pubmlst.org/bpseudomallei/).
The present study involved MLST analysis of 64 clinical isolates of B. pseudomallei obtained during the time span of seven years (2008–2014) from India. These isolates were from states of Karnataka (84%), Kerala (6.25%), Goa (6.25%) and Puducherry (1.5%). Among the 64 isolates, 31 (48.4%) were obtained from patients with bacteremic melioidosis and 27 (42.2%) from patients with localized form of melioidosis. All B. pseudomallei isolates belonged to the Asian biogeographic YLF variant. Isolate study based on the clinical condition, geographical location, year of isolation, YLF/BTFC genotype and sequence type are enlisted in S1 Table.
Among obtained STs,—the numbers of alleles per locus varied from 2 to 4 and SNPs ranged from 1 to 7. The level of locus sequence diversity for each of the 7 loci was found to be around 2% (Table 1). Nucleotide diversity (π) among Indian isolates of this study was found to be 0.00137 and changed to 0.00329 for total Indian isolates in the database.
Among 64 isolates of this study, a total of 32 STs were identified. The frequencies of STs among isolates ranged from 1–11 with ST1368 (n = 11), ST42, ST1373, ST1478 (n = 4), ST124, ST1512, ST1507, ST1375 (n = 3) being the most predominant STs. Seventeen STs identified in this study (ST1373, ST1374, ST1506, ST1507, ST1508, ST1509, ST1510, ST1511, ST1512, ST1513, ST1514, ST1515, ST1516, ST1517, ST1518, ST1519, ST1520) were found to be novel that were not reported previously and 15 STs were previously documented, five of which were reported as novel in our previous study [11].
When analyzed by eBURST, almost all STs of this study formed in one large group with ST1368 as predicted founder; only ST1141, ST1374, ST1506 and ST859 were found to be singleton. ST1368 had 7 SLV, 4 DLV, 7 TLV and 9 satellites STs. ST550, ST124, ST42, ST1511, ST1512, and ST1518 were found to be subgroup founders. Interestingly, all novel 17 STs of this study except ST1506 and ST1374 also clustered in the same group (Fig 1, S2 Table).
A total of 130 isolates are now available (as on 16th October, 2018) in B. pseudomallei MLST database from India which include data from 64 isolates of this study. All isolates from India except one, which was isolated from soil, were clinical isolates. Of the 65 STs, 47 STs (109 isolates) cluster in one group with the remaining 18 STs (21 isolates) being singleton. ST1368 is dispersed all over the four Indian states of Karnataka, Kerala, Goa and Puducherry (S1 Table), [11]. It had a frequency of 26 with 7 SLV, 10 DLV, 11 TLV and 18 satellites STs. The present study has expanded the clonal cluster of Indian isolates by adding more branching STs. eBURST analysis of total Indian isolates revealed ST1513 and ST1372 to be as additional sub-group founders (Fig 2).
Measure of genetic differentiation (FST) between Indian and Australian or Asian isolates was determined and was found to be 0.1561 and 0.09082 respectively, suggesting that Indian isolates are closer to Asian clade rather than Australian clade. When Indian STs were analyzed by PHYLOViZ with the global collection of 5541 isolates in the B. pseudomallei MLST isolates database (as on 19th May, 2018), majority of Indian isolates grouped in four groups (Fig 3). Group A (ST42, ST550, ST1051, ST1370, ST1555) and B (ST405, ST856, ST1637) clustered with Asian clade and accounted for about 43% (28/65) of the known STs. Indian isolates appeared to be different from the main Thailand cluster and grouped closure to isolates from South Asia e.g. Sri Lanka, Bangladesh. There was overlap of Indian STs with STs of other Asian countries e.g. Bangladesh (ST42, ST43, ST300); Vietnam (ST550, ST858, ST1051); Sri Lanka (ST293, ST1141); Thailand (ST300, ST405) and China (ST405).
About 57% of Indian STs grouped with Australian STs in two groups (Group C–ST124, ST1375, ST1507, ST1512, ST1552, ST1554 and Group D—ST1368, ST1372, ST1373). ST468 and ST1051 were the only Indian STs that overlapped with Australia. However, all isolates of this study were YLF positive, a gene cluster found predominantly among isolates of Southeast Asian origin. Some of the Indian STs also had overlapping STs with USA; these included ST1427, ST550 and ST960. B. pseudomallei has never been found in the US environment, thus all of these USA patient melioidosis infections would have been acquired overseas.
We also wanted to find out whether predominant Indian STs had SLVs with STs of other countries. SLV of ST1368, ST1373 (ST293), ST1478 (ST1147, ST1137), ST1507 (ST1138, ST1152) and ST1512 (ST1152) were reported from Sri Lanka. A few Indian STs had SLVs with STs from Thailand (ST42-ST405, ST501, ST371), Bangladesh (ST42-ST43) and Vietnam (ST1373- ST1568). Only a few predominant STs of other countries had SLVs with Indian STs which included STs from Sri Lanka (ST1132-ST1514, ST1517, ST1630), Thailand (ST371-ST1514) and Vietnam (ST41- ST1051).
The topology and grouping of all STs from India was displayed with selected STs from Asia and Australia and results are shown in Fig 4. This type of cluster analysis presents several advantages, such as the ease of interpretation and the creation of an hierarchical grouping of the isolates that can provide a global overview of the relatedness of the isolates under study and how the defined clusters are connected to each other. It was found that majority of the Indian isolates grouped in ‘Group 1’along with STs from Sri Lanka, Vietnam, Bangladesh, Cambodia, Malaysia, Thailand, China and Laos. ‘Group 4’ was another group in which STs from India and other Asian countries clustered together. Both of these groups had one ST each from Australia. ‘Groups 2, 6 & 7’ which were relatively smaller groups, had STs only from India and Australia. ‘Groups 8, 9 & 10’ had STs only from India, which probably represent some distantly related STs from main Indian cluster.
Attempts were also made to find association between the STs and the clinical conditions. In this study, ST1368 was found to be predominantly associated with localized infections (8 of 27), however this association was not statistically significant (P = 0.913). We also did not observe any association of STs with geographical location and year of isolation.
MLST is an unambiguous and powerful procedure to study the bacterial populations and global epidemiology [14]. It has been extensively utilized to characterize B. pseudomallei and analysis of MLST data by different phylogenetic tools such as eBURST, goeBURST and PHYLOViZ has succeeded in making the genetic relatedness of B. pseudomallei from distinct geographical locations. The present study was undertaken to characterize 64 clinical isolates of B. pseudomallei by MLST from various geographical locations in the southwestern coastal region of India. Here, we wanted to understand the diversity of genotypes of India and their relatedness to STs from Asia or Australia.
In this study, 32 STs were identified and overall diversity of isolates was 0.5 STs/isolate. The low level of locus sequence diversity and nucleotide diversity (π) for STs indicates low diversity among Indian isolates. When analysed by eBURST, an algorithm which provides a hypothetical pattern of descent for the analyzed strains, most of the isolates of this study (or all Indian isolates available in MLST database) formed one clonal complex suggesting limited genotypic diversity. One of a predominant ST (ST1630) reported as singleton in a recent study [20], which was isolated from southern states of Tamil Nadu and Andhra Pradesh, also appeared to be part of this clonal complex. Only 18 Indian STs were found to be singleton and were outliers. These singleton STs also included ST300 which was reported to be group founder in above study [20]. Topology and grouping of all Indian STs with other Asian and Australian STs by MEGA showed that majority of Indian isolates grouped together in one large group, however few STs were distant and unique particularly ST1636. ST1368 was found to be the predominant ST representing about 17% of the total isolates. Predominance of particular genotype among single population communities have previously been reported from various endemic areas of Malaysia [21], Sri Lanka [22], Thailand, and Australia (http://pubmlst.org/bpseudomallei/).Occurrence of ST1368 over a period of time, which appeared from an isolate of 2006 [11] and in all 4 states of India suggests high level of genetic uniformity among B. pseudomallei isolates from southwestern region of India.The results of this study suggest that the Indian isolates are closely related with lesser heterogeneity among them and belong to single group. However, it appears that there are a few outliers which are distantly related to the majority of Indian STs.
Attempts were made to find out the association between common STs (e.g. ST1368, ST42, ST124) of this study and disease presentation, however like our previous work [11], no such association could be observed. One of the reasons for this lack of association could be small sample size. Preliminary results of this study could only be strengthened if larger STs data is available and correlated with disease presentation. Except a few studies [22], most studies, however, could not find any association between STs and disease presentation.
Measure of genetic differentiation (FST), presence of YLF genetic marker in B. pseudomallei isolates, PHYLOVIZ analysis and phylogenetic clustering by MEGA suggested Indian isolates to be closer to Asian clade rather than Australian clade. PHYLOViZ analysis further revealed that within Asian clade, Indian isolates grouped were more closely with isolates from Sri Lanka, Vietnam, Bangladesh and Thailand. There were many STs which overlapped with these countries. Some of the overlapping STs (ST300, ST405) were first isolated in Thailand in 1965, while other STs (e.g. ST42, ST293, ST550) were isolated in Bangladesh, Sri Lanka and Vietnam only after late 1990s suggesting continued migration of bacteria among India and Southeast Asian countries. These results are however, contrary to our previous findings[11], which suggested Indian strains to be distinct from Southeast Asian or Australasian isolates. This may be due to better analysis of available (and larger) data by multiple and refined (e.g. goeBURST) phylogenetic tools, which provided better linkage among STs. Recently, association of Indian B. pseudomallei isolates with Southeast Asian isolates have been reported by others as well [20].
MLST has been widely applied to determine the clonal relationship among strains of B. pseudomallei and other various clinically relevant bacterial species. The online MLST databases facilitated the sharing and analysis of MLST data. However, MLST is based on the sequencing of only seven (housekeeping) genes, therefore, has low resolution. MLST has added limitations in phylogeographic analysis of B. pseudomallei, because this bacterium has high rates of recombination and instances of homoplasy. Therefore, it is possible that the isolates may share the same ST despite being distinct genetically. Many Indian STs including a few predominant Indian STs like ST1368 and ST1373 were found to cluster with Australian isolates. Grouping of Asian isolates with isolates of Australian origin by eBURST analysis has previously been shown and two STs shared between Australia and Cambodia were found to be genetically unrelated on the whole-genome level [23]. Whole genome sequencing (WGS) based SNP typing provides higher resolution and can be more robust and accurate to identify the origin of strains. WGS based methods are being used to study the phylogeny and evolution of bacteria in the newer studies [24,25]. It would be interesting to find out in future how WGS can be used to elucidate the clustering of Indian isolates with Asian (or Australian) isolates and also to find out (or rule out) the clustering of those Indian STs that overlapped/grouped with Australia.
Overall, results of this study suggest that Indian B. pseudomallei isolates are closely related with lesser heterogeneity among them and belong to single group. However, it appears that there are a few outliers which are distantly related to the majority of Indian STs. The results further suggest that Indian isolates are closely related to B. pseudomallei isolates from Southeast Asia, particularly from South Asia, and there seems to be migration of bacteria between India and other Southeast Asian countries. However, due to low resolution of MLST, future studies should focus on WGS based SNP typing to find out (or rule out) the clustering of Indian B. pseudomallei with Asian or Australian isolates.
|
10.1371/journal.ppat.1006973 | Limited immune surveillance in lymphoid tissue by cytolytic CD4+ T cells during health and HIV disease | CD4+ T cells subsets have a wide range of important helper and regulatory functions in the immune system. Several studies have specifically suggested that circulating effector CD4+ T cells may play a direct role in control of HIV replication through cytolytic activity or autocrine β-chemokine production. However, it remains unclear whether effector CD4+ T cells expressing cytolytic molecules and β-chemokines are present within lymph nodes (LNs), a major site of HIV replication. Here, we report that expression of β-chemokines and cytolytic molecules are enriched within a CD4+ T cell population with high levels of the T-box transcription factors T-bet and eomesodermin (Eomes). This effector population is predominately found in peripheral blood and is limited in LNs regardless of HIV infection or treatment status. As a result, CD4+ T cells generally lack effector functions in LNs, including cytolytic capacity and IFNγ and β-chemokine expression, even in HIV elite controllers and during acute/early HIV infection. While we do find the presence of degranulating CD4+ T cells in LNs, these cells do not bear functional or transcriptional effector T cell properties and are inherently poor to form stable immunological synapses compared to their peripheral blood counterparts. We demonstrate that CD4+ T cell cytolytic function, phenotype, and programming in the peripheral blood is dissociated from those characteristics found in lymphoid tissues. Together, these data challenge our current models based on blood and suggest spatially and temporally dissociated mechanisms of viral control in lymphoid tissues.
| CD4+ T cells have classically been divided into different subsets based on their different abilities to help and regulate specific parts of the immune system. Recent work in the HIV field has demonstrated that HIV-specific CD4+ T cells with unique effector functions, such as cytolytic activity and β-chemokine production, can play a direct role in control of HIV replication. However, HIV infection is generally considered to be a disease centered in lymphoid tissues, where unique CD4+ T helper cell subsets are present to orchestrate the maturation and priming of adaptive immunity. In this study, we identify that two specific transcription factors, T-bet and Eomes, mark cytolytic and β-chemokine producing CD4+ T cells. While this effector CD4+ T cell population is part of immunosurveillance mechanisms in blood, we find that lymph nodes largely lack this effector population–independent of HIV infection or disease progression status. These results indicate that current effector CD4+ T cell mediated correlates of HIV control are limited to blood and not representative of potential correlates of control in lymphoid tissues.
| CD4+ T cells are classically known to orchestrate immunity by providing helper functions to other arms of the immune system. However, CD4+ T cells can also exercise direct cell-to-cell mediated effector functions to control pathogens and tumors. Effector CD4+ T cells with cytolytic activity are generated during many acute viral infections and represent a front-line defense in the gut-intraepithelial compartment [1]. Cytolytic CD4+ T cells directly recognize tumor cells and are involved in host protection against chronic viral infections, such as EBV and CMV [2]. As HIV has evolved numerous ways to escape recognition by CD8+ T cells and neutralizing antibodies, it remains important to understand what role effector CD4+ T cells play to control HIV replication and limit disease progression.
Several studies have demonstrated broad and vigorous responses of peripheral blood HIV-specific CD4+ T cells in untreated individuals with low HIV viremia [3–8]. A growing body of evidence suggests that increased cytolytic and non-cytolytic mechanisms mediated by highly differentiated CD4+ T cells are linked to better HIV control. The emergence of HIV-specific CD4+ T cells with cytolytic properties during early infection has been associated with slower subsequent disease progression [7, 9]. Furthermore, HIV and SIV elite controllers demonstrate strong Nef- and Gag-specific CD4+ T cell responses in vivo, that can suppress viral replication in vitro in both macrophages and CD4+ T cells, potentially through cytolytic activity [9–13]. Late differentiated CD4+ T cells can also mediate non-cytolytic functional mechanisms to limit CCR5-tropic HIV infection via autocrine production of β-chemokines CCL3 (MIP-1α), CCL4 (MIP-1β) and CCL5 (RANTES), through either blocking the interaction between gp120 and CCR5 or downregulating CCR5 from the cell surface [14]. In this regard, CMV-specific CD4+ T cells are particularly known to be efficient producers of β-chemokines and are notably preserved in late HIV infection [15–17]. Moreover, elite controllers possess CD4+ T cells resistant to CCR5-tropic HIV, potentially through the impact of increased MIP-1α and MIP-1β production by these cells [4, 18].
The T-box transcription factors T-bet and eomesodermin (Eomes) regulate effector T cell differentiation and have been closely associated with the programming of effector functions of CD8+ T cells (reviewed in [19]) and cytolytic CD4+ T cells [11, 20]. These transcription factors also have a role in driving CD4+ T cell polarization, where T-bet is particularly known as the classical Th1 lineage defining transcription factor. Interestingly, T-bet directly regulates many of the genes encoding CD4+ T cell functions associated with HIV control, including prf1, gzmb, ccl3, ccl4, and ccl5 are directly regulated by T-bet [21]. Eomes can compensate for the loss of T-bet to retain effector functions such as IFNγ production [22] and as such can function, to some extent, as a “paralog” to T-bet. We have previously shown that cytolytic CD8+ T cells have high levels of T-bet and intermediate expression of Eomes in peripheral blood [23]; however, it remains less clear if production of cytolytic granules or β-chemokine by human CD4+ T cells is similarly coupled to the expression levels of T-bet and Eomes at the single-cell level.
While considerable evidence suggests that cytolytic and autocrine β-chemokine producing HIV-specific CD4+ T cells might be involved in control of HIV infection, it remains unclear whether these cells can mediate antiviral responses within lymphoid tissues, the primary site of HIV/SIV replication [24]. Here, we sought to directly evaluate the role of effector CD4+ T cell-mediated control of viral replication in HIV-infected lymph nodes (LNs) over the course of HIV infection. We identify a T-bethiEomes+ population that almost exclusively marks cytolytic and β-chemokine producing CD4+ T cells in blood of healthy and HIV-infected subjects. However, this population of CD4+ T cells is nearly absent in lymphoid tissues of chronically infected HIV-infected subjects, suggesting that cytolytic HIV-specific CD4+ T cells are not a major component of long term viral control in lymphoid tissues.
We have previously shown that cytolytic molecule (perforin and Granzyme B) expression in peripheral blood CD8+ T cells is strongly associated with high expression levels of T-bet (T-bethi) and intermediate expression of Eomes (Eomes+) [23]. Independent of HIV infection status, we found that high levels of T-bet and Eomes expression in CD4+ T cells were also strongly linked to perforin and Granzyme B expression (Fig 1A and S1 Fig). Expression of perforin was highly enriched within the T-bethiEomes+ CD4+ T cell population (Fig 1B). Furthermore, the frequency of T-bethi CD4+ T cells correlated strongly with the frequency of perforin+ CD4+ T cells (Fig 1C). Using Imagestream analysis, we found, similar to CD8+ T cells [25], that T-bet was primarily localized in the nucleus in T-bethi CD4+ T cells, while T-betdim cells had T-bet more localized in the cytoplasm (Fig 1D), suggesting that T-bet may be transcriptionally active in T-bethi CD4+ T cells. A previous murine study on gut intra-epithelial CD4+ T cells demonstrated that cytolytic CD4+ T cells can have a ThPokloRunx3hi phenotype [26]. We found no association between low levels of ThPok expression and the frequency of Granzyme B+perforin+ CD4+ T cells in blood (S1 Fig). Furthermore, not all Granzyme B+perforin+ CD4+ T cells had a Runx3hi phenotype (S1 Fig).
In order to generate a spatial CD4+ T cell differentiation map, we next used non-linear dimensional reduction t-SNE analysis by embedding multi-parametric single cell information from blood CD4+ T cells. The t-SNE analysis generated a plot where the cells with high CCR7 expression were located on one (top) side of the scale, and T-bet and Eomes expressing cells were on the other (bottom) side of the non-linear dimensional space along with Granzyme A, Granzyme B and perforin-expressing cells (Fig 1E). Furthermore, single Eomes+ CD4+ T cells were more early-differentiated (S2 Fig), whereas T-bethiEomes+, as well as single expressing T-bethi, cells were enriched in more terminally-differentiated CD4+ T cells (S2 Fig). Together, these data demonstrate that particularly high levels of T-bet marks cytolytic CD4+ T cells.
Autocrine production of β-chemokines has been described to protect CD4+ T cells from CCR5-tropic HIV infection [15, 16]. Terminally-differentiated CD4+ T cells have generally been known to produce β-chemokines; however, whether T-bethiEomes+ CD4+ T cells are specifically the subset capable of producing β-chemokines remains unclear. MIP-1α and MIP-1β production were highly co-expressed in CD4+ T cells (S3A Fig), and almost exclusively produced by T-bethiEomes+ CD4+ T cells irrespective of polyclonal (SEB) or antigen (CMV-pp65) stimulation (S3B Fig). HIV-specific CD4+ T cells produced low levels of β-chemokines in most subjects and were therefore not analyzed further [15]. Stimulation with αCD3-CD28 similarly showed that primarily the T-bethiEomes+ cells had the potential to produce MIP-1α (Fig 1F), and individuals with no T-bethiEomes+ population showed low production of MIP-1α (S3C Fig). Likewise, the frequency of T-bethiEomes+ cells correlated strongly with the aCD3-CD28-activated CD4+ T cells producing MIP-1α (Fig 1F). Both T-betdim and T-bethi cells expressed IFNγ after SEB and CMV stimulations (S3B and S3C Fig). Although T-betdim cells can produce IFNγ, individuals that lack T-bethi cells had poor overall IFNγ responses after αCD3-CD28 stimulations (S3C Fig). Together these data demonstrate that T-bethi cells have the highest capacity to express cytolytic molecules, β-chemokines and IFNγ and thus represent the major effector CD4+ T cell population in blood.
The presence of effector CD4+ T cells during acute and chronic HIV infection has been associated with slower disease progression [9]. We first determined the direct impact of HIV-1 infection in vivo on the frequency of T-bet expressing CD4+ T cells in blood. HIV seronegative individuals (n = 10) that subsequently seroconverted were followed from before infection and longitudinally up to almost 800 days after the first positive HIV test using the RV217 cohort [27]. All subjects were untreated during this period of time. Between the initial visit prior to infection and >1-year post-infection, the median frequency of the T-bethi CD4+ T cell subset increased by almost 100% (Fig 2A). Further analysis from before infection and longitudinally during the acute phase of HIV-1 infection revealed that the frequency of T-bethi CD4+ T cells sharply increased at the first sampled time-points after infection, transiently decreased after peak viremia, and subsequently slowly increased again (Fig 2B). The expression pattern of both T-bet and Eomes was further evaluated in a larger cross-sectional European cohort of HIV-seronegative and -positive subjects, where we found that the median frequency of T-bethiEomes+ cells in memory CD4+ T cells was 6.7 times higher in HIV-infected chronic progressors compared to healthy subjects (Fig 2C). Furthermore, T-bethiEomes+ cells were significantly less frequently infected in vitro than other conventional memory (CD45RO+) CD4+ T cells (S4 Fig), supporting previous studies showing that terminally-differentiated and β-chemokine producing CD4+ T cells contain less HIV-DNA molecules than more early-differentiated CD4+ T cell subsets [15, 28]. We further conducted longitudinal assessments after ART was initiated and found that the frequency of T-bethiEomes+ cells decreased after ART. However, this drop was associated with an accumulation of naïve cells following ART initiation (S5 Fig), suggesting that the change in frequency is a consequence of naïve cells redistributing in blood after ART. Indeed, other data demonstrated that the absolute counts of T-bethiEomes+ cells did not change longitudinally after ART (S5 Fig), indicating that HIV replication and rebound have a low impact on the effector CD4+ T cell population in blood. Altogether, these data suggest that T-bethi CD4+ T cell levels are preserved in blood during HIV infection.
We next used the RV217 acute cohort to monitor the frequency of T-bet and effector characteristics of HIV-Gag-specific CD4+ T cells from before infection and with close intervals during acute infection. Early after infection, we found a sharp increase in HIV-specific IFNγ+ CD4+ T cell responses that later declined (S6 Fig). The very early response at peak viremia was associated with an effector HIV-specific CD4+ T cell response with elevated levels of T-bet (Fig 2D), perforin (Fig 2E) and MIP-1α (Fig 2F) production. Following peak viremia, however, T-bet expression decreased rapidly, concordant with a subsequent decline of perforin and MIP-1α production by HIV-specific CD4+ T cells (Fig 2D–2F) demonstrating a close temporal relationship between effector CD4+ T cell responses and T-bet expression following HIV infection.
While peripheral blood CD4+ T cells with an effector profile are preserved in chronic HIV infection (Fig 2), it remains uncertain if similar cells are present in lymphoid tissues, such as lymph nodes (LNs), where they could interact with HIV-infected target cells. To determine this, we first assessed the maturation status (CCR7, CD27, CD45RO) of bulk CD4+ T cells in blood and LNs (S7A Fig) from HIV-infected and–uninfected individuals. We found that HIV chronic progressors (CP) and ART-treated (ART+) subjects particularly had elevated levels of transitional memory (TM) cells in LNs (Fig 3A), which correlated with the expansion of T follicular helper cells (Tfh) (S7B Fig) as previously described for HIV-infected subjects [29, 30]. In contrast, HIV-seronegative subjects had higher levels of TM cells in blood (Fig 3A). Independently of HIV-infection status, both effector memory (EM) cells and terminally-differentiated (TD) cells were significantly reduced in LNs compared to blood (Fig 3A).
Given the decreased levels of terminally-differentiated CD4+ T cells in LNs, we next assessed if this was further associated with fewer T-bethiEomes+ effector CD4+ T cells in lymphoid tissues compared to blood. Accordingly, we found very few T-bethiEomes+ CD4+ T cells in LNs for all groups (Fig 3B). The few Eomes+ cells present in LNs had greatly reduced T-bet expression in contrast to blood-derived CD4+ T cells (Fig 3B). We generally found significantly lower frequencies of perforin+ and Granzyme B+ CD4+ T cells in LNs compared to peripheral blood independent of HIV-infection status (Fig 3C). Perforin expression was consistently lower for both LN and blood CD4+ T cells compared to matched LN and blood CD8+ T cells (S7C Fig). Furthermore, Granzyme B showed low co-expression with perforin and these cells were, instead, skewed towards a CD27+ profile in LNs (Fig 3D). We next employed tSNE analysis by merging single-cell CD4+ T cell data together with matched blood and LN for an HIV+ CP with high levels of Tfh cells in LNs and cytolytic CD4+ T cells in blood. Based on a set of memory and cytolytic markers, the tSNE analysis confirmed a dissociation between effector and Tfh cells in the blood and LN. (Fig 3E). Independent of HIV-infection status, these data together demonstrate an inherent lack of the blood-like T-bethiEomes+ CD4+ T cell population, and thereby cytolytic cells, in LNs.
Based on these premises, we next assessed the functional properties of CD4+ T cells following stimulation with overlapping Gag and Env peptide stimulations. Paired blood and LN samples from HIV+ CPs and ART+ subjects were collected for these assessments. We found higher frequencies of Gag-specific CD4+ T cells in LNs compared to blood for both CPs and ART+ subjects (Fig 4A). Additionally, the magnitude of the Env-specific CD4+ T cell response was only higher in LNs compared to blood for HIV CPs (Fig 4A). Although the LN CD4+ T cell responses were higher compared to blood, LN Gag-specific CD4+ T cells tended to be dominated by either CD107α/TNF alone, or in combination, and less commonly produced IFNγ (Fig 4B). In contrast to LN CD4 T cells, blood Gag-specific CD4+ T cells had more polyfunctional responses (Fig 4B). Likewise, SEB stimulated CD4+ T cells exhibited similar characteristics and showed an elevated polyfunctional repertoire in blood compared to LN (S8 Fig).
Next, we examined whether LN CD4+ T cells could upregulate perforin after stimulation as a measure of their cytolytic potential [31, 32]. In general, few LN Gag-specific CD4+ T cells upregulated perforin after peptide stimulation, whereas Gag-specific CD4+ T cells in the blood expressed perforin to a higher degree (Fig 4C). The limitation of LN Gag-specific CD4+ T cells to express perforin correlated with lower levels of T-bethi cells in LN compared to blood (Fig 4D). No difference between LN and blood was found in ART+ subjects for either perforin or T-bethi cells suggesting that cytolytic Gag-specific CD4+ T cell responses are only present in blood during viremic episodes.
We previously observed that β-chemokines are predominantly produced by the T-bethiEomes+ population (Fig 2), suggesting that they may not be expressed by LN CD4+ T cells given the general lack of high T-bet expression. We further explored this phenomenon on SEB- and CMV-specific CD4+ T cells and found significantly lower frequencies of MIP-1α+ SEB stimulated (Fig 4E) and CMV-specific (Fig 4F) CD4+ T cells in LNs compared to blood. Notably, LN CMV-specific CD4+ T cells had poor co-expression of IFNγ and TNF in contrast to their counterparts in blood (S9A Fig), indicating an inherent lack of multiple effector functions also by CMV-specific CD4+ T cells in LNs.
HIV-1 elite controllers usually demonstrate a vigorous and highly polyfunctional effector HIV-specific CD4+ T cell response in peripheral blood [3–6, 8]. We obtained LNs from a cohort of HIV elite controllers (n = 9) and assessed whether those subjects had any evidence of cytolytic CD4+ T cell activity in LNs. Most Gag-specific CD4+ T cells in blood from elite controllers tended to express moderate levels of T-bet but still showed an enrichment in the T-bethiEomes+ subset compared to LN Gag-specific CD4+ T cells (Fig 5A). The majority of LN HIV-specific CD4+ T cells were instead T-betdimEomes- (Fig 5A). In addition, the frequency of perforin+ (Fig 5B) and MIP-1α+ (Fig 5C) Gag-specific CD4+ T cells in HIV elite controllers was significantly higher in blood compared to LN. Similarly, we found very few perforin/Granzyme B+ CD4+ T cells in acute/early (Fiebig IV-VI) HIV seroconverter LNs (Fig 5D). Furthermore, early cycling (Ki-67+) CD4+ T cells, indicative of HIV-specific T cells [33, 34], in blood showed tendencies of higher perforin, Granzyme B and T-bethi expression than LN Ki-67+ CD4+ T cells in the acute/early HIV seroconverters (Fig 5D). Taken together, our data provide evidence of dissociated effector-like HIV-specific CD4+ T cell responses in LN compared to blood suggesting spatial and temporal dissociated mechanisms of viral control in LNs of HIV elite controllers and early HIV seroconverters.
Cytolytic T cell activity requires granule release through a stable immunological synapse [38]. As such, we directly visualized the structure of CD4+ T cell immunological synapses formed by degranulating LN and blood CD4+ T cells in real time by exposing CD4+ T cells to planar lipid bilayers containing fluorescently labelled soluble ICAM-1 and -CD3 antibodies. In order to maximize the vertical resolution at the T cell/bilayer interface, we used total internal reflection fluorescence (TIRF) microscopy to examine the structure of the T cell contact surface and the appearance of CD107a in real time [39–41]. Analysis of LN and blood CD4+ T cells isolated from either chronically HIV-infected or uninfected individuals revealed four different groups of synapse formation: CD4+ T cells capable of establishing mature cytolytic synapses, but do or do not release granules, CD4+ T cells able to release granules without establishing cytolytic synapses, and, finally, CD4+ T cells that neither form the synapses nor release granules (Fig 7A). CD4+ T cells from both blood and LN of chronically infected individuals contained a considerable fraction of cells that are capable of releasing granules without formation of mature synapses. However, the fraction of blood CD4+ T cells demonstrating both synapse formation and granule release was considerably larger than for LN CD4+ T cells in HIV-infected individuals (Fig 7B). Furthermore, the difference between LN and blood CD4+ T cells in the kinetics of granule release, a parameter linked to the efficiency of T cell cytolytic activity [42], was even more striking. Regardless of the infection status, blood-derived CD4+ T cells were able to release granules almost twice as fast as LN-derived CD4+ T cells (Fig 7C and 7D). Not surprisingly, the HIV-specific CD4+ T cell clone AC25 showed superior ability to establish mature immunological synapse and rapid degranulation (Fig 7A–7D). Altogether, these data suggest that while some LN CD4+ T cells are able to degranulate, few have the necessary functional qualities to launch an efficient antiviral effector response at the significant site of HIV replication.
The progressive decline of memory CD4+ T cells is a general hallmark of chronic HIV disease in most individuals. Accumulating evidence suggests that HIV-specific CD4+ T cells play an important role in host defense mechanisms against the virus. Late stage maturation and increased effector functions, including β-chemokine production and cytolytic activity, by peripheral blood CD4+ T cells have been linked to a lower degree of viral susceptibility and slower disease progression in HIV-infected individuals [9–18]. However, whether CD4+ T cells derived from lymphoid tissues from HIV-infected individuals possess such effector functions has heretofore remained unclear. Indeed, lymphoid tissues are known from studies on HIV-uninfected subjects to harbor CD4+ T cells with entirely different plasticity and functional characteristics than CD4+ T cells in blood [43, 44]. In this study, we demonstrate that CD4+ T cells expressing high levels of T-bet and Eomes represent the major producers of classical T cell effector functions, such as cytolytic molecules and β-chemokines. However, this effector CD4+ T cell subset is rare in HIV-infected or -uninfected LNs; a finding similar to that recently described for CD8+ T cells [32]. Importantly, HIV elite controllers also have lower frequencies in vivo effector CD4+ T cell responses in LNs during chronic disease, implicating a spatial or temporal displacement in maintaining control of HIV replication in LNs. Together these results indicate that effector CD4+ T cells are unlikely to play a major direct role in control of HIV replication within lymphoid tissue.
Synergy between the transcription factors T-bet and Eomes has been intensively studied in the context of effector CD8+ T cell differentiation, and studies have suggested that the interplay between T-bet and Eomes in CD4+ T cell differentiation drives the cytolytic CD4 T cell program [11]. T-bet was originally defined as the master regulator of CD4+ T cell Th1 polarization and expression of effector cytokines such as IFNγ and TNF [21]. Eomes is also expressed in CD4+ T cells, where it can compensate for loss of T-bet to retain IFNγ production and thereby Th1 polarization [22]. Recent studies have implicated that Eomes is particularly important for driving cytolytic CD4+ T cell responses in vivo [45]. In addition, other transcription factors, such as ThPok, have been associated with cytolytic CD4+ T cell activity in mice [26]. Similar to our previous observations in CD8+ T cells [23, 46], we found a very strong correlation between high expression levels of T-bet and perforin within Eomes+ peripheral blood human CD4+ T cells. However, Eomes+T-betdim/- CD4+ T cells had low perforin expression and poor β-chemokine production, indicating that Eomes alone is not sufficient to maintain effector functions in human CD4+ T cells. Together these data suggest that CD4+ T cells with high levels of T-bet are the preeminent producers of cytolytic molecules and β-chemokines, representing a multifunctional effector CD4+ T cell population in human blood.
Increased frequencies of cytolytic CD4+ T cells have been described previously in the context of HIV infection [10], but it remains unknown to what degree they are preserved. Through access to cohorts with time points from before and after HIV infection we now show that the frequencies of effector CD4+ T cells increase longitudinally after HIV infection and that in peripheral blood these cells are not depleted during chronic HIV infection. The reason for their preservation is probably multifaceted, but our data revealed that T-bethiEomes+ CD4+ T cells were less susceptible to in vitro HIV infection compared to conventional CD45RO+ CD4+ T cells, potentially due to autocrine β-chemokine production [15]. Another explanation however could simply be that effector CD4+ T cells are preserved during HIV infection because they are not present in lymphoid tissues where the vast majority of viral replication takes place [24]. Most T cells have been thought to recirculate between blood and lymphoid or non-lymphoid tissues [47, 48], but recent human data on multiple organs implicate that terminally-differentiated (effector) CD4+ T cells are primarily present in peripheral blood and highly-vascularized tissues [44, 49]. Our data support these previous studies on human organ donors, demonstrating that few CD27- terminally-differentiated CD4+ T cells are present in LNs, independent of HIV infection. As such, it remains possible that CD4+ T cell tissue compartmentalization and trafficking properties of unique subsets could partly explain why certain CD4+ T cells are depleted or preserved during acute and chronic HIV disease.
Peripheral blood only contains 2% of the total number of lymphocytes, where a predominant fraction resides in lymphoid tissues [50, 51]. However, most of our understanding on CD4+ T cell function, phenotype, and transcriptional characteristics in HIV infection comes from peripheral blood T cells. Knowledge of LN CD4+ T cell responses against HIV is limited, despite the fact that lymphoid tissues serve as key sites for the dissemination and long-term maintenance of HIV replication and the viral reservoir [52]. Previous studies have established that CD8+ T cells in HIV-infected LNs generally express low-levels of cytolytic molecules [53, 54], but similar studies have not been conducted on effector LN CD4+ T cells. While CD8+ T cells seem to express some effector molecules in HIV-infected LNs [32], we found few T-bethiEomes+ CD4+ T cells, which is associated with minimal expression of cytolytic molecules, β-chemokines and IFNγ. The presence of cytolytic Gag-specific CD4+ T cell responses was only present in viremic subjects, which indicates that ongoing viral replication is necessary to maintain the presence of cytolytic CD4+ T cells against HIV in blood. Furthermore, we also identify that LN CD4+ T cells form impaired immunological synapses and release cytolytic granules with slower kinetics, consistent with inefficient potency of target cell destruction. HIV-specific CD4+ T cells were frequently present in LNs; however, these responses appeared to be monofunctional for TNF production or CD107a upregulation and less expression of IFNγ. The lack of β-chemokine production in LNs is of particular interest as several studies have proposed that such functional properties provide resistance to HIV infection [15, 16]. Autocrine β-chemokine production by CD4+ T cells could still be a protective mechanism prohibiting productive infection by HIV in peripheral blood, but likely not in LNs.
The magnitude of HIV-specific CD4+ T cell responses was highest in blood during peak-viremia and coincided with increased T-bet, perforin and MIP-1α production. In contrast, very few cytolytic CD4+ T cells were present in LNs during acute/early HIV infection. Notably, previous studies have shown that LN T cells egress from tissues to enter the blood stream after the acquisition of cytolytic activity [55]. Despite the finding that LCMV-specific LN CD8+ T cells can degranulate during acute infection, they do kill target cells less efficiently [55]. This notion is also in line with our chronic HIV data, showing a disassociation between LN and blood for different subsets of CD4+ T cells that can degranulate. LN CD4+ T cells can degranulate, but these cells express low levels of cytolytic markers and form impaired immunological synapses. Instead they seem to be skewed towards a CXCR5+ phenotype, suggesting that unique Tfh-subsets degranulate and may secrete factors of unknown nature. Future studies should clarify the role and content of granules in LN CD4+ T cells, as such responses seem to be differently regulated and not contain cytolytic molecules or β-chemokines.
Numerous studies have identified that HIV elite controllers possess CD4+ T cell responses with higher polyfunctionality, Gag specificity, proliferative capacity, cytolytic activity and β-chemokine production [4, 5, 8, 9, 18, 56]. Surprisingly, we found a distinct dissociation of effector CD4+ T cell responses in LNs compared to PB in HIV elite controllers. The higher abundance of effector CD4+ T cell responses in peripheral blood could suggest that such responses, together with other factors, are maintaining control in the blood circulation, but not in lymphoid tissues. Furthermore, previous studies have established that HIV-specific CD4+ T cells are generating pressure on the virus, based on the fact that escape mutations emerge within MHC-II-restricted epitope regions [57–59]. Given the limited expression of cytolytic molecules for both HIV-specific CD4+ and CD8+ T cells in LNs [32], it remains possible that non-cytolytic factors are involved in control mechanisms of HIV and could generate viral escape mutants in LNs [60]. Indeed, elegant studies combining CD8+ T cell depletion and ART to study the lifespan of SIV-infected cells longitudinally found that depletion of CD8+ T cells had minimal effect on the death rate of virus infected cells, indicating that CD8+ T cells must act via other mechanisms than direct lysis of cells [61, 62]. Similar studies have also demonstrated that the death-rate of cells infected with wild type and escape mutant viruses are not different, indicating that non-cytolytic functions can drive viral escape and is associated with control of lentiviruses [63]. Modeling of escape has also shown that non-cytolytic functions can select for escape variants, although more slowly than cytolytic responses [64]. Further studies should clarify if cell antiviral functions [65] produced by HIV-specific T cells can generate selective pressure on the virus and lead to elite control of HIV in lymphoid tissues.
In conclusion, we have determined that CD4+ T cells with elevated levels of T-bet and Eomes represent a pleiotropic effector population that is present and preserved in HIV-infected blood. This unique CD4+ T cell population is however almost excluded from HIV-infected LNs. As a consequence, LN CD4+ T cells generally possess lower effector functions, independent of HIV disease status or infection. Our data provide evidence of lack of association between effector-like HIV-specific CD4+ T cell responses in LNs and blood, suggesting dissociated mechanisms of viral control in LNs.
Written informed consent was obtained from all study participants and blood samples were acquired with institutional review board approval at each collecting institution: University of Pennsylvania (IRB#809316, IRB# 815056), Human Subjects Protection Branch (RV217/WRAIR#1373), The United Republic of Tanzania Ministry of Health and Social Welfare (MRH/R.10/18/VOLL.VI/85), Tanzanian National Institute for Medical Research (NIMR/HQ/R.8aVol.1/2013), Royal Thai Army Medical Department (IRBRTA 1810/2558), Uganda National Council for Science and Technology–National HIV/AIDS Research Committee (ARC 084), Uganda National Council of Science and Technology (HS 688), The Swedish Regional Ethical Council (Stockholm, Sweden 2009/1485-31, 2013/1944-31/4, 2014/920-32, 2012/999-32 and 2009/1592-32), INER-CIENI Ethics Committee and the Federal Commission for the Protection against Sanitary Risk (COFEPRIS), the Institutional Review Boards of the Case Western Reserve University (CWRU) and Cleveland Clinic Foundation (CCF). All human subjects were adults. This study was conducted in accordance with the Declaration of Helsinki.
Mesenteric, iliac, inguinal and cervical lymph node (LN) biopsies and peripheral blood were collected from individuals classified as HIV−(n = 51), HIV+ chronic and naïve to ART (n = 71), HIV+ chronic on ART (n = 25), HIV+ elite controllers (n = 18), and HIV+ acute/early seroconverters (n = 17). Recruitment occurred at six sites: University of Pennsylvania (Penn) (HIV−blood and iliac LNs; HIV+ ART+/ART−blood and matched iliac LNs); INER-CIENI in Mexico City (HIV+ ART+/ART−blood and matched cervical LNs); Case Western Reserve University (HIV−mesenteric LNs); University of California, San Francisco (UCSF) (HIV+ EC inguinal LNs); Karolinska Instituet (HIV−and HIV+ ART+/ART−blood) and the RV217 early capture HIV cohort (longitudinal HIV−and HIV+ ART−blood samples). Subject grouping, tissue types, and clinical parameters are summarized in S1 Table. All clinical characteristics of the RV217 cohort has been described previously [66]. HIV−LN samples were obtained from following procedures/conditions: mesenteric LNs (patients undergoing abdominal surgery) and iliac LNs (kidney transplant donors). Sample size was based on the availability of biological samples rather than a pre-specified effect size and was not blinded to the persons executing experiments.
Peripheral blood mononuclear cells (PBMCs) were collected from whole blood or leukapheresis products using Ficoll-Hypaque (GE Healthcare) density gradient centrifugation. Lymph node mononuclear cells (LNMCs) were isolated by manual disruption or using a gentleMACS tissue dissociator. PBMCs and LNMCs were cryopreserved and stored at -140°C for further usage in all experiments.
For all experiments, human cryopreserved PBMCs and LNMCs were thawed and rested for at least 1 hour in complete media (R10), containing RPMI-1640 media supplemented with 10% FBS, 1% L-glutamine, and 1% penicillin/streptomycin. Cells were then washed with PBS, pre-stained for chemokine receptors/ adhesion molecules at 37°C for 10 minutes, stained with LIVE/DEAD Aqua (Invitrogen) for 10 minutes, and surface stained with an optimized antibody cocktail for 20 further minutes. Cells were then washed with FACS buffer (PBS containing 0.1% sodium azide and 1% BSA), fixed, and permeabilized using the Cytofix/Cytoperm Buffer Kit (BD Biosciences) or the FoxP3 Transcription Factor Buffer Kit (eBioscience). An optimized antibody cocktail was then added for 1 hour to detect intracellular/-nuclear markers. Cells were fixed in PBS containing 1% paraformaldehyde (Sigma-Aldrich) and stored at 4°C. All samples were acquired within 3 days using an LSRII (BD Biosciences) and data was analyzed with FlowJo software (version 9.8.8 or higher, TreeStar).
For stimulations, cells were stimulated with 5 μg/mL Brefeldin A (Sigma Aldrich), and overlapping HIV Gag-p55 (NIH), HCMV IE1/pp65 peptides (NIH), SEB (Sigma Aldrich), α-CD3 (BioRad) and α-CD28/CD49d (BD) or medium alone (negative controls). When anti-CD107a was added at the start of the stimulation protocol [67], monensin (0.7 μg/mL, BD Bioscience) was also supplemented.
For sorting experiments, PBMCs and LNMCs were thawed and rested overnight. Cells were next day stained in 15-mL conical tubes following the procedure described above with higher concentrations of antibodies (i.e. not diluted 1:50 in FACS buffer). Cells were then washed with PBS and suspended in R10 media. Sorting was carried out using a FACSAriaII (BD Biosciences) instrument.
The following antibodies were used for human FACS experiments: α-CCR7 APC-Cy7 or BV711 (clone G043H7, BioLegend), α-CD14 BV510 or V500 (clone M5E2, BioLegend), α-CD14 PE-Cy5 (clone 61D3, Abcam), α-CD19 BV510, V500 or PE-Cy5 (clone HIB19, BioLegend), α-CD3 BV570, AF700, APC-H7, APC-Cy7 or APC-R700 (clone UCHT1, BD Biosciences), α-CD4 PeCy5.5 (clone S3.5, Invitrogen), α-CD8a BV570, BV605, or BV711 (clone RPA-T8, BioLegend), α-CD27 BV650 or BV785 (clone O323, BioLegend), α-CD45RO BV650 or PE-CF594 (clone UCHL1, BD Biosciences), α-CD45RO ECD (clone UCHL1, Beckman Coulter), α-CXCR5 AF488 or AF647 (clone RF8B2, BD Biosciences), α-CXCR5 PE-Cy7 (clone J252D4, BioLegend), α-PD-1 APC-Cy7 or BV421 (clone EH12.2H7, BioLegend), α-HLA-DR BV605 or BV650 (clone G46-6, BD Biosciences), α-CD38 APC or BV711 (clone HIT2, BioLegend), α-IFNγ FITC or AF700 (clone B27, Invitrogen), α-TNF PE-Cy7 or BV605 (clone MAb11, BD Biosciences), α-CD107a PE-Cy5 or PE-Cy7 (clone H4A3, eBioscience), α-MIP-Iα FITC (clone MAB271, R&D Systems), α-MIP-Iβ PE-Cy7 or PE (clone D21-1351 BD), α-perforin BV421 or PE-Cy7 (clone B-D48, BioLegend), α-granzyme B Alexa700 (clone GB11, BD Biosciences), α-T-bet PE (clone 4B10, eBioscience), α-T-bet PE-Dazzle 594 or FITC (clone 4B10, BioLegend), α-Eomes AF647 or EF660 (clone WD1928, eBioscience), and α-Ki67 FITC (clone B56, BD Biosciences). LIVE/DEAD Aqua (Invitrogen) or DAPI was used to discriminate dead cells.
PBMCs and LNMCs from HIV–uninfected subjects (n = 2) were thawed, rested overnight and stained as described above. SEB stimulated CD107+ CD4+ T cells were single-cell index sorted into individual wells of a 96-well PCR plate according to the gating strategy depicted in S10A Fig. Each well contained 5 μL lysing buffer, consisting of 4.725 μL of DNA Suspension Buffer (10 mM Tris, pH 8.0, 0.1 mM EDTA; TEKnova), 0.025 μL of 20 U/μL SUPERase (Ambion) and 0.25 μL of 10% NP40 (Thermo Scientific). After FACS sorting, PCR plates were frozen and kept in -80°C until usage.
Similar to before [35], PCR plates were thawed and pre-heated for 90 seconds at 65°C. Reverse Transcription Master Mix (Fluidigm) was added to each well and plate was placed into a thermocycler for reverse transcription (25°C for 5 min, 42°C for 30 min, 85°C for 5 min). Next, pre-amplification mix, consisting of pooled mixture of all primer assays (500nM), 5× PreAmp Master Mix (Fluidigm) and H2O, was added to each well and run on a thermocycler (95°C for 5 min followed by 18 cycles: 96°C for 5 sec 60°C for 6 min). Exonuclease mixture, containing of Exonuclease I (New England BioLabs), 10× Exonuclease I Reaction Buffer (New England BioLabs) and H2O was added to each well to remove excessive primers and the plate was run on a thermocycler (37°C for 30 min, 80°C for 15 min). Each well was diluted (1:4) with DNA Suspension Buffer (10 mM Tris, pH 8.0, 0.1 mM EDTA; TEKnova). Distinct primer assays were generated by adding individual primer pairs (5μM) together with a mix of 2x Assay Loading Reagent (Fluidigm) and 1x DNA suspension buffer to each well of a new plate. A “sample PCR plate” was created by dispensing a sample master mix, containing 2x Sso Fast EvaGreen Supermix with Low ROX (Bio-Rad), 20x DNA Binding Dye Sample Loading Reagent (Fluidigm) and H2O to each well. Pre-amplified samples were added to each well of the sample PCR plate. Control line fluid (Fluidigm) was injected to the 96.96 Dynamic Array chip (Fluidigm) and the chip was primed using an IFX Controller HX. After priming the chip, primer assays and sample mix was added to the unique detector inlets of the chip and transferred to the IFX Controller HX for loading the mixtures. The chip was then transferred to a Biomark HD instrument (Fluidigm) and run using the GE Fast 96x96 PCR+Melt v2.pcl program. All primers were purchased from IDT and assay efficiency as well as melting and amplification curves for each assay were evaluated beforehand on separate Biomark HD runs and using qPCR.
All data were pre-analyzed with the Real-time PCR analysis software (Fluidigm), and Linear (Derivative) and User (Detectors) were used as settings to generate Ct values. A Ct value of 25 was used as limit of detection (LOD). The relative gene expression was defined as a log2 value using the formula: log2 = LOD–Ct. All downstream analysis of the gene expression data, including tSNE analysis, were performed using R Studio or Graph Pad Prism v7.0 (GraphPad).
Frozen PBMCs or LNMCs of uninfected or HIV-infected individuals (n = 2) were thawed and incubated overnight in complete media (RPMI supplemented with 10% of fetal bovine serum, penicillin/streptomycin and glutamine) at 37°C. CD4+ T cells were then purified by negative immunomagnetic sorting using MACS technology CD4+ T cell purification kit according to the manufacturer’s instructions. The cells were transferred into the assay buffer (20 mM HEPES, pH 7.4, 137 mM NaCl, 2 mM Na2HPO4, 5 mM D-glucose, 5 mM KCl, 1 mM MgCl2, 2 mM CaCl2, and 1% human serum albumin) and kept at +4°C (1–2 hours) prior to the analysis. The human HIV-specific CD4+CTL clone AC-25 that recognizes PEVIPMFSALSEGATP (PP16) peptide from HIV Gag protein was used as a positive control [39, 68].
1,2-Dioleoyl-sn-glycero-3-phosphoethanolamine-N-(cap biotinyl) (sodium salt) (Biotinyl cap PE), 1,2-di-(9Z-octadecenoyl)-sn-glycero-3-[(N-(5-amino-1-carboxypentyl)iminodiacetic acid)succinyl] (nickel salt) (NiNTA-DOGS) and 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC) lipids were supplied by Avanti Polar Lipids. Planar lipid bilayers were prepared as previously described [39–41, 69]. Liposomes that contained biotinyl-CAP functionalized lipids (Avanti Polar Lipids) at 4 mol% and liposomes containing NiNTA-DOGS lipids at 35 mol% were used to prepare bilayers. The final concentrations of biotinyl-CAP and NiNTA-DOGS in the bilayers were 0.01 mol% and 17.5 mol%, respectively. Streptavidin (2 μg/ml) and monobiotinylated anti-CD3 (OKT3) monoclonal antibody labeled with Alexa Fluor 488 (2 μg/ml) were reacted sequentially with the biotinylated bilayers to produce the antibody density of 50 molecules/μm2. Cy5-ICAM-1-His6 molecules were incorporated into the bilayers at the density 300 molecules/μm2. Densities of Cy5-ICAM-1 and anti-CD3 antibody on the bilayers were determined as described elsewhere [70].
TIRF images were acquired with an Andor Revolution XD system equipped with Nikon TIRF-E illuminator, 100/1.49 NA objective, Andor iXon X3 EM-CCD camera, objective heater, and a piezoelectric motorized stage with Perfect Focus. The cells were combined with Alexa 568-labeled anti-CD107a antibody Fab fragments at a final concentration of 4 μg/ml and injected into the channel of ibidi slides containing a lipid bilayer. The images of the bilayers were then recorded for 30 minutes at a rate of one image per minute. The resulting images were analyzed using the MetaMorph imaging suite.
To determine the parameters of cell-bilayer interactions we chose only CD4+ T cells productively interacting with the bilayers. That was determined by accumulation of anti-CD3 antibodies at the interface and confirmed by morphology analysis of the cells observed in the transmitted light images. Clustered cells and visibly damaged or apoptotic cells were excluded from analysis.
The efficiency of ICAM-1accumulation for selected cells was measured by determining the average fluorescence intensity of accumulated Cy5-labeled ICAM-1 molecules at the cell-bilayer interface over background fluorescence outside of the contact area in close proximity to the cell. Cell was discerned to accumulate ICAM-1, if the signal-to-background ratio was at least 1.3. If accumulated ICAM-1 molecules formed a ring structure that was observed on at least two consecutive images, we determined that the cell was developing pSMAC. The granule release was evaluated by measuring the average fluorescence intensity of accumulated Alexa 568-labeled anti-CD107a antibody Fab fragment at the T cell/bilayer interface over the background outside the contact area but in close proximity to the cell. Cells with the ratio of Alexa 568 signal to background of at least 1.3 were designated as degranulating cells. All analyzed cells demonstrated one of the following four degranulation patterns: forming or not forming a pSMAC and degranulating or not at the same time. For every sample, we determined the fractions of the total cells corresponding to each pattern of degranulation. To quantify the kinetics of granule release at cell-bilayer interface for all of the degranulating cells, we determined the earliest time point when degranulation was observed.
For imaging flow cytometry, human PBMC were stained with DAPI (5 ug/mL) for 5 min and 100,000 events were imaged on an ImageStreamX (Amnis Corp). Images were captured using a 60x lens with an extended depth of field option using Inspire software (Amnis Corp). Antibody capture beads (BD Biosciences) were used as individual compensation tubes for each fluorophore. Masking functions within IDEAS 5.0 were used to define nuclear and cytoplasmic T-bet and Eomes as previously described [25].
DHIV3 plasmid was provided by Dr. Edward Barker [71] (Rush University, Chicago, IL). VSV-G pseudotyped viruses were produced as previously described [72]. PBMCs were stimulated with SEB for 3 days before addition of the pseudotyped virus. Gag-p24 (clone KC57, Beckman Coulter) expression was determined by flow cytometry on day 5 post-infection.
Mann-Whitney or unpaired t-tests were used to compare differences between unmatched groups, and Wilcoxon matched-pairs single rank or paired t-tests were used to compare differences between matched samples. Spearman or Pearson tests were used for correlation analyses. Non-parametric or parametric tests were conducted based on normal distribution of the data points (Shapiro-Wilk normality test). All analyses were performed using R studio, Graph Pad Prism v7.0 (GraphPad). FlowJo and Cell ACCENSE (Automatic Classification of Cellular Expression by Nonlinear Stochastic Embedding) analyses were used to conduct multivariate tSNE analysis on the single-cell flow data sets [73].
|
10.1371/journal.pcbi.1003500 | MRFalign: Protein Homology Detection through Alignment of Markov Random Fields | Sequence-based protein homology detection has been extensively studied and so far the most sensitive method is based upon comparison of protein sequence profiles, which are derived from multiple sequence alignment (MSA) of sequence homologs in a protein family. A sequence profile is usually represented as a position-specific scoring matrix (PSSM) or an HMM (Hidden Markov Model) and accordingly PSSM-PSSM or HMM-HMM comparison is used for homolog detection. This paper presents a new homology detection method MRFalign, consisting of three key components: 1) a Markov Random Fields (MRF) representation of a protein family; 2) a scoring function measuring similarity of two MRFs; and 3) an efficient ADMM (Alternating Direction Method of Multipliers) algorithm aligning two MRFs. Compared to HMM that can only model very short-range residue correlation, MRFs can model long-range residue interaction pattern and thus, encode information for the global 3D structure of a protein family. Consequently, MRF-MRF comparison for remote homology detection shall be much more sensitive than HMM-HMM or PSSM-PSSM comparison. Experiments confirm that MRFalign outperforms several popular HMM or PSSM-based methods in terms of both alignment accuracy and remote homology detection and that MRFalign works particularly well for mainly beta proteins. For example, tested on the benchmark SCOP40 (8353 proteins) for homology detection, PSSM-PSSM and HMM-HMM succeed on 48% and 52% of proteins, respectively, at superfamily level, and on 15% and 27% of proteins, respectively, at fold level. In contrast, MRFalign succeeds on 57.3% and 42.5% of proteins at superfamily and fold level, respectively. This study implies that long-range residue interaction patterns are very helpful for sequence-based homology detection. The software is available for download at http://raptorx.uchicago.edu/download/. A summary of this paper appears in the proceedings of the RECOMB 2014 conference, April 2–5.
| Sequence-based protein homology detection has been extensively studied, but it remains very challenging for remote homologs with divergent sequences. So far the most sensitive methods employ HMM-HMM comparison, which models a protein family using HMM (Hidden Markov Model) and then detects homologs using HMM-HMM alignment. HMM cannot model long-range residue interaction patterns and thus, carries very little information regarding the global 3D structure of a protein family. As such, HMM comparison is not sensitive enough for distantly-related homologs. In this paper, we present an MRF-MRF comparison method for homology detection. In particular, we model a protein family using Markov Random Fields (MRF) and then detect homologs by MRF-MRF alignment. Compared to HMM, MRFs are able to model long-range residue interaction pattern and thus, contains information for the overall 3D structure of a protein family. Consequently, MRF-MRF comparison is much more sensitive than HMM-HMM comparison. To implement MRF-MRF comparison, we have developed a new scoring function to measure the similarity of two MRFs and also an efficient ADMM algorithm to optimize the scoring function. Experiments confirm that MRF-MRF comparison indeed outperforms HMM-HMM comparison in terms of both alignment accuracy and remote homology detection, especially for mainly beta proteins.
| Sequence-based protein alignment and homology detection has been extensively studied and widely applied to many biological problems such as homology modeling [1]–[4], phylogeny inference [5]–[7] and protein function prediction [8]–[10]. Although extensively studied, remote homology detection still remains very challenging, especially for homologs with divergent sequences. So far the most sensitive method for homology detection is based upon comparison of protein sequence profiles, which are usually derived from multiple sequence alignment (MSA) of sequence homologs in a protein family. That is, instead of aligning two primary sequences, homologs can be detected by aligning protein sequence profiles. To facilitate comparison and alignment, an MSA is usually represented as a position-specific scoring matrix (PSSM) [11] or an HMM (Hidden Markov Model) [12], [13]. HMM is more sensitive than PSSM because 1) HMM contains position-specific gap information; and 2) HMM also takes into account correlation among sequentially adjacent residues. Sequence signature libraries [14] and intermediate sequence based methods [15], [16] are also developed to make use of evolutionary information of a protein. All these methods are sensitive to close homologs, but not good enough for remote homologs. The main issue of existing profile-based methods lies in that they make use of only position-specific amino acid mutation patterns and very short-range residue correlation, but not long-range residue interaction. However, remote homologs may have very divergent sequences and are only similar at the level of (long-range) residue interaction pattern, which is not encoded in current popular PSSM or HMM models.
To significantly advance homology detection, this paper presents a Markov Random Fields (MRFs) modeling of a multiple sequence alignment (MSA). Compared to HMM, MRFs can model long-range residue interactions and thus, encodes information for the global 3D structure of a protein family. In particular, MRF is a graphical model encoding a probability distribution over the MSA by a graph and a set of preset statistical functions. A node in the MRF corresponds to one column in the MSA and one edge specifies correlation between two columns. Each node is associated with a function describing position-specific amino acid mutation pattern. Similarly, each edge is associated with a function describing correlated mutation statistics between two columns. With MRF representation, alignment of two proteins or protein families becomes that of two MRFs. To align two MRFs, a scoring function or alignment potential is needed to measure the similarity of two MRFs. We use a scoring function consists of both node alignment potential and edge alignment potential, which measure the node (i.e., amino acid) similarity and edge (i.e., interaction pattern) similarity, respectively.
It is computationally challenging to optimize a scoring function containing edge alignment potential. To deal with this, we formulate the MRF-MRF alignment problem as an integer programming problem and then develop an ADMM (Alternative Direction Method of Multipliers) algorithm to solve it efficiently to a suboptimal solution. ADMM divides the MRF alignment problem into two tractable sub-problems and then iteratively solve them until they reach consistent solutions.
Experiments show that our MRF-MRF alignment method, denoted as MRFalign, can generate more accurate alignments and is also much more sensitive than others in detecting remote homologs. MRFalign works particularly well on mainly-beta proteins.
Cowen has developed a program SMURFLite for fold recognition based upon the MRF representation of a protein family [17]. Nevertheless, our MRFalign method is significantly different from SMURFLite in a couple of aspects: 1) SMURLite builds an MRF based upon multiple structure alignment instead of multiple sequence alignment (MSA). As such, it cannot apply to sequence-based homology detection in the absence of native structures. In contrast, our method builds MRFs purely based upon MSA and thus, applies to sequence-based protein alignment and homology detection; and 2) SMURLite can only align a single primary sequence to an MRF, while our method aligns two MRFs to yield higher sensitivity. This difference requires us to develop totally new methods to build MRFs from MSA, measure similarity of two MRFs, and optimize the MRF-MRF alignment potential.
Quite a few PSSM-based profile comparison methods for homology detection have been developed, including [11], [18]–[23]. Some studies such as [20] also combine phylogeny information with PSSM-based profile comparison. Homology detection can also be done without aligning proteins. For example, we can represent a protein sequence or profile as a feature vector and then search for homologs by comparing feature vectors. Early methods such as [24] usually conduct straightforward comparison of feature vectors, but are not very sensitive [25]. Improvement in these alignment-free methods results from the application of discriminative learning approaches such as SVM–Fisher [26], SVM-pairwise [27], SVM with the spectrum kernel [28] and SVM with the mismatch kernel [29]. These SVM-based methods are reported to outperform the simple feature comparison methods. Comparing to alignment-based homology detection, alignment-free methods are usually faster but less sensitive.
To train the node alignment potential, we constructed the training and validation data from SCOP70. The sequence identity of all the training and validation protein pairs is uniformly distributed between 20% and 70%. Further, two proteins in any pair are similar at superfamily or fold level. In total we use a set of 1400 protein pairs as the training and validation data, which covers 458 SCOP folds [30]. Five-fold cross validation is used to choose the hyper-parameter in our machine learning model. In particular, every time we choose 1000 out of the 1400 protein pairs as the training data and the remaining 400 pairs as the validation data such that there is no fold-level redundancy between the training and validation data. A training or validation protein has less than 400 residues and contains less than 10% of residues without 3D coordinates. The reference alignment for a protein pair is generated by a structure alignment tool DeepAlign [31]. Each reference alignment has fewer than 50 gap positions in the middle and the number of terminal gaps is less than 20% of the alignment length.
The data used to test alignment accuracy has no fold-level overlap with the training and validation data. In particular, we use the following three datasets to test the alignment accuracy, which are subsets of the test data used in [4] to benchmark protein modeling methods.
Set3.6K: a set of 3617 non-redundant protein pairs. Two proteins in a pair share <40% sequence identity and have small length difference. By “non-redundant” we mean that in any two protein pairs, there are at least two proteins (one from each pair) sharing less than 25% sequence identity.
Set2.6K: a set of 2633 non-redundant protein pairs. Two proteins in a pair share <25% sequence identity and have length difference larger than 30%. This set is mainly used to test the performance of one method in handling with domain boundary.
Set60K: a very large set of 60929 protein pairs, in most of which two proteins share less than 40% sequence identity. Meanwhile, 846, 40902, and 19181 pairs are similar at the SCOP family, superfamily and fold level, respectively, and 151, 2691 and 2218 pairs consist of only all-beta proteins, respectively.
We use the following benchmarks to test remote homology detection success rate.
We run PSI-BLAST with 5 iterations to detect sequence homologs and generate MSAs for the first three datasets. The MSA files for the three SCOP benchmarks are downloaded from the HHpred website (ftp://toolkit.genzentrum.lmu.de/pub/). Pseudocounts are used in building sequence profiles. Real secondary structure information is not used since this paper focuses on sequence-based homology detection.
To evaluate alignment accuracy, we compare our method, denoted as MRFalign, with sequence-HMM alignment method HMMER [12] and HMM-HMM alignment method HHalign [13]. HHMER is run with a default E-value threshold (10.0). HHalign is run with the option “-mact 0.1”. To evaluate the performance of homology detection, we compare MRFalign, with FFAS [11] (PSSM-PSSM comparison), hmmscan (sequence-HMM comparison) and HHsearch and HHblits [33] (HMM-HMM comparison). HHsearch and hmmscan use HHalign and HMMER, respectively, for protein alignment.
Three performance metrics are used including reference-dependent alignment precision, alignment recall and homology detection success rate. Alignment precision is defined as the fraction of aligned positions that are correctly aligned. Alignment recall is the fraction of alignable residues that are correctly aligned. Reference alignments are used to judge if one residue is correctly aligned or alignable. To reduce bias, we use three very different structure alignment tools to generate reference alignments, including TM-align [34], Matt [35], and DeepAlign [31].
As shown in Tables 1 and 2, our method MRFalign exceeds all the others regardless of the reference alignments on both dataset Set3.6K and Set2.6K. MRFalign outperforms HHalign by ∼10% on both datasets, and HHMER by ∼23% and ∼24%, respectively. If 4-position off the exact match is allowed in calculating alignment recall, MRFalign outperforms HHalign by ∼11% on both datasets, and HHMER by ∼25% and ∼33%, respectively.
On the very large set Set60K, as shown in Table 3, our method outperforms the other two in each SCOP classification regardless of the reference alignments used. MRFalign is only slightly better than HHalign at the family level, which is not surprising since it is easy to align two closely-related proteins. At the superfamily level, our method outperforms HHalign and HMMER by ∼6% and ∼18%, respectively. At the fold level, our method outperforms HHalign and HHMER by ∼7% and ∼14%, respectively.
As shown in Tables 4 and 5, our method MRFalign exceeds all the others regardless of the reference alignments on both data sets Set3.6K and Set2.6K. MRFalign outperforms HHalign by ∼8% and ∼5%, respectively, and HMMER by ∼15% and ∼13%, respectively. If 4-position off the exact match is allowed in calculating alignment precision, MRFalign outperforms HHalign by ∼8% and ∼9%, and HMMER by ∼14% and ∼18% on Set3.6K and Set2.6K, respectively.
On the very large set Set60K, as shown in Table 6, our method outperforms the other two in each SCOP classification regardless of the reference alignments used. At the family level, our method outperforms HHalign and HMMER by ∼3% and ∼4%, respectively. At the superfamily level, our method outperforms HHalign and HMMER by ∼4% and ∼5%, respectively. At the fold level, our method outperforms HHalign and HHMER by ∼5% and ∼8%, respectively.
To evaluate homology detection rate, we employ three benchmarks SCOP20, SCOP40 and SCOP80 introduced in [32]. For each protein sequence in one benchmark, we treat it as a query, align it to all the other proteins in the same benchmark and then examine if those with the best alignment scores are similar to the query or not. We also conducted homology detection experiments using hmmscan, FFAS, HHsearch and HHblits with default options. The success rate is measured at the superfamily and fold levels, respectively. When evaluating the success rate at the superfamily (fold) level, we exclude those proteins similar to the query at least at the family (superfamily) level. For each query protein, we examine the top 1-, 5- and 10-ranked proteins, respectively.
As shown in Table 7, tested on SCOP20, SCOP40 and SCOP80 at the superfamily level, our method MRFalign succeeds on ∼6%, ∼4% and ∼4% more query proteins than HHsearch, respectively, when only the first-ranked proteins are considered. As shown in Table 8, at the fold level, MRFalign succeeds on ∼11%, ∼11% and ∼12% more proteins than HHsearch, respectively, when only the first-ranked proteins are evaluated. At the superfamily level, SCOP20 is more challenging than the other two benchmarks because it contains fewer proteins similar at this level. Nevertheless, at the fold level, SCOP80 is slightly more challenging than the other two benchmarks maybe because it contains many more irrelevant proteins and thus, the chance of ranking false positives at top is higher.
Similar to alignment accuracy, our method for homology detection also has a larger advantage on the beta proteins. In particular, as shown in Table 9, tested on SCOP20, SCOP40 and SCOP80 at the superfamily level, MRFalign succeeds on ∼7%, ∼5% and ∼7% more proteins than HHsearch, respectively, when only the first-ranked proteins are evaluated. As shown in Table 10, at the fold level, MRFalign succeeds on ∼13%, ∼16% and ∼17% more proteins than HHsearch, respectively, when only the first-ranked proteins are evaluated. Note that in this experiment, only the query proteins are mainly-beta proteins, the subject proteins can be of any types. If we restrict the subject proteins to only beta proteins, the success rate increases further due to the reduction of false positives.
To evaluate the contribution of our edge alignment potential, we calculate the alignment recall improvement resulting from using edge alignment potential on two benchmarks Set3.6K and Set2.6K. As shown in Table 11, our edge alignment potential can improve alignment recall by 3.4% and 3.7%, respectively. When mutual information is used, we can further improve alignment recall by 1.1% and 1.9% on these two sets, respectively. Mutual information is mainly useful for proteins with many sequence homologs since it is close to 0 when there are few sequence homologs. As shown in Table 11, if only those proteins with at least 256 non-redundant sequence homologs are considered, the improvement resulting from mutual information is ∼3%.
Figure 1 shows the running time of MRFalign with respect to protein length. As a control, we also show the running time of the Viterbi algorithm, which is used by our ADMM algorithm to generate alignment at each iteration. As shown in this figure, MRFalign is no more than 10 times slower than the Viterbi algorithm. To speed up homology detection, we first use the Viterbi algorithm to perform an initial search without considering edge alignment potential and keep only top 200 proteins, which are then subject to realignment and rerank by our MRFalign method. Therefore, although MRFalign may be very slow compared to the Viterbi algorithm, empirically we can do homology search only slightly slower than the Viterbi algorithm.
We conducted two experiments to show that our MRFalign is not overtrained. In the first experiment, we used 36 CASP10 hard targets as the test data. Our training set was built before CASP10 started, so there is no redundancy between the CASP10 hard targets and our training data. Using MRFalign and HHpred, respectively, we search each of these 36 test targets against PDB25 to find the best match. Since PDB25 does not contain proteins very similar to many of the test targets, we built a 3D model using MODELLER from the alignment between a test target and its best match and then measure the quality of the model. As shown in Figure 2, MRFalign can yield much better 3D models than HHsearch for most of the targets. This implies that our method can generalize well to the test data not similar to the training data.
In the second experiment, we divide the proteins in SCOP40 into three subsets according their similarity with all the training data. We measure the similarity of one test protein with all the training data by its best BLAST E-value. We used two values 1e-2 and 1e-35 as the E-value cutoff so that the three subsets have roughly the same size. As shown in Table 12, the advantage of our method in remote homology detection over HHpred is roughly same across the three subsets. Since HHpred is an unsupervised algorithm, this implies that the performance of our method is not correlated to the test-training similarity. Therefore, it is unlikely that our method is overfit by the training data.
This paper has presented a new method for sequence-based protein homology detection that compares two protein sequences or families through alignment of two Markov Random Fields (MRFs), which model the multiple sequence alignment (MSA) of a protein family using an undirected general graph in a probabilistic way. The MRF representation is better than the extensively-used PSSM and HMM representations in that the former can capture long-range residue interaction pattern, which reflects the overall 3D structure of a protein family. As such, MRF comparison is much more sensitive than HMM comparison in detecting remote homologs. This is validated by our large-scale experimental tests showing that MRF-MRF comparison can greatly improve alignment accuracy and remote homology detection over currently popular sequence-HMM, PSSM-PSSM, and HMM-HMM comparison methods. Our method also has a larger advantage over the others on mainly-beta proteins.
We build our MRF model of a protein family based upon multiple sequence alignment (MSA) in the absence of native structures. The accuracy of the MRF model depends on the accuracy of an MSA. Currently we rely on the MSA generated by PSI-BLAST. In the future, we may explore better alignment methods for MSA building or even utilize solved structures of one or two protein sequences to improve MSA. The accuracy of the MRF model parameter usually increases with respect to the number of non-redundant sequence homologs in the MSA. Along with more and more protein sequences are generated by a variety of sequencing projects, we shall be able to build accurate MRFs for more and more protein families and thus, detect their homologous relationship more accurately.
An accurate scoring function is essential to MRF-MRF comparison. Many different methods can be used to measure node and edge similarity of two MRFs, just like many different scoring functions can be used to measure the similarity of two PSSMs or HMMs. This paper presents only one of them. In the future we may explore more possibilities. It is computationally intractable to find the best alignment between two MRFs when edge similarity is taken into consideration. This paper presents an ADMM algorithm that can efficiently solve the MRF-MRF alignment problem to suboptimal. However, this algorithm currently is about 10 times slower than the Viterbi algorithm for PSSM-PSSM alignment. Further tuning of this ADMM algorithm is needed for very large-scale homology detection.
Given a protein primary sequence, we run PSI-BLAST [36] with 5 iterations and E-value cutoff 0.001 to find its sequence homologs. PSI-BLAST also generates an MSA of the sequence homologs. Let be a finite discrete random variable representing the amino acid at column i in the MSA, taking values from 1 to 21, corresponding to 20 amino acids and gap. Then we can use a multivariate random variable , where N is the number of columns, to model the MSA. We use an MRF to define the probability distribution of X. MRF is an undirected graph that can be used to model a set of correlated random variables. As shown in Fig. 3, an MRF node represents one column in the MSA and an edge represents the correlation between two columns i and k when . We ignore very short-range correlation (i.e., ) since it is not very informative. The MRF consists of two types of functions: and , where is an amino acid preference function for node i and is a pairwise amino acid preference function for edge (i, k) that reflects interaction between two nodes. Then, the probability of observing a particular protein sequence X can be calculated as follows.(1)where Z is the normalization factor.
We use two kinds of information in MRFs for their alignment. One is the occurring probability of 20 amino acids and gap at each node (i.e., each column in MSA), which can also be interpreted as the marginal probability at each node. The other is the correlation between two nodes, which can be interpreted as interaction strength of two MSA columns and calculated by several different ways. For example, we can use a contact prediction program such as PSICOV [37] and PhyCMAP [38] for this purpose. PSICOV assumes that is a Gaussian distribution function and calculates the correlation between two columns by inverse covariance matrix. PhyCMAP takes sequence information (including mutual information) as input and predicts the probability of two residues forming a contact, which can be used to indicate the interaction strength of two columns. However, it takes time to run these programs, in current implementation we calculate the mutual information (MI) and its power series of two columns as interaction strength. That is, we use MI, MI2, …, MI11 to quantify all the pairwise interaction strength where MI is the mutual information matrix. The MI power series are much more informative than the MI alone, as tested in our contact prediction program PhyCMAP.
Our scoring function for MRF-MRF alignment is a linear combination of node alignment potential and edge alignment potential with equal weight. Let T and S denote two MRFs for the two proteins under consideration. There are three possible alignment states M, and where M represents two nodes being aligned, denotes an insertion in T (i.e., one node in T is not aligned), and denotes an insertion in S (i.e., one node in S is not aligned). As shown in Fig. 4, each alignment can be represented as a path in an alignment matrix, in which each vertex can be exactly determined by its position in the matrix and its state. For example, the first vertex in the path can be written as (0, 0, dummy), the 2nd vertex as and the 3rd vertex as . Therefore, we can write an alignment as a set of triples, each of which has a form like where represents the position and u the state.
As mentioned before, an alignment can be represented as a path in the alignment matrix, which encodes an exponential number of paths. We can use a set of binary variables to indicate which path is chosen, where and are the lengths of the two MSAs, is an entry in the alignment matrix and u is the associated state. is equal to 1 if the alignment path passes with state u. Therefore, the problem of finding the best alignment between two MRFs can be formulated as the following quadratic optimization problem.(P1)where and are node and edge alignment potentials as described in previous section. Meanwhile, is equal to 0 if either u or v is not a match state. L is the alignment length and is used to make the accumulative node and edge potential have similar scale. Note that L is unknown and we will describe how to determine it later in this section. Finally, the solution of P1 shall be subject to the constraint that all those with value 1 shall form a valid alignment path. This constraint shall also be enforced to all the optimization problems described in this section.
It is computationally intractable to find the optimal solution of P1. Below we present an ADMM (Alternating Direction Method of Multipliers) method that can efficiently solve this problem to suboptimal. See [42] for a tutorial of the ADMM method. To use ADMM, we rewrite P1 as follows by making a copy of z to y, but without changing the solution space.(P2)Problem P2 can be augmented by adding a term to penalize the difference between z and y.(P3)P3 is equivalent to P2 and P1, but converges faster due to the penalty term. Here is a hyper-parameter influencing the convergence rate of the algorithm. Some heuristics algorithms were proposed for choosing at each iteration, such as [43], [44]. Empirically, setting to a constant ( = 0.5) enables our algorithm to converge within 10 iterations for most protein pairs.
Adding the constraint using a Lagrange multiplier to Eq. (7), we have the following Lagrangian dual problem:(P4)It is easy to prove that P3 is upper bounded by P4. Now we will solve P4 and use its solution to approximate P3 and thus, P1.
Since both z and y are binary variables, the last term in (P4) can be expanded as follows.(5)For a fixed , we can split P4 into the following two sub-problems.(SP1)where (SP2)where
The sub-problem SP1 optimizes the objective function with respect to y while fixing z, and the sub-problem SP2 optimizes the objective function with respect to z while fixing y. SP1 and SP2 do not contain any quadratic term, so they can be efficiently solved using the classical dynamic programming algorithm for sequence or HMM-HMM alignment.
In summary, we solve P4 using the following procedure.
Initialize z by aligning the two MRFs without the edge alignment potential, which can be done by dynamic programming. Accordingly, initialize L as the length of the initial alignment.
Solve (SP1) first and then (SP2) using dynamic programming, each generating a feasible alignment.
If the algorithm converges, i.e., the difference between z and y is very small or zero, stop here. Otherwise, we update the alignment length L as the length of the alignment just generated and the Lagrange multiplier using subgradient descent as in Eq. (6), and then go back to Step 2).(6)
Due to the quadratic penalty term in P3 this ADMM algorithm usually converges much faster and also yields better solutions than without this term. Empirically, it converges within 10 iterations for most protein pairs. See [42] for the convergence proof of a general ADMM algorithm.
A summary of this paper appears in the proceedings of the RECOMB 2014 conference, April 2–5 [45]. |
10.1371/journal.ppat.1004303 | Distinct APC Subtypes Drive Spatially Segregated CD4+ and CD8+ T-Cell Effector Activity during Skin Infection with HSV-1 | Efficient infection control requires potent T-cell responses at sites of pathogen replication. However, the regulation of T-cell effector function in situ remains poorly understood. Here, we show key differences in the regulation of effector activity between CD4+ and CD8+ T-cells during skin infection with HSV-1. IFN-γ-producing CD4+ T cells disseminated widely throughout the skin and draining lymph nodes (LN), clearly exceeding the epithelial distribution of infectious virus. By contrast, IFN-γ-producing CD8+ T cells were only found within the infected epidermal layer of the skin and associated hair follicles. Mechanistically, while various subsets of lymphoid- and skin-derived dendritic cells (DC) elicited IFN-γ production by CD4+ T cells, CD8+ T cells responded exclusively to infected epidermal cells directly presenting viral antigen. Notably, uninfected cross-presenting DCs from both skin and LNs failed to trigger IFN-γ production by CD8+ T-cells. Thus, we describe a previously unappreciated complexity in the regulation of CD4+ and CD8+ T-cell effector activity that is subset-specific, microanatomically distinct and involves largely non-overlapping types of antigen-presenting cells (APC).
| HSV-1 is a widely distributed pathogen causing a life-long latent infection associated with periodic bouts of reactivation and severe clinical complications. Adaptive immune responses encompassing CD4+ and CD8+ T-cell activities are key to both the clearance of infectious virus and the control of latent infection. However, precisely how such T-cell responses are regulated, particularly within acutely infected peripheral tissues, remains poorly understood. Using a mouse model of HSV-1 skin infection, we describe a complex regulation of T-cell responses at the site of acute infection. These responses were subset-specific and anatomically distinct, with CD4+ and CD8+ T-cell activities being directed to distinct anatomical compartments within the skin. While IFN-γ-producing CD4+ T cells were broadly distributed, including skin regions a considerable distance away from infected cells, CD8+ T-cell activity was strictly confined to directly infected epithelial compartments. This unexpected spatial segregation was a direct consequence of the involvement of largely non-overlapping types of antigen-presenting cells in driving CD4+ and CD8+ T-cell effector activity. Our results provide novel insights into the cellular regulation of T-cell immunity within peripheral tissues and have the potential to guide the development of T-cell subset-specific approaches for therapeutic and prophylactic intervention in antimicrobial immunity and autoimmunity.
| Infection results in the priming of pathogen-specific T-cell responses in LNs draining the site of infection. Depending on the nature of the pathogen, this critical step in generating adaptive immunity involves the interaction of naive T cells with various types of migrating and LN-resident DCs [1], [2]. During skin infection with herpes simplex virus (HSV)-1, LN-resident CD8α+ DCs and skin-derived CD103+ DCs can activate naïve CD8+ T-cells, presumably through the cross-presentation pathway involving the acquisition of noninfectious antigen [1]–[4]. By contrast, all subsets of skin-derived migratory DCs, including epidermal Langerhans cells, dermal CD11b+ and dermal CD103+ DCs, in addition to LN-resident CD8α+ DCs acquire the ability to stimulate naive HSV-specific CD4+ T cells [1], [2], [4]. Following appropriate activation by DCs, T cells undergo a program of clonal expansion, which is accompanied by the acquisition of effector functions and the induction of migration molecules that facilitate their infiltration of infected tissues.
While CD4+ helper T cells support the generation of antibody and CD8+ T-cell responses in lymphoid tissues, both CD4+ and CD8+ T-cells also contribute directly to pathogen control at sites of infection [5], [6]. The latter is achieved through two principle effector functions: the contact-dependent elimination of infected tissue cells and the local production of inflammatory and antimicrobial cytokines [5], [6]. The extent to which these T-cell activities contribute to immunity depends on the nature of the infection. For instance, control of non-cytopathic viruses, such as lymphocytic choriomeningitis virus, strictly requires cytolytic T-cell activity [7]. By contrast, immunity against cytolytic viruses, such as vaccinia and vesicular stomatitis virus, does not rely on target cell elimination by T cells [8]. Instead, under circumstances where infection will ultimately result in lytic cell death regardless of T-cell killing, pathogen containment and clearance is dependent on the production of cytokines by effector CD4+ and CD8+ T cells [9]–[11]. Together these diverse effector T-cell (TEFF) activities are essential for efficient immune protection, however, they may also cause the destruction of uninfected tissues, as seen in the context of immunopathology, autoimmunity or transplant rejection. Therefore, a detailed understanding of T-cell-mediated immunity in peripheral tissues forms an essential basis for therapeutic interventions to modulate T-cell responses against both harmful and innocuous antigens. Nevertheless, the cellular and molecular mechanisms controlling T-cell effector activities in nonlymphoid organs remain poorly defined [2], [10].
At its simplest, T-cell effector functions are regulated by T-cell receptor (TCR) stimulation through peptide-MHC complexes on APCs. Importantly in this respect, disengagement of the TCR from antigen-MHC complexes results in the immediate cessation of T-cell cytokine production [10], [12]. This “on-off cycling” of effector activity provides a sophisticated level of antigen specificity and places important temporal and spatial constraints on TEFF-cell responses [10]. As a consequence, effector T cells circulating through the blood or uninfected tissues are thought to shutdown cytokine production and to regain this effector function only upon reencounter with antigen in infected tissues [10]. In addition, noncognate signals delivered through inflammatory mediators and costimulatory molecules, such as interleukin (IL)-18, IL-12, type I IFNs or CD80 and CD86, may also trigger or further modulate T-cell cytokine production and cytotoxic activity [13]–[16]. Thus, the presence of appropriate APCs providing antigen stimulation together with accessory signals is critical in regulating T-cell immunity in situ and targeting effector activities to pathogen-containing tissues [2], [17]. Indeed, various types of professional and nonprofessional APCs, including monocyte-derived inflammatory DCs, B cells, neutrophils and parenchymal cells, have been suggested to elicit T-cell effector functions within nonlymphoid tissues [15], [18]–[21]. Key aspects in this regulation, however, particularly those pertaining to the infection status of APCs and the role of distinct APC subtypes in driving CD4+ versus CD8+ TEFF-cell responses, remain poorly understood.
Here, we define the cellular interactions that control TEFF-cell activity during the course of skin infection with HSV-1. We focus our analysis on the production of IFN-γ, a central component of adaptive immune responses. IFN-γexerts proinflammatory and regulatory effects on a variety of target cells, including the stimulation of antimicrobial activity and the induction of MHC molecules and inflammatory chemokines [5], [6]. Protection from HSV infection strictly requires TEFF-cell activities, with both CD4+ and CD8+ T cells contributing to virus control in skin, mucosa and sensory ganglia [22]–[26]. Moreover, efficient immunity against HSV infection requires IFN-γ [27] and, interestingly, it has been proposed that production of this cytokine rather than cytolytic activity is the major CD8+ T-cell mechanism for virus control in neuronal tissues [28], [29] as well as during lytic infection in genital mucosa [26]. The shared role of IFN-γ as a key effector molecule produced by both CD4+ and CD8+ T cells allowed us to directly compare the regulation of these TEFF-cell subsets side by side. We further took advantage of the tropism of HSV-1 for epithelial tissues [30] to document a distinct anatomical distribution of IFN-producing TEFF-cell subsets in relation to the presence or absence of infectious virus in different microanatomical compartments. Importantly, this unexpected spatial segregation of TEFF-cell effector activity was a direct result of the involvement of largely non-overlapping subsets of professional and nonprofessional APCs in driving CD4+ and CD8+ TEFF-cell responses.
To determine population kinetics and cytokine production by TEFF cells in lymphoid and peripheral tissues, we utilized a skin infection with HSV-1 in combination with adoptive transfer of TCR-transgenic T cells specific for determinants derived from the HSV glycoproteins gB (CD8+ gBT-I cells) [31] and gD (CD4+ gDT-II cells) [4], respectively. Consistent with the tropism of HSV for epithelial tissues, immunofluorescence microscopy (IFM) of skin revealed that infection was largely confined to the epidermal layer and hair follicle epithelium (Fig. S1A). Separation of epidermal and dermal tissue (Fig. S1B,C) revealed that HSV-specific TEFF cells began to infiltrate infected skin around 5 days post-infection, albeit with fewer cells in the smaller epidermal compartment (Fig. 1A). TEFF-cell numbers peaked around 8 days after inoculation and declined thereafter (Fig. 1A).
To analyze the production of IFN-γ by TEFF cells in situ, we adopted protocols that facilitate intracellular cytokine staining following exposure to the Golgi inhibitor brefeldin A (BFA), either in vivo after intravenous injection [15], [32] and/or ex vivo immediately after tissue harvest and during enzymatic digestion [21]. Of note, in order to focus our analysis on IFN-γ production in situ, neither of these approaches involved overt restimulation with high concentrations of peptide antigen ex vivo. A considerable portion of gBT-I and gDT-II TEFF cells in the epidermis and dermis of infected skin produced IFN-γ 5–6 days post-infection (Fig. 1B–D). Concomitant with clearance of infectious virus from skin [25], IFN-γ production by TEFF cells ceased around day 7, with virtually no IFN-γ+ TEFF cells present 8 days post-infection (Fig. 1D). Similar kinetics of IFN-γ production were also observed for endogenous CD8+ and CD4+ TEFF cells (Fig. S2A,B). While roughly equal portions of gDT-II cells produced IFN-γ in the epidermal and dermal layers of skin 5 days after infection, the fraction of IFN-γ+ gBT-I cells was approximately 3-fold higher in the epidermis as compared to the dermis (Fig. 1C,D). Note, that dermal preparations contained hair follicles of epithelial origin and therefore also harbored some replicating virus (Fig. S1A–C). The broader distribution of IFN-γ+ gDT-II cells in the skin also extended to lymphoid tissues, with skin-draining axillary and brachial LNs, but not spleen, containing an appreciable fraction of IFN-γ-producing gDT-II cells (Fig. 1E,F). By contrast, IFN-γ+ gBT-I cells were virtually absent from all lymphoid tissues. Together, these results suggested a distinct anatomical distribution of IFN-γ-producing CD4+ and CD8+ TEFF cells in both peripheral and lymphoid tissues.
To gain further insight into the microanatomical localization of IFN-γ-producing TEFF-cell subsets, we obtained skin tissue for IFM analysis. Staining of skin sections with anti-IFN-γ antibodies confirmed the presence of IFN-γ-producing cells during the acute phase of infection (Fig. 2A,B), with both endogenous CD4+ and CD8+ (Fig. S2C,D), as well as transgenic gBT-I and gDT-II TEFF cells (Fig. 2C,D) contributing to this response. Interestingly, although gBT-I TEFF cells were broadly distributed throughout the skin, IFN-γ+ gBT-I cells were strictly confined to the epidermis and hair follicle epithelium (Fig. 2C). By contrast, the majority of IFN-γ+ gDT-II cells localized to the dermal layer, where they were found either in association with hair follicles or in considerable distance to the epithelium (Fig. 2D). Thus, in contrast to the strict confinement of IFN-γ+ CD8+ TEFF cells to the epithelium, the dermal layer was the predominant site of the CD4+ T-cell IFN-γ response.
The kinetics of IFN-γ production by TEFF cells suggested that the presence of infectious virus was likely to play a role in the induction of cytokine production. This was indeed the case, as gBT-I and gDT-II TEFF cells primed by HSV infection did not produce IFN-γ in non-specifically inflamed skin after treatment with 1-fluoro-2,4-dinitrobenzene (DNFB) (Figs. 3A and S3A). Likewise, in vitro activated gBT-I TEFF cells transferred into HSV-infected mice lacking H-2Kb molecules (H-2Kb−/−) did not produce significant amounts of IFN-γ in the skin (Figs. 3B,C and S3B). Transfer of activated TEFF cells was necessary as H-2Kb−/− mice cannot prime naïve gBT-I cells due to lack of the relevant MHC-I restriction element. IFN-γ production by transferred gBT-I TEFF cells was completely restored in similar experiments following HSV-1 infection of bone marrow chimeric mice in which H-2Kb molecules were expressed exclusively in radioresistant cells, but were absent from the radiosensitive hematopoietic compartment (Figs. 3D and S3C). By contrast, we observed a significant reduction in the frequency of in vivo primed IFN-γ+ gBT-I cells in chimeric mice in which H-2Kb molecules were selectively missing from radioresistant cells, when compared with fully MHC-I-sufficient control chimeras (Figs. 3E and S3D). Interestingly, the overall frequencies of IFN-γ+ gBT-I TEFF cells appeared to be increased in this particular experimental set-up, possibly related to altered immune activation thresholds in previously irradiated recipient mice. Regardless, these results indicated that presentation of viral antigens by radioresistant epithelial cells, such as keratinocytes, Langerhans cells [33] and dendritic epidermal T cells (DETC) [34], was necessary and sufficient for optimal IFN-γ responses by gBT-I TEFF cells.
IFN-γ production by gDT-II cells was also a consequence of antigen recognition, as in vitro activated gDT-II TEFF cells transferred into infected MHC-II-deficient mice (I-A/E−/−) failed to produce IFN-γ (Figs. 3F,G and S3E). Transfer of activated gDT-II cells was necessary to overcome the inability of I-A/E−/− mice to support CD4+ T-cell priming. By contrast, gDT-II TEFF cells primed in vivo produced IFN-γ in chimeric mice in which only radiosensitive, but not radioresistant cells expressed MHC-II molecules, although we observed a moderate, yet significant reduction in the frequency of IFN-γ+ gDT-II cells in this situation (Figs. 3H and S3F). These results implied that bone marrow-derived MHC-II+ APCs were largely responsible for presentation of viral antigens and eliciting cytokine production by CD4+ TEFF cells.
Consistent with an involvement of MHC-II-expressing APCs in driving CD4+ TEFF-cell responses, MHC-IIhi cells accumulated in the skin during the first week post-infection [35] (Fig. S4A,B). The majority of these cells had a CD11cintCD11b+ phenotype and further expressed CD64 and MAR-1, identifying them as monocyte-derived inflammatory DCs [36], [37] (Fig. S4C). IFM of infected skin revealed a broad distribution of MHC-II+ cells, with the vast majority localizing to the dermal layer, where they were found in close proximity to IFN-γ+ gDT-II cells (Fig. 4A). Indeed, partial depletion of CD11c+ DCs upon diphtheria toxin (DT) treatment of CD11c.DTR mice resulted in a significant reduction in the frequency of IFN-γ+ gDT-II cells in infected skin (Figs. 4B,C and S5A). Furthermore, treatment of mice with antibodies blocking the costimulatory molecules CD80 and CD86, typically expressed by professional APCs such as DCs, also abrogated the CD4+ TEFF-cell IFN-γ in skin and draining LNs (Figs. 4D and S5B). In stark contrast, costimulation blockade had no bearing on IFN-γ production by gBT-I TEFF cells (Fig. 4D). These experiments suggested that MHC-II+ DCs were the main drivers of cytokine production by CD4+ TEFF cells. Nevertheless, in Ccr2−/− mice, an absence of monocyte-derived DCs [20], [38], which numerically dominated the cutaneous DC network during acute infection, rather increased than decreased the frequency of IFN-γ+ gDT-II and gBT-I cells (Fig. S5C,D), potentially related to impaired virus control in these mice [20]. Furthermore, we observed normal IFN-γ production by gDT-II and gBT-I TEFF cells in the absence of Langerhans cells, CD103+ dermal DC and CD8+ LN-resident DCs upon DT treatment of Langerin.DTR mice [39], [40], and similarly, also in B-cell-deficient μMT mice [41] (Fig. S5E–G). These results suggested a level of redundancy regarding the involvement of different types of professional APCs in regulating CD4+ TEFF activities in infected skin.
To more directly establish a role for DCs in CD4+ T-cell responses, we utilized an ex vivo stimulation assay in which APCs purified from infected mice were cocultured with in vitro generated TEFF cells. Initial experiments revealed that maximal IFN-γ production occurred 18 hours after antigen-dependent restimulation for gDT-II and 5 hours for gBT-I TEFF cells (Fig. S6A,B). We purified CD11chi DCs from HSV-infected skin and divided them into CD11bhi and CD11blo subsets (Fig. 5A), with the former expected to contain monocyte-derived and dermal DCs and the latter expected to contain Langerhans cells and CD103+ dermal DCs [4], [37], [40]. Notably, both subsets induced robust IFN-γ production by gDT-II TEFF cells, whereas monocytes (CD11b+CD11c−Ly6Chi) and neutrophils (CD11b+CD11c−Ly6Cint) failed to do so (Fig. 5B,C). Unexpectedly, none of these APCs induced IFN-γ production by gBT-I TEFF cells (Fig. 5B,C). This was not related to potentially compromised expression of H-2Kb molecules after APC isolation, since all subtypes triggered IFN-γ production by gBT-I TEFF cells when pulsed with high doses of gB-peptide prior to cell sorting (Fig. S6C,D). In separate experiments, we specifically sorted CD103+ dermal DCs, as this DC subset is capable of cross-presenting viral antigens to CD8+ T cells during skin infection [4]. Once again, while both CD103+ and CD103− CD11chi DCs triggered IFN-γ production by gDT-II TEFF cells, neither of the two subsets activated gBT-I TEFF cells (Fig. 5D). Remarkably, this disparate response also extended to DCs isolated from LNs draining the site of infection with CD103+, CD11b+, CD8α+ and Langerhans cell-containing CD103−CD11b−CD8α− subsets inducing IFN-γ production by gDT-II, but not gBT-I TEFF cells (Fig. 5E). Of note, gBT-I TEFF cell unresponsiveness towards DC stimulation was observed irrespective of the culture period for 5–18 hours (not shown). Together, these results demonstrated that various skin-derived and LN-resident DC subsets acquired and presented viral antigen for the activation of CD4+ TEFF cells. By contrast, none of these DCs elicited IFN-γ-production by CD8+ TEFF cells, even though at least some of them, namely the CD8α+ and CD103+ subsets, have the ability to present viral antigens for the activation of naïve CD8+ T cells [3], [4].
Given that IFN-γ+ gBT-I cells were found exclusively in skin epithelium (Figs. 1B–D and 2B), we reasoned that this compartment contained APCs capable of stimulating CD8+ TEFF cells. Therefore, we purified CD45.2+ hematopoetic and CD45.2− parenchymal and stromal cells from epidermal sheets of infected mice by cell sorting and tested their ability to activate CD8+ TEFF cells. Note that here we used expression of CD45.2 to distinguish between hematopoeitc and non-hematopoeitc epidermal cells, whereas in other analyses we used this molecule as a marker for CD45.1+CD45.2+ gBT-I and gDT-II cells. Both fractions induced IFN-γ production by gBT-I TEFF cells, although the keratinocyte-containing CD45.2− subset appeared to be slightly more potent in this regard (Fig. 6A,B). Induction of IFN-γ production was antigen-specific, since co-cultured OT-I TEFF cells of an irrelevant specificity did not respond to either of the APC subsets (Fig. S7A,B). Furthermore, stimulation of IFN-γ production by CD45.2− epidermal APCs was specific for CD8+ TEFF cells since these APCs failed to activate gDT-II TEFF cells (Figs. 6B and S7C). In line with this, the vast majority of CD45.2− epidermal cells from infected skin lacked expression of MHC-II molecules (Fig. S7D). To better define the nature of epidermal APCs capable of stimulating CD8+ TEFF cells, we sorted these cells into keratinocytes (CD45.2−EpCAM+), DCs (CD45.2+CD11chi), DETCs (CD45.2+Vγ3+), as well as residual CD45.2+CD11c− cells. As expected, keratinocytes induced moderate levels of IFN-γ production by gBT-I TEFF cells, and so did epidermal DCs (Fig. 6C,D), in contrast to their counterparts isolated from total skin preparations and LNs (Fig. 5C–E). While residual CD45.2+ epithelial cells had only a weak stimulatory capacity, remarkably, DETCs were by far the most potent APCs triggering IFN-γ production by gBT-I TEFF cells (Fig. 6C,D).
Given that the epidermis was the predominant site of viral replication in vivo, we hypothesized that the stimulatory capacity of epidermal APCs resulted from their direct infection. In agreement, intravital two-photon microscopy of skin infected with a cyan fluorescent protein (CFP)-expressing HSV-1 strain revealed that slow-moving gBT-I TEFF cells were swarming around virally infected epidermal cells (Movie S1). By contrast, gBT-I TEFF cells more distal to infection foci displayed significantly higher mean velocities (Fig. 7A,B). Importantly, using IFM, we observed that IFN-γ+ gBT-I cells co-localized with HSV-infected cells in the epidermis and hair follicle epithelium (Fig. 7C). To further identify infected cells, we inoculated mice with a recombinant strain of HSV-1 expressing green fluorescent protein (HSV.GFP) and analyzed epidermal cells 5 days post-infection using flow cytometry. We observed small numbers of GFP-expressing cells amongst various populations of epidermal cells, including keratinocytes (CD45.2−EpCAM+), DETCs (CD45.2+Vγ3+), Langerhans cells (CD45.2+EpCAM+MHC-IIhi), other DCs (CD45.2+EpCAM−MHC-IIhi), as well as undefined CD45.2+MHC-II− cells (Fig. 7D). As expected, GFP+ cells were absent after infection with the wild-type HSV-1 KOS strain. Next, we sorted GFP+ and GFP− DETCs, MHC-IIhi DCs and keratinocytes from epidermal sheets (Fig. 7E) and tested their ability to stimulate TEFF cells ex vivo. Strikingly, all GFP+, presumably infected, APCs were able to trigger IFN-γ production by gBT-I TEFF cells (Fig. 7F,H). In contrast, IFN-γ production was not elicited by their GFP− counterparts. Finally, DCs, but not DETCs, elicited IFN-γ production by gDT-II TEFF cells, irrespective of their GFP-expression status (Fig. 7G,H). Together, these results indicated that various types of epidermal APCs induced cytokine production by CD8+ TEFF cells. Importantly, direct viral infection was a strict requirement for their stimulatory capacity. Overall, our data highlight a previously unappreciated complexity in the regulation of T-cell effector activity that was subset-specific, microanatomically distinct and involved largely non-overlapping subsets of professional and nonprofessional APCs for CD4+ and CD8+ T-cell responses (Fig. S8).
Our results highlight a stringent and complex regulation of TEFF-cell responses that targets effector activities strictly to the site of infection and related lymphoid tissues. Thus, IFN-γ production was limited to the time of acute infection and occurred in an antigen-dependent fashion, requiring in situ restimulation via peptide-MHC complexes on bone marrow-derived professional APCs for CD4+ TEFF cells and on directly infected tissue cells for CD8+ TEFF cells. Consistent with a critical involvement of DCs in peripheral CD4+ TEFF-cell responses, we observed a pronounced accumulation of monocyte-derived inflammatory DCs in infected skin. Such DCs are thought to exert multiple functions, including the local production of inflammatory mediators [42], trafficking of antigen to lymph nodes [43] and replenishment of peripheral DC populations following the resolution of infection [35], [44]. In addition, our experiments employing genetic approaches, ex vivo stimulation assays and costimulation blockade provide compelling evidence that DCs in inflamed skin play a key role in stimulating CD4+ T-cell effector activity. These results reinforce the concept that DCs regulate various aspects of peripheral T-cell responses [2], [15], [20], [21], [45], [46]. In fact, the presence of DCs appeared to be essential for CD4+ TEFF-cell responses as non-professional APCs, such as keratinocytes and DETCs, largely lacked MHC-II expression and failed to elicit IFN-γ production in ex vivo assays. It is possible that certain APC subsets may dominate the regulation of peripheral CD4+ TEFF-cell activities, as suggested for CD11chiCD11bhi dermal DCs after skin injection of model antigens [21] or CCR2-depedent monocyte-derived inflammatory DCs during mucosal HSV-2 infection [20]. Nevertheless, our study revealed a considerable degree of redundancy in this regard, with various DC populations from skin and LNs displaying strong stimulatory capacities for CD4+ TEFF cells. These results imply that all DC subsets, relative to their abundance in infected skin, contribute to CD4+ TEFF activation in vivo. Consistent with this, we observed normal IFN-γ+ production in mice deficient in specific APC subsets, such as monocyte-derived DCs, Langerhans cells, CD103+ dermal DCs or B cells. According to our analysis, monocyte-derived inflammatory DCs are by far the most abundant DC subtype in HSV-infected skin [35] and therefore, may be the major drivers of peripheral CD4+ TEFF-cell responses during HSV-1 skin infection. Nevertheless, our results demonstrate that they may not be essential in this regard as other DC subsets may compensate for their absence.
The IFN-γ response by CD4+ T cells occurred in both draining LNs and the epithelial and dermal layers of infected skin, including regions a considerable distance away from infection foci in the epithelium. This remarkably broad distribution echoes the diverse functions of CD4+ TEFF cells in infection control, ranging from the initiation of antibody class-switching in LNs to the regulation of inflammatory cell infiltration and activity as well as direct antimicrobial effects within infected tissues [5]. Interestingly in this respect, IFN-γ can exert long-range effects on target cells located as far as 80 µm from CD4+ TEFF-cell-APC conjugates, as recently shown for skin infection with Leishmania major [47]. Thus, DC-mediated CD4+ TEFF-cell activation in the dermis may be an essential component of the host defense that restricts infection to the skin epithelium and limits its spread after virus reemergence in sensory nerve endings. The importance of the CD4+ TEFF-cell response is further illustrated by the lack of CD8+ TEFF-cell IFN-γ production in the dermis, as shown here. Supporting this notion, CD4+ TEFF-cell responses are thought to dominate the clearance of HSV-1 from the skin [22], [23], most likely via antibody-independent functions such as direct inflammatory and antiviral activities [5], [48], [49].
In striking contrast to the stimulation requirements for CD4+ TEFF cells, we identified nonprofessional APCs, such as keratinocytes and DETCs, as the main drivers of IFN-γ production by CD8+ TEFF cells. According to our histological analysis, keratinocytes are the most abundant cell type in infected epidermis suggesting that they may be largely responsible for activating CD8+ TEFF cells. In addition, various types of inflammatory cells that infiltrate the epithelial layer during infection may contribute to this response. Importantly, both keratinocytes and DETCs are highly susceptible to direct infection by HSV-1 in vivo [30], [50]. Indeed, their stimulatory capacity, and surprisingly also that of epidermal DCs, was strictly dependent on direct infection. DETCs are invariant γδ-T cells that form a dense network in the epidermis of mice and have been implicated in both innate and adaptive immune responses [51]. Interestingly, it has been speculated that human γδ-T cells may act as professional APCs capable of stimulating naïve CD4+ T cells [52], although we did not find evidence supporting a similar role for DETCs, as they failed to activate CD4+ TEFF cells. Nevertheless, their contribution to CD8+ T-cell responses may be particularly relevant at lesion borders where high numbers of infected DETCs [50] could elicit strong IFN-γ responses required to curb the lateral spread of infection. In addition to DETCs, other epidermal-infiltrating T cells may also act as potent APCs for local CD8+ TEFF cells upon infection with virus [53]. A role for infected T cells and DCs in triggering local CD8+ TEFF-cell activity is further supported by our observation that chimeric mice, in which only radiosensitive APCs could activate CD8+ T cells, had a residual IFN-γ response.
Despite the fact that all DC subsets could activate CD4+ TEFF-cells, only HSV-infected epidermal DCs were capable of activating CD8+ TEFF cell to produce IFN-γ. Strikingly, uninfected DCs from the same location failed to stimulate CD8+ TEFF cells, highlighting the importance of direct infection in determining the outcome of CD8+ TEFF-cell-DC interactions. Previous studies have suggested that various types of professional and nonprofessional APCs, including inflammatory DCs and neutrophils, can trigger IFN-γ production by CD8+ TEFF cells during pulmonary infection with influenza virus [15], [18]. Given the ability of influenza virus to infect a broad range of target cells in addition to its primary tropism for lung epithelial cells [17], it is tempting to speculate that in this situation, direct infection may also be required for CD8+ T-cell activation. Supporting this notion, a large number of inflammatory DCs and neutrophils from influenza virus-infected lungs express viral antigens, most likely as a consequence of direct infection [15], [18], and infected neutrophils display a far superior ability to elicit CD8+ TEFF-cell cytokine production than their uninfected counterparts [18].
One of the more surprising findings from our study was that uninfected DCs failed to elicit IFN-γ production by CD8+ TEFF cells, even though they had access to viral antigen and efficiently activated CD4+ TEFF cells. We have previously shown that LN DCs are able to activate naïve CD8+ T cells at various stages during skin infection, unequivocally demonstrating that they present viral antigens in the context of MHC-I molecules [3], [4], [54]. The CD8α+ and CD103+ DC subsets are of particular interest in this regard, because those DCs cross-present viral antigens in draining LNs 5 days after HSV-1 infection [4], corresponding to the time point analyzed here. Despite this, CD103+ DCs from skin and LNs failed to trigger IFN-γ production by CD8+ TEFF cells. This finding parallels observations with CD8+ T-cell responses to migratory DCs after influenza virus infection, where naïve but not memory T cells proliferate in response to antigen presented on migratory DCs [55].
Although not directly addressed in our study, it is tempting to speculate that the inability of uninfected DCs to activate CD8+ TEFF cells may be a means to prevent their elimination by triggering cytotoxic effector functions. While T-cell killing of DCs has been proposed to represent a negative feedback regulation on T-cell priming [56], [57], it should be noted that prolonged antigen presentation is a common feature in a variety of infections [58]–[61]. Therefore, CD8+ T-cell-mediated elimination of DCs in vivo may be inefficient at best, which is also consistent with our preliminary data demonstrating the failure of gBT-I TEFF cells to lyse CD103+ DCs from LN of HSV infected mice during short-term co-culture. The DC-dependent production of IFN-γ by CD4+ TEFF cells observed in our study further supports this assumption. Of note in this respect, minute amounts of surface antigen can trigger T-cell cytotoxicity [62], [63]. Furthermore, cross-presenting nonprofessional APC, such as liver sinusoidal endothelial cells, can drive CD8+ T-cell TNFα production during viral hepatitis [64]. Thus, it appears unlikely that quantitative differences in antigen presentation between directly infected and cross-presenting DCs alone can explain the CD8+ TEFF-cell unresponsiveness described here. In the light of this, and given that infected epidermal DCs were indeed able to trigger CD8+ TEFF-cell IFN-γ production, we speculate that direct infection may alter the functional status of DCs, for instance through interference with putative inhibitory pathways [65], to allow for the activation of CD8+ TEFF cells. Such modulation of DC stimulation thresholds may be particularly relevant when low levels of peptide-MHC-I complexes are presented to CD8+ TEFF cells. Further studies will be required to elucidate the precise molecular mechanisms operating in DCs and/or T cells to prevent the activation of CD8+ TEFF cells by uninfected cross-presenting DCs.
Overall, the stringent temporal, cellular and molecular constraints on TEFF-cell responses identified in our study are likely in place to prevent collateral damage and autoimmune inflammation initiated by TEFF-cell activation in tissues not involved in infection. Our results are compatible with a scenario where CD8+ T-cell responses are strictly focused on infected tissue compartments, whereas CD4+ responses may induce a more regional state of antimicrobial protection in tissues surrounding infection foci. Having identified dramatically distinct requirements for CD4+ and CD8+ T-cell effector activity, our study has provided novel insights into the regulation of cellular immune responses in nonlymphoid tissues. Such knowledge has the potential to guide the development of T-cell subset-specific approaches for therapeutic and prophylactic intervention in antimicrobial immunity and autoimmunity.
All experiments were done according to Australian NHMRC guidelines contained within the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes and under approvals ID1112038 and ID1112345 from the University of Melbourne Animal Ethics Committee.
C57BL/6, B6.SJL-PtprcaPep3b/BoyJ (B6.CD45.1), gBT×B6.CD45.1 (gBT-I.CD45.1), gBT-I.EGFP, gBT-I.DsRed, OT-I×B6.CD45.1 (OT-I.CD45.1), gDT-II×B6.CD45.1 (gDT-II.CD45.1), gDT-II.EGFP, C57BL/6Ji-Kbtm1N12 (H-2Kb−/−), B6.129S2-H2dlAb1-Ea/J (MHC-II−/−), CD11c.DTR, B6129.CCR2(Ccr2−/−), Langerin.DTR.EGFP (Lg.DTR.EGFP) and μMT mice were bred in the Department of Microbiology and Immunology. gBT-I.CD45.1 and gDT-II.CD45.1 are F1 generation offspring expressing both CD45.1 and CD45.2.
HSV-1 KOS, KOS pCMV/EGFP Cre (HSV.GFP) and 17 HSVgDUL47ΔYFP (HSV.CFP) were grown and titrated as previously described [46]. HSV.GFP expresses an EGFP/Cre fusion gene under the control of the CMV-IE promoter from the intergenic space between UL3 and UL4 resulting in EGFP expression by infected cells. HSV.CFP was derived from HSV-1 gDUL47 (strain 17 expressing YFP from UL47 and CFP-tagged gD protein) [66], from which the YFP was removed and sequences from UL47 restored. HSV.GFP and HSV.CFP were made by homologous recombination between parent viral genomes with appropriate transfer plasmids in 293A cells after which green/yellow fluorescence were selected for and against, respectively. Final recombinants were verified by PCR and sequencing of relevant parts of the genome after at least three rounds of plaque purification.
Mice were irradiated with two of doses 550 cGy 3 hours apart followed by reconstitution with 5×106 T-cell-depleted donor bone marrow cells and treatment with purified anti-Thy1 antibody (clone T24/31.7 from hybridoma supernatants) 1 day later. Chimeric mice were allowed to reconstitute for at least 8 weeks before experiments. WT→I-A/E−/− chimeras were not treated with anti-Thy1 antibody and received 5×106 enriched splenic CD4+ T cells from wild-type mice 1 day and 4 weeks after irradiation.
Mice were infected on their flanks with 1×106 plaque-forming units of HSV-1, as previously described [25]. For DNFB treatment, 15 µL of 0.5% (w/v) DNFB was applied to flank skin, as previously described [67]. For depletion of CD11c+ cells, CD11c.DTR mice were injected with 200 ng diphtheria toxin (DT) or PBS intraperitoneally and intradermally 4 d post-infection. Langerin.DTR mice were injected with 500 ng DT or PBS control intraperitoneally. For costimulation blockade, mice were injected with 0.25 mg anti-CD80 (16-10A1) and -CD86 (GL1) blocking antibodies or rat IgG2 (B81-3) and IgG2a (R35-95) control antibodies (BD Pharmingen) intraperitoneally 4 d post-infection.
gBT-I and gDT-II cells were isolated from lymphoid tissues and gDT-II cells were further enriched by positive and negative selection using magnetic beads, as described previously [4]. 5×104 gBT-I or 1×104 gDT-II cells were transferred into naïve mice intravenously via the tail vein, respectively. For transfer of in vitro activated T cells, 1.5×106 transgenic cells were injected.
Splenic gBT-I or OT-I cells were activated with peptide-pulsed splenocytes, as previously described [25]. Purified gDT-II cells were activated by co-culture for 5 days with 4×107 irradiated wild-type splenocytes pulsed with 10 µM of gD315–327 in the presence of 2 µg LPS (Sigma). Cells were cultured in 20 ml RPMI 1640 (Department of Microbiology and Immunology) supplemented with 10% FCS (CSL), 5 mM HEPES (Gibco), 2 mM glutamine (Gibco), 5×10−5 M 2-β-mercaptoethanol (Sigma), antibiotics (Gibco, CSL) (RP-10) and 4 µg lipopolysaccharide. On days 2–4, respectively, all cultures were diluted 1∶2 in fresh medium containing 20 U/mL recombinant human IL-2 (PeproTech).
As indicated, skin tissue was chopped and digested with 3 mg/ml collagenase type 3 (Worthington Biochemicals, USA) and 5 µg DNase (Roche, Germany) for 90 min at 37°C. Alternatively, skin was digested in 2.5 mg/mL dispase II (Roche) diluted in PBS for 90 min at 37°C. Then, the epidermis and dermis were separated mechanically and epidermal sheets were incubated in trypsin/EDTA (0.25%/0.1%) (SAFC Biosciences), while the dermis was chopped and incubated in collagenase type 3 and DNase, as previously described [49]. When analyzing ex vivo IFN-γ production, 10 µg/mL Brefeldin A (Sigma) was included during each enzymatic digestion step. For some experiments, mice were additionally injected with 0.25 mg BFA intravenously 6 hours prior to sacrifice.
Axillary lymph nodes of HSV-1-infected mice were desiccated with a scalpel blade and digested with continual mixing in RPMI 1640 containing 1 mg/mL collagenase type 3 and 2 µg/mL DNase for 20 min, prior to the addition of 600 µL 0.1 M EDTA and continual mixing for 5 minutes further. DCs were subsequently enriched for by magnetic beads, as previously described [4].
APC subsets were stained with the appropriate monoclonal antibodies, purified by cell sorting using a FACSAria III (BD Pharmingen) and then washed and resuspended in RP-10. Increasing concentrations of the APCs were cultured with 1.25×104 in vitro activated transgenic T cells in round bottom plates for 5 to 18 hours, as indicated, in the presence of 10 µg/mL BFA for the last 5 hours.
Antibodies were from BD Pharmingen: anti-CD3 (145-2C11), -CD4 (RM4-5), -CD8α (53-6.7), -CD11b (M1/70), -CD19 (ID3), -CD45.1 (A20), -CD80 (16-10A1), -CD86 (GL1), -IFN-γ (XMG1.2), -Ly6C (AL21), -NK1.1 (PK136), -Vα2 (B20.1) and -Vβ8 (MR5-2); from eBioscience: anti-CD45.2 (104) and -CD11c (N418); or from BioLegend: anti-CD326 (g8.8) and -Vα3.2 (RR2-16). For intracellular staining, cells were fixed with a Cytofix/Cytoperm kit (BD Pharmingen). A FACSCanto II (BD Pharmingen) and FlowJo software (TreeStar) were used for analysis. Propidium iodide (Sigma Aldrich) and SPHERO calibration particles (BD Pharmingen) were added for identification of live cells and enumeration.
Skin was fixed at room temperature for 30 min in PLP buffer (0.2 M NaH2PO4, 0.2 M Na2HPO4, 0.2 M L-lysine and 0.1 M sodium periodate with 2% paraformaldehyde), washed twice with PBS and incubated for 30 min in 20% sucrose, prior to being embedded, frozen, cut and stained as previously described [48]. IFN-γ staining (AlexaFluor647, BD Pharmingen) was performed overnight at 4°C (1∶75 in PBS containing 2.5% [w/v] donkey serum). Anti-keratin-5 and -14 polyclonal antibodies were from Jomar Biosciences; anti-CD4 (RM4-5) from BioLegend; anti-CD8 (53.67) from BD Pharmingen and polyclonal anti-HSV from Dako North America. Slides were mounted with ProLongGold antifade media (Invitrogen), air-dried and viewed using a Zeiss LSM700 confocal microscope and Imaris 7.1 software (Bitplane).
Mice were anesthetized and HSV-1-infected flank skin was mounted on an imaging platform and acquired with an upright LSM710 NLO multiphoton microscope as described previously [48]. Imaging data was processed and automatic cell tracking aided by manual corrections was performed with Imaris 7.1 software. For movies, image sequences were composed in Adobe After Effects CS5.
Graphs were plotted using Prism 5 (Graphpad) and comparison of data sets was performed by one-way analysis of variance followed by Tukey post-test, or Mann-Whitney or student t tests, as indicated. All graphs depict means ± s.e.m..
|
10.1371/journal.ppat.1000075 | Intraspecies Transmission of BASE Induces Clinical Dullness and Amyotrophic Changes | The disease phenotype of bovine spongiform encephalopathy (BSE) and the molecular/ biological properties of its prion strain, including the host range and the characteristics of BSE-related disorders, have been extensively studied since its discovery in 1986. In recent years, systematic testing of the brains of cattle coming to slaughter resulted in the identification of at least two atypical forms of BSE. These emerging disorders are characterized by novel conformers of the bovine pathological prion protein (PrPTSE), named high-type (BSE-H) and low-type (BSE-L). We recently reported two Italian atypical cases with a PrPTSE type identical to BSE-L, pathologically characterized by PrP amyloid plaques and known as bovine amyloidotic spongiform encephalopathy (BASE). Several lines of evidence suggest that BASE is highly virulent and easily transmissible to a wide host range. Experimental transmission to transgenic mice overexpressing bovine PrP (Tgbov XV) suggested that BASE is caused by a prion strain distinct from the BSE isolate. In the present study, we experimentally infected Friesian and Alpine brown cattle with Italian BSE and BASE isolates via the intracerebral route. BASE-infected cattle developed amyotrophic changes accompanied by mental dullness. The molecular and neuropathological profiles, including PrP deposition pattern, closely matched those observed in the original cases. This study provides clear evidence of BASE as a distinct prion isolate and discloses a novel disease phenotype in cattle.
| For approximately two decades, bovine spongiform encephalopathy (BSE), now termed classical BSE (BSE-C), has been regarded as the only and exclusive prion disorder affecting cattle. However, over the last 4 years, two additional bovine prion strains, bovine amyloidotic spongiform encephalopathy (BASE, also named BSE-L) and BSE-H, have been discovered and characterized in Canada, the United States, Japan, and nine European countries, which applied an active surveillance program on slaughtered cattle. Although a total of 20 BSE-L and 16 BSE-H have been reported to date, the disease phenotype of these conditions remains largely unknown. Intriguingly, recent evidence has been provided that the BSE-C and BASE strains disclose converging properties after transmission to inbred mice. Here, we show that intraspecies transmission of BASE induces a disease phenotype characterized by dullness and progressive amyotrophy, the latter highly suggestive of a motor neuron disorder. This is at variance with the over-reactivity and hypersensitivity, but not muscle changes, observed in BSE-transmitted cattle. This study confirms that BASE and BSE represent two distinct prion disorders in cattle with diverging molecular features and disease phenotypes.
| Prion diseases, or transmissible spongiform encephalopathies (TSEs), are mammalian neurodegenerative disorders of sporadic, genetic, or infectious origin characterized by accumulation and deposition of an abnormal isoform (PrPTSE) of the cellular prion protein (PrPC) in the brain [1]. TSEs include a wide range of animal and human disorders, such as BSE in cattle, scrapie in sheep and goats, chronic wasting disease in deer and elk, and Creutzfeldt-Jakob disease (CJD) in humans [1].
First identified in 1986 in the UK, BSE has been confirmed in over 180,000 cases, although more than one million cattle have been estimated to be infected [2]. Evidence of the spread of the BSE agent across certain mammalian species, including humans, indicates that this disease is a major animal and human public health issue [3]–[5]. Common neurological signs in cattle include apprehension, hyperaesthesia, kicking, and pelvic limb ataxia, accompanied by general signs such as reduced milk yield and loss of conditions. In all cases, progression to behavioural, sensory and posture/movement alterations led to death within a few months [6],[7].
Early transmission studies showed that isolates from field BSE cases and variant CJD (vCJD), its human counterpart, were all caused by a single prion strain [4]. In addition, PrPTSE from BSE and vCJD cases exhibited a distinctive glycotype signature, with high glycosylation site occupancy and similar electrophoretic mobility of the unglycosylated protease-resistant PrPTSE fragment [8]. These PrPTSE traits have been used as biochemical indicators of the BSE prion strain.
Until recently, monitoring of BSE in cattle was accomplished by passive surveillance and pathological confirmation of suspected clinical cases. In 2001, the European Community imposed an active surveillance system based on biochemical tests of brain tissues from all slaughtered cattle over 30 months of age. This strategy led to the recent identification of new PrPTSE types, provisionally termed as “type-H” and “type-L” according to the electrophoretic migration of the unglycosylated proteinase K-resistant PrPTSE, which is higher (BSE-H) or lower (BSE-L) than classical BSE (BSE-C) [9]–[11]. An additional distinctive signature of type-H and type-L is the even representation of di-, mono-, and unglycosylated PrPTSE species.
In 2004, we described two aged asymptomatic Italian cattle of Piemontese and Alpine brown breeds neuropathologically characterized by the presence of PrP-amyloid plaques [12]. This new pathological phenotype, named BASE, was characterized by marked involvement of olfactory areas, hippocampus, and thalamus, with relative sparing of the brainstem. The molecular signature of BASE PrPTSE was similar to that later detected in BSE-L cases [11]. A feature shared between BASE, a condition both well-defined molecularly and pathologically, and “L-type” cases, defined only on a molecular basis, is the older age of the affected animals (approximately 12 years) as compared to BSE cases (5–6 years).
Recent studies have shown that BASE and “L-type” isolates exhibit similar biological properties upon transmission to Tgbov XV, and have shorter incubation period and survival time than BSE; these findings are suggestive of a single prion strain for BASE and BSE-L [10],[13]. In contrast, the H-type phenotype showed an unusually long incubation period in Tgbov XV [10].
To date, all of the available demographic and molecular evidence strongly suggests that H-type BSE and BASE-L represent sporadic forms of bovine spongiform encephalopathies [14].
Human susceptibility to BASE has been suggested by experimental transmission to primates and to PrP humanized transgenic mice [14].
Here we inoculated cattle of different breeds with brain homogenates from Italian BASE and BSE cases, in order to assess the strain attributes and disease phenotype of the above isolates in their natural hosts.
All Friesian cattle were homozygous for six octapeptide repeat copies, and three cattle carried a silent mutation at codon 78 (CAG/CAA). Four out of six Alpine brown cattle were homozygous for six octapeptide repeat copies; one animal carried 6/7 and another 5/7 octapeptide repeat copies. Four different silent mutations were found at codons 78 (CAG/CAA), 23 (CTC/CTT), 95 (CCA/CCC), and 77 (GGT/GGC) in four cattle. Homozygosity for 23 bp and 12 bp deletion alleles was present in three Friesian cattle. Results of these genetic studies are summarized in Table 1.
A total of twelve cattle, two groups of three Alpine brown and three Friesian, intracerebrally inoculated with either BSE or BASE, developed neurological signs and were killed at the terminal stage of disease (Table 2). In contrast, two saline inoculated Friesian cattle are free of clinical signs at the time of writing, i.e. 42 months post-inoculation. In BASE-treated cattle the clinical disease duration was shorter than in BSE-inoculated animals; however, caution in evaluating these differences is dictated by the low number of experimental animals in addition to the undetermined infectivity titre of the inocula.
In BSE-inoculated cattle, clinical signs at onset consisted of behavioural changes and hypersensitivity. As the disease progressed, major clinical signs included aggressiveness, frequent bellowing and head shaking, postural abnormalities, exaggerated blink reflex, generalized cutaneous hyperaesthesia, and stimulus-induced myoclonic jerks (Table 2 and Figure 1A and Video S1).
Conversely, early neurological signs in both Friesian and Alpine brown cattle inoculated with BASE consisted of fasciculations of gluteal muscles, a dull coat and postural and behavioural signs of depression, including low head carriage, mild kyphosis, and decreased alertness. With progression, muscle atrophy, beginning in the gluteal region and progressing to the paravertebral region and to other hind limb musculature became apparent (Figure 1B and 1C). Fore-limb muscles were relatively spared (Video S2). With the exception of the “downer” cattle, neither gait ataxia nor difficulties in rising were observed throughout the disease course.
Cattle showed an exaggerated response to facial touch or pinch, but not to light and sound stimuli. Observations via night filming showed that BASE cattle were prone to sudden falls. One Friesian cow (code # 254) showed a “downer” syndrome at onset.
Immunoblot analysis of proteinase K-treated (PK) brain homogenates from each BSE- and BASE-infected cattle revealed the presence of PrPTSE in all sampled areas. However, all BSE-challenged animals showed a di-glycosylated-dominant PrPTSE type, whereas in all BASE-inoculated cattle a mono-glycosylated-dominant PrPTSE type was detected. In addition, the molecular mass of the PK-resistant unglycosylated fragment was identical to that of the original inoculum in each animal (Figure 2A–2C). In cattle infected with BSE, the highest amounts of PrPTSE were observed in the thalamus, basal ganglia, obex, olfactory areas and hippocampus, whereas very low amounts were seen in cerebral cortices and cerebellum (Figure 2D and 2E). Differently from BSE, in BASE-infected cattle consistently high amounts of PrPTSE were observed in cerebral cortex, hippocampus, and cerebellum (Figure 2F and 2G). In both groups, low amounts of PrPTSE were found in the spinal cord.
In all experimentally infected animals, no PrPTSE was detected in peripheral tissues, including cervical and mesenteric lymph nodes, spleen, thymus, liver, lung, peripheral nerves and forelimb and hind limb muscles, either by standard Western blot analysis or following phosphotungstic acid precipitation.
Typical neuropathological changes, including spongiosis and gliosis were detected in all cattle (Figure S1 and S2). The conventional lesion profile, based on vacuolation score, was similar in BSE- and BASE-infected cattle; however, a more severe involvement of central grey matter (periaqueductal grey) and rostral colliculus but not the vestibular nuclear complex were observed in BASE-inoculated cattle as compared to BSE-challenged animals, which showed severe involvement of the putamen (Figure S1). Additional brain areas, including the olfactory areas, amygdalae, hippocampi and dorsal horns of spinal cords, were severely involved in both groups. Ventral and dorsal roots did not show major pathological changes.
Friesian and Alpine brown muscle tissue was normal in BSE-infected cattle (Figure 3A, 3C, 3E and 3G), whereas groups of atrophic muscle fibers were observed in the gluteus medius (Figure 3B, 3D and 3F) and, to a decreasing extent, in major psoas, longissimus dorsi, and triceps brachii of BASE-infected cattle (Figure 3H).
In BSE cattle, a synaptic-punctate and “glial-associated” stellate pattern of PrP deposition was observed in different brain areas, including olfactory areas, cerebral cortex, basal ganglia, thalamus, cerebellum, medulla, and spinal cord (Figure 4A and inset, 4E, 4G, 4I and inset). Conversely, in BASE-inoculated cattle, abundant amyloid PrP plaques were observed in subcortical white matter and in deep grey nuclei, as observed in natural BASE cases (Figure 4B and inset, 4F; and Figure S2 and S3). No PrP plaques were seen in the olfactory glomeruli, the cerebellum or the spinal cord (Figure 4I and 4J). Neurons from BSE cattle showed intracellular PrP deposition in contrast to the membrane-associated deposits observed in neuronal cells of BASE cattle (Figure 4C and 4D). These patterns of PrP neuronal staining were also seen in ventral horn neurons of the spinal cord (Figure 4I and J insets) and in the dorsal root ganglion cells (data not shown). No PrP staining was detected in the peripheral nerves and muscles.
In the present work, we demonstrate that BSE and BASE isolates maintain distinct biological properties and induce different disease phenotypes after transmission in their natural host. The similarity of the molecular typing differences between BASE and BSE PrPTSE in Friesian and Alpine Brown cattle also supports the notion that the two conditions are caused by different prion strains.
Cattle inoculated with BASE developed a syndrome characterized by progressive muscle atrophy and behavioural changes. Amyotrophic changes were preceded by fasciculations, findings denoting a lower motor neuron deficit. The absence of anorexia or difficulty in feeding and swallowing suggests that amyotrophy may be caused by motor neuron dysfunction and, therefore, not indicative of a generalized wasting syndrome, such as that observed in chronic wasting disease [15]. Consistent with clinical findings of lower motor neuron involvement, pathological examination of muscle tissues disclosed groups of atrophic fibers more frequently detected in proximal than distal hind limb muscles. However, there was no convincing loss of ventral horn neurons.
Pathogenic mechanisms leading to motor neuron dysfunction remain unknown; however, the role of pathological PrP deposition at the plasma membrane of motor neurons or a loss of PrP function, as observed in experimental models of amyotrophic lateral sclerosis, cannot be ruled out [16].
Clinical signs of motor neuron dysfunction, including stiffness, posterior paresis with “clonic spasms of muscle bundles” [17] and generalized weakness, accompanied by severe lethargy and ataxia, were previously reported in cattle experimentally infected with American strains of sheep scrapie (either at first or at second passage). Cattle inoculated with pre-1975 and post-1990 sources of sheep scrapie from the UK presented similarly with ataxia and weakness and most showed dullness with low head carriage and did not over react to external stimuli [17]–[19]. While the clinical characterization described previously in cattle infected with scrapie is suggestive of upper and lower motor neuron involvement, results obtained in BASE cattle point to lower motor neuron dysfunction or to peripheral neuropathy as the cause of amyotrophic changes.
In contrast to the amyotrophic changes observed in BASE-inoculated cattle, animals inoculated with BSE presented a disorder characterized by apprehension and hypersensitivity to external stimuli, overlapping clinical features described in early accounts of UK BSE [20].
Molecular features of PrPTSE from BASE and BSE donor cattle were preserved with high fidelity in recipient animals. In particular, the conformation of PrPTSE, as assessed by the electrophoretic motility of the core fragment, and the glycosylation status were indistinguishable in recipient animals of different breeds compared to the original inocula. These PrPTSE traits were maintained in all cortical and subcortical investigated brain regions.
Taken together, PrPTSE molecular traits and PrPTSE regional distribution showed distinct patterns in the two groups of animals, supporting the notion of two different prion strains, as also suggested by results from experimental transmission to transgenic mice expressing bovine PrP [10],[13].
Differences between BASE- and BSE-inoculated animals were also observed at the neuropathological and immunohistochemical levels. At variance with original BASE cases, where no spongiform changes were observed, as a likely effect of early disease stage, marked vacuolation was seen in BASE-inoculated cattle with a lesion profile divergent from that seen in BSE-treated animals in at least four regions. Indeed, extensive vacuolar pathology was seen in the hindbrain of BASE-inoculated cattle, whereas severe involvement of the putamen was a distinguishing feature in the BSE group.
The divergent biological properties of the two strains were further confirmed by the different patterns of PrP deposition, indistinguishable from those patterns seen in naturally occurring BSE and BASE. Moreover, the distinct neural and microglial cells involved in the two groups and the subcellular sites of PrP accumulation denote a different trafficking and propagation of PrP.
Taken together, intraspecies transmission of BASE and BSE recapitulated the key neuropathological hallmarks observed in these naturally occurring cattle TSEs. This is at variance with the alternate patterns of PrPTSE depositions seen in inoculated Tgbov XV mice, i.e., uni- and multicentric plaques in BSE-challenged animals and diffuse/focal PrP deposition, but not amyloid plaques, in BASE-inoculated mice [10],[13].
We recently showed that in TgBov XV mice challenged with the same inocula used in the present study, BASE-inoculated mice had significantly shorter incubation periods and survival times than BSE-inoculated mice, consistent with results from another laboratory [10],[13]. This effect was not influenced by any species barrier phenomenon and is therefore likely to be strain-dependent. An exception to diverging phenotypic characteristics observed in BSE- and BASE-inoculated cattle was the incubation time observed for BSE-inoculated Alpine brown (but not BSE-treated Friesian animals) which did not significantly differ from times assessed for BASE-inoculated Friesian and Alpine brown cattle.
However, the small number of investigated animals and the individual variability in incubation times dictate caution in the interpretation of these data. In contrast, the breed-associated effect in BASE-inoculated cattle, with significantly shorter incubation periods and survival times in Friesian than in Alpine brown, suggests that disease-modifier genetic loci other than known PRNP polymorphisms could be relevant to both of these parameters.
While it is now clear that vCJD originated from human exposure to BSE, it is still uncertain whether emerging cattle TSEs, including BASE, or L-BSE, and H-BSE have infected humans or to which extent they can be potentially dangerous for human and animal health. Recent experimental data show that the BASE strain is efficiently transmitted to Tgbov XV mice and to TgOv mice; in the latter, BASE transmits at first passage with a 100% attack rate, as opposed to cattle BSE that transmits with a low attack rate [21]. Moreover, transmitted BASE shows shorter incubation periods than BSE in Cynomolgus monkeys [14]. Paradoxically, while BASE is efficiently transmitted at first passage and with a high attack rate to 129 Met/Met Tg humanized mice [22], human transgenic lines of all genotypes at codon 129 are resistant to BSE transmission [22],[23]. Taken together, these data might suggest that the BASE agent could transmit to humans more efficiently than the BSE agent.
All procedures involving animals and their care were conducted in conformity with national and international laws and policies (EEC Council Directive 86/609, OJL358, 1, 12 December 1987; Italian Legislative Decree 116/92, Gazzetta Ufficiale della Repubblica Italiana 10, 18 February 1992; and Guide for the Care and Use of Laboratory Animals, U.S. National Research Council, 1996), and the study was approved by the authors' Institutional Review board.
PRNP ORF amplification, sequencing and determination of the octapeptide repeat copy number was performed as previously described [12]. Polymorphisms of the 12-bp indel, located within intron 1, and 23-bp indel, located in the promoter region, were determined as previously reported [24]. Primer pairs 5′-CCTGTTGAGCGTGCTCGT/5′-ACCTGCGGCTCCTCTACC-3′ and 5′-GAAGTCACGTGAAGGCACT-3′/5′-CAAAGAGTTGGACAGGCACA-3′ were used to amplify the 12-bp indel (202 bp/214 bp) and 23-bp indel (167 bp/190 bp), respectively, as described above. PCR was performed as 30 cycles of 30 sec at 94°C, 30 sec at 55°C and 45 sec at 72°C. High resolution agarose (3.5%) gel electrophoresis was used to visualise the allelic PCR products whose specificity and length was also confirmed by direct sequencing with the same primer used for the PCR described above.
10% brain homogenates from the thalamus of a BSE-affected Friesian (code #128204) and a BASE-affected Piemontese (code #1088) were prepared in phosphate-buffered saline. These cattle carried the same PrP genotype with six octapeptide repeats, and were extensively studied in our previous work [12]. BSE and BASE inocula were prepared to obtain a comparable amount of PrPTSE as assessed by Western blot analysis with the 6H4 anti-PrP monoclonal antibody (Prionics).
Eight Friesian and six Alpine brown cattle (4 months old) were purchased from Italian herds in which no cases of BSE had ever been recorded. All calves were free of neurological signs. Prior to inoculation, animals remained in the new environment for one month for adaptation. Inoculation was carried out using a semi-stereotaxic technique in surgical aseptic conditions. Calves were anesthetized with xylazine (50 μg/kg), a midline incision was made at the junction of the parietal and frontal bones, and a 1-mm hole was drilled through the calvarium. The inoculum was injected into the frontal lobe via a 22-gauge 9-cm-long disposable needle while the needle was withdrawn. Two groups of animals, each comprising three Friesians and three Alpine brown cattle, were inoculated with 1 milliliter of 10% brain homogenate from BSE-and BASE-affected animals, respectively. Conversely, two Friesians cattle were challenged with 1 milliliter of phosphate buffered saline and used as controls. To avoid potential cross-contamination, BSE and BASE transmission experiments were performed on different days and the facility was decontaminated with 10% sodium hypochlorite solution after each inoculation.
Clinical evaluations were comprised of a bi-weekly observation by the veterinarian and two daily observations by animal husbandry staff who reported any observed motor and/or behavioural changes. For assessment of the gait cows were walked along the corridor outside the pen. The cattle were filmed nightly with closed circuit television monitoring to record signs of disease that may not have been observed during the day. Once a month, a veterinarian trained in neurology examined each cattle by conventional neurological scale evaluations [25],[26]. Animals were considered symptomatic when they showed two of the following neurological signs observed in two separate consecutive examinations: abnormal behaviour, abnormal posture, aberrant reactions, or hyperreactivity to sensitive stimuli, light and sound.
Cattle at the terminal stage were euthanized with pentobarbital administered intravenously. Peripheral organs, including cervical and mesenteric lymph nodes, spleen, thymus, liver, lung, peripheral nerves and forelimb and hind limb muscles (m. triceps brachii, m. longissimus dorsi, m. gluteus medius and m.major psoas), were sampled and each sample was divided equally; one portion was fixed in 4% buffered formaldehyde for H&E stain and PrP immunohistochemistry, and the other was frozen. Serial 10-μm-thick muscle cryosections were stained with H&E and adenosine triphosphatase (ATPase), after pre-incubation at pH 4.3, 4.6 and 10.4. Nervous tissue was removed in a separate area to avoid cross-contamination. The fixed half of the brain sample was used for neuropathological examination, while the frozen brain sample was stored at −80°C for biochemical analyses. The spinal cord was sampled at cervical, thoracic, lumbar and sacral levels and sections were fixed in 10% buffered formaldehyde. The remaining tissue was frozen at −80°C for further studies.
From each neural tissue sample, including optic nerve, olfactory bulb, frontal cortex, occipital cortex, hippocampus, nucleus caudatus, putamen, globus pallidus, thalamus, cerebellum, obex, and cervical, thoracic, lumbar and sacral spinal cord, 100 mg of wet tissue was homogenized in 9 volumes of lysis buffer (100 mM sodium chloride/10 mM EDTA/0.5% Nonidet P-40/0.5% sodium deoxycholate/10 mM Tris, pH 7.4) and digested with 50 μg/ml of proteinase K (Boehringer Mannheim) for 1 h at 37 °C. Digestion was blocked by the addition of phenylmethylsulfonyl fluoride at 2 mM. For deglycosylation, proteinase K-digested samples were deglycosylated with recombinant peptide N-glycosidase F (PNGase F) according to manufacturer's instructions (Boehringer Mannheim). Samples, equivalent to 400 μg of wet tissue, were resolved on 13% polyacrylamide gels and then transferred onto PVDF membrane (Immobilon P; Millipore, Bedford MA) for 2 hours at 60V. Membranes were blocked with 1% non-fat dry milk in TBST (10 mM Tris/150 mM sodium chloride/0.1% Tween-20, pH 7.5) for 1 hour at 37°C and incubated overnight at 4°C with anti-PrP monoclonal antibody 6H4 (Prionics) diluted to 1/5,000. Blots were developed using the Amersham enhanced chemiluminescence (ECL) system, as described by the supplier and visualized on an autoradiographic film. Films were scanned by using a densitometer (GS-710, Biorad), calculating the relative amounts of PrPSc in a semiquantitative manner.
To enhance PrPTSE detection, extraneural tissues, including cervical and mesenteric lymph nodes, spleen, thymus, liver, lung, peripheral nerves and forelimb and hind limb muscles, that were negative on a standard immunoblot were subjected to phosphotungstic acid (PTA) precipitation and analyzed by Western blot, as previously described [27]. Briefly, 100 mg of wet tissue were homogenized in nine volumes of 2% sarkosyl in phosphate-buffered saline, pH 7.4. Cellular debris were removed by centrifugation at 1,000 rpm for 2 minutes and samples were incubated for 30 minutes at 37°C with constant agitation in phosphate-buffered saline containing 50 units/mL Benzonase and 1 mmol/L magnesium chloride. Subsequently, samples were adjusted to 0.3% sodium phosphotungstic acid, incubated at 37°C for 30 minutes and centrifuged at 14,000 rpm for 30 minutes. The supernatant was saved, and the pellet dissolved in 20 μl phosphate-buffered saline, pH 7.4, containing 0.1% sarkosyl. The supernatant and the pellet were adjusted to a final concentration of 20 μg of proteinase K per milliliter and incubated at 37 °C for 30 minutes.
Three-mm thick samples were embedded in paraffin after decontamination with 98% formic acid for 1 hour. The paraffin-embedded blocks selected for the study included coronal sections at the level of the olfactory bulb, the frontal, parietal and occipital cortices, the pyriform lobus, hippocampus, striatum, thalamus, brainstem, sagittal sections through the cerebellum and spinal cord at cervical, thoracic, lumbar and sacral levels. Histological sections were deparaffinized, rehydrated, and stained with hematoxylin and eosin. Additional sections were stained with thioflavin S. The distribution of spongiosis, was determined by using a conventional lesion profile, which allows to characterize strain tropism and to compare the present results with those of previous studies on field and experimental BSE [28]. The definition of each score was performed by three independent observers blinded to the animal identification, as follows: 0 no vacuolation, 1 a few vacuoles (minimum 3 per field×10 objective), 2 several vacuoles evenly distributed, 3 moderate numbers or many vacuoles evenly distributed, and 4 numerous vacuoles some of which coalescing, as previously described [28].
For immunohistochemical study, sections obtained form nervous and extraneural tissues, were rehydrated and treated with 98% formic acid for 20 min at room temperature, followed by hydrated autoclaving in distilled water at 121°C for 30 min. After rinsing, sections were incubated overnight at 4°C with anti-PrP monoclonal antibody F99/97.6.1 (VMRD, inc.; diluted to 1/1,000), recognizing a conserved epitope (QYQRES) on the cattle PrP. Subsequent antibody detection was carried out using a biotinylated goat anti-mouse secondary antibody diluted to 1/200 for 20 min (Vector Laboratories, Burlingame, CA) at room temperature, followed by the avidin-biotin-peroxidase complex (Vectastain ABC kit; Vector Laboratories) according to manufacturer's protocol. Immunoreactivity was visualized using 3,3′-diaminobenzidine as chromogen. |
10.1371/journal.ppat.1003157 | Therapeutic Efficacy of Antibodies Lacking FcγR against Lethal Dengue Virus Infection Is Due to Neutralizing Potency and Blocking of Enhancing Antibodies | Dengue hemorrhagic fever and dengue shock syndrome (DHF/DSS) are life-threatening complications following infection with one of the four serotypes of dengue virus (DENV). At present, no vaccine or antiviral therapies are available against dengue. Here, we characterized a panel of eight human or mouse-human chimeric monoclonal antibodies (MAbs) and their modified variants lacking effector function and dissected the mechanism by which some protect against antibody-enhanced lethal DENV infection. We found that neutralizing modified MAbs that recognize the fusion loop or the A strand epitopes on domains II and III of the envelope protein, respectively, act therapeutically by competing with and/or displacing enhancing antibodies. By analyzing these relationships, we developed a novel in vitro suppression-of-enhancement assay that predicts the ability of modified MAbs to act therapeutically against antibody-enhanced disease in vivo. These studies provide new insight into the biology of DENV pathogenesis and the requirements for antibodies to treat lethal DENV disease.
| The four dengue virus serotypes (DENV1-4) cause the most prevalent mosquito-transmitted viral disease globally, infecting 50–100 million people annually in tropical and sub-tropical regions worldwide, yet no vaccine or therapy has been licensed to prevent or treat dengue. The greatest risk factor for severe dengue disease is a previous infection with a different serotype, which is thought to be due in part to a phenomenon known as antibody-dependent enhancement (ADE) whereby anti-DENV antibodies from a prior infection augment DENV infection of target Fcg receptor (FcgR)-expressing cells. We previously developed a mouse model that demonstrates antibody-enhanced lethal DENV disease and showed that genetically-modified antibodies incapable of interacting with the FcgR eliminate ADE in vitro and in vivo. In this study, we studied a larger panel of modified MAbs that recognize different regions of the DENV envelope protein. While all modified MAbs acted therapeutically to prevent a lethal, virus-only DENV infection, only certain MAbs effectively protected mice following an antibody-enhanced lethal infection. We determined that therapeutically effective MAbs following an ADE infection worked by competing for binding of enhancing antibodies on the DENV virion. Based on this, we designed an in vitro suppression-of-enhancement assay that predicted the ability of modified MAbs to act therapeutically in vivo.
| The four serotypes of dengue virus (DENV) are transmitted by Aedes aegypti and Ae. albopictus mosquitoes and are endemic predominantly in tropical and sub-tropical regions of the world [1], [2]. Syndromes associated with DENV infection range from inapparent infection to classic dengue fever (DF), a debilitating self-limited disease, to life-threatening dengue hemorrhagic fever/dengue shock syndrome (DHF/DSS), characterized by vascular permeability and hypotensive shock [3]. Due to several factors, including geographic expansion of the DENV mosquito vectors and increased global urbanization, trade, and travel [4], [5], there has been a substantial increase in both the incidence of dengue epidemics and co-circulation of the four DENV serotypes in the same region [6]. This has resulted in an increased number of severe cases in dengue-endemic regions previously known for epidemics of only mild disease [1], [7]–[10]. While several tetravalent dengue vaccines are currently in various stages of clinical evaluation [11]–[14], no vaccine or therapy has been licensed to prevent or treat DENV-induced disease.
DENV is a member of the Flavivirus genus and is closely related to other medically important arboviruses including West Nile (WNV), Japanese encephalitis, tick-borne encephalitis, and yellow fever viruses [15], [16]. DENV has a 10.7-kb, positive-sense RNA genome with 5′ and 3′ untranslated regions flanking a polyprotein that encodes three structural and seven non-structural proteins [17]. Among the three structural proteins, the pre-membrane (prM/M) and envelope (E) proteins are the primary antigenic targets of the humoral immune response in humans [18]–[20]. The E protein is comprised of three domains (I (EDI), II (EDII) and III (EDIII) [21]–[24]), with EDII and EDIII containing the fusion peptide [25] and putative viral receptor binding site(s) [26], [27], respectively. For DENV, the most potently neutralizing antibodies generated in mice thus far target two sites on EDIII, corresponding to epitopes on the lateral ridge and A-strand [26], [28]–[31]. However, in human dengue-immune serum after primary DENV infection, highly neutralizing type-specific antibodies appear to be directed to quaternary epitopes on adjacent E proteins present only on virons [32]. A large proportion of human anti-DENV antibodies appear to be cross-reactive and to target the fusion loop or prM [18], [19].
Epidemiological analysis has established that a previous DENV infection is the greatest risk factor for the development of severe disease [33]–[37]. Infection with one serotype is believed to provide life-long immunity against re-infection with the same serotype but does not provide sustained protection against re-infection with a different serotype [38], [39]. Indeed, adaptive B and T cell responses may be poorly inhibitory against re-infection with a second serotype, and in a small percentage (∼1%) of cases, even exacerbate disease. One hypothesis, termed antibody-dependent enhancement, is that antibodies from a previous infection facilitate virus entry into Fcγ-receptor (FcγR)-bearing target cells, thereby increasing viral load and ultimately disease severity [40]. Experimental evidence in cell culture and in animal models supports this concept [41]–[44]. In a mouse model of ADE, passive transfer of monoclonal antibodies (MAb) or polyvalent serotype-cross-reactive serum, when administered at sub-neutralizing concentrations, was sufficient to enhance infection and cause lethal disease with DENV2 strain D2S10 in interferon α/β and γ-receptor deficient (AG129) mice [42], [43]. Recently, we showed that passive transfer of genetically engineered MAbs lacking binding to FcγR and C1q was sufficient to reduce viral load and TNF-α levels and to prevent lethal disease in vivo, even when administered one or two days after infection. Here, we evaluated the therapeutic activity of a larger panel of MAbs targeting different epitopes on the E protein following both a virus-only as well as an antibody-enhanced lethal infection. We determined that the two most potent therapeutic MAbs acted by competitively displacing either fusion-loop specific MAbs or enhancing polyclonal serum antibodies targeting a proximal epitope. Using this information, we designed a novel suppression-of-enhancement assay in human FcγRIIA-expressing K562 cells that predicts the ability of modified MAbs to act therapeutically against antibody-enhanced disease in vivo. Our observations provide new insight into the mechanism by which therapeutic MAbs prevent an antibody-enhanced lethal DENV infection.
Severe forms of DENV infection, including DHF/DSS, can be fatal, as no specific antiviral therapy is currently available. As such, we extended previous observations of the prophylactic and therapeutic efficacy of the EDII fusion loop-specific MAb E60 [42] by studying a larger panel of neutralizing MAbs targeting additional E protein epitopes, including the dimer interface (E44) on EDII and the C-C′ loop (E87) and A-strand (E76 and 87.1) on EDIII (Figure 1A).
Although secondary infection with a different DENV serotype is the greatest risk factor for severe DENV disease, DHF/DSS also has been reported following primary infection [45]. Thus, for a genetically-modified MAb to be a viable therapeutic option, it must protect following both a virus-only and an antibody-enhanced lethal DENV infection. To assess the ability of MAbs to protect in a direct model of lethal DENV infection, AG129 mice were infected with a lethal dose (4×106 PFU) of DENV2 D2S10 and 24 hours later, administered 20 µg of individual genetically modified MAbs lacking the ability to bind FcγR or C1q. Notably, all of the modified MAbs tested prevented development of overt disease and protected against death in this model (P<0.05 for all MAbs as compared to untreated mice, Figure 1B, Table S1). We subsequently assessed whether these MAbs also protected against antibody-enhanced lethal DENV infection. Anti-DENV1 serum was administered 24 hours prior to a sub-lethal infection (105 PFU) of DENV2 D2S10, and animals were treated 24 hours post-infection with 20 µg of genetically-modified MAbs lacking effector functions, where the N297Q variant MAbs are fully aglycosylated and the LALA variant MAbs remain glycosylated but incapable of binding either FcγR or C1q [42], [46]. Of those tested, only the EDII fusion loop-specific E60 N297Q and EDIII A strand-specific 87.1 LALA MAbs completely prevented mortality (P<0.01, Figure 1C and Table S1). In comparison, the EDIII-A-strand-specific MAb E76 N297Q showed partial protection (P<0.05), whereas MAbs E44 N297Q (EDII dimer interface) and E87 N297Q (EDIII C-C′ loop) provided no protection against lethal disease (Figure 1C, Table S1).
To determine why some MAbs had therapeutic activity in the virus-only lethal infection model but not in the context of antibody-enhanced infection, we examined several properties including epitope specificity, neutralization potency, and avidity. We first assessed whether neutralization potency correlated with in vivo therapeutic potential. The neutralizing activity against DENV2 D2S10 of each of the MAbs was assessed using a flow cytometry-based assay with human monocytic U937 cells expressing DC-SIGN, a known attachment factor for DENV [47]. The potency of each intact and modified MAb was assessed and expressed as the 50% neutralization titer (NT50 in ng/ml of MAb). No significant difference was observed between each intact MAb and its modified variant. The NT50 of therapeutically effective MAbs E60 N297Q (EDII fusion loop) and 87.1 LALA (EDIII A strand) were 72 ng/mL and 24 ng/mL, respectively. In comparison, the NT50 of MAbs E44 N297Q (EDIII C-C′ loop) and E87 N297Q (EDII dimer interface), which bound other epitopes in EDII and EDIII and lacked therapeutic activity, were similar (68 ng/mL and 95 ng/mL, respectively) (Table 1). Thus, NT50 values among MAbs targeting different epitopes failed to demonstrate a clear relationship between neutralizing potency and in vivo therapeutic efficacy in the context of antibody-enhanced lethal infections (Spearman ρ 0.47, P = 0.45).
We hypothesized that MAb avidity, the strength of binding between a bivalent antibody and two ligands on a single virion or across virions, might correlate better with therapeutic efficacy following an antibody-enhanced lethal infection. To test this hypothesis, we measured the avidity (Kdapp) of binding, using a direct, virion-coated ELISA [30]. While we noted a correlation between MAb neutralization titer and avidity (Spearman ρ 0.9, P<0.083), analogous to our neutralization data, we did not observe a relationship between MAb avidity and therapeutic efficacy (Spearman ρ 0.32, P<0.68) by MAbs targeting non-fusion loop epitopes (Table 1).
As neutralization potency and avidity failed to correlate directly with the therapeutic efficacy of our modified MAbs across different epitopes, we investigated whether epitope specificity had greater predictive potential. As MAb E60 N297Q was highly protective even 48 hours following antibody-enhanced DENV infection [42], we hypothesized that the fusion loop epitope might be an important target for therapeutic MAbs. Therefore, we tested the therapeutic activity following either a virus-only or an antibody-enhanced lethal DENV infection of three additional modified MAbs that also target the EDII fusion loop but displayed between 2- to 8-fold reduced neutralization potency compared to MAb E60 N297Q: 82.11 LALA (NT50 131 ng/ml), E18 N297Q (NT50 363 ng/mL) and E28 N297Q (NT50 544 ng/mL) (Table 1). Whereas all of the animals treated with MAb 82.11 LALA, E18 N297Q or E28 N297Q survived infection after a virus-only lethal challenge (P<0.01 compared to PBS-treated mice, Figure 2A and Table S2), MAbs E18 N297Q and E28 N297Q failed to confer a therapeutic benefit following an antibody-enhanced lethal infection (Figure 2B, Table S2). MAb 82.11 LALA protected 50% (3/6) of animals following an antibody-enhanced infection, though this difference trended but did not attain statistical significance compared to non-treated control animals (Figure 2B, Table S2). In contrast to the experiments with non-fusion loop-specific MAbs, studies with MAbs targeting the same fusion loop epitope suggest that neutralization potency can predict therapeutic efficacy following an antibody-enhanced infection (Spearman ρ 0.9487, P<0.083).
To explain the correlation between in vitro neutralizing potency and in vivo therapeutic efficacy within fusion loop-specific MAbs, we generated a model of competitive displacement. Recent work has suggested that a significant fraction of the human anti-flavivirus E protein antibody response is directed against the fusion loop epitope in EDII [20], [48]–[50]. We hypothesized that these cross-reactive antibodies found in DENV-immune serum are of intermediate or low affinity and bind to the heterologous virus at a stoichiometry insufficient for neutralization but adequate for enhancement of infection [51]. However, after administration of a therapeutic, high-affinity, genetically-modified fusion loop-specific MAb, natural dissociation of the enhancing antibody occurs, and the more avid therapeutic MAb binds to the fusion loop epitope, effectively preventing enhancing antibodies from binding again to the virion. Additionally, highly avid modified MAbs would compete favorably with enhancing antibodies for binding to nascently-produced virions. In this scenario, modified MAbs lacking effector functions either coat the virion allowing for direct neutralization or compete against cross-reactive fusion-loop enhancing antibodies in serum, such that the stoichiometry required for enhancement [51] is not reached. To test this hypothesis, we used 4G2, a weakly neutralizing (NT50 of 393 ng/mL) mouse MAb that binds to the fusion loop epitope [49] to enhance an otherwise sub-lethal DENV2 D2S10 infection and administered 20 µg of the modified MAbs 24 hours post-infection. E60 N297Q, the most therapeutic fusion loop-specific MAb in the context of a polyvalent serum-enhanced infection, again achieved 100% protection (P<0.01) when administered after a 4G2-enhanced infection, whereas MAb 82.11 LALA was less therapeutic (P<0.05), protecting 4 of 6 treated animals (Figure 2C, Table S3). None of the animals treated with MAb E18 N297Q succumbed to infection (P<0.05), although all demonstrated signs of illness (P<0.05 as compared to E60 N297Q-treated mice, Table S3). However, mice treated with MAb E28 N297Q all succumbed to 4G2-enhanced DENV2 D2S10 infection. These data support a model in which modified fusion loop-specific MAbs of sufficient avidity and neutralizing potency compete effectively for binding sites in the context of enhancing polyvalent DENV-immune serum or other fusion loop-specific MAbs to prevent disease.
We next evaluated directly whether MAb E60 (EDII fusion loop-specific) could effectively compete for binding with less potent fusion loop-specific MAbs, as compared to either therapeutic MAb 87.1 (EDIII A strand-specific) or non-therapeutic MAb E87 (EDII C-C′ loop-specific) that both target distinct epitopes. After directly coating DENV2 virions on microtiter plates, we added the moderately neutralizing mouse MAb 4G2 at 1 µg/mL mixed with increasing concentrations (0.1, 1 and 10 µg/mL) of modified human MAbs followed by an anti-mouse, Fc-specific secondary MAb. Binding of mouse MAb 4G2 was not affected by the amount of bound E87 (non-therapeutic, C-C′ loop-specific) (P = 0.64 by Friedman's analysis of data combined from seven experiments). In contrast, both MAb E60 (therapeutic, fusion loop-specific) and, surprisingly, MAb 87.1 (therapeutic, A strand-specific) altered binding of MAb 4G2; higher concentrations of MAb E60 and MAb 87.1 resulted in lower amounts of MAb 4G2 bound (P≤0.001 for both E60 and 87.1 by Friedman's analysis of data combined from seven experiments, Figure 3A). Less potently neutralizing and non-therapeutic fusion loop-specific MAbs competed as or less effectively against MAb 4G2 for binding to the fusion loop epitope (Figure S1A).
We next tested whether modified MAbs targeting the same or different epitopes with respect to the enhancing MAb 4G2 (fusion loop-specific MAb) could suppress enhancement in vitro. We mixed serial dilutions of both 4G2 and the modified MAbs E60 N297Q (therapeutic EDII fusion loop-specific, Figure S2A), 87.1 LALA (EDIII, A strand-specific, Figure S2B) and E87 N297Q (EDII dimer interface-specific, Figure S2C) in the following ratios: 100% 4G2; 95% 4G2 and 5% modified MAb; 85% 4G2 and 15% modified MAb; 75% 4G2 and 25% modified MAb. Each MAb combination was incubated with DENV2 D2S10 virus and used to infect K562 cells, a human erythroleukemic cell line that expresses FcγRIIA (CD32A) and is non-permissive in the absence of enhancing anti-DENV antibodies. Infection was monitored after 48 hours by intracellular DENV antigen staining and quantified by flow cytometry. Peak enhancement with MAb 4G2 occurred at 466 ng/mL and resulted in ∼7 to 15% of the cells becoming infected (Figure S2). However, when E60 N297Q or 87.1 LALA, MAbs that were effective therapeutically and lack the ability to engage Fcγ receptors, comprised only 5% of the MAb population, peak enhancement by 4G2 was reduced by 33% and 65%, respectively (Figure 3B, Figure S2). Remarkably, when 4G2 and the modified MAbs were mixed in a 75%: 25% ratio, E60 N297Q and 87.1 LALA both reduced infection by 85% and 86%, respectively (Figure 3D, Figure S2). In comparison, mixture of 5% of the non-therapeutic modified MAb E87 N297Q, reduced enhancement by only 21% (P<0.04, compared to E60 N297Q and P<0.02 compared to 87.1 LALA), and by 57% (P<0.01, compared to E60 N297Q and P<0.02 compared to 87.1 LALA) when a 75%: 25% mixture was used (Figure 3B–3D). Similarly to E87 N297Q, MAbs E18 N297Q and E28 N297Q, both fusion loop-specific but non-therapeutic MAbs, reduced enhancement by 19% and 6% when E18 N297Q and E28 N297Q were 5% of the MAb population, respectively, and by 46% when either E18 N297Q or E28 N297Q comprised 25% of the MAb population (Figure S1B–D). Thus, highly avid, fusion-loop specific MAb E60 N297Q and A-strand-specific MAb 87.1 LALA minimized in vitro enhancement, presumably by preventing binding of the enhancing fusion loop-specific MAb 4G2. Furthermore, the ability of the modified MAbs to prevent 4G2-mediated enhancement in vitro correlated with in vivo therapeutic activity in the context of anti-DENV polyvalent serum-enhanced infection.
Given the results in K562 cells with mixtures of modified MAbs and decreased 4G2-mediated enhancement, we wanted to evaluate further this relationship in vivo. We administered 20 µg of MAb 4G2 24 hours prior to infection with DENV2 D2S10 and then treated mice one day post-infection with 20 µg of MAb E60 N297Q (EDII fusion loop-specific)), 87.1 LALA (EDIII A strand-specific), or E87 N297Q (EDIII C-C′-loop-specific). While MAb E87 N297Q was not therapeutically protective against a 4G2-enhanced lethal infection, E60 N297Q and 87.1 LALA protected against mortality in 6 of 6 and 5 of 6 animals, respectively (P<0.05, Figure 3E, Table S3). Thus, potently neutralizing, fusion loop-specific (E60 N297Q) and A strand-specific (87.1 LALA) MAbs both prevent antibody-enhanced disease, likely by displacing binding of enhancing MAbs that target the fusion loop epitope.
We next assessed how different ratios of intact and genetically modified variants affected enhancement in K562 cells in vitro. We selected the two most therapeutically effective, modified MAbs (E60 N297Q (EDII fusion loop) and 87.1 LALA (EDIII A strand)), and mixed them with the intact parent MAbs in the following proportions: 100% intact MAb, 90% intact and 10% modified MAb, 75% intact and 25% modified MAb, 50% of each MAb, 25% intact and 75% modified MAb, 10% intact and 90% modified MAb, and 100% modified MAb. While several mixtures of E60 parent:E60 N297Q showed reduced enhancement, only the combination of 10% intact:90% modified was non-enhancing in vitro, suggesting that the majority of the antibody mixture must not bind to Fcγ receptors in order to abolish enhancement of DENV infection when MAb pairs are of comparable neutralizing potency and avidity (Figure 4A). The combination of intact 87.1 and 87.1 LALA also demonstrated a complete reduction in enhancement, but this occurred under conditions where a lower ratio of intact to modified mAb was required (ratios of 25% intact:75% aglycosylated (Figure 4B)). The relative differences in enhancement profiles observed between the E60:E60 N297Q and 87.1:87.1 LALA MAb pairs could be due to the small difference in the avidity and neutralization potency of the intact and modified MAb. Similar relationships between modified and intact MAb pairs were observed when studying MAbs that were moderately (E76/E76 N297Q) and poorly (E18/E18 N297Q) therapeutic (data not shown).
Using combinations of intact E60 and modified E60 N297Q (EDII fusion loop-specific MAb), we evaluated whether the requirement for 90% of the MAb mixture to lack FcγR binding for suppression of enhancement in vitro translated into therapeutic efficacy in vivo. The same ratios were mixed in a total of 20 µg and administered therapeutically 24 hours after serum-enhanced DENV2 infection of AG129 mice. Notably, and consistent with our data in K562 cells, complete therapeutic protection in vivo required 90% of the E60 mixture to be present in the modified form (P<0.02, Figure 4C). Mixtures that were combined in a ratio of less than 9∶1 showed reduced or no therapeutic efficacy (Figure 4C). This in vivo data suggests that when the same MAb is used for enhancement and therapy (intact versus modified), the majority of the mixture must lack the capacity for binding FcγR to avoid enhancement. Thus, a low stoichiometric threshold of binding is likely sufficient for enhancement of infection and disease.
Given the results with intact and modified MAbs, we evaluated whether we could use this in vitro relationship to predict the ability of modified MAbs to be therapeutically effective in vivo in the context of immune serum-enhanced DENV infection. Initially, using DENV1-immune mouse serum, we identified the serum dilution (1∶180) responsible for peak enhancement of DENV2 D2S10 in K562 cells (Figure 5A). We then tested the ability of modified MAbs to reduce enhancement in K562 cells by pre-incubating D2S10 with the peak enhancing dilution of DENV1-immune serum for 30 minutes, then adding increasing amounts of modified MAbs for 30 minutes, followed by incubation with K562 cells for 48 hours. Importantly, the concentrations of DENV-immune serum and virus used in the in vitro assay were comparable to those used in the in vivo infections. At concentrations of 2,000 ng/mL and 1,000 ng/mL, modified MAbs with moderate to strong (>60% protection) therapeutic activity in vivo were more efficient (P<0.05) at suppressing ADE in K562 cells than MAbs that were less therapeutically active (Figure 5B and 5C). The three most therapeutically effective MAbs (87.1 LALA, E60 N297Q and E76 N297Q) reduced enhancement on average by 88%, 70% and 65%, respectively, when added at 1,000 ng/mL while less protective MAbs (82.11 LALA, E44 N297Q, E87 N297Q, E18 N297Q and E28 N297Q) reduced enhancement less efficiently (Figure 5C). This trend also was observed when DENV1-immune serum was added at a different enhancing concentration (1∶540 dilution) (data not shown). Based on these data that differentiate in vitro therapeutic from non-therapeutic MAbs, we established ∼50% reduction at 1,000 ng/mL as the criterion for predicting therapeutic efficacy using the suppression-of-enhancement assay.
As the K562 cell-based assay with mouse polyclonal anti-DENV1 serum and modified MAbs appeared to predict in vivo outcomes, we repeated the experiments with DENV-immune human serum; this was important as humans and mice produce overlapping yet distinct antibody repertoires against flavivirus epitopes [19], [48], [52], [53]. We evaluated whether modified MAbs reduced enhancement in K562 cells using human DENV-immune serum collected years after a primary DENV4 infection. The peak serum enhancement dilution again was identified as between 1∶180 and 1∶540 (Figure 6A). In contrast to the limited protection provided by the modified MAbs following a mouse DENV1-serum enhanced infection, most modified MAbs suppressed enhancement by DENV4 human immune serum below the 50% cut-off at the higher (P<0.05; E18 N297Q and E28 N297Q, P<0.08), yet physiologically relevant concentrations (1 and 2 µg/mL) of modified MAb (Figure 6B, Figure S3), while non-binding, DENV4-specific MAb 22.3 LALA did not (Figure 6B). Similar results were obtained when primary DENV1 or DENV3 human immune serum was tested (Figure S3).
To determine whether the enhancement data with human serum predicted protection in vivo, we administered normal human serum (NHS) or enhancing amounts of anti-DENV4 human immune serum 24 hours prior to a sub-lethal infection with DENV2 D2S10, and tested the therapeutic efficacy of the modified MAbs. As expected, all mice pre-treated with NHS survived infection without any signs of morbidity. All mice receiving enhancing anti-DENV4 human immune serum and treated with a modified MAb (E60 N297Q, 87.1 LALA, 82.11 LALA, E87 N297Q, and E28 N297Q) survived lethal enhanced infection with minimal signs of disease (P<0.05 for all modified MAbs compared to PBS-treated controls, Figure 6C), whereas mice treated with modified, DENV4-specific MAb 22.3 LALA did not (Figure 6C). Thus, the suppression-of-enhancement assay in K562 cells correlated with the therapeutic efficacy of modified MAbs in vivo in an antibody-enhanced lethal DENV model in the context of both mouse and human DENV immune serum. Moreover, and for reasons that likely relate to the distinct repertoire of cross-reactive enhancing antibodies in human serum, modified MAbs against the EDIII A-strand, EDIII C-C′ loop, and EDII fusion loop all efficiently suppressed antibody enhancement in cell culture and in vivo.
In this report, we analyzed a panel of eight MAbs that bind to several epitopes on the dengue virion, including the fusion loop and dimer interface on EDII and the A strand and C-C′ loop on EDIII. We determined that differences exist between the ability of modified MAbs lacking the capacity to engage FcγR and C1q to act therapeutically following a virus-only lethal infection and an antibody-enhanced lethal infection. Analysis of MAb characteristics such as binding avidity and neutralization potency did not clearly define an in vitro correlate of in vivo efficacy across different epitopes, but were more predictive when studying MAbs targeting a specific class, such as those binding the fusion loop epitope. Further analysis suggested that modified, fusion loop- and A-strand-specific MAbs act therapeutically by competing against enhancing antibodies in polyvalent serum that recognize the same or proximal epitopes. By studying these relationships, for the first time, we established a novel in vitro suppression-of-enhancement assay with polyclonal mouse and human anti-DENV immune serum that appears to predict the ability of modified MAbs to act therapeutically against ADE in vivo. Thus, we provide in vivo data that support in vitro observations about the mechanism of ADE as well as a means to suppress ADE in vivo.
Multiple parameters, including neutralization potency, avidity and epitope specificity, affect whether a modified MAb is therapeutic against an antibody-enhanced DENV infection. In our panel, in addition to binding to either the fusion loop or A-strand epitope, a therapeutic MAb needed to be strongly neutralizing (NT50<100 ng/mL), which itself is a function of epitope accessibility on the virion, mechanism of inhibition, and avidity of binding [51]. Four of the MAbs tested (E60, 82.11, E18, and E28) recognize similar residues within the EDII fusion loop ([54] and S. Sukupolvi-Petty and M.S. Diamond, unpublished data), but two (E18 and E28) had lower neutralizing potency and avidity of binding to the virion, and, correspondingly, showed less or no therapeutic activity in vivo following DENV enhancement by polyvalent mouse serum. While the avidity of binding to solid-phase DENV2 for 82.11 LALA and E60 N297Q was comparable, E60 N297Q is ∼2.5 fold more neutralizing, suggesting that the two MAbs might bind overlapping yet slightly distinct epitopes, or that the ensemble of viral conformations in solution [55] allows for enhanced recognition of E60 relative to 82.11. Analogously, when comparing two modified MAbs targeting the A-strand in EDIII, MAb 87.1 LALA showed higher avidity of binding and therapeutic efficacy in vivo compared to MAb E76 N297Q. Although further study is warranted, our data suggest that within an epitope class, there is a direct relationship between MAb avidity and neutralization potential in vitro and therapeutic efficacy in vivo.
Studies comparing therapeutic efficacy following virus-only and mouse antibody-enhanced lethal DENV2 infections revealed that all modified MAbs tested were therapeutic following a virus-only infection, but only two (E60 N297Q and 87.1 LALA) were completely protective following antibody-enhanced infection with DENV1-immune mouse serum. This observation suggests a direct interplay between the enhancing antibodies in polyvalent serum and the neutralizing therapeutic MAbs that determines outcome. Thus, a second parameter affecting therapeutic efficacy is the ability of a modified MAb to out-compete the enhancing antibodies in polyvalent immune serum for binding to the virion. This concept is supported by functional data in vitro and in vivo using the weakly neutralizing, fusion loop-specific MAb 4G2 and a panel of modified fusion loop-specific MAbs. In cellular assays, the more avid and strongly neutralizing MAb E60 N297Q was more effective at suppressing 4G2-enhanced infection in K562 cells than the less potent E18 N297Q and E28 N297Q MAbs. Consistent with this, E60 N297Q but not E28 N297Q prevented mortality as a therapeutic when MAb 4G2 was used to enhance a sub-lethal DENV2 D2S10 infection. Our data support a model in which therapeutic activity occurs when high-affinity, modified MAbs can bind to virions and neutralize infection by competing with and/or displacing enhancing antibodies for binding to similar epitopes.
An additional parameter that likely affects therapeutic efficacy of a MAb against antibody-enhanced DENV infection is its mechanism of action: whether the MAb binds prior to or following attachment of the virion to the target cell. However, the interpretation is not straightforward, as mechanism of neutralization of a given MAb may be affected by several variables: (a) stoichiometry and relative fractional occupancy at a given concentration [51]; (b) cell type and repertoire of attachment ligands or receptors [51], [54]; (c) virus particle maturation [56]; and (d) dynamic state of the virion [55]. From in vitro ADE experiments using K562 cells and non-modified MAbs, we can conclude that all MAbs (excluding E44 as it was not available for these studies) have the capacity to neutralize infection via a post-attachment mechanism at saturating concentrations (Figure 4 and Figure S2). Following uptake via FcγR, antibody-enhanced DENV may still be neutralized by MAbs that block post-attachment – this phenomenon of trans-dominant neutralization of ADE by MAbs was described previously with the anti-WNV MAb E16 [57], a MAb that neutralizes WNV infection by blocking the structural changes required for viral fusion [58], [59]. Indeed, at saturating concentrations, all of the non-modified WT MAbs in our panel reduce K562 infection to background levels (Figures 4A and B, and Figure S2, right side of the curve), most likely by blocking fusion, a critical step required for release of the DENV genome into the cytosol following receptor-mediated endocytosis. To distinguish this post-attachment neutralization pattern from a MAb that blocks via a pre-attachment mechanism and cannot prevent infection in a K562 assay, we can compare these data to the effects of anti-fusion loop MAb E60 on WNV infection. MAb E60 cannot prevent WNV infection in K562 cells even at saturating concentrations (E. Mehlhop and M.S. Diamond, unpublished data). The inability of MAb E60 to neutralize WNV in K562 cells occurs because WNV virions are present in a mature state to a far greater degree than DENV and therefore have a lower stoichiometry of binding for the fusion loop epitope [56], [60]. For WNV, MAb E60 fails to achieve a stoichiometry sufficient to block fusion of virus that has entered via FcγR-dependent enhancement. In contrast, in the current study, all MAbs appear to act in a post-attachment mechanism in K562 cells at saturating concentrations. However, in vivo, it is unlikely that the modified MAbs circulate at saturating concentrations, given the large amount of DENV virus and antigen present. Thus, the relevant question becomes which MAbs can reduce enhancement (generated by either polyvalent DENV-immune sera or MAbs such as fusion loop-specific 4G2) most efficiently when the modified MAbs are at sub-saturating concentrations. Under these conditions, therapeutically effective MAbs (87.1 LALA and E60 N297Q) show greater efficacy than the other MAbs evaluated, likely due to their ability to compete for binding with enhancing antibodies, which interferes with FcγR crosslinking and limits DENV uptake and infection.
Remarkably, MAb 87.1 LALA, which maps to the A strand of EDIII, also was effective against anti-fusion loop MAb 4G2-mediated lethal DENV infection. While not as potent as MAb E60, MAb 87.1 appeared to compete with fusion loop-specific MAb 4G2 in the solid-phase DENV2 ELISA and also suppressed 4G2-induced enhancement in K562 cells. One possible explanation is that the A-strand epitope on EDIII is located next to the EDII fusion loop on adjacent DENV E proteins within a dimer [23], such that on the virion in solution, high avidity binding of 87.1 LALA prevents lower-avidity fusion loop-specific enhancing MAbs (e.g., 4G2) from binding. While all MAbs tested appear to block DENV in a post-attachment mechanism at saturating concentrations, MAb 87.1 may be more potent at blocking fusion at sub-saturating concentrations than MAb E60. This hypothesis may explain why 87.1 LALA is more efficient at reducing 4G2-enhanced DENV infection in K562 cells, but less efficient at competing with MAb 4G2 in a fixed-virion ELISA than MAb E60 N297Q. Another possible explanation is that binding of the A-strand MAb 87.1 LALA alters the conformation of the mature DENV virion [61], [62], enhancing exposure of the fusion loop epitope and increasing binding and neutralization. Even though E87 N297Q is unable to compete for binding with fusion loop-specific MAb 4G2 in the solid phase assay, it can still bind to the virion and contribute to the stoichiometry required to neutralize DENV, thus accounting for the ∼50% reduction in in vitro enhancement when E87 N297Q comprises 25% of the antibody mixture. Despite this, E87 N297Q did not have therapeutic activity in vivo when 4G2 was used as the enhancing MAb. In addition, MAb E87 N297Q was less efficient at reducing MAb 4G2-enhanced infection than the therapeutically effective MAbs E60 N297Q and 87.1 LALA at any of the three conditions tested (5%, 15% or 25% modified MAb). Although more study is warranted, we speculate that the MAbs which bind epitopes that do not displace 4G2 enhancing MAbs did not protect in vivo because they failed to reach a stoichiometry that was sufficient for neutralization or do not block a post-attachment step (e.g., viral fusion).
Previous studies have established that the epitope repertoire of anti-flavivirus neutralizing antibody in mouse and human serum is different. Mice were found to generate neutralizing antibody responses against epitopes in EDIII (∼30%) [24], [26], [28], [30], [63], [64] that can be serotype-specific [29], [30], [65] or cross-reactive [28], [29], [31], [66]. In comparison, DENV-immune human serum preferentially targets the fusion loop epitope in EDII [52], [53] as well as complex quaternary epitopes near the EDI-DII hinge that span adjacent E proteins within a dimer [32], with little EDIII-specific neutralizing antibody generated (10–15%) [48], [64], [67], [68]. While it has not been explicitly studied, it seems plausible that the epitope repertoire for enhancing antibodies against DENV in human and mouse serum also vary. In support of this, we observed differences in the ability of modified MAbs to prevent antibody-enhanced lethal DENV infection when DENV-immune mouse or human serum was used. Only E60 N297Q, 87.1 LALA and E76 N297Q were therapeutically effective against infection enhanced with anti-DENV1-immune mouse serum. In contrast, all modified DENV2-reactive MAbs were therapeutic following an infection enhanced with DENV4-immune human serum. These data likely imply one of two non-mutually exclusive hypotheses: (a) cross-reactive enhancing MAbs present in DENV-immune human serum are weakly avid, such that higher affinity modified MAbs can bind and/or displace the enhancing antibodies, resulting in therapeutic protection in vivo; (b) cross-reactive enhancing MAbs present in DENV-immune human serum bind distinct epitopes, which do not interfere with binding and neutralization by modified MAbs targeting the EDII fusion loop, EDII dimer interface, EDIII A strand, or EDIII C-C′ epitopes. In possible support of this, recent studies of the human antibody repertoire against DENV suggest that anti-prM antibodies are a major component of the cross-reactive response and promote enhancement in vivo [18], [20]. Future studies that test the therapeutic efficacy of modified E protein MAbs in the presence of enhancing concentrations of prM-specific MAbs will be important to perform.
One limitation of this study is the passive transfer model used to develop lethal DENV disease; we tested limited concentrations of enhancing polyvalent immune serum to distinguish between therapeutic and non-therapeutic modified MAbs. In the future, a more detailed dose-response study with different enhancing sera or MAbs will be needed to determine the range of efficacy of modified MAbs in mediating protection. In addition, while the passive transfer model of enhancement and protection may be relevant for infant DHF/DSS where potentially enhancing antibodies are received passively in utero, it remains uncertain if similar principles apply during natural secondary DENV infection.
In summary, our results suggest a model in which neutralization, avidity, and epitope specificity contribute to the therapeutic efficacy of modified MAbs. Despite the differences between mouse and human polyvalent antibody repertoires, the suppression-of-enhancement assay in K562 cells accurately predicted in vivo therapeutic efficacy in both situations. While further study is needed, this assay could be used to screen additional modified MAbs for potential use as DENV therapeutics. Overall, given these promising results, we suggest that further exploration of the utility of modified MAbs as therapy for DENV infections is warranted.
This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The protocol was approved by the Institutional Animal Care and Use Committee at the University of California Berkeley (R252-1012B).
All viruses were propagated in Aedes albopictus C6/36 cells (American Type Culture Collection) and titered by plaque assay on baby hamster kidney cells (BHK21, clone 15) [69]. DENV2 D2S10 was generated as previously described [70]. All in vitro neutralization and enhancement assays and sub-lethal in vivo infections were performed with non-concentrated virus. DENV2 D2S10 virus was concentrated by ultra-centrifugation for use in the virion ELISA and by centrifugation using 100,000 MWCO Amicon filters (Millipore) for lethal, virus-only in vivo infections. U937-DC-SIGN (gift from A. de Silva, University of North Carolina, Chapel Hill) and K562 cells were used for flow cytometry-based in vitro neutralization [71] and enhancement assays [64], respectively.
Mouse MAbs E60, E18 and E28 were generated against WNV E protein, but are cross-reactive with DENV E protein [54]. Anti-DENV2 MAb E44, E76, and E87 were generated against DENV2 and described previously [28]. All mouse MAbs were purified by protein A affinity chromatography (Invitrogen, Carlsbad, CA) and have been mapped previously [29], [72]. The generation of a chimeric human-mouse E60 MAb with the human IgG1 constant region and the mouse VH and VL region was performed as described previously [72]. The generation of chimeric E18, E28, E44, E76, and E87 MAbs was performed similarly. Point mutations in the Fc region (N297Q) that abolish Fcγ receptor and C1q binding were introduced by QuikChange mutagenesis (Stratagene). All recombinant MAbs were produced after transfection of HEK-293T cells, harvesting of supernatant, and purification by protein A affinity chromatography. The accession numbers for the sequences of the VH-VL regions of the recombinant MAbs are as follows: E18_VL KC254882 E18_VH KC254883 E28_VL KC254884 E28_VH KC254885 E44_VL KC254886 E44_VH KC254887 E60_VL KC254888 E60_VH KC254889 E76_VL KC254890 E76_VH KC254891 E87_VL KC254892 E87_VH KC254893.
MAbs 87.1 and 82.11 are fully human MAbs, and their generation has been described previously [20]. Production of the LALA variants was performed according to a previously published protocol [20]. Recombinant MAbs were produced in HEK-293T cells and purified by sequential protein A affinity chromatography and size-exclusion chromatography. The accession numbers for the sequences of the VH-VL regions of the recombinant human MAbs are as follows: DV82VH KC294013, DV82VL KC294014, DV87VH KC294015, DV87VL KC294016.
The single DENV1- and DENV3-immune human sera samples used in the in vitro suppression-of-ADE assay were de-identified and pre-collected as part of the Nicaraguan Pediatric Dengue Hospital-based Study [73]. Both serum samples were collected three months post-symptom onset and were obtained from individuals with a primary DENV infection [74], [75]. The protocol for the study was reviewed and approved by the Institutional Review Boards (IRB) of the University of California (UC), Berkeley, and of the Nicaraguan Ministry of Health. Parents or legal guardians of all subjects provided written informed consent, and subjects 6 years of age and older provided assent.
The primary DENV4-immune human serum sample used for in vitro suppression-of-ADE and in vivo ADE experiments was a gift from Dr. Aravinda de Silva (University of North Carolina (UNC), Chapel Hill) and were received as de-identified and pre-collected samples, and as such were not considered part of human subjects research by the IRB at UC Berkeley. Convalescent DENV-immune sera were obtained at UNC Chapel Hill from volunteers who had experienced natural DENV infections during travel abroad. The protocol for recruiting and collecting blood samples from returned travelers was approved by the IRB of UNC Chapel Hill. Written informed consent was obtained from all subjects before collecting blood.
All procedures were pre-approved and conducted according to UC Berkeley Animal Care and Use Committee guidelines. AG129 mice [76] were bred at UC Berkeley.
AG129 mice were infected intraperitoneally (i.p.) with 105 PFU of DENV1 strain 448 (gift from S. Kliks). Six to eight weeks post-infection, mice were sacrificed, and whole blood was collected by terminal cardiac puncture. Serum was isolated from whole blood by centrifugation, heat inactivated, and stored at −80°C.
DENV2 D2S10 enhanced disease: AG129 mice were administered either 25 µl mouse anti-DENV1-immune serum or 200 µl human anti-DENV-4 immune serum or 20 µg of MAb 4G2 (in 200 µl final volume) i.p. 24 hours prior to infection with an intravenous (i.v.) sub-lethal, 105 PFU dose of DENV2 D2S10. DENV2 D2S10 virus-only lethal disease: AG129 mice were infected i.v. with 4×106 PFU of DENV2 D2S10. Treatment model: Mice were administered 25 µl of DENV1-immune serum on Day −1, 105 PFU of DENV2 D2S10 on Day 0, and 20 µg of modified MAbs in a final volume of 100 µl i.v. 24 hours following infection (Day +1). All animals were monitored carefully for morbidity and mortality for 10 days following infection.
The neutralization titer of each parent and modified MAb was measured using the U937-DC-SIGN flow cytometry-based neutralization assay as described [71]. NT50 titers were calculated as described previously [42]. Each NT50 titer is the average of between 3 and 5 individual experiments, with the exception of E44 N297Q (1 experiment).
Polyvalent serum: Enhancement curves were generated as described in [64]. Briefly, eight, three-fold dilutions of DENV-immune serum beginning at 1∶10 were pre-mixed with DENV2 D2S10 prior to addition to K562 cells. A Gaussian distribution was used to fit each enhancement curve, where percent infection was recorded on the y-axis and log-reciprocal serum dilution on the x-axis. The area under the curve (AUC) for each enhancement infection was calculated using Prism software. MAb competition: Intact and modified MAbs were pre-mixed in different ratios at a starting concentration of 40 µg/mL. Eight 3-fold dilutions were incubated in either duplicate or triplicate with DENV2 D2S10 at an MOI of 0.1 in a 1∶1 ratio before being added to K562 cells for an enhancement of infection assay [64].
DENV-immune serum was diluted to the concentration that resulted in the greatest enhancement of DENV2 infection in K562 cells. DENV2 D2S10 virus at an MOI of 0.1 and serum were mixed together in equal volumes for 30 to 45 minutes at 37°C. Modified MAbs were prepared in five 2-fold dilutions beginning at 2,000 ng/mL, added to the polyvalent serum/virus mixture, and incubated for an additional 30–45 minutes prior to the addition of 5×104 K562 cells. The cells were washed 2 hours following infection and fixed and stained for viral antigen. Relative infection was expressed as the average percent infection for each duplicate divided by the percent infection measured without modified antibody (positive control, between 7 and 15%).
Avidity ELISA: DENV2 D2S10 virus was isolated by ultra-centrifugation at 53,000× g for 2 hours at 4°C and resuspended in cold PBS with 20% FBS. Concentrated virus was diluted to 5×104 pfu in carbonate coating buffer, pH 9.6, and 50 µl was added to each well of a 96-well flat-bottomed plate as described previously [30]. The plate was coated overnight at 4°C and washed thoroughly with PBS with 0.1% Tween-20 (PBS-T) prior to blocking (5% milk w/v in PBS-T) for one hour. Both the non-modified and modified MAbs were diluted to 120 µg/mL in blocking buffer and titrated two-fold for a total of 12 serial dilutions. Each MAb dilution was added in duplicate to the coated plate for one hour. The plates were washed with PBS-T and incubated with an alkaline phosphatase (AP)-conjugated goat anti-human secondary antibody (Meridian) and AP substrate PNPP (Sigma) for one hour each, with additional PBS-T washes in between each step. The reaction was developed for 45 minutes, and the absorbance was read at 405 nm on a UV-plate reader (Bio-Tek) using KC Junior software. Competition ELISA: Ninety-six-well flat-bottomed plates were coated with DENV2 as described above. Mouse MAb 4G2 was diluted to 1 µg/mL and mixed with human mAb diluted to 10, 1 and 0.1 µg/mL in a separate 96-well plate, and 100 µl of the mixture was added. After one hour, the plates were washed and incubated with goat anti-mouse Fcγ-specific biotinylated secondary (Jackson) followed by Streptavidin-AP (Invitrogen) and the plates were developed with PNPP as described above.
All graphs were produced using GraphPad Prism 5 software (La Jolla, CA). Statistical analysis was performed using Stata v10 (College Station, TX) and Prism 5 software. Comparison between NT50 titers of non-modified and modified MAb pairs was conducted using a Wilcoxon rank-sum analysis. Comparison of survival rates was conducted using a non-parametric log rank test. A Spearman rho (ρ) was calculated to assess correlations between modified MAb NT50 titer, avidity, and therapeutic efficacy (0–100% survival). A Kruskal-Wallis test was used to compare 4G2 binding across increasing concentrations of human MAb, and a Friedman's analysis (matched pairs Kruskal-Wallis) was conducted by combining all data for each MAb tested. A Wilcoxon rank-sum test was used to compare differences in the percent reduction of 4G2-enhanced D2S10 infection in K562 cells with different mixtures of modified MAb as well as in the enhancement-suppression assay using mouse DENV1-immune serum to compare relative infection between therapeutic MAbs and non-therapeutic MAbs. A sign rank test was used to determine whether 1,000 ng/mL of modified MAb could reduce an infection enhanced with DENV4-immune serum significantly lower than 50%.
|
10.1371/journal.pgen.1004446 | DNA Topoisomerase 1α Promotes Transcriptional Silencing of Transposable Elements through DNA Methylation and Histone Lysine 9 Dimethylation in Arabidopsis | RNA-directed DNA methylation (RdDM) and histone H3 lysine 9 dimethylation (H3K9me2) are related transcriptional silencing mechanisms that target transposable elements (TEs) and repeats to maintain genome stability in plants. RdDM is mediated by small and long noncoding RNAs produced by the plant-specific RNA polymerases Pol IV and Pol V, respectively. Through a chemical genetics screen with a luciferase-based DNA methylation reporter, LUCL, we found that camptothecin, a compound with anti-cancer properties that targets DNA topoisomerase 1α (TOP1α) was able to de-repress LUCL by reducing its DNA methylation and H3K9me2 levels. Further studies with Arabidopsis top1α mutants showed that TOP1α silences endogenous RdDM loci by facilitating the production of Pol V-dependent long non-coding RNAs, AGONAUTE4 recruitment and H3K9me2 deposition at TEs and repeats. This study assigned a new role in epigenetic silencing to an enzyme that affects DNA topology.
| DNA topoisomerase is an enzyme that releases the torsional stress in DNA generated during DNA replication or transcription. Here, we uncovered an unexpected role of DNA topoisomerase 1α (TOP1α) in the maintenance of genome stability. Eukaryotic genomes are usually littered with transposable elements (TEs) and repeats, which pose threats to genome stability due to their tendency to move or recombine. Mechanisms are in place to silence these elements, such as RNA-directed DNA methylation (RdDM) and histone H3 lysine 9 dimethylation (H3K9me2) in plants. Two plant-specific RNA polymerases, Pol IV and Pol V, generate small and long noncoding RNAs, respectively, from TEs and repeats. These RNAs then recruit protein factors to deposit DNA methylation or H3K9me2 to silence the loci. In this study, we found that treatment of plants with camptothecin, a TOP1α inhibitor, or loss of function in TOP1α, led to the de-repression of RdDM target loci, which was accompanied by loss of H3K9me2 or DNA methylation. The role of TOP1α in RdDM could be attributed to its promotion of Pol V, but not Pol IV, transcription to generate long noncoding RNAs.
| DNA methylation and histone H3 lysine 9 (H3K9) methylation are two chromatin modifications widely employed by eukaryotes to maintain genome stability [1], [2]. H3K9 methylation and DNA methylation are targeted via small interfering RNAs (siRNAs) to repeats and transposable elements (TEs) and are required for their transcriptional silencing [1], [2].
In plants, cytosine methylation is established through a process known as RNA-directed DNA methylation (RdDM), which involves small and long noncoding RNAs produced by plant-specific RNA polymerases, Pol IV and Pol V, respectively [2]. Pol IV is thought to transcribe RdDM target loci and generate long precursor RNAs. These are eventually processed into 24-nucleotide (nt) siRNAs that are loaded into the Argonaute protein AGO4 [3], [4], [5], [6], [7]. In parallel, Pol V generates long non-coding RNA transcripts from RdDM target loci, and these transcripts recruit siRNA-AGO4 to chromatin [8], [9]. Through the concerted action of these two polymerases, siRNA-AGO4 becomes localized to target loci, and this ultimately recruits the methyltransferase DRM2, which effects de novo DNA methylation. In plants, DNA methylation occurs in three sequence contexts, CG, CHG, and CHH. In contrast to CG and CHG methylation, which can be maintained through the DNA methyltransferases MET1 and CMT3, respectively, CHH methylation is propagated by constant de novo methylation through RdDM [2], [10].
In plants, H3K9 dimethylation (H3K9me2) is another repressive chromatin mark associated with TE and repeat silencing [11], [12], [13]. H3K9me2 and CHG methylation act in a self-reinforcing loop to promote the maintenance of these marks by histone methyltransferases KRYPTONITE (KYP or SUVH4), SUVH5 and SUVH6 and the DNA methyltransferase CMT3 [14]. How H3K9me2 is initially deposited is less well understood, but the RdDM pathway plays a role, as mutations in RdDM pathway genes cause marked reductions in H3K9me2 levels at RdDM target loci [7], [8], [15]. In fact, a recent study revealed a strong genome-wide inter-dependence between non-CG (CHG and CHH) DNA methylation and H3K9 dimethylation [16].
DNA topoisomerases are enzymes that maintain proper DNA topology [17]. During replication or transcription, the DNA helical structure opens to form the replication or transcription fork, and the DNA in front of the fork becomes positively supercoiled, while the DNA behind the fork becomes negatively supercoiled. Topoisomerases bind these regions, nick the DNA to relieve the torsional stress, and re-ligate the DNA. Topoisomerases are divided into two major types, I and II, and further subtypes depending on their mode of action and structure [17], [18].
In Arabidopsis, there are two genes encoding type IB topoisomerases, TOP1α and TOP1β, which are tandemly arrayed in the genome. top1α mutants exhibit gross morphological defects, while top1β mutants are phenotypically normal [19]. RNAi-mediated knockdown of TOP1β in a top1α background is lethal [19]; thus these two genes are functionally redundant.
Here, we uncover a role of TOP1α in transcriptional silencing of TEs. We exploited a luciferase-based reporter (LUCL) that undergoes transcriptional silencing by DNA methylation [20] to perform a chemical genetics screen. We found that camptothecin (CPT) released the DNA methylation of LUCL and de-repressed its expression. CPT is a well-studied natural quinoline alkaloid that targets type 1B topoisomerases [21], [22]. Both the addition of CPT and loss-of-function in TOP1α led to the de-repression of RdDM target loci accompanied by a release of DNA methylation and/or a decrease in H3K9me2 levels. TOP1α is dispensable for Pol IV-mediated siRNA biogenesis but is required for the production of Pol V-dependent, long non-coding RNA transcripts. Consistent with the current model that these transcripts recruit siRNA-AGO4 to chromatin, inactivation of TOP1α resulted in reduced AGO4 occupancy at these loci. Taken together, through the identification of TOP1α as a player in RdDM, we have assigned new roles to a protein affecting DNA topology.
To identify genes involved in DNA methylation, we performed a chemical genetics screen with LUCL, a transcriptionally-silenced luciferase (LUC)-based reporter line [20]. In LUCL, LUC is driven by a dual 35S promoter and both the 35S promoter and the LUC coding region harbor DNA methylation [20]. The DNA methylation at LUCL, and consequently its transcriptional silencing, is controlled by MET1, and to a lesser extent, by the RdDM pathway [20].
Over 3,000 compounds were screened against LUCL seedlings for their effects on LUC expression. A hit compound, camptothecin (CPT) (Figure 1A), was found to release LUC silencing in a concentration- and time-dependent manner (Figure 1B and C). Interestingly, CPT released LUC silencing in a bi-phasic manner, with optimal levels at 10 µM. Further, the release of LUC activity was not observed until one day of chemical addition in a time course assay (Figure 1B). Consistently, continuous live imaging revealed that an increase in LUC activity occurred at about 15 hr after the addition of the chemical (Figure 1C). The slow kinetics suggested that cell division is likely necessary for the de-repression of the reporter. The effects of CPT on LUC protein activity reflected a release of LUCL silencing, as the addition of CPT led to an increase in LUC transcript levels (Figure 1D). Consistent with the dose-dependent effects of CPT on LUC activity, LUC transcript levels were most de-repressed at 10 µM of CPT (Figure 1D).
Previous experiments with LUCL ruled out that it reports miRNA activity, even though it contains the miR172 binding sequence [20]. Consistently, we found that the addition of CPT did not release the LUC activity of a miRNA reporter line, Pro35S::LUC Pro35S::miR-LUC (Figure 1C; [23]). Thus, CPT released the LUC activity of LUCL through a miRNA-independent mechanism.
To determine whether CPT increased LUC transcript levels by reducing DNA methylation, we performed McrBC-PCR to examine the methylation status of LUCL. After digestion of genomic DNA with McrBC, an enzyme that only cuts methylated DNA [24], 35S promoter sequences were amplified by PCR. In the DMSO-treated control sample, little product was observed, indicating that this region was highly methylated in LUCL. However, after CPT treatment, the amount of PCR products increased (Figure 2A), suggesting that CPT treatment led to a reduction in 35S promoter methylation. In addition, the DNA methylation status of the 35S promoter and the LUC coding region was examined by bisulfite sequencing (Figure 2B). The addition of 10 µM CPT resulted in a drastic reduction of CHH methylation, and to some extent CHG methylation, in region #1 (Figure 2C). CG methylation was largely unaffected upon CPT treatment, with the exception of region #4 (Figure 2C).
Due to their potent anti-cancer properties, CPT and its analogs have been intensely studied. The cellular target of CPT is topoisomerase I and the mechanism by which CPT inhibits topoisomerase I is well understood [25]. Given this knowledge, our finding that CPT de-represses LUCL implicated TOP1α in transcriptional gene silencing. A top1α mutant allele, top1α-2, had been found in an unrelated project (Xigang Liu and Xuemei Chen, unpublished results). The top1α-2 mutant carried a C→T point mutation in the second exon, which generates a premature stop codon (Figure S1A). top1α-2, which had been isolated in the Landsberg erecta background, was introgressed into Col-0 through five backcrosses to derive top1α-2Col. top1α-2Col was then crossed to LUCL in the Col-0 background. Unlike CPT, which released LUC activity, the top1α-2Col mutation was not able to release LUC activity (Figure S1B), probably due to activity of the partially redundant TOP1β gene.
We next asked whether TOP1α inactivation or CPT treatment affected DNA methylation of endogenous RdDM loci. 5S rDNA is present with thousands of copies in the genome and is under RdDM regulation [5]. We digested genomic DNA with HpaII, an enzyme that cuts unmethylated DNA in a CG context, to determine the status of 5S rDNA methylation. We found that, like nrpe1-11, a Pol V mutant, top1α-2 and CPT-treated seedlings had less methylated DNA, as indicated by the increase in intensity of the lower molecular weight restriction fragments (Figure 2D). top1α-7 (also known as mgo1-7 [26]; Figure S1A) has weaker developmental defects than top1α-2Col. The top1β-1 loss-of-function mutant in the Col-0 background (Figure S1A) has no obvious morphological defects (Xigang Liu and Xuemei Chen, unpublished results). CG methylation at 5S repeats was only weakly reduced in top1α-7 mutants and unaffected in top1β-1 mutants. Similarly, DNA blot analyses were conducted to examine CHG methylation at MEA-ISR and 180 bp repeats (Figure S1C and D), and CHH methylation at 5S rDNA repeats. Only a slight reduction in CHG methylation at the 180 bp repeats was detected in top1α-2 (Figure S1D). top1α-2 was indistinguishable from the isogenic Ler parental line in terms of CHG methylation at MEA-ISR (Figure S1C) or CHH methylation at 5S repeats (Figure S1D).
The studies above on a small number of loci revealed a limited role of TOP1α in DNA methylation. In order to obtain a global view of the function of TOP1α in DNA methylation, we performed whole genome bisulfite sequencing (MethylC-seq) on Ler, top1α-2, Col-0, top1α-7, nrpd1-3 (a Pol IV mutant) and nrpe1-11 seedlings. A total of 10 libraries representing one to three biological replicates of the genotypes (Table S1) were sequenced. Acceptable bisulfite conversion efficiency (Table S1) and read coverage (Table S2) were achieved for each library.
We identified differentially methylated regions (DMRs) using established procedures in the literature (see Material and Methods and Text S1). We compared each mutant to its wild-type control in the same biological replicate. We also called DMRs among the three Col-0 replicates to establish the background of spontaneous DMRs in wild type. Despite the high degree of reproducibility of the biological replicates (Table S3), when the three Col-0 replicates were subjected to the DMR analysis, we found thousands of CHH DMRs, but very few CG and CHG DMRs, between any two Col-0 replicates (Table S4A). In MethylC-seq data of three Col-0 replicates from a published study [27], we also identified thousands of CHH DMRs between any two replicates (Table S4B). This suggested that CHH methylation is considerably variable. In light of such variability, we took a conservative approach towards the identification of robust DMRs by considering only the overlap between two biological replicates or mutant alleles. For example, to derive DMRs between wild type and nrpd1-3 or nrpe1-11, we first compared the mutant to wild type within each biological replicate and then retained only DMRs that overlapped in both biological replicates (Table S5B and C). To derive DMRs between wild type and top1α, we first compared top1α-7 to Col-0 and top1α-2 to Ler, and then obtained the overlapped DMRs between the two alleles (Table S5A). In addition, hypervariability (HV) regions that are prone to changes in DNA methylation over generations [28], [29] were subtracted from the overlapped DMRs. The final set of CHH DMRs between wild type and nrpd1-3 (or nrpe1-11) consisted of over 7,500 loci showing reduced DNA methylation in the mutants (Table S5B and C), consistent with the known roles of Pol IV and Pol V in CHH methylation [27], [30]. The final set of DMRs between wild type and top1α consisted of the following: reduced in methylation in top1α — 97 (CHH), 35 (CG), and 0 (CHG); and increased in methylation in top1α — 10 (CHH), 9 (CG), and 1 (CHG) (Figure 3A, Table S5A and Table S6). The overall change in CHH methylation in top1α was very limited in comparison to that in nrpd1-3 or nrpe1-11 (Table S5). Most of the 97 WT-top1α CHH DMRs are in TEs or intergenic regions (Figure 3B). 91% of the WT-top1α CHH DMRs require Pol IV or Pol V for their CHH methylation (Figure 3C). This suggested that TOP1α promotes DNA methylation at a small number of RdDM loci.
Since the methylation-sensitive DNA blot analyses only revealed an effect of top1α alleles on DNA methylation at the 5S and 180 bp repeats and the methylome profiling studies did not support a global role of TOP1α in DNA methylation, we sought to evaluate whether TOP1α is required for the transcriptional silencing of endogenous RdDM loci. qRT-PCR was performed to determine transcript levels from seven well-known RdDM loci. In both wild-type seedlings treated with CPT as well as top1α (both top1α-2 and top1α-7) seedlings, these endogenous siRNA target loci were de-repressed (Figure 4A). This confirmed a role of TOP1α in silencing the RdDM target loci.
We asked whether the release of transcriptional silencing of endogenous RdDM target loci (Figure 4A) in top1α or CPT-treated seedlings was accompanied by a loss of DNA methylation. We performed McrBC-qPCR assays to quantify the levels of DNA methylation amongst different genotypes/treatments at six endogenous RdDM loci. At most of the loci, DNA methylation was reduced in the two top1α mutants, but the reductions were small in top1α-7 (Figure 4B). Treatment of wild-type (Ler) plants with CPT resulted in reductions in DNA methylation at four of the six tested loci (Figure S1E). Although the overall trend of reduced DNA methylation in the two top1α mutants and CPT treated plants agreed with the observed de-repression of these loci, there were also inconsistencies whereby de-repression was not accompanied by reductions in DNA methylation, such as at siR02 in top1α-2 and CPT-treated plants.
This incomplete correlation between TE de-repression and a reduction in DNA methylation prompted us to ask whether TOP1α silences TEs through another mechanism. Previous studies have shown that H3K9me2 is a major repressive mark for transposon silencing and that H3K9me2-dependent silencing acts in concert or in parallel with RdDM [31], [32], [33]. Like DNA methylation, H3K9me2 is targeted to specific TEs through siRNA-AGO4 [7]. Thus, we investigated whether loss of TOP1α function or CPT treatment altered H3K9me2 levels at TEs. Chromatin immunoprecipitation (ChIP)-qPCR showed that H3K9me2 levels at AtSN1, sir02, cluster4, AtGP1, and AtMuI were reduced in both top1α-7 and nrpe1-11 (Figure 4C). We also performed ChIP-qPCR on LUCL seedlings treated with DMSO or CPT. CPT treatment was found to cause a strong reduction in H3K9me2 levels at four TE loci (Figure 4D). As CPT was initially isolated through a chemical genetics screen with LUCL, we asked whether the LUC transgene in LUCL also harbored H3K9me2 and, if so, whether CPT treatment reduced its H3K9me2 levels. Indeed, ChIP-qPCR showed that the d35S of the LUC transgene (region #1 in Figure 2B) harbored H3K9me2, with CPT treatment reducing H3K9me2 levels (Figure 4D).
As H3K9me2, which is introduced by KYP and its paralogs, and CHG methylation, which is deposited by CMT3, act in a self-reinforcing loop, and both H3K9me2 and CMT3 contribute to CHH methylation [14], [16], we asked whether the role of TOP1α in DNA methylation depends on KYP or CMT3. To address this question, we treated Ler (wild-type), kyp-2 and cmt3-7 plants with CPT to inhibit topoisomerase I activity and then assayed DNA methylation at six TE loci. CPT treatment of wild-type plants resulted in reduced DNA methylation at four of the six loci (Figure S1E). The reduction in DNA methylation caused by CPT treatment was minimal at these four loci in either cmt3-7 or kyp-2 (Figure S1E). This suggested that the effects of TOP1α in DNA methylation require CMT3- and KYP-mediated H3K9 dimethylation.
The promotion of DNA methylation and/or H3K9me2 deposition at TEs implicates a role of TOP1α in RdDM, a process that involves Pol IV and Pol V. As topoisomerases are required to release DNA topological tension generated by transcription [17], it would be reasonable to expect that TOP1α is required for the activities of either Pol IV or Pol V. We first tested whether TOP1α is required for the activities of Pol IV, the output of which is the accumulation of 24-nt siRNAs from RdDM target loci. RNA blot analysis showed that siRNA accumulation at several loci was similar in Ler and top1α-2 (Figure S3A). To gain a global view on the potential relationship between TOP1α and Pol IV, we compared deep sequencing profiles of small RNAs from Ler, top1α-2, Col-0, nrpd1-3, and nrpe1-11. The size distributions of all small RNA reads in Ler and top1α-2 were almost identical (Figure S3B). To determine whether TOP1α affects siRNA accumulation at specific regions of the genome, we identified differential small RNA regions (DSRs). While large numbers of DSRs were found in nrpd1-3 or nrpe1-11 relative to the wild-type control, consistent with the essential role of Pol IV and the auxiliary role of Pol V in siRNA biogenesis [3], [5], [6], very few were found in top1α-2 (Table S7). Furthermore, analysis of small RNA abundance throughout the genome did not support a global role of TOP1α in small RNA accumulation (Figure S3C). Therefore, Pol IV activity does not appear to require TOP1α.
Given that we had found 71 WT-top1α DSRs (Table S7), we asked whether the reduced CHH methylation at the 97 WT-top1α DMRs was associated with reduced siRNA levels. We found that only 11 of the 97 DMRs overlapped with WT-top1α DSRs (Figure 3D). A representative of such a locus is shown in Figure S2A. Most of the 97 DMRs did not overlap with the 71 WT- top1α DSRs; two such loci are shown in Figure S2B and C. Therefore, the reduced CHH methylation in top1α could not be explained by reduced siRNA levels. On the other hand, more than 60% of the 97 WT-top1α DMRs overlapped with WT-nrpd1 DSRs (Figure 3D; Figure S2A and B), suggesting that these regions, which require TOP1α for CHH methylation, undergo Pol IV-dependent siRNA production. Therefore, TOP1α must promote CHH methylation at these RdDM loci independently of siRNA biogenesis.
We next tested whether TOP1α promotes the production of Pol V-dependent transcripts. We performed qRT-PCR and RT-PCR to detect Pol V-dependent transcripts from eight loci, MEA-ISR, AtSN1, and six IGN loci that produce such transcripts [9], [30]. At all eight loci, the levels of the Pol V-dependent transcripts were reduced in top1α-2 as compared to Ler (Figure 5A and B). We previously showed that Pol II generates long noncoding transcripts at the soloLTR locus [34]. The accumulation of these transcripts at soloLTR was also reduced in top1α-2 (Figure 5B). Therefore, TOP1α contributes to the production of Pol V-dependent or Pol II-dependent long noncoding transcripts.
As the Pol V- or Pol II-dependent long noncoding transcripts facilitate the recruitment of siRNA-AGO4 to chromatin to ultimately result in RdDM or H3K9me2 deposition, we asked whether TOP1α promotes AGO4 occupancy at these RdDM target loci. ChIP-qPCR was conducted with anti-Myc antibodies in Myc-AGO4 [35] and Myc-AGO4 top1α-2 plants. At four well-known RdDM target loci, AGO4 occupancy was reduced in top1α-2 (Figure 5C).
To determine whether TOP1α might act directly at these RdDM loci, we examined TOP1α occupancy at these loci. We first generated a TOP1α-HA fusion driven by the TOP1α promoter (TOP1α-HA) and introduced it into top1α-2. The morphological phenotypes of top1α-2 plants were completely rescued by TOP1α-HA, indicating that the transgene was functional. We then performed ChIP-qPCR using anti-HA antibodies. TOP1α was found at all six loci examined (Figure 5D).
Beginning with a forward chemical genetics screen with a transcriptionally silenced reporter, LUCL, we have discovered that the well-studied anti-cancer compound CPT can de-repress loci undergoing transcriptional silencing by releasing H3K9 methylation and/or DNA methylation. As topoisomerase I is the cellular target of CPT, this implicates topoisomerase I in transcriptional silencing. Indeed, two top1α alleles, top1α-2 and top1α-7, mimic CPT treatment in de-repressing the expression of endogenous RdDM target loci and reducing H3K9me2 or DNA methylation levels at these loci.
Here, we first consider whether TOP1α acts through RdDM or independently of RdDM to silence TEs. RdDM requires Pol IV and Pol V, which generate siRNAs and long noncoding RNAs, respectively. We show that TOP1α is dispensable for siRNA accumulation, but is required for the production of Pol V-dependent long noncoding RNAs, which are known to recruit siRNA-AGO4 to chromatin. Consistently, TOP1α promotes the recruitment of AGO4 to RdDM target loci. Moreover, 88 out of 97 WT-top1α CHH DMRs with reduced methylation in top1α also require Pol IV or Pol V for CHH methylation (Figure 3C). 5S rDNA loci lose CG methylation in top1α-2 and nrpe1-11 mutants, and provide an example of a genomic region where CG methylation requires TOP1α, Pol IV, and Pol V. These data suggest that TOP1α acts at least in part through RdDM to silence TEs and repeats. However, MethylC-seq analyses revealed that TOP1α has a limited role in DNA methylation. We envision two possibilities for the limited role in DNA methylation observed for TOP1α. First, TOP1α may have a much broader role in DNA methylation in the genome, and the limited effects of top1α mutants on DNA methylation could be due to the redundant functions of TOP1β. So far, our efforts to knock down TOP1β in the top1α-2 background have been unsuccessful. Second, TOP1α's primary functions may lie in the promotion of H3K9 dimethylation, with DNA methylation being a secondary effect of H3K9 dimethylation. From our studies of a limited number of RdDM loci, we found that reduced H3K9me2 levels, but not necessarily reduced DNA methylation, always accompany the de-repression of these loci by CPT treatment or by mutations in TOP1α. Therefore, it is likely that the primary function of TOP1α lies in facilitating H3K9me2 deposition. Consistent with this model, the observed effects of CPT treatment on DNA methylation at four loci require CMT3 and KYP, both of which promote H3K9 dimethylation. Another observation consistent with this hypothesis is that CPT treatment had no effect on LUCH (Figure 1C), a reporter gene that is strictly repressed by CHH methylation and is insensitive to loss of function in CMT3 [36]. As CMT3-mediated DNA methylation requires H3K9me2 [14], we presume that LUCH is not repressed by H3K9me2. The lack of an effect of CPT treatment on LUCH would be consistent with TOP1α acting in TGS through H3K9me2 deposition.
Our finding that TOP1α promotes the production of Pol V-dependent transcripts is consistent with what is known about the function of topoisomerases in bacteria and yeast. Topoisomerases are thought to facilitate transcription elongation by relaxing supercoils [37]. Consistent with this model, loss of Top1 in Schizosaccharomyces pombe results in the accumulation of Pol II in gene bodies [38], [39]. The parallels of Pol V- and Pol II-mediated transcription have recently been highlighted [40], and we propose that TOP1α promotes transcription elongation by Pol V as it does for Pol II.
Although we prefer a model in which TOP1α acts in RdDM by facilitating the production of long noncoding RNAs by Pol V or Pol II, an alternative model cannot be overlooked. Studies in other systems have shown that topoisomerases interact with SMC-containing proteins acting in chromosome compaction [41], [42]. DMS3, a player of the RdDM machinery, contains an SMC domain [43]; therefore, there is a possibility that TOP1α may facilitate RdDM through DMS3.
In summary, we have discovered a role for DNA topoisomerase I in H3K9 methylation and DNA methylation in Arabidopsis. Another study showed that chemical inhibitors of topoisomerases I and II release the epigenetic silencing of an imprinted gene in mouse [44]. Together, these studies point to a role of topoisomerases in epigenetic silencing. Given that CPT is a canonical anti-cancer compound and several of its derivatives are presently used in cancer therapy [25], the emerging role of topoisomerase I in epigenetic gene silencing allude to the mode of carcinogenesis.
Raw data from Illumina sequencing were filtered to remove reads that failed to pass the Illumina quality control and to condense multi-copy reads to a single copy. Hereafter, the reads were mapped to TAIR 10 Arabidopsis genome as well as a C-to-T converted genome using BS_Seeker [45] with default settings. Only perfectly and uniquely mapped reads were retained. For Ler and top1α, which are in the Landsberg ecotype, the reads were mapped to a pseudo-Ler genome generated by incorporating the Ler polymorphisms into the Tair10 Columbia genome (ftp://ftp.arabidopsis.org/Polymorphisms/Ecker_ler.homozygous_snp.txt). This enables the direct comparison of DMR regions between the Columbia and Landsberg samples.
DMRs were identified following a published method [27] with some modifications. In brief, the genome was split into continuous 100 bp windows. The Cs or Ts were counted in each window in the three different contexts (CG, CHG or CHH) separately. Only windows with least 4 Cs each sequenced at least 4 times in the wild-type sample were kept for the DMR analysis. The methylation level for a window was determined as:in which ai denotes the number of read “C”s and bi denotes the number of read “T”s mapping to the ith cytosine site. The methylation level in each window in wild type is then compared to the corresponding window in a mutant. A methylation difference of 0.4, 0.2, and 0.1 for CG, CHG, and CHH, and an adjusted p-value (FDR)<0.01 (Fisher's exact test) were used as the cutoff for defining DMRs.
Additional measures were taken to reduce experimental noise. First, two or three biological replicates/alleles were examined. In deriving initial DMRs, we compared each wild type/mutant pair from the same biological replicate (Table S5). Then, DMRs located within 200 bp of each other were merged. Next, the overlap in DMRs from the two biological replicates/alleles was identified (Table S5). Finally, we removed the DMRs that overlapped with the hypervariability (HV) regions found to be prone to changes in DNA methylation [28], [29] (Table S5).
See Text S1 for Supplemental Methods and Table S8 for oligonucleotides used in this study.
The gene accession numbers used in this study are At5g55310 (TOP1α), At5g55300 (TOP1β), At1g05460 (NRPD1), and At2g40030 (NRPE1). MethylC-seq and small RNA-seq read data have been deposited into NCBI GEO under the identification numbers GSE50691 and GSE50720, respectively.
|
10.1371/journal.pbio.1000151 | Inherent Dynamics of the Acid-Sensing Ion Channel 1 Correlates with the Gating Mechanism | The acid-sensing ion channel 1 (ASIC1) is a key receptor for extracellular protons. Although numerous structural and functional studies have been performed on this channel, the structural dynamics underlying the gating mechanism remains unknown. We used normal mode analysis, mutagenesis, and electrophysiological methods to explore the relationship between the inherent dynamics of ASIC1 and its gating mechanism. Here we show that a series of collective motions among the domains and subdomains of ASIC1 correlate with its acid-sensing function. The normal mode analysis result reveals that the intrinsic rotation of the extracellular domain and the collective motions between the thumb and finger induced by proton binding drive the receptor to experience a deformation from the extracellular domain to the transmembrane domain, triggering the channel pore to undergo “twist-to-open” motions. The movements in the transmembrane domain indicate that the likely position of the channel gate is around Leu440. These motion modes are compatible with a wide body of our complementary mutations and electrophysiological data. This study provides the dynamic fundamentals of ASIC1 gating.
| The acid-sensing ion channels (ASICs) are key receptors for extracellular protons and are becoming increasingly important drug targets. However, their gating mechanism is still not fully understood. The crystallographic structure of the ASIC1 protein provides a clue, but the dynamics of the channel remains to be elucidated. Using computational biology, site-directed mutagenesis, and electrophysiological recordings, we investigated the dynamics of ASIC1 gating. Through “normal mode analysis,” we detected a series of collective motions between the beta turn and transmembrane domain, and between the thumb and finger domains, suggesting a deformation pathway related to channel gating. The intrinsic rotation of the extracellular domain and the collective motions between the thumb and finger domains that are induced by proton binding serve to deform the channel from the extracellular to the transmembrane domain, triggering a “twist-to-open” motion of the channel pore. The relationship between the dynamics and the gating mechanism was experimentally confirmed by a series of complementary mutations in ASIC1 and electrophysiological measurements. Our study also indicated that the likely position of the channel gate is around Leu440 within the ASIC1 protein. We propose a clear model correlating the structural dynamics of ASIC1 and its gating mechanism.
| Extracellular acidosis has profound effects on neuronal function, and acid-sensing ion channels (ASICs) are the key receptors for extracellular protons [1],[2]. ASICs are members of the degenerin/epithelial channel family, which transport Na+ through the cell membrane [1],[3], and serve as a paradigm for all proton-gated channels. Six ASIC isoforms, 1a, 1b, 2a, 2b, 3, and 4, have been identified, among which 1a, 2a, and 2b are expressed in the central nervous system (CNS) [2],[4]. In the CNS, ASICs are tightly connected to synaptic plasticity as well as learning and memory in the brain [5],[6]. In addition, it has been demonstrated that activation or sensitization of Ca2+-permeable ASIC1a channels are responsible for acidosis-mediated ischaemic brain injury [7],[8] and neuroinflammatory damage [2],[9]. ASICs are therefore becoming increasingly important drug targets [2],[10].
While studies have led to the characterization of ASICs and have furthered the role that they play in neurological diseases, one of the remaining challenges is to fully elucidate their gating mechanisms, which are critical for understanding their biological functions and for developing effective therapeutics [2]. These studies are challenged by the complicated process of ASIC gating: it is proton concentration-dependent, can be blocked by amiloride, and its sodium selectivity and variations of desensitization differ from subtype to subtype [2]. In addition, investigations of the ASIC1 gating mechanism have advanced slowly because of the lack of detailed structural information at atomic resolution. The recent low-pH crystal structure of the chicken ASIC1 (cASIC1) at 1.9 Å resolution has revealed the overall organization of the ASIC1, which provides a framework for probing the mechanism underlying the gating of ASICs [11].
The crystal structure of cASIC1 revealed that receptors in the superfamily are homo- or heterotrimers [11]. Structurally, the ASIC1 has three subunits with a stoichiometry α3, forming a chalice-like architecture. Each subunit is composed of two domains, a large extracellular (EC) domain, and a transmembrane (TM) domain. The EC domain resembles a clenched hand, which can be further divided into finger, thumb, palm, knuckle, and β-turn subdomains. The TM domain comprises two transmembrane helices, TM1 and TM2, in a “forearm” arrangement (Figure 1). This structure has provided insight into the architecture of ASICs and raises intriguing questions about its gating mechanism. For example, what is the function of the large EC domain? Where is the gate located? How is proton concentration sensed by the channel and how does this process trigger opening and closing of the channel? In particular, the intrinsic dynamics of the receptor underlying the gating mechanism is still unclear.
Computational simulation has been a promising tool to address the dynamic behavior of biological molecules. Recently, Shaikh and Tajkhorshid carried out molecular dynamics (MD) simulations on cASIC1, which provided useful information for the potential binding sites of cations and protons in ASIC1 [12]. However, the current MD methods are limited to address the local movements of proteins. As a complementary approach, normal mode analysis (NMA) [13]–[15] is efficient for predicting the collective dynamics and inherent flexibilities in biological macromolecules. This method has been widely applied in studying the structure (dynamics)-function relationship for several important ion channels, such as the prokaryotic large conductance mechanosensitive channel (MscL) [16], the potassium ion (K+) channe1 [17]–[19], and the nicotinic acetylcholine receptor (nAChR) [20]–[23].
In the present study, the dynamic behavior of ASIC1 has been studied on the basis of the crystal structure of cASIC1 using NMA along with complementary mutagenesis and electrophysiological experiments. The NMA revealed complementary twisting motions throughout the receptor, with which the ion channel may undergo an open motion. Further analyses on the motion modes detect a series of collective movements among the subdomains of EC domain, which control and induce the motions of channel pore. Furthermore, the twisting motion modes of the TM domain indicate the probable position of the channel gate. Electrophysiological assays on the human ASIC1a (hASIC1a) mutants of a series of key residues associated with the motions support the computational results. This study provides new information on the intrinsic dynamic behavior of ASIC1 motions associated with the channel opening, enabling us to construct a new model for the gating mechanism of the channel. This model, supported by a number of key experimental observations from others as well as our own, for the first time to our knowledge, provides a clear picture of the correlation between the structural dynamics of ASIC1 and its gating mechanism.
NMA is a computational approach that can efficiently predict the collective dynamics and inherent flexibilities in biological macromolecules. Accordingly, we used NMA to detect the intrinsic motion modes of ASIC1. First, we examined the low-frequency modes of ASIC1 produced by NMA because they may reflect the global motions of the ASIC1 channel and are often potentially related to biological functions [24],[25]. The 100 lowest frequency modes resulted from NMA were used to describe the overall motions of the entire channel, since these normal modes are sufficient to capture all the collective motions of the ASIC1, as revealed in other proteins [25],[26]. The detailed motions between the essential subdomains (e.g., thumb and finger) are discussed in the following sections and the motions of some important modes are presented in Figures S1 and S2 and Videos S1, S2, S3, S4, S5. In brief, these modes revealed conformational changes that may involve the whole receptor including a rocking motion of the EC domains around the wrist region which connects the EC and TM domains (Figure 2A) and coinstantaneous rotations of the EC and TM domains around subunit C (Figure 2B).
NMA revealed that the most relevant modes of ASIC1 to its gating mechanism might be modes 1 and 3, which undergo similar motions. In addition to the rock and rotation of the EC domain (Figure 2), the six TM helices underwent a concerted global rotation in both clockwise and anticlockwise directions, as indicated by the rotating angles along the harmonic period (Figure 3A). However, the whole TM domain did not rotate in a simple manner; instead, it may adopt a twisting rotation during the conformational change. As illustrated by the motion of the TM domain in mode 1 (Figure 3B), the direction of the motion as well as the displacement of the movement for different regions along the pore axis (e.g., top and bottom) were different. As shown in Figure 3C, with the exception of TM1(A), the other five helices may undergo twisting motions, changing the direction of their motion around the hinges: L60 for TM1(B), L58 for TM1(C), Q437 and L440 for TM2(A) and TM2(B), and L440 for TM2(C). As a result, the twisting motions of the five helices synthesize a twisting rotation for the whole TM domain as represented in Figure 3B. This result implies that the region near the hinge for the twisting rotation of TM domain is important for controlling the gating of the ASIC1. This hinge is located between G436 and L440 of the TM2 helices (Figure 3D). Because of the asymmetric nature of the ASIC1 structure, the innermost residues of three TM2 helices around the hinge form two rings (designated ring-I and ring-II hereinafter). Ring-I is composed of G436(A), Q437(A), Q437(B), and L440(C), and ring-II consists of L440(A), L440(B), and A444(C) (Figure 3D). Similar to the nAChR [23], the position of the gate of ASIC1 is possibly located near the hinge of the pore. Accordingly, we monitored the radius profiles along the pore axis for each motion mode. Indeed, the channel pore has a bottleneck around the hinge (∼125 Å) (Figure 3E). Rather, the radius profiles for modes 1 and 3 display a visible opening of the channel pore (Figure 3E), suggesting that the twisting motion of TM tends to increase the width of the entire pore. This result indicates that the structure of the closed, desensitized state of ASIC1 intrinsically tends to undergo a twisting motion to open the gate. The hinge sharply divides the pore into two portions: a top segment (110–125 Å) and a bottom segment (125–150 Å). The displacements of the channel diameters for the bottom segment are larger than those of the top, suggesting that the top segment may play a larger role in ion selectivity. Moreover, the motion of mode 3 indicates that the diameter of the bottleneck may maximally increase ∼2 Å (Figure 3E). This displacement in pore diameter is significant for the function of the channel, because it has been suggested that even relatively minor displacements in the gate area can trigger the functional transition of the pore from closed to open [27].
After predicting the positions of the ASIC1 channel gate using NMA, we investigated the molecular basis of the ASIC1 gating further using mutagenesis and electrophysiological experiments. Using human ASIC1a (hASIC1a), we constructed eight mutants around the putative gate position of ring-I and ring-II (Figure 3D): G436A, G436P, Q437A, Q437E, Q437N, Q437R, L440A, and A444G. The profiles of the inward currents elicited by acidic solution in the wild-type (WT) ASIC1 and its mutants are shown in Figure S3. Three mutants (G436A, G436P, and Q437R) resulted in nonfunctional channels as no current was detected in response to pH 5.0 in these mutants, suggesting that G436 and Q437 are key determinants for ASIC1 function. In addition, several mutations (Q437N, L440A, and A444G) altered the reversal potentials of acid-induced currents (Figure S4; Table S1). Consistent with the altered reversal potentials, the relative Na+, Li+, and K+ permeability of these mutated channels were also significantly shifted (Figures 3F and S4). Interestingly, whereas the reversal potentials of the Q437A, Q437E, and Q437N mutants were relatively unchanged (i.e., reflecting unaltered Na+/K+ selectivity) with respect to the WT channel under standard intra- and extracellular ion compositions (Table S1), the selectivity of Na+ over Li+ (pNa/pLi-value) of Q437N mutant decreased significantly (Figure 3F). Furthermore, replacement of Q437 with a positively charged residue Arg (Q437R) caused the loss of function of the mutated channel presumably because of the repulsion between monocations and the Arg side chain in the pore region (Figure S3). These results suggest that Q437 is critical for ion passage. Besides G436 and Q437, we found that two additional residues L440 and A444 affected the ASIC1 gating markedly because the reversal potential and the selectivity of Na+ over K+ (pNa/pK-value) for both L440A and A444G mutants decreased dramatically (Figure 3F). We attribute this phenomenon to the enlargement of ring-I or ring-II due to L440A and A444G replacement, and the mutated channels pass more K+ than Na+ with respect to WT channel. These results indicate that L440 and A444 play important roles in ion selectivity and gating of ASIC1.
It should be emphasized that the identified key residues contributing to the possible gate of ASIC1 are mostly conserved among the superfamily of ASICs as indicated by the sequence alignment (Figure 3G). On the basis of electrophysiological studies, previous studies suggested that G587 and S589 in the α-subunit are key residues to define the selectivity filter of epithelial sodium channel (ENaC) [28]–[31]. In ASIC1 these two residues correspond to G443 and S445, respectively. Mutations of G443C, G443V, G443P, S445C, S445T, and S445V all resulted in a nonfunctional ASIC1 (Figure S3). Interestingly, as mentioned above, mutating A444, a residue located between G443 and S445, to Gly, resulted in an intact ASIC1 with significantly altered ion selectivity (Figure 3F). However, the corresponding residue of A444 in ENaC (S588 of α-subunit) does not affect the ion selectivity of ENaC [29]. Thus the region for ion selectivity in ASIC1 may be the same as that in ENaC, but the residue contribution is different between these two ion channels (Figure 3G). In addition, we found that L440 also played a critical role in the ion selectivity of ASIC1 (Figure 3F). Nevertheless, the corresponding residue of L440 in ENaC (L584 of α-subunit) does not contribute to the ENaC ion selectivity [31]. We attribute this difference to the distinct “gating” mechanisms of ASIC1 and ENaC. In fact, one striking difference between these two channels is that ENaC is constitutively active (i.e., without a functional “gate”) [3], whereas ASIC1 opens following channel gating by agonist (proton) binding. Taking together, we conclude that the region around ring-I and ring-II (Figure 3D) may undergo a substantial conformational change that is coupled to channel gating and constitutes an important regulatory region of ASIC1 function.
After detecting the motions of the pore and the probable gate position of the channel, we asked how the channel pore undergoes the twisting-to-open motion. For this purpose, we examined the relationship between conformational changes of other parts of the EC and the TM domain. Structurally, the β1 and β12 strands are connected to TM1 and TM2, respectively; the β9 and β10 strands are linked to the thumb subdomain; β1, β12, β9, and β10 form a metacarpal plane (M-plane) (Figure 1). There is a loop between β9 and α5, and the β turn of the loop interacts with the TM domain via directly interactions of Y72 with W288 and P287 (Figure 4B). Accordingly, the motions of M-plane and the loop should be essential to the conformational changes of the TM domain. To test this idea, we closely examined the motions of the M-plane and the loop between β9 and α5. Indeed, the motions of the M-plane in most of the modes are associated with the motions of the channel pore. For example, the shearing vibration of the M-plane in mode 3 induces TM1 and TM2 helices to undergo twisting movements, and the bending of the M-plane in mode 11 triggers a swinging motion to the TM domain (see Videos S1 and S2). Intriguingly, the motions of the TM domain are coupled with the motions of the β turn (Figure 4A). The synchronous motions of the β turn along with the TM helices are possibly due to the strong hydrophobic interactions of Y72 with W288 and P287 (Figure 4B), indicating the importance of these residues in the gating of the channel. In addition, on the basis of the NMA modes, we also derived the cross-correlation map, which displays the correlations between the movements of different residues (Figure S1). The cross-correlation map also shows that the movement of Y72 correlates strongly with that of W288 and P287 with correlation coefficients of 0.92 and 0.82, respectively (Figure 4C).
Key interactions between the β turn and the TM domain important for the gating mechanism were further characterized by electrophysiological experiments. To this end, we designed mutations to respectively disrupt the π–π stacking interaction between Y72 and W288, the C-H···π hydrogen bonding [32] between P287 and Y72, and the disulfide bond between C291 and C366 (Figure 4B). We found that the W288A mutation abolished the opening of the channel (Figure 4D), and the Y72A mutation greatly decreased the pH sensitivity of the channel. These results suggest the importance of the hydrophobic (mainly π–π stacking) interaction between W288 and Y72 for controlling the channel gating (Figure 4B). To further test this hypothesis, we investigated the pH sensitivity of the Y72F mutant, which may keep the π–π stacking intact. As expected, the electrophysiological characteristic of this mutant was unchanged (Figure 4D).
The C-H···π hydrogen bond between P287 and Y72 represents another important interaction that may be responsible for the collective motion of the β turn with the TM domain. In addition, because P287 is located at the lip of the β turn, it may play a structural role in stabilizing the β-turn conformation and likely affect channel gating. As predicted, the P287G mutation abolished the channel opening capacity (Figure 4E). On the other hand, P285G and P286G mutants were unaffected, suggesting that these two residues are not important for the channel gating. Finally, we studied the disulfide bond (S–S bond) interactions between C291 and C366 (Figure 4B). In addition to stabilizing the conformation of the β turn, this disulfide bond is a linkage between the β turn and the β10 strand, which is a bridge that conducts the motions of the finger and knuckle to the β turn as will be discussed in next section. Hence, we hypothesized that this disulfide bond may also play a role in the channel gating. To test this idea, electrophysiological characterizations were performed on C291A and C366A mutants. Both C291A and C366A mutations resulted in termination of the channel opening activity (Figure 4D).
Of note, the importance of the interaction between Y72 and W288 to the channel gating was also addressed by Li et al. [33]. The result was reported online during the reviewing process of this manuscript.
The crystal structure of cASIC1 suggests that one end of the β turn links to the thumb domain and the other end connects to the knuckle domain through the β9 strand; the S–S bond between C291 and C366 also plays an important role to the interaction between the β turn and the M-plane. In addition, the TM1 and TM2 helices connect with the finger and knuckle domains through the β1 and β12 strands, respectively (Figure 1). This structural arrangement indicates that the EC domain may communicate with the TM domain and β turn through a series of collective motions, implying that the collective motions of different regions in the EC domain are possibly relevant to the channel gating. Accordingly, we analyzed the motion modes of the EC domain. Interestingly, all motion modes revealed that the thumb always moves correlatively with the finger (Figure 5A). This result suggests that channel gating may be facilitated by the attractive forces between the thumb and finger domains. Essential hydrogen bonding and hydrophobic interactions between the thumb and finger are shown in Figure 5B and 5C, respectively. To test this hypothesis, we first disrupted several pairs of hydrogen bonds (H-bonds) and electrostatic interactions between the thumb and finger via site-directed mutagenesis, including D238···D350, E239···D346, and R191···D350 pairs, which are identified by Jasti et al. as tentative proton binding sites (Figure 5B) [11]. Abolishment of the R191···D350 and E239···D346 H-bonds by substituting R191 with Ala, E239 with Gln or Lys, and D346 with Asn, and the D238···D350 H-bond by substituting D238 with Ala or Asn had profound effects on the pH50 (pH of half-maximal activation) of the acid-induced currents (Figure 5D); all of these mutations reduced the pH50 values of the ASIC1 (Figure 5D). Similar effects had been observed with the D346N mutation by Jasti et al. [11]. Binding free energy (ΔGbinding) calculations indicate that all these mutations decrease the binding affinity between the thumb and finger (see Discussion, Figure 5E, and Table S2). To firmly establish the importance of the attractive interaction between the thumb and finger to the opening of the gate, we designed another two mutants, D238K and D238S, which might enhance the interaction between the thumb and finger. Binding free energy calculations are consistent with this notion, as the ΔGbinding values between the two subdomains for these two mutations are respectively reduced by ∼10.0 and 3 kcal/mol relative to WT (Figure 5E; Table S2). Consistently, these mutations also led to higher pH50 values than that of the WT channel (Figure 5D).
Hydrophobic interactions are another dominant component to the collective motion between the thumb and finger. NMA results indicate that the pairs of hydrophobic interaction between these two subdomains move together with a high correlation as their cross-correlation coefficients (Cij) are larger than 0.9 (Figure 5C). This suggests that mutations that decrease the hydrophobic interaction between these two subdomains would cause the channel to respond to a lower pH. We thus mutated H328 and P338 to Ala to weaken the hydrophobic interaction between the thumb and finger domains. The electrophysiological results are consistent with the computational predictions, i.e., both mutations decreased the pH50 values (Figure 5D). Remarkably, the calculated binding free energies between the thumb and finger for the WT channel and all its mutants correlate well with the pH50 values with a high correlation coefficient, R2 = 0.61 (Figure 5E). These results, together with our NMA analysis, strongly support that the collective motions between the thumb and finger are of significance to the channel gating.
Our studies revealed collective motions that occur amongst the subdomains of the EC domain, so we sought to map the deformation pathway related to channel gating because of its importance in understanding the overall function of ASIC1. In addition to the collective motions between the β turn and TM domain and between the thumb and finger domains mentioned above, bending and swing vibrations between the finger and knuckle were also revealed by NMA (Figure S2). These motions lead to bending and twisting motions of the M-plane (β1, β12, β9, and β10), which further evoke different motions of the TM domain (Videos S3, S4, S5). Accordingly, the NMA modes clearly show the deformation pathway for domain motions: collective motions of thumb with finger (class I motions) couple with the vibrations between finger and knuckle (class II motions), which further associate with the bending and twisting motions of the M-plane (class III motions) (Figure 6A). Both class I and class III motions are able to trigger the TM domain to undergo rotation and twisting motions and the β turn to engage in swinging motions; the latter movement enhances the motions of the TM domain through noncovalent interactions (Figure 4B). This deformation pathway demonstrates the inherent structural flexibility of ASIC1 for implementing their functions and also implies the functional importance of the EC domain of the receptor.
To further understand the dynamic behavior of each subunit in isolation versus its homotrimer conformation and their relation to gating, NMA was also performed on the monomer using the structure of the subunit taken from the X-ray crystal structure of cASIC1 [11]. Again, the first 100 lowest frequency normal modes were obtained. Most of the motions of the monomer are, in general, similar to those of each subunit within the trimer, suggesting that the intrinsic properties of each subunit determine the receptor's motions, which further control the gating of the whole channel. Figure 6B shows the differences in flexibility between the monomer treated as an isolated subunit (red trace) versus part of the trimer (black trace), as shown in the profile of the root-mean-square-fluctuations (RMSFs) from the 100 lowest frequency normal modes. Here, we only use subunit A to discuss the differences, because similar results were obtained for subunits B and C. As shown in Figure 6B, the flexibility of each domain is restricted in the trimer compared to the monomer. In both the monomer and as part of the trimer, the tips of the thumb and finger are two of the most flexible portions. This result is consistent with the harmonious motions of these two domains, demonstrating again the importance of their motions in gating. In fact, Jasti et al. have hypothesized that the cleft between the thumb and finger contains an acidic pocket for sensing acidic conditions (i.e., proton levels) [11]. Moreover, our electrophysiological experiments on the mutants of the residues along the acidic pocket also suggest that the motions of these two domains are linked to the physiological function of this protein (Figure 5). Another region that displays pronounced flexibility in the individual monomer is the knuckle domain. Although the mobility of the knuckle is restricted in the trimer, its tip still shows high fluctuation (Figure 6B), enabling vibration motions between the knuckle and finger. Other regions showing high fluctuation are the TM1 and TM2 helices, but their flexibility is also restrained in the trimer, especially the TM2 helix (Figure 6B). Remarkably, the RMSF profiles for both TM1 and TM2 helices in the trimerized channel form inverted parabola-like curves, indicating that the two ends are more mobile than the middle, and the flexibility of the hinge position is seriously restricted (Figure 6B). This result is consistent with the twisting motion modes and site-directed mutagenesis results (Figures 3 and S3; Table S1). When existing as part of the trimer, most of the β strand's flexibility is also reduced, but they still undergo local motions induced by the motions of thumb, finger and knuckle. The collective evidence obtained by the NMA results and the flexibility map suggest that the deformation pathway involves the following domain motions: thumb, finger, and knuckle are activists of the receptor, their dynamic behaviors concomitantly propagate to the palm, leading β1, β9, β10, and β12 to undergo bending and twisting motions and the β turn to undergo a swinging vibration. These motions are further transmitted to the TM domain, triggering a twisting motion that opens the channel pore.
While X-ray crystal structures of the closed, desensitized-like state of the cASIC1 channel has been determined recently [11], some important issues of the gating mechanism of the ASIC channels are still unknown. Here, we use a combination of computational simulations together with electrophysiological measurements to investigate the relationship between the inherent dynamics of ASIC1 and its gating mechanism. The deformation pathway derived from the NMA calculations is compatible with our site-directed mutagenesis and electrophysiological characterizations, providing a plausible model for the conformational changes underlying the gating mechanism of ASIC1.
We have carried out normal mode analyses on both trimer and monomer forms of cASIC1 by using the X-ray structure, representing the closed-form, as the starting model. The first 100 lowest frequency modes were obtained and analyzed for each of the starting models. The NMA modes are of particular interest because they reveal intrinsically dynamic motions of ASIC1 that are essential to the gating process and function of the receptor.
The lowest frequency modes of each subunit in isolation (monomer) and in the trimer revealed the most flexible portions of cASIC1. Each subunit is more flexible in the isolation state than in the trimer, suggesting that the trimerization of the receptor restrains the mobility and plasticity of the subunits (Figure 6B). Still, the thumb, finger, and knuckle of the EC domain maintain enough flexibility to carry out the receptor's function in gating (Figure 6B). The TM domain shows distinct flexibility. In general, TM1 is more flexible than TM2, and the two ends of both TM helices are more flexible than their middle parts, resulting in parabola-like curves of RMSF profiles for these two helices, of which the valleys are located around the hinge point of the channel pore (Figure 6B). This indicates that the flexibility of the receptor is amenable to the requirement of gating.
We investigated the holistic motions of the EC domain by analyzing the motion modes. The EC domain as an entire body may undergo rotation around the pore axis and swing round the wrist region of the receptor. For example, the EC domain in mode 2 displays a rocking vibration (Figure 2A), and in modes 1 and 3 experiences a gyroscopic precession like a rotating peg-top, as viewed by superposing the TM domain (Figure 2B). These motions trigger the TM domain to undergo a twisting motion, which is possibly related to the channel gating as demonstrated by the mutagenesis and electrophysiological characterization. In addition, the rigid body motions of EC not only can trigger the conformational changes of the TM domain, but also drives the subdomains within EM to undergo further motions specific to gating. In particular, the thumb and finger domains may undergo swinging and rocking motions (Figure 5A); in addition, bending and swinging vibrations were also found between the finger and knuckle domain (Figure S2). Collectively, these motions provoke the M-plane and β turn to undergo vibration and swinging motions, which directly stimulate the functional motions of the TM domain (Figure 6A). Accordingly, the inherent flexibility and dynamics of ASIC1 are tightly associated with the channel gating.
Two important findings that were significant to the gating mechanism were uncovered amongst the normal modes of EC domain. The first is the consistent motions between thumb and finger domains (Figure 5A). This result indicated the vital role of the attractive interaction between these two subdomains in the gating process. Indeed, a series of mutations and corresponding electrophysiological measurements have confirmed this hypothesis (Figure 5D and 5E). Another point is that the β turn always moves concomitantly with the motions of the TM domain, demonstrating that the interaction between the β turn and TM domain dominates the gating (Figure 4A–4C). This conclusion has also been verified by our mutagenesis and electrophysiological experiments (Figure 4D and 4E).
NMA also revealed a notable twisting to open pore motion located in the TM domain, which is directly associated with channel gating (Figure 3). Most importantly, the twisting motions of the TM helices are evoked by the motions of the EC domain (Figure 6A). Meanwhile, the hinge of twisting motions positions around the bottleneck of the channel pore (ring-I and ring-II, Figure 3D) suggests that this region (Q437–A444) might be the gate of the channel. This computational prediction has also been validated by using a series of mutagenesis experiments (Figures 3F and S3; Table S1). Clearly, the architecture of ASIC1 is organized to allow for communication between the EC and TM domains through a deformation pathway, which triggers the gating process of the channel (Figure 6A).
The function of the ASIC large EC domain in the gating mechanism remains to be elucidated. Results of this study indicate that motions of all subdomains and regions of the EC domain may collectively stimulate the motions and conformational changes of the TM domain, affecting the shape of the channel pore (Figure 6A). The dynamic pathway seems to be associated with the function of EC domain, which raises an intriguing question: what is the driving force that triggers these motions? A large body of evidence has indicated that the channel opens in response to hydrogen ions, allowing sodium ions to pass into the cell [1],[2]. The recently determined X-ray crystal structure of cASIC1 [11] and MD simulation [12] suggest that H+ ion may bind to an acidic pocket between thumb and finger subdomains. This process provides the energy to trigger movement of the EC domains; however, how the H+ binding would drive such events is still unclear. Here, we provide a possible explanation based on our NMA and mutagenesis experiments.
On the basis of the holistic motion modes of the EC domain and the consistent motion of thumb and fingers to the channel gating described above, we hypothesize that the initial driving force for the EC domain movements is the attraction between thumb and finger (Figure 6A). H+ binding to the acidic pocket enhances the interaction between these two domains, which heightens their intrinsic motion during gating process. This hypothesis has been verified by a series of mutations and electrophysiological determinations: mutations that either abolish the H-bonds or weaken the hydrophobic interaction between these two subdomains shift the dose–response curve to a lower pH region and decrease the pH50 values; and the mutations of D238K and D238S that might enhance hydrogen bonding between the thumb and finger shift the dose–response curve to the higher pH region and increase the pH50 (Figure 5D and 5E). Theoretical calculations are consistent with these results: removing the H-bonding interactions and decreasing the hydrophobic interactions between thumb and finger domains decrease the binding free energy between these two domains whereas the D238K and D238S mutants, which strengthen the hydrogen bonding interaction, increase the binding affinity between the two subdomains. Moreover, the binding free energies correlate well with the pH50 values (Figure 5E). This computational result demonstrates again that the attractive interaction between thumb and finger might be a driving force to channel gating.
The current study shows that ASIC1 exhibits an intimate connection between the intrinsic structural dynamics and the gating process. On the basis of the NMA results and related mutagenesis and electrophysiological experiments, we propose a dynamic mechanism for the proton-activated gating of ASIC1. The first step of the mechanism is the binding of H+ to the acidic pocket [11]. In contrast to an earlier hypothesis that was raised, on the basis of inspection of the crystal structure of cASIC1 [34], we believe that H+ binding does not displace the thumb during gating, but instead enhances the binding affinity between thumb and finger through strengthening the H-bonds formed between acidic residues. This conclusion is based on the collective evidence gathered from the mutagenesis and electrophysiological measurements as well as binding free energy calculations on the WT channel and mutants (Figure 5). Thus, H+ binding induces thumb and finger domains to move close to each other, thereby initiating and magnifying a series motions along the intrinsic deformation pathway of the receptor (Figure 6A). These motions trigger conformational changes of the TM domain, which provoke the TM domain to undergo a twisting motion to open the gate. This mechanism is clearly of general evolutionary significance of ASIC. The hand-like structure of the monomer and the chalice-like architecture of the entire receptor provide an elegant solution for controlling the gating mechanism of ASIC.
The atomic coordinates for the crystal structure of cASIC1 (Protein Data Bank [http://www.rcsb.org/pdb] entry 2QTS) [11] was used as the starting structure in a series of computational simulations and calculations. NMA was conducted using the web server developed by Delarue et al. (http://lorentz.immstr.pasteur.fr/nomad-ref.php) [35],[36]. During the NMA simulations, the single-parameter Hookean potential, a simplified all-atom potential [35], was used (Equation 1),(1)where dij is the distance between two atoms i and j, is the distance between the atoms in the 3-D structure, c is the spring constant of the Hookean potential (assumed to be the same for all interacting pairs), and Rc is an arbitrary cut-off. In this study, Rc was set to be 10 Å.
RMSF of each atom from the nontrivial modes and frequencies was calculated using the method of [37],(2)where mi is the mass for atom i; ωk is the vibration frequency of mode k; aik is the ith components of the kth eigenvector.
Cross-correlations (Cij) of atomic motion were computed with the modes and frequencies derived from the NMA using Equations 3 and 4 [37],(3)(4)where mi and mj are the masses for atoms i and j; ωk is the vibration frequency of mode k; aik and ajk are the ith and jth components of the kth eigenvector.
All constructs were expressed in CHO cells. Transient transfection of CHO cells was carried out using Lipofectamine 2000 (Invitrogen). Electrophysiological measurements were performed 24–48 h after transfection. The cDNA encoding hASIC1a was a generous gift from Jun Xia (The Hong Kong University of Science and Technology, Hong Kong, China).
The incubation solution contained the following components (in mM): 124 NaCl, 24 NaHCO3, 5 KCl, 1.2 KH2PO4, 2.4 CaCl2, 1.3 MgSO4, and 10 glucose, aerated with 95% O2/5% CO2 (to a final pH of 7.4). The standard external solution contained (in mM): 150 NaCl, 5 KCl, 1 MgCl2, 2 CaCl2, and 10 glucose, buffered to various pH values with either 10 mM HEPES (pH 6.0–7.4), or 10 mM MES (pH<6.0). The standard patch pipette solution for whole-cell patch recording was (in mM): 120 KCl, 30 NaCl, 1 MgCl2, 0.5 CaCl2, 5 EGTA, 2 Mg-ATP, 10 HEPES. The internal solution was adjusted to pH 7.2 with Tris-base. Unless otherwise noted, the electrophysiological recordings were carried out under standard conditions.
For measurement of the relative permeability of Li+, K+, and Na+, the internal solution contained (in mM): 150 NaCl, 10 EGTA, and 10 HEPES, and the external solution contained (in mM): 150 test monovalent cation (X), 10 HEPES (replaced with MES when pH is 5.0), 10 glucose, and 2 CaCl2. The relative permeability of Li+ and K+ over Na+ was measured by comparing the reversal potentials when the external solution contained LiCl, KCl, or NaCl with internal NaCl in each case. The osmolarities of all these solutions were maintained at 300–325 mOsm (Advanced Instruments).
Solutions with different pH values were applied using a rapid application technique termed the “Y-tube” method throughout the experiments [38]. This system allows a complete exchange of external solution surrounding a neuron within 20 ms.
The human ASIC1a cDNA was subcloned into the pEGFPC3 vector (Promega Corporation). Each mutant was generated with the QuikChange mutagenesis kit (Stratagene) in accordance with the manufacturer's protocol.
The electrophysiological recordings were performed in the conventional whole-cell patch recording configuration under voltage clamp conditions. Patch pipettes were pulled from glass capillaries with an outer diameter of 1.5 mm on a two-stage puller (PP-830, Narishige Co., Ltd.). The resistance between the recording electrode filled with pipette solution and the reference electrode was 4–6 MΩ. Membrane currents or potentials were measured using a patch clamp amplifier (Axon 700A, Axon Instruments) and were sampled and analyzed using a Digidata 1320A interface and a computer with the Clampex and Clampfit software (version 9.0.1, Axon Instruments). In most experiments, 70%–90% series resistance was compensated. Unless otherwise noted, the membrane potential was held at −60 mV throughout the experiment under voltage clamp conditions. All the experiments were carried out at room temperature (22–25°C).
Results were expressed as the mean±SEM. Statistical comparisons were made with the Student's t-test. The permeability ratios of pNa/pLi and pNa/pK were determined by the modified Goldmann-Holdgkin-Katz equation: pX/pNa = exp (ΔVrevF/RT) due to the equimolar cations in the external and internal solution, where X represents the test cation, ΔVrev is the change in reversal potential when Na+ was replaced by the tested cation, F is the Faraday constant, R is the gas constant, and T is the absolute temperature.
|
10.1371/journal.pgen.1002470 | USF-1 Is Critical for Maintaining Genome Integrity in Response to UV-Induced DNA Photolesions | An important function of all organisms is to ensure that their genetic material remains intact and unaltered through generations. This is an extremely challenging task since the cell's DNA is constantly under assault by endogenous and environmental agents. To protect against this, cells have evolved effective mechanisms to recognize DNA damage, signal its presence, and mediate its repair. While these responses are expected to be highly regulated because they are critical to avoid human diseases, very little is known about the regulation of the expression of genes involved in mediating their effects. The Nucleotide Excision Repair (NER) is the major DNA–repair process involved in the recognition and removal of UV-mediated DNA damage. Here we use a combination of in vitro and in vivo assays with an intermittent UV-irradiation protocol to investigate the regulation of key players in the DNA–damage recognition step of NER sub-pathways (TCR and GGR). We show an up-regulation in gene expression of CSA and HR23A, which are involved in TCR and GGR, respectively. Importantly, we show that this occurs through a p53 independent mechanism and that it is coordinated by the stress-responsive transcription factor USF-1. Furthermore, using a mouse model we show that the loss of USF-1 compromises DNA repair, which suggests that USF-1 plays an important role in maintaining genomic stability.
| UV is responsible for DNA damage and genetic alterations of key players of the Nucleotide Excision Repair (NER) machinery promote the development of UV-induced skin cancers. The NER is the major DNA–repair process involved in the recognition and removal of UV-mediated DNA damage. Different factors participating in this DNA repair are essential, and their mutations are associated with severe genetic diseases such as Cockayne Syndrome and Xeroderma Pigmentosum. Here, we show for the first time that the specific regulation of expression in response to UV of two NER factors CSA and HR23A is required to efficiently remove DNA lesions and to maintain genomic stability. We also implicate the USF-1 transcription factor in the regulation of the expression of these factors using in vitro and in vivo models. This finding is particularly important because UV is the major cause of skin cancers and dramatically compromises patients with highly sensitive genetic diseases.
| Maintaining the integrity of the genome through cell generations is critical to ensure accurate cell function and to avoid tumor formation. Cells are continuously challenged by environmental insults and they are equipped with specific and efficient defense machinery to remove any DNA alterations. The importance of these processes is underscored by genetic disorders, such as Bloom, Werner, Cockayne Syndromes and Xeroderma Pigmentosum (XP) that result from their impaired function. Despite an enormous amount of progress in identifying the protein complexes and their detailed function in DNA repair pathways, very little is still known about whether these complexes are regulated at a gene expression level.
The skin is a good model in which to address this question because it is the organ most exposed to environmental stresses. The principal cause of DNA damage in the skin is solar irradiation, which induces cyclobutane pyrimidine dimers (CPD) and 6-4 photoproducts in the epidermal cell layers and which, if not removed, can promote skin cancers. The Nucleotide Excision Repair (NER) is the most versatile DNA repair system and is responsible for specifically and constantly eliminating any distorted DNA lesions, including these dimers [1]–[6]. NER can be divided into at least two sub-pathways, Global Genome Repair (GGR) [4] and Transcription Coupled Repair (TCR) [3], [5], [7]. Which one is triggered depends on where the distorted DNA is localized on the genome. GGR, as its name implies, is responsible for removing DNA lesions across the genome including the non-coding part, silent genes and the non-transcribed strands of active genes. The TCR sub-pathway, on the other hand, is dedicated to repairing only DNA lesions detected during transcription and is responsible for removing bulky DNA lesions from the transcribed strands of active genes [2], [3]. The sequence of events implicated in the GGR and TCR DNA repair pathways include: DNA lesion-recognition (the rate limiting step), DNA-unwinding, excision and repair synthesis and except for the damage recognition step, they share common processes and protein machineries for the remaining events [2]. In the GGR sub-pathway, the XPC-HR23 complex is responsible for the recognition of DNA lesions. The DNA-binding protein, XPC, has a strong affinity for damaged DNA [6], [8], [9]. However, its interaction with the evolutionarily conserved HR23 proteins (homologues of the yeast RAD23) is critical for its function. HR23 increases the physiological stability of XPC and thereby its damage recognition activity [10]. In the TCR sub-pathway, lesion recognition occurs through the arrest of the elongating RNA Pol II (RNAPII) when it encounters DNA damage. This essential step initiates the subsequent recruitment of the repair factors CSA and CSB, which are required for the removal of the lesion [5].
While it is well accepted that the functional activity of proteins responsible for the removal of DNA-lesions are regulated and indeed crucial to ensure an orchestrated cascade of events [6], it is not known whether this involves modulation in gene expression. This study addresses this question by using an intermittent UV-irradiation protocol and investigates the gene expression profile of key players in the NER DNA-damage recognition step. We show that UV-induced DNA photo-lesions initiate a specific program of gene expression with the stress responsive transcription factor Upstream Stimulatory Factor 1 (USF-1) playing a central role [11]–[13]. Using a combination of in vivo and in vitro assays we demonstrate, in our system, that there is a specific and coordinated regulation of HR23A, HR23B, CSA and CSB genes and their protein levels in response to UV-mediated DNA damage. We show that up-regulation of both HR23A and CSA is driven by a common p53 independent mechanism involving USF-1. Furthermore, we provide novel evidence that while HR23A and HR23B share a similar function in DNA-damage recognition, their temporal expressions are different, which may imply that they function at different times, in response to UV-induced DNA-damage.
Results from this study have important implications for our understanding of the role of gene expression regulation in the DNA-damage repair pathways and reveal a role for USF-1 in DNA-repair and in maintaining genome integrity.
Very little is known about how genes that encode key components of the NER recognition step are regulated at a transcriptional level, to mediate their role in DNA lesion recognition. We thus performed a UV-induced DNA-lesion protocol (Figure 1A), which generates immediate DNA photo-lesions through repetitive doses of short wavelength UV pulses rather than delivery of a single high dose [14], [15]. Using RT-qPCR, we then followed the expression of genes specifically involved in the recognition events of TCR (CSA and CSB) and GGR (HR23A and HR23B), immediately post-irradiation. Cultured mouse and human keratinocytes (XB2, HaCaT) were irradiated with four to eight 10 J/m2 UV pulses (254 nm) at 15 min intervals and collected at the indicated times (from 30 min to 5 h) after the last pulse (Figure 1A). We first checked for the presence of CPD post UV-irradiation (Figure S1) and for cell viability over 24 h confirming that the irradiation procedure was inducing DNA-damage without compromising cell numbers (90% and 75% cell survival at respectively 3 and 24 h) (Figure 1B). The irradiation protocol resulted in a significant increase of CSA mRNA levels (6-fold after 30 min), while the abundance of CSB gene transcripts was not affected (Figure 2A). CSA mRNA levels remained elevated at 1 h and decreased from 2 hours. Comparable results were obtained in p53-deficient human HaCaT keratinocytes [16] (Figure S1A). Up-regulation of CSA gene expression was accompanied by a significant increase in CSA protein levels (Figure 2B), peaking at 3 hours compared to un-stimulated cells, where CSA protein is almost undetectable. The increase of CSA protein levels following UV-irradiation was also observed by immunofluoro-staining in XB2 keratinocytes (Figure S1B). This increase in CSA protein levels is significantly reduced over time when cells were pre-treated with α-amanitin, an agent that disrupts transcription. These results indicate that the increase in protein levels results in part from transcriptional regulation.
We next investigated the regulation of the GGR pathway-specific mediators, HR23A and its homologue HR23B, following the same irradiation protocol (Figure 1A). While no significant effect was observed on HR23B mRNA levels, the irradiation protocol resulted in a mild but reproducible (6 independent experiments) and significant increase of HR23A mRNA levels (1.5-fold at 4 h) (Figure 2C). Comparable results were obtained in p53-deficient human HaCaT keratinocytes (Figure S1C). In parallel with UV-induced HR23A transcripts, protein levels increased progressively, reaching a 4-fold increase at 8 hours. This effect was abrogated when cells were pre-treated with α-amanitin (Figure 2D). The increase of HR23A protein levels following UV-irradiation was also observed by immunofluoro-staining in XB2 keratinocytes and correlates with an increase in CPD (Figure S1D). In contrast, HR23B protein levels decreased over time after UV-irradiation suggesting that it is regulated post-transcriptionally since there was no change in its mRNA levels (Figure 2C). These results indicate that HR23A and HR23B are regulated differently.
UV-induced transcription is a tightly regulated process that involves both cis and trans UV-responsive elements. We thus explored potential cis/trans factors involved in UV-induced regulation of CSA and HR23A expression by in silico analysis of their respective proximal promoter sequences using Consite and Zpicture (Rvista 2) softwares [17], [18]. We found that both promoters belong to the TATA-less class and that their proximal regions contain consensus E-box motifs (CACGTG) upstream from the transcription start site (TSS) at −246 for CSA (Figure 3A) and −154 and −37 for HR23A promoter (Figure 3D), which are highly conserved across species. By contrast, no such conserved E-box motif was found in the CSB and HR23B promoter regions (data not shown and Figure 3D). Given that USF-1 acts as a key player of UV-regulated gene expression by interacting specifically with E-box cis-regulatory elements (CACGTG) as homodimers or as heterodimers with its partner USF-2 [19]–[21], we suspected that CSA and HR23A may be USF-1 target genes. To test this hypothesis, we performed chromatin immunoprecipitation (ChIP) assays using antibodies specific for either USF-1 or USF-2. DNA recovered from the HaCaT cell line was amplified by PCR using primers targeting distinct promoter sequences (Figure 3B). Results showed specific amplification products corresponding to the binding of USF-1 and USF-2 factors to the CSA proximal promoter (−246 bp), whereas no PCR product was observed for the distal region (−2 kb), the proximal CSB promoter or with non-specific IgG antibodies (Figure 3B). We next investigated the impact of UV-mediated DNA-damage (8×10 J/m2) on the recruitment of USFs to the CSA proximal promoter over time. UV-irradiation specifically and rapidly (15 min) promoted an 8-fold enrichment of USF-1, but not USF-2, at the CSA proximal promoter (Figure 3B). Using in vitro binding assays (EMSA), we tested the ability of USFs to bind the identified conserved E-box motif (−246 bp), which was also present in the ChIP amplified product. Specific DNA-protein complexes were obtained with a probe spanning the E-box motif at −246 (Figure 3C), which were efficiently competed by homologous cold wild type, but not mutant probe. These DNA-protein complexes were super-shifted by antibodies against either USF-1 or USF-2 but not by non-specific antibodies (IgG or Tbx2). No DNA-protein complex was formed with probes carrying mutated E-box sequences.
In vivo DNA-binding assays revealed also that USF factors interact specifically with the HR23A proximal promoter but not the distal promoter or HR23B promoter and that UV-irradiation promotes the interaction of the USF-1 transcription factor by a 3-fold and USF-2 by a 2.5-fold enrichment (Figure 3E). As shown previously for CSA, EMSA assays confirmed the DNA-protein complexes spanning the conserved −154 and −36 E-box motifs, present also in the ChIP amplified products (Figure 3F). Competition between the two E-box probes did not reveal any preferential binding site (data not shown). In addition to the E-box sites present in the HR23A proximal promoter, in silico analysis identified conserved GC-rich regions (−131 and −18 from TSS) (Figure 3D) known to interact with members of the SP1/SP3 transcription factor family [11]. To examine their respective contribution to the regulation of HR23A expression, we performed in vitro and in vivo DNA-binding assays as described above. Specific protein-DNA complexes were formed only in the presence of the −131 intact GC box that interacts with SP1 and SP3 transcription factors (Figure 3G). Also, under the experimental conditions used, only SP3 was able to bind the HR23A proximal promoter in vivo and SP3 loading was not affected by UV-irradiation. Interestingly, a comparable SP3 binding profile was obtained with the HR23B proximal promoter that shares homologous GC rich sequences with HR23A (Figure 3H) but whose mRNA levels were not modulated by UV, suggesting that the GC motifs might not be UV-inducible. Taken together these results provided compelling evidence that, in response to UV-irradiation, USF-1 interacts directly with the CSA and HR23A proximal promoters, suggesting it may be responsible for the UV-induced CSA and HR23A expression observed in this study.
The relevance of the E-box motifs in mediating USF regulation of the CSA and HR23A promoters was next assessed by luciferase assays. We first transiently co-transfected XB2 cells with a wild type (WT) and E-box mutated CSA promoter (−847/+1) cloned upstream of a luciferase reporter (pGL3-Luc) (Figure 4A) and USF-1 or USF-2 expression vectors (pCMV) [12], [20]. Both USF-1 and USF-2 expression vectors led to significant increases of CSA-luciferase activity (Figure 4B). Following UV-irradiation, WT CSA promoter activity demonstrated a rapid, 6-fold significant increase (30 min after the last UV pulse) (Figure 4C). Furthermore, this intact E-box cis-regulatory element proved to be required for UV-induced activation and to mediate the binding of USF trans-activators (Figure 4B–4C).
We next transiently co-transfected XB2 cells with a WT and E-box mutated HR23A promoter (−186/+73) construct cloned upstream of a luciferase reporter (Figure 4D). USF-1 and USF-2 expression vectors led to mild but significant increases of HR23A-luciferase activity (Figure 4E). In response to UV-irradiation, HR23A promoter activity increased slightly but significantly only in the presence of the USF-1 expressing vector (Figure 4G).
Interestingly, when the two E-box motifs were mutated, we observed a 4-fold reduction of the basal HR23A-luciferase activity (Figure 4F) and the USF-1 mediated UV-response was abrogated (Figure 4G). Mutation of the −131 GC-rich motif did not significantly affect HR23A basal activity and did not impair the USF-mediated UV-response (Figure 4F–4G), supporting the idea that the UV response is driven by the USF/E-box protein/DNA complexes.
The physiological significance of the regulation of CSA and HR23A by USF-1 in response to UV-induced DNA damage was established using genetic approaches with XB2 USF-1 knock-down (KD) cells (Figure 5) and USF-1 knock-out (KO) mice (Figure 6) [13].
Firstly, we quantified the level of CPDs in cells in which either USF-1 or CSA or HR23A mRNA was targeted with two different and independent siRNA. While the level of CPDs in the un-stimulated USF-1-KD cells (siUSF-1 N°1 and N°2) remained low and comparable to the control cells (siCtrl N°1 and N°2), the level increased dramatically 4 hours following UV exposure and was significantly higher than the control cells exposed to UV. Surprisingly, although confirmed by two independent siRNAs, levels of CPDs in CSA-KD cells (siCSA N°1 and N°2) and in HR23A-KD cells (siHR23A N°1 and 2) were both significantly elevated in the absence of UV-irradiation compared to USF-1-KD and control cells. Following UV-irradiation, there was a mild increase in levels of CPDs in CSA-KD and HR23A-KD cells which was probably due to the initial high level of CPDs in KD-cells coupled to quantification limits. Nonetheless, these increases remained significantly higher compared to irradiated control cells (Figure 5). Secondly, using skin punch biopsies prepared from USF-1 KO mice and the WT littermates, we analyzed the UV-response by comparing gene transcription efficiency and levels of CPD. RNA analysis comparing irradiated versus non-irradiated WT skin punch biopsies showed that CSA and HR23A mRNA increased 3.5-fold and 2.5-fold at 1 and 5 h post-irradiation, respectively (Figure 6A–6B). CSA and HR23A transcript levels remained at basal levels in USF-1 KO mice and CSB and HR23B mRNA were not affected by UV-irradiation in both WT and KO mice (Figure 6A–6B). By contrast, UV-inducible but USF-1-independent genes, such as the Gadd45α prototype displayed UV-induced transcript profiles in WT and KO USF-1 mice (Figure 6C) [22], [23]. However, we detected a 2 h delay of the mRNA increase in the USF-1 KO mice (Figure 6C), which is consistent with RNAPII being arrested to permit DNA-repair of transcribed genes before the commencement of transcription supporting that TCR is compromised in USF-1 KO mice. Moreover, because HR23 proteins are crucial to stabilize XPC at the DNA-photolesion sites to permit removal of damage, we quantified the level of CPD by ELISA immediately after 3 UV-pulses, and after 4 UV-pulses over 36 h (Figure 6D). While basal levels of CPD were comparable in both WT and USF-1 KO mice as expected from siRNA results, UV-irradiation led to rapid increases of DNA-damage that were comparable immediately after 3 UV-pulses but remained higher over time in KO mice compared to WT mice after 4 UV-pulses. Importantly, calculating the rate of CPD-clearance over 36 h, we observed a difference between WT and KO mice (Figure 6D). Whereas CPDs were removed in WT mice at 36 h, CPDs remained elevated in USF-1-KO mice at this time point. Taken together these results provide compelling evidence that in response to UV-induced DNA-damage, loss of USF-1 compromises the tight regulation of the NER resulting in altered removal of UV-induced DNA-damage.
DNA carries the genetic instruction required for the development and functioning of all living organisms. This information must be transmitted to daughter cells with high fidelity, and therefore specific DNA-repair programs are present to eliminate DNA-lesions produced by regular threats. The NER pathway is dedicated to repair distorted DNA, and for decades studies have focused on elucidating the molecular mechanisms involved in the recognition, signaling and removal of these DNA-lesions [2], [24]. Using a multiple dose UV-irradiation protocol with repetitive lower UV-doses that more accurately mimics our exposure to solar irradiation compared to a single high dose, our study identifies an early and coordinated gene expression regulation program of the CSA and HR23A genes in mammals that relies on the presence of the USF-1 transcription factor.
CSA and CSB proteins have been shown to have dedicated and specific functions in the TCR pathway [5]. It has indeed been clearly established, even in the absence of DNA damage, that a large part of the CSB protein is found associated with chromatin and that RNAPII even in the absence of DNA damage, and this association increases upon UV-irradiation [25], [26]. CSA has however been shown to interact indirectly with RNAPII [25], but it is required in cooperation with CSB for the recruitment of XAB2, HMGN1 and TFIIS, to trigger DNA repair mediated by XP complexes and PCNA protein [5], [24]. The importance of CSA in the early DNA damage response might also reside in the timing of its specific gene expression as its levels are low in resting cells but increase dramatically immediately after UV-irradiation. One possible explanation would be that because CSA acts as a unique player in the initial step of TCR, appropriate levels of the protein is required almost immediately after UV-induced DNA damage and before RNAPII gets arrested by de novo DNA photo-lesions. No increase in CSA protein leads to a delay in transcription likely by an impairment of its associated function: recruitment and stabilization of the initiation complex on the chromatin [25]. This is also supported by deficient CSA being directly linked to the Cockayne syndrome type A genetic disorder [27] and by siRNA results. However, these patients are not prone to developing skin-cancers like XP patients, presumably due to 1- the presence of additional DNA-repair machinery operating post DNA-replication [28], 2- increased cell-death after DNA-damage [28] and to an average life-span for these patients generally being limited to 12 years [29]. Interestingly, specific mutations in the repair-enzyme genes XPB, D and G produce phenotype reflecting a combination of traits present with XP and CS syndromes. This suggests that simultaneous alteration of GGR and TCR will promote mutagenesis in certain cells [29].
HR23A and HR23B proteins share common domains and are both able to form a complex with XPC [30], [31]. The XPC-HR23B complexes were however reported to be more abundant than the XPC-HR23A complexes and have been shown to participate almost exclusively in DNA-photolesions recognition in vivo [32]. The XPC-HR23A complexes were consequently regarded as having a functionally redundant role to XPC-HR23B. This is therefore the first study to report conditions under which HR23A and B protein levels are modulated differently, which suggest that HR23A may have a function distinct from HR23B in the UV-induced DNA damage pathway. We show that in response to repetitive UV-irradiation there is a 4-fold increase in the level of HR23A protein which is associated with a concomitant loss of HR23B and we propose that this may favor XPC-HR23A complex formation which leads to sustained XPC-stabilization for appropriate recognition of DNA lesions [32], [33]. Indeed, while HR23A and HR23B KO mice are NER proficient, double HR23A and HR23B KO derived cells show an XPC-like phenotype [34]. We propose that differential regulation of these two HR23 homologues may provide a safety mechanism to ensure the stability of XPC and its function in response to multiple UV-exposure. This possibility is supported by our data that show (i) a reduction of DNA–lesion removal in HR23A-KD cells, (ii) a reduction of DNA-lesion removal in UV-irradiated USF-1 KO tissue and KD cells, which occurs presumably in part due to an abrogation of HR23A gene expression in response to UV-rays and (iii) a diminution of HR23A protein when UV-induced HR23A transcription is abrogated with α-amanitin. We thus believe that our study reveals a difference in the DNA damage response to a single high dose of UV-irradiation compared to repetitive lower doses and that our conditions mimic the accumulation of DNA-damage over a short period of time which is more applicable to every day life. These results are particularly interesting in the light of the Saccharomyces cerevisiae RAD23 gene, the ortholog of both HR23A and HR23B, which also presents with an UV-inducible phenotype [35]. Our results show that the UV-induced function has been conserved through evolution and restricted to one member for specific regulatory purposes.
USF-1 is activated by the stress-dependent p38 kinase and then operates as a transcriptional rheostat of the stress response [20], [21]. Combined regulation of HR23A and CSA gene expression by USF-1 thus allows a tight and sequential regulation of these two genes. The observation that there is first an increase in USF-1 occupancy on the CSA promoter followed by its occupancy on the HR23A promoter suggests a sequential and dynamic recruitment of USF-1 to fulfill specific steps of a common task. USF-1 as a stress responsive factor is also proposed to be a key player in regulating pigmentation gene expression in response to UV-irradiation [12], [21], [36]. USF-1 may thus elicit a skin protection program against UV-induced DNA damage by controlling two independent and complementary pathways: the DNA-photolesions repair process and the UV-induced tanning response. More importantly, USF-1 functions independently of p53 but both pathways are expected to be coupled [37]. Since USF-1 mediates an independent and crucial DNA-repair program as highlighted by our USF-1 KO and KD assays, we propose that impairment of this pathway will promote genome instability in response to environmental insults, which is a hallmark of cancer. This hypothesis is supported by the reported loss of USF activity in breast cancer cells [38], and impairment of the recruitment of USF factors to specific E-box elements due to SNPs, as observed in the variant rs1867277 FOXE1 gene, conferring thyroid cancer susceptibility [39]. Furthermore, CpG methylation can also impair USF interaction with core E-box motifs and subsequently alter gene expression, as for the metallothionein-I gene which is silenced in mouse lymphosarcoma [40].
Our findings indicate that, in response to repetitive environmental threats that lead to the accumulation of UV-induced DNA damage, the NER pathway undergoes a program of gene expression that correlates with the DNA repair processes and that the USF-1 transcription factor is central to this program. These results may thus have important implications for our global understanding of how genome instability is promoted.
HaCaT (human - p53 deficient) and XB2 (mouse) keratinocytes were maintained in D-MEM (Invitrogen) supplemented with 10% FBS (Sigma) and 1% Penicillin-Streptomycin (Invitrogen) at 37°C in 5% CO2 atmosphere.
Skin biopsies (0.8 cm diameter) were recovered from the backs of WT and USF-1 knockout mice (8 weeks) [13] and maintained in culture for up to 24 h in RPMI (Invitrogen) supplemented with 1% Penicillin-Streptomycin at 37°C in a 5% CO2 atmosphere.
Specific DNA photo-lesions were generated with ultraviolet bulbs (254 nm) [14], using the Stratalinker apparatus (Stratagene) as previously described [12], [20], [21]. The day before UV exposure, cells were plated at 50–70% confluence, depending on their doubling time, in 10 cm Petri dishes. Twelve to twenty-four later, the medium was replaced with fresh medium supplemented with 2% FBS and 1% antibiotics. The following day, cells were UV irradiated (2× to 8× 10 J/m2). UV pulse set at 10 J/m2 lasted 3 seconds. The medium was completely removed before and replaced after irradiation. At the time point indicated, cells were washed twice in cold PBS, harvested by scraping, centrifuged and resuspended in appropriate buffer. For transcription inhibition experiments, cells were pre-treated with α-amanitin (5 µg/ml; Sigma) 30 min prior to UV-irradiation.
Mouse skin biopsies were irradiated with four successive pulses of 50 J/m2 UV, recovered at the indicated time points by placing the skin biopsy directly in RNA later buffer (Qiagen) and stored at −20°C for subsequent RNA extraction.
Cell viability in response to UV (254 nm) was analysed in 96 well plates. Briefly, cells were plated at 1×104 cells/well, 10 h before UV induction, tetrazolium salt (MTT, 0.5 mg/ml (Sigma) was added to culture medium. After 3 h of incubation (37°C), the medium was removed and 150 µl of DMSO was added to each well. Percentage of cell viability was then analysed by measuring the DMSO-optical density (OD), at 690 and 540 nm with a Multiskan spectrophotometer.
RNA was extracted using Nucleospin RNA II kit (Macherey Nagel) and quantified using the Nanodrop device. For skin explants, an extra Trizol/chloroform purification step was needed to remove protein. cDNA was obtained by reverse transcription using a High-Capacity cDNA Reverse Transcription Kit (Applied Biosystem) from 1 µg of total RNA. Gene expression was analyzed by qPCR in sealed 384-well microtiter plates using the SYBR Green TM PCR Master Mix (Applied Biosystem) with the 7900HT Fast Real-Time PCR System (Applied Biosystem). Relative amounts of transcripts were determined using the delta Ct method. The mRNA levels at each time point following stimulation are expressed as fold increase, relative to non-irradiated cells. Data were normalized independently to at least two housekeeping genes HPRT and GAPDH. Because comparable data were obtained only the HPRT ones are presented. Each experiment was carried out at least twice and each time point was repeated in triplicate. Forward (F) and reverse (R) primers were designed using the Universal Probe Library Assay Design Center (Roche) and have been previously tested for their efficiency (Sequences available on request).
Harvested cells were immediately lysed by incubation for 30 min in ice-cold RIPA buffer (supplied in protease and phosphatase inhibitors). Equal amounts of protein were denaturated in Laemmli buffer for 5 min at 95°C and resolved by 15% SDS-PAGE. Membranes were probed with appropriate antibodies and signals detected using the LAS-3000 Imaging System (Fujifilm) were quantified with ImageJ (http://rsbweb.nih.gov/ij/).
Gel electrophoresis DNA binding assays were performed with crude HaCaT keratinocyte nuclear extracts under conditions previously described [12], [41], [42], with modifications. Double-stranded oligonucleotides were labeled with T4 polynucleotide kinase in the presence of P32-γATP (3000 Ci/mmol) and purified in columns (Mini Quick Spin Oligo Columns, Roche Diagnostic). Reaction mixtures contained 2–4 µg of total protein and 0.03 pmol of P32 end-labelled probe in binding buffer (Hepes 25 mM, KCl 150 mM, 10% Glycerol, DTT 10 mM, 1 µg of poly(dIdC), 1 µg salmon sperm DNA). After 20 min of incubation, samples were loaded onto a low ionic strength 6% polyacrylamide gel (29∶1 cross-linking ratio) containing Tris Borate Na EDTA buffer pH 8.3.
Supershift and competition assays were performed by adding competitor probes (1× to 100×) or antibodies (0.2 µg) prior to incubation with labelled probes (Sequences available on request). Radioactive bands were quantified with a STORM 840 PhosphorImager (Molecular Dynamics).
ChIP assays, using 1.5–2×106 HaCaT cells, were performed as previously described [41], [43], with specific adaptations. The cells were cross-linked (1.5% formaldehyde), washed twice and collected in 1 ml cold PBS. Cells were lysed and the samples were then sonicated for DNA fragmentation (Sonifier Cell Disruptor, Branson) in 1 ml lysis buffer (10 mM EDTA, 50 mM Tris-HCl (pH 8.0), 1% SDS, 0.5% Empigen BB) and diluted 2.5-fold in IP buffer (2 mM EDTA, 100 mM NaCl, 20 mM Tris-HCl (pH 8.1), 0.5% Triton X-100). This fraction was subjected to immunoprecipitation overnight with 3 µg of the appropriate antibody. These samples were then incubated for 3 h at 4°C with 50 µl of protein A-Sepharose beads slurry. Precipitates were washed several times, cross-linking reversed and DNA purified using a Nucleospin Extract II kit (Macherey Nagel).
PCR or qPCR analyses were carried out with primers spanning HR23A, HR23B and CSA proximal promoters or, as a reference, with primers targeting an unrelated promoter region (HSP70 promoter region) or unspecific regions of target promoter genes (sequences available on request). End-point PCR was performed in semi-quantitative conditions for ChIP (30 amplification Cycles). For qPCR analysis, fold enrichment was determined using the ΔΔCt method: Fold enrichment = 2−(Δct1−ΔCt2), where ΔCt 1 is the ChIP of interest and ΔCt2 the control ChIP.
−744/+73 and −185/+73 HR23A promoter region were obtained by PCR and inserted into the luciferase reporter plasmid pGL3-basic (Promega). E boxes and the GC box were mutated using a QuickChange Site-Directed Mutagenesis Kit (Stratagene). The same protocol was used for the CSA promoter sequence lying −847/+1.
CSA and HR23A promoter regulation was studied in mouse XB2 keratinocytes. Cells were plated at 60–70% confluence in 12-well plates in medium supplemented with 10% SVF without antibiotics and were maintained for 12 h. Cells were co-transfected or not with pGL3 reporter vector and pCMV (empty, USF-1 or USF-2), as previously described [12], [20], [21]. The transfection mix, containing up to 500 ng of plasmid DNA, was prepared in Optimem medium (Invitrogen) and used to transfect cells for 3 h using Lipofectamin 2000 (Invitrogen). 3 h after transfection, the medium was replaced with fresh medium supplemented with 10% SVF and 1% antibiotics. 48 h later, cells were irradiated with UV, as described above and harvested up to 5 h following UV. Cells were then passively lysed and luciferase activity was quantified in a Microplate Luminometer Centro LB 960 (Berthold) using the Luciferase Reporter Assay System (Promega).
XB2 cells were seeded and transfected in 10 cm-diameter dishes (1×106 cells per dish) in DMEM medium complemented with 10% FBS, with 40 pmol of siRNA. Two different siRNA (N°1 and 2) were used independently for each target gene tested (CSA, HR23A and USF-1) as for control (siOTP1, siNT1) (Sigma-Genosys, St Louis, MO) using Lipofectamine 2000 (Invitrogen, Paisley, UK). Transfections were performed following provider's instructions. 72 hours later, the cells were UV irradiated as previously described and recovered 4 hours after the irradiation protocol for CPD quantification and western blot analysis. siRNA sequences are available on request.
Quantifications of CPD in skin explants following UV (254 nm) (4×50 J/m2) were performed by ELISA, accordingly to Cosmo bio recommendations. DNA purification was performed by phenol/chloroform extraction and ethanol precipitation. Briefly, 200 ng of denatured DNA was distributed onto protamine sulfate precoated 96 well plates (Polyvinylchloride flat-bottom). Detection of DNA-lesion was performed using specific mouse anti-CPD antibodies, and revealed with the biotin/peroxidase-streptovidin assay. Quantification was obtained by the absorbance at 492 nm. Each experiment was performed independently with punch biopsies of three independent WT and USF-1 KO mice.
XB2 (mouse keratinocytes) cell lines were cultured in D-MEM at 37°C on glass coverslips in 35-mm dishes. 24 hours later, cells were UV-irradiated with 6×10 J/m2 in serum free medium following as previously described. Cells were then fixed and permeabilized after different times of induction accordingly to Cosmo bio Co protocol. Previously to CPD immunostaining in cells, we denatured DNA with HCl 2 M for 30 min at room temperature. Indirect immunofluorescence was then performed using specific recommendations of Cosmo bio Co protocol with specific primary antibodies mouse anti-CPD (TDM2 clone, MBL) (1∶3000). Fluoro-staining was performed with labeled donkey anti-mouse IgG (Alexa Fluor 488). CSA immunostaining was performed with specific anti-rabbit antibody from Santa Cruz.
Anti USF-1 (C:20), USF-2 (N-18), Sp1 (PEP 2), Sp3 (D-20), TBX-2 (C-17), HR23B (P-18), HSC70 (B-6) were purchased from Santa Cruz. Anti HR23A (ARP42211) was purchased from Aviva. Anti CSA was purchased from Abcam (ab96780). Anti CPD (TDM2) was purchased from MBL. Anti α-Tubulin (ARP42211) was purchased from Sigma.
Errors bars represent standard deviation, stars indicate statistically significant differences (two-tailed Student's t-test) between control and irradiated samples * P<0.05; ** P<0.01; *** P<0.001.
The present animal study follows the 3R legislation (Replace-Reduce-Refine). It has been declared and approved by the French Government Board. Animal welfare is a constant priority: animals were thus sacrificed under anesthesia.
|
10.1371/journal.pbio.1002053 | FAX1, a Novel Membrane Protein Mediating Plastid Fatty Acid Export | Fatty acid synthesis in plants occurs in plastids, and thus, export for subsequent acyl editing and lipid assembly in the cytosol and endoplasmatic reticulum is required. Yet, the transport mechanism for plastid fatty acids still remains enigmatic. We isolated FAX1 (fatty acid export 1), a novel protein, which inserts into the chloroplast inner envelope by α-helical membrane-spanning domains. Detailed phenotypic and ultrastructural analyses of FAX1 mutants in Arabidopsis thaliana showed that FAX1 function is crucial for biomass production, male fertility and synthesis of fatty acid-derived compounds such as lipids, ketone waxes, or pollen cell wall material. Determination of lipid, fatty acid, and wax contents by mass spectrometry revealed that endoplasmatic reticulum (ER)-derived lipids decreased when FAX1 was missing, but levels of several plastid-produced species increased. FAX1 over-expressing lines showed the opposite behavior, including a pronounced increase of triacyglycerol oils in flowers and leaves. Furthermore, the cuticular layer of stems from fax1 knockout lines was specifically reduced in C29 ketone wax compounds. Differential gene expression in FAX1 mutants as determined by DNA microarray analysis confirmed phenotypes and metabolic imbalances. Since in yeast FAX1 could complement for fatty acid transport, we concluded that FAX1 mediates fatty acid export from plastids. In vertebrates, FAX1 relatives are structurally related, mitochondrial membrane proteins of so-far unknown function. Therefore, this protein family might represent a powerful tool not only to increase lipid/biofuel production in plants but also to explore novel transport systems involved in vertebrate fatty acid and lipid metabolism.
| Fatty acid synthesis in plants occurs in chloroplasts—the organelle more commonly known for conducting photosynthesis. For subsequent lipid assembly to be possible in the endoplasmatic reticulum (ER), export of these fatty acids across the chloroplast envelope membranes is required. The mechanism of this transport until now has not been known. We isolated FAX1 (fatty acid export 1), a novel membrane protein in chloroplast inner envelopes. FAX1 function is crucial for biomass production, male fertility, and the synthesis of fatty acid-derived compounds like lipids, waxes, or cell wall material of pollen grains. Whereas ER-derived lipids decreased when FAX1 was missing, levels of plastid-produced lipids increased. FAX1 over-expressing mutants showed the opposite behavior, including an increase of triacyglycerol oils. Because FAX1 could complement for fatty acid transport in yeast, we concluded that FAX1 mediates the export of free fatty acids from chloroplasts. In vertebrates, FAX1 relatives are structurally related proteins of so-far unknown function in mitochondria. This protein family may thus represent a powerful tool not only to increase lipid oil and biofuel production in plants but also to explore novel transport systems in animals.
| Fatty acids (FAs) are building blocks for the majority of cellular lipids, which are essential throughout life of organisms. Besides their role as constituents of biological membranes, plant acyl-lipids are used for diverse functions at different destinations and tissues (reviewed in [1]). For example, triacylglycerols (TAGs) in seeds of oilseed plants represent the major form of carbon and energy storage. Cuticular waxes at the surface of plants restrict loss of water and provide protection against pathogen attack. Furthermore, the formation of pollen cell walls is strictly dependent on delivery of modified FAs from tapetum cells in anthers (reviewed in [2]). De novo FA synthesis in plants occurs in plastids (for overview, see [1,3]). Growing alkyl chains in the plastid stroma are attached as acyl moieties to acyl carrier protein (ACP), and in seed plants become available for lipid assembly mainly in the form of palmitoyl (16:0)- and oleoyl (18:1)-ACP. Part of these long-chain FAs will be integrated into lipids inside plastids (prokaryotic pathway); the majority, however, is exported to the endoplasmic reticulum (ER) for further elongation, acyl editing, and lipid synthesis (eukaryotic pathway).
Although it is generally agreed that free FAs are shuttled across plastid envelope membranes, the mode of export still remains enigmatic [4] since until now, no membrane-intrinsic transporter protein could be associated with a direct function in plastid FA export (for overview, see [1,3]). On the one hand, a facilitated diffusion of free FAs through the lipid environment of membranes is suggested, which is supported by the recent finding that an acyl-ACP synthase in the cyanobacterium Synechocystis sp. PCC6803 is necessary and sufficient for FA transfer across membranes [5]. On the other hand, several ATP-binding cassette (ABC) transporter proteins for lipids, FAs, or acyl-coenzyme A (CoA), and for import of FAs into peroxisomes [6], as well as FA-transport systems from Escherichia coli, yeast, or mammals, provide evidence for an active mode of transport in plastids. Nevertheless, before transport, acyl-ACP thioesterases at the inner plastid envelope membrane catalyze the hydrolysation of fatty acyl-ACP to free FAs. After crossing both inner and outer plastid envelope membranes (IE, OE), free FAs are re-activated to acyl-CoAs by long-chain acyl-CoA synthetases (LACSs). As demonstrated for the protein LACS9, these enzymes can attach to the cytosolic face of the plastid OE [7–9]. At the ER membrane, the ABC transporter ABCA9 has recently been described to be involved in FA-uptake, most likely in the form of acyl-CoA, thereby being important for TAG synthesis during seed filling [10]. Once arrived in the ER lumen, plastid-derived FAs are utilized for synthesis of specific lipid classes via the so-called eukaryotic pathway, where phosphatidic acid (PA) represents an important intermediate, phosphatidyl-choline (PC) is a major membrane phospholipid, and TAGs are the energy storage lipids produced. Subsequently, these eukaryotic lipids are distributed to various subcellular locations. For re-import of eukaryotic lipids into plastids, most likely in the form of ER-derived PA, an ABC transporter system (TGD1, 2, 3) at the IE [3] and the PA-binding ß-barrel lipid transfer protein TGD4 in the OE [11] are required. In plastids, the diacylglycerol backbone from these eukaryotic precursors is used for synthesis of the galactolipids MGDG, DGDG (monogalactosyl-, digalactosyl-diacylglycerol), and the sulfolipid SQDG (sulfoquinovosyl-diacylglycerol). In addition, however, a prokaryotic-type pathway also produces MGDG, DGDG, SQDG, and the phospholipid phosphatidyl-glycerol (PG) directly from newly synthesized FAs and thus does not require previous FA-export from plastids (for overview, see [1]).
Here we describe FAX1, a novel protein in the IE of plastids that belongs to the Tmemb_14 superfamily of membrane proteins with so-far unknown function. Functional studies in yeast as well as FAX1 mutant analysis in Arabidopsis thaliana clearly demonstrate that FAX1 mediates FA-export from plastids and thus, to our knowledge, represents the first membrane-intrinsic protein described to be involved in this process. In mammals, FAX1 relatives are structurally related mitochondrial membrane proteins, for which the biological task is not yet clear [12–14]. Thus, FAX1 not only is a missing link to explain the mode of plastid FA-export and to improve plant lipid/biofuel production but might also propel the understanding of Tmemb_14 protein performance in general.
The Arabidopsis protein At-FAX1 (At3g57280, for fatty acid export 1) was previously annotated as potential plastid-targeted and plant-specific solute transporter by proteomic and phylogenetic analysis [15,16]. Furthermore, we identified transcripts of At-FAX1 to be up-regulated upon induction of early leaf senescence [17]. To analyze protein function, we isolated the cDNA of FAX1 genes from Arabidopsis and pea (Pisum sativum). For both proteins, chloroplast targeting peptides and four hydrophobic α-helices are predicted (Fig. 1A). By the latter, plant FAX1 clearly groups to the so-called Tmemb_14 superfamily of proteins with so-far unknown function. The Tmemb_14 family is ubiquitous, with members in nearly all eukaryotes and some bacteria (InterPro|UPF0136). In Arabidopsis, four proteins (FAX1–FAX4) are predicted to be targeted to plastids, while three (FAX5–FAX7) most likely are directed to other, non-plastid membranes via the secretory pathway (Fig. 1B). The plastid-intrinsic FAX1 is restricted to the chlorophyll-containing plant kingdom, with representatives in mono- and dicotyledons as well as in mosses and green algae (compare InterPro|UPF0136, [15]). Relatives of non-plastid predicted At-FAX proteins, however, can be found in eukaryotes such as mammals, insects, or yeast, and in some bacteria and cyanobacteria (e.g., Chlamydiae or Nostocales).
For all Tmemb_14 proteins, four hydrophobic α-helical domains are predicted (Fig. 1). However, nuclear magnetic resonance (NMR) structure determination of the human Tmemb_14 proteins TMEM14A and TMEM14C [14] showed that only three of these helices are membrane-spanning. TMEM14A contains an amphiphilic N-terminal helix, presumably located at the lipid micelle-water interface, while for TMEM14C an amphiphilic helix that orients perpendicular to the lipid bilayer, is placed between the second and third membrane domain. Amino acid sequences of the plastid FAX1 and the non-plastid At-FAX6 nicely align to both TMEM14A and 14C (Fig. 1C), but structural modeling revealed that the mature At-FAX1 and Ps-FAX1 are more similar to TMEM14C (Fig. 2A). Here, three membrane-spanning and one amphiphilic helix are likely, while the additional N-terminal amino acids of FAX1 proteins might form another α-helical domain not present in TMEM14C. In contrast, the structure of At-FAX6 resembles that of TMEM14A with an N-terminal amphiphilic helix followed by three transmembrane domains (Fig. 2B).
With its membrane-spanning domains, FAX1 inserts into the inner envelope membrane (IE) of plastids as could be shown by in vivo GFP-targeting and immunoblot analysis. At-FAX1-GFP signals in Arabidopsis protoplasts, which can be detected as rings around chloroplasts (Fig. 2C), point to an envelope localization. This could be confirmed and specified to IE by immunoblot analysis using sub-fractionated pea chloroplasts. In pea IE membranes as well as in Arabidopsis chloroplast envelopes, FAX1 signals appear as a band of about 25kDa (Fig. 2D). In agreement, FAX1 peptides in proteomic analyses of plastid membranes were exclusively detected in IE preparations (see [16] and references therein). To exclude ER localization, we further probed against ER-enriched Arabidopsis microsomal membranes (see [20]), where no FAX1 signals could be detected (Fig. 2D).
To study the in vivo function of FAX1, we analyzed loss-of-function and over-expressing mutant lines in Arabidopsis. We selected fax1–1 and fax1–2 with T-DNA insertions in the first intron and first exon of the FAX1 gene, respectively (S1A Fig.). Reverse transcriptase-polymerase chain reaction (RT-PCR) analysis showed that both homozygous alleles represent knockouts for FAX1 (S1B Fig.). To complement this loss-of-function, At-FAX1 cDNA under control of the 35S promoter was introduced into heterozygous fax1–2 plants. Subsequently, two lines homozygous for the fax1–2 allele (Co#7 and Co#54) were selected for further analysis. To stable over-express FAX1 in wild-type plants, the 35S::FAX1 construct was introduced into Col-0, and two independent lines named ox#2 and ox#4 were identified as FAX1 over-expressors. Quantitative real time RT-PCR revealed that FAX1 transcripts in line Co#7 are at wild-type levels, whereas line Co#54 contains about 12 times more FAX1 mRNA. Over-expression in ox#2 seedlings was mild (about 2-fold), but strong in line ox#4 (about 200 times more mRNA than in wild type; S1C Fig.). Immunoblot analysis confirmed the strength of FAX1 expression in these lines and the knockout in fax1–2 on the protein level (S1D Fig.).
Homozygous fax1–1 and fax1–2 knockout mutants both were characterized by reduced biomass at mature rosette stages (Fig. 3A, Table 1). Full flowering fax1 knockouts were significantly smaller than wild type, had thinner inflorescence stalks, and flowers producing short siliques that contained almost no seeds (Fig. 3B, C). Detailed analysis of different tissues and organs revealed that the decrease in biomass of fax1 lines was detectable throughout the entire plant body, including root, leaf, and stem tissues (Table 1). Because differences in stem dry weight were slightly more pronounced than in fresh-weight (FW) samples, most likely cell wall synthesis was affected. This could be confirmed by ultrastructural analysis of stem tissue (S2 Fig.). Here fax1 knockouts showed small vascular bundles with reduced secondary cell walls (S2A, D, G Fig.). Since the same phenotype was detected in both independent T-DNA insertion lines fax1–1 and fax1–2, and could be reverted by complementation in lines Co#7 and Co#54 (Fig. 3, Table 1), we conclude that the reduced biomass is caused by the loss of FAX1 function.
Remarkably, FAX1 over-expressing lines ox#2 and ox#4 were slightly bigger and produced more biomass as well as thicker inflorescence stalks than wild type (Fig. 3, Table 1), thus behaving opposite to fax1 knockouts. In stems, this led to about one more hypodermal cell layer and to extended vascular strands, including an increased amount of xylem and phloem vessels, as well as a multi-layered procambuim (S2C, F, I Fig.). Because fresh weight of ox#2 and ox#4 stems was significantly higher than in wild type, but—in contrast to fax1 knockouts—stem dry weight of FAX1ox lines was similar to wild type (Table 1), the increased biomass of FAX1 over-expressors is most likely mainly due to enhanced production of cells. However, since tracheid walls of ox#2 appeared to be slightly thicker than in Col-0 (S2H, I Fig.), we can’t fully exclude an additional effect on the size of secondary cell walls. Interestingly, the rate of FAX1 overproduction—i.e., 2-fold for ox#2, 200-fold for ox#4—did not quantitatively affect the strength of biomass phenotypes, indicating a rather non-linear effect of protein function.
To understand the peculiar loss-of-function phenotype of homozygous fax1 knockouts during flower and silique development, segregation analysis of mutant alleles was performed. Self-fertilization of heterozygous fax1–1 and fax1–2 revealed that the ratio of homozygous progeny was 7% and 4%, respectively, pointing to defect male and/or female gametophytes (Table 2). However, when stigmata from homozygous fax1–2 flowers were fertilized with wild-type fax1–2 pollen, normal seeds with 100% heterozygous fax1–2 mutant alleles were produced, indicating fertile fax1 knockout female gametophytes and sporophyte organs. In contrast, pollination of wild-type stigmata with homozygous fax1–2 anthers, produced short siliques, as observed during selfing of homozygous fax1–2 mutants (see Fig. 3C), and led to an estimated seed yield <0.1% of wild type. Furthermore, during manual crossing it became evident that pollen grains of homozygous fax1–2 flowers were improperly released from anthers. To minimize potential anther defects from the paternal sporophyte, we thus pollinated homozygous fax1–2 stigmata with heterozygous fax1–2 anthers, thereby producing 12% progeny homozygous for fax1–2 (Table 2). In summary, these results point to impaired transmission of male gametophytes (pollen) and defects of the male sporophyte (anther) in fax1 knockouts, finally leading to the observed male sterility.
To further analyze flower development, in particular that of male parts, we examined the morphology of flower tissue from FAX1 mutant lines (Fig. 4; S3 Fig.). In FAX1 wild-type and over-expressors (ox#2, ox#4), pollen release by anther dehiscence, transfer to the stigma, and fertilization as indicated by high yield of viable seeds was normal. However, flowers of fax1 knockout mutants showed stigmata with non-pollinated papillae. In addition, fax1 anthers released only very few pollen grains (Fig. 4A, B; S3A, B, G Fig.). In flowers of complemented fax1–2 lines (Co#7, Co#54) in comparison, more free pollen grains than in fax1 knockouts but less than in wild type were visible, indicating incomplete recovery of pollen release (S3D, G Fig.). In contrast to the rest of the plant organs, where regeneration of fax1 knockout defects in Co#7, Co#54 lines was 100% (see Fig. 3; Table 1), complementation of fax1–2 pollen phenotypes was incomplete. This effect was best visualized by the colorless pollen of fax1 knockout and complementation lines (S3B, D, G Fig.), due to the absence (fax1–2) or incomplete (Co#54) assembly of a pollen coat (compare Fig. 4D–F; S3I–K Fig.) that normally includes yellow flavonoid and carotenoid deposits (for overview, see [2]). The incomplete complementation was restricted to pollen grains and most likely is due to the 35S promoter system, which in Arabidopsis shows no or reduced activity in pollen grains and anther tissue, respectively [21].
Subsequently, the detailed structure of anther tissue and pollen grains of fax1–2 knockout, Col-0 wild type, and the complementation line Co#54 was visualized by light- and transmission electron microscopy (TEM) at the mature, tricellular pollen stage (Fig. 4C–F; S3H–K Fig.). In general, anthers of fax1–2 were smaller than in wild type and the surface of pollen grains appeared to be wrinkled (Fig. 4C). Cross sections revealed an impaired release of fax1–2 pollen, although pollen sacks were wide open, indicating full dehiscence of anthers. Tapetum cells seemed to be degraded as expected for the developmental stage analyzed, however, the locule of fax1–2 anthers was covered by an electron-dense material, which stuck to pollen grains and thus most likely was responsible for improper pollen delivery (Fig. 4C–E). Ultrastructural resolution demonstrated that the well-defined structures of the outer pollen cell wall—i.e., the exine layer and the pollen coat, which covers the exine surface and its cavities—were absent in fax1–2 knockouts (Fig. 4E, F). The intine, representing the innermost layer of the pollen cell wall and composed of cellulose, pectin, and various proteins, secreted by the microspore (gametophytic origin, see [22]), however, looked intact. In contrast, mature wild-type pollen showed a complete exine, consisting of a flat nexine layer and the sculpted sexine parts tectum and bacula. Furthermore, the latter were filled and covered with the tryphine pollen coat (Fig. 4E, F). Pollen cell walls of Co#54 displayed an intermediate state of biogenesis with visible nexine layers, but incomplete arrangement of tectum and bacula structures as well as pollen coat material (S3J, K Fig.). As described above, these findings point to an incomplete complementation of fax1–2 knockouts in pollen.
In conclusion, structural analysis of anthers and mature pollen grains showed that FAX1 is essential for biogenesis of the outer pollen cell wall, in particular for the assembly of exine and pollen coat. Both layers consist of complex biopolymers, such as sporopollenin (exine) and tryphine (pollen coat), that are mainly made of FA-derivatives and lipids originating from the tapetum tissue of anthers (sporophytic origin, see [2]). Thus, FAX1 might play a role in delivery of these compounds by mediating FA-export from tapetum cell plastids. Most likely, the electron-dense, sticky material in fax1 knockout anthers that prevents release of pollen grains represents cellular debris of degenerated tapetum cells and/or not-incorporated sporopollenin or tryphine material.
Because during analysis of FAX1 mutants an altered surface of epidermal cells was apparent, we investigated structure as well as wax and cutin coverage of epidermis cells from primary inflorescence stalks of FAX1 mutants (Fig. 5). Microscopic analysis revealed that the width of epidermal cell walls in fax1–1 was strongly reduced when compared to wild type (Fig. 5A, B). As for cell walls in xylem vessels (S2 Fig.), a strong effect was only found for knockout and not for FAX1 over-expressing lines. However, an electron-dense cover at the extracellular side of the cell walls, most likely representing the cutin matrix of the cuticular layer, appeared to be more intense in ox#2, but reduced in fax1–1 (Fig. 5B, C). To examine the constitution of the cuticular layer, we therefore determined wax and cutin coverage from stem epidermis cells. Surprisingly, the total loads of different wax and cutin species were not altered for all lines analyzed (fax1–1, fax1–2, Col-0, WT2, Co#7, Co#54, ox#2, ox#4, see S1A Data). Furthermore, no change in composition regarding aliphatic chain length or functional groups (e.g. ketones, acids or aldehydes) could be detected. The only significant difference we found was for C29-ketone wax components, which were reduced in both fax1 knockout lines by more than 50% when compared to stems from all other plants (Fig. 5D).
Since cutin contents were unchanged, the different strengths of the outer layer of epidermal cell walls observed by TEM most likely are due to stronger (ox#2) and weaker (fax1–1) crosslinking of the cutin matrix with cell walls. The wax composition of cuticular layers, however, is dependent on plastid FA-synthesis as well as excretion of modified FAs via the plasma membrane of epidermis cells (for overview, see [1]). In parallel to the assembly of sporopollenin and tryphine material in pollen cell walls (see above), FAX1 might thus be necessary for plastid FA-export, previous to synthesis and release of ketone wax components.
Because fax1 knockouts revealed a lack of FA- and lipid-derived compounds in pollen as well as stem epidermis cells (see above), we measured free FAs and polar lipid species such as phospho-, sulfo-, galacto-lipids, and triacylglycerols in leaves and flowers of mature FAX1 mutant plants (S1 Table). To spotlight changes in FAX1 mutants compared to wild type, we determined relative values and summarized representatives of significantly different levels, as well as abundant species from each molecule class in the next two figures. For comparison to the overall FA/lipid composition of each tissue, we listed contents in mol% of all significantly different species in S2–S4 Tables, and further estimated the impact of changes in mol% of each molecule class in the next table.
In leaves of fax1 knockout plants, levels of 30 FA and polar lipid species (irrespective of TAGs) were significantly different from wild type (S2 Table). For free FAs, we observed an increase of plastid-produced FA 18:2 (Fig. 6A, S2A Table) and a decrease for FAs 20:0, 24:0, 26:0 (Fig. 6C, S2C Table), which are elongated at the ER. Whereas aggregate levels of 34:x glycolipids (MGDG, DGDG, SQDG) were only slightly elevated (Table 3, S2A Table), the highly abundant MGDG 36:6 (11.7 mol% in wt) with an ER-made diacylglycerol backbone was considerably less (2.7 mol%) than in wild type (Fig. 6C, S2C Table). The eukaryotic-type DGDG 36:6, however, increased contributing about 0.7 mol% more to the overall lipid content (Table 3, S2C Table). Strong upward changes were observed for phosphatidyl-glycerol (PG) species (2.7- to 5-fold; Fig. 6A, S2A Table), leading to an entire gain of up to 3.2 mol% (Table 3) of these mainly plastid-derived phospholipids. In contrast, the ER-produced phospholipids phosphatidyl-choline (PC) and -ethanolamine (PE) mostly decreased in fax1 knockout leaves (Fig. 6C, S2C Table). Here the effect, in particular of highly abundant PC 34:3, PC 36:6 (9.3, 7.4 mol% in wt), was especially strong and is estimated to primarily contribute to a total reduction of PCs by 8.8 mol% (Table 3). Whereas the overall decrease of PE was about 0.5 mol%, phosphatidyl-inositol (PI) contents showed a pronounced upward fold change, which, however, only very slightly contributed to the overall lipid composition, and thus leaf PI might rather be involved in signaling (Table 3, S2C Table).
In leaves of FAX1 over-expressing lines (Fig. 6B, D; Table 3), we found an opposite distribution of free FAs and lipids as in fax1 knockouts. Here, without counting TAGs, 28 molecule species were significantly different from wild type (S2B, D Table). Contents of all differentially regulated and mainly plastid-derived FAs, 34:x glycolipids (MGDG, DGDG, SQDG) as well as PG 34:2 dropped (Fig. 6B, S2B Table). A considerable impact on the total lipid content came from reduction of highly abundant molecules such as MGDG 34:5, 34:6; DGDG 34:2, 34:3; or SQDG 34:3, all with levels higher than 2 mol% in wild type. In consequence, the estimated overall reduction was about 2.8, 1.9, and 0.9 mol% for 34:x MGDG, DGDG, and SQDG, respectively (Table 3). For lipids produced by the eukaryotic pathway at the ER, we found a mild decrease of MGDG 36:5 (0.4 mol%) and only very minor changes (0.04–0.08 mol%) for SQDG 36:4, 36:5, and PE 34:3 (Table 3, S2D Table). The effect on PC contents, however, again was quite strong (total increase of about 3.0 mol%, see Table 3), including elevated levels of the abundant PC 34:1, 34:3, 36:2, and 36:3 (Fig. 6D, S2D Table).
When compared to leaves, flower tissue of fax1 ko and FAX1ox lines showed a similar differential pattern of free FAs and lipids, which are presumably mainly produced via the prokaryotic pathway (Table 3, S3A, B Table). Whereas in fax1, FAs that after synthesis have to be exported from chloroplasts (i.e. 16:0, 18:0, 18:1) increased when compared to wild type (largest change for 16:0 = 0.24 mol%), the plastid internal FA 18:3 and the plastid external FA 24:0 decreased by about 0.1 mol% each (S3A, C Table). In FAX1 ox flowers only a minor increase of FA 18:0 was detected (S3B Table). As found in leaves, overall levels of 34:x MGDG, DGDG, and SQDG increased in knockouts but decreased in over-expressors (Table 3). The most prominent changes were for MGDG 34:6 (increase of 0.6 mol% in fax1, S3A Table) and for MGDG 34:5 (decrease of 0.3 mol% in FAX1ox, S3B Table).
For several lipid species, which are assembled at the ER, however, patterns in flowers were different and more diverse than in leaves. These included a rise in 36:x MGDG levels (dominated by +0.8 mol % of MGDG 36:6) in fax1 knockouts (S3C Table); an increase and a decrease of about 0.45 mol% PE in fax1 and FAX1ox, respectively (Table 3); as well as a strong reduction of PC in FAX1ox flowers (up to 5.6. mol%, Table 3). In fax1 knockout flowers in contrast, PC (mostly PC 36:6 by-1.0 mol%; S3C Table) and also PI species (-0.26 mol%) significantly dropped (Table 3).
The most robust differential distribution in both mutant lines and tissues, however, was found for triacylglycerol oils (Fig. 7; Table 3). Here we determined significant changes for more than half of the molecules measured (S4 Table). More important, however, was a massive decrease of high and low abundant TAGs in fax1 knockout leaves (Fig. 7A, S4A Table) and flowers (Fig. 7C, S4C Table) as well as a strong increase in FAX1ox leaves (Fig. 7B, S4B Table) and flowers (Fig. 7D, S4D Table). Fold changes were highest for low abundant TAGs (e.g., 8.3-fold decrease for TAG 56:7 in fax1 leaves, S4A Table). As expected, the biggest impact on overall TAG content was by significant changes in high abundant species, resulting in a drop of-4.3 and-7.2 mol% in leaves and flowers of fax1 knockouts and an accumulation of +3.2 and +6.6 mol% for leaf and flower tissue from FAX1ox lines, respectively (Table 3).
In summary, determination of free FAs and lipids in FAX1 mutants clearly shows that the function of FAX1 in the IE membrane of chloroplasts impacts cellular FA and lipid homeostasis. Overall we found significant differences compared to wild type for more than 50% of all species determined (Table 3). Together with the observed lack of FA- or lipid-derived compounds in fax1 knockout pollen cell walls and cuticular waxes (see above), these findings point to a function of FAX1 in FA-export from plastids (for details, see Discussion).
In baker’s yeast (Saccharomyces cerevisiae) import of exogenous long-chain FAs by the so-called vectorial acylation process requires a multiprotein complex, which consists of Fat1p (the membrane-spanning transport protein) and Faa1p or Faa4p, acyl-CoA synthetases for intracellular FA activation [25]. In order to test a potential FA-transport function of FAX1, we thus analyzed growth complementation of the yeast fat1 and faa1/faa4 mutants, which represent knockouts for Fat1p and Faa1p/Faa4p, respectively [26]. Therefore, we transformed the coding sequence of the mature At-FAX1 protein into fat1 and faa1/faa4 cells. Since previous studies revealed that the uptake of the polyunsaturated FA α-linolenic acid (C18:3) into yeast cells was toxic for wild-type but not for fat1 cells [5], we challenged growth of FAX1-containing yeast mutants by addition of high α-linolenic acid concentrations (3.6 mM) to the media (Fig. 8A–C). In drop tests on agar plates, all cells showed normal growth under control conditions (Fig. 8A, left). In addition, yeast mutant cells, transformed with the empty vector only, were resistant to excess α-linolenic acid (Fig. 8A, right). However, fat1 cells expressing the mature At-FAX1 protein died in the presence of α-linolenic acid overload (Fig. 8A, right), indicating that FAX1 is able to restore FA-uptake in fat1 mutants. In contrast, α-linolenic acid induced cell death was not observed in faa1/faa4 cells, neither with nor without FAX1, pointing to a FAX1 function in FA-transport and not in FA-activation.
Furthermore, we monitored a very similar behavior for growth kinetics of the respective yeast cells in liquid media (Fig. 8B, C). Here, a FAX1-mediated toxicity of α-linolenic acid was significant after 18 h when compared to empty vector cells. While this effect was highly significant and strong in fat1 mutants as indicated by a reduction of cell density to about 54% after 29 h (Fig. 8B), only a mild growth inhibition was detected in Δfaa1/faa4 (density of FAX1 cells was about 78% of control cells after 29 h; Fig. 8C). In addition, when compared to vector-only cells grown without α-linolenic acid, we observed a slight growth reduction by addition of α-linolenic acid itself as well as for FAX1-transformed cells in absence of α-linolenic acid, independent of yeast mutant strains (Fig. 8B, C). Whereas the former observation can be explained by unspecific, background uptake of α-linolenic acid provided at excess external concentrations, the latter effect might be due to a general, but minor, toxic effect of FAX1 expression in yeast.
To assess specificity of FAX1 for FAs, which have to be exported from chloroplasts in vivo, i.e., palmitic (C16:0), stearic (C18:0), and oleic acid (C18:1), we performed additional yeast growth complementation assays in the presence of the FA-biosynthesis inhibitor cerulenin and supply of moderate external FA concentrations (i.e., 100 μM; Fig. 8D, S4 Fig.). Results with rapidly (S4A–C Fig.) and non-exponentially growing cells (S4D Fig.) allowed definition of a potential substrate specificity of FAX1, preferring C16:0 over C18:1 and C18:0 FAs (for details see S4 Fig. and Discussion). When we tested α-linolenic acid (C18:3), which in planta is not exported from plastids (see [1]), as a control in this assay, FAX1 specificity was in the range as for stearic/oleic acid, but significantly lower than for palmitic acid (Fig. 8D, S4B Fig.).
In summary, our results show that the protein FAX1 in the IE of plastids is able to mediate FA-export as supported by the following findings: (i) FA-transport function of FAX1 in yeast; (ii) differential distribution of ER- and plastid-derived FAs/lipids in FAX1 mutant plants; (iii) male sterility of fax1 knockout lines, caused by impaired delivery of FA-derived compounds; (iv) decreased ketone wax compounds in cuticular layers of fax1 knockout stems; (v) a focus on differential expression of genes for acyl lipid as well as carbohydrate and cellular/cell wall metabolism in FAX1 mutant lines (see S1 Text).
Complementation of the yeast fat1, but not of the faa1/faa4 mutant, indicates that FAX1 is acting only in membrane transfer of FAs and not in FA-activation. This is in contrast to, for example, yeast and human FA-transporters such as Fat1p and FATPs, which in addition have acyl-activating functional domains [28,29]. FAX proteins group into the Tmemb_14 family and thus most likely contain three hydrophobic, membrane-spanning α-helical domains and one amphiphilic helix at the lipid bilayer/water interface. Thus, it is tempting to speculate that the latter might be responsible for binding and transfer of FA-chains across the IE membrane. Once loaded with a FA produced in the plastid stroma, this α-helix might become lipophilic enough to fold into the lipid bilayer and flip FAs over the IE. Furthermore, FAX1 and also FAX2 (see Fig. 1B) contain an extended N-terminal region (gray helix in Fig. 2A). Structural modeling indicates that these stretches fold into additional, most likely non-membrane associated α-helices: one for FAX1, two for FAX2, respectively. Interestingly, the two anti-parallel helices of the FAX2 N-terminus fit to sequence and structure of a ‘four-helical up-and-down bundle’ of the human apolipoprotein apoE3, which is involved in lipid transport and binding during formation of lipoprotein particles. Amphiphilic α-helices in the C-terminus of apoE3 are described to bind to lipids and thereby induce a conformational change in the N-terminal helix bundle that allows detergent-like solubilization of lipids and formation of lipoprotein particles (for overview, see [30,31]). Therefore, the N-terminal helices of plastid FAX proteins might be involved in similar functions during FA-transport. The different apparent molecular weights observed for FAX1 (S1D Fig.), most likely resulting from discrete conformations and/or packing of membrane domains, support these hypotheses for a transport mode. Once at the intermembrane space, FA-handover from FAX1 to substrate binding proteins, and transport across the OE membrane via a ß-barrel protein might be possible. For plastid re-import of eukaryotic lipids for example, the latter two proteins are represented by TGD2 (substrate binding) and TGD4 (OE ß-barrel, [3,11]). Furthermore, in E. coli, a similar system has been described for export of lipopolysaccharides, including an ABC transporter that flips the lipid moiety across the inner membrane, transfer proteins in the periplasm, and a ß-barrel protein in the outer membrane [32]. For plastids, subsequently an acyl-CoA synthetase (ACS) at the cytosolic face of the OE might finally drive FA-transfer in a passive, carrier-like process. Co-expression of At-FAX1 with LACS4 (ATTED-II coexpression networks), and regulation of LACS1, 3, and 5 expression in FAX1 mutants (S5 Table, S6 Table) underline a possible cooperation with ACS.
The close structural similarity of FAX proteins to the human TMEM14A and TMEM14C, which both localize to mitochondrial membranes, in the future might enable explanation of TMEM14 protein function in vertebrates. Whereas TMEM14C was identified to coexpress with the core machinery of heme biosynthesis and its knockdown causes anemia in zebrafish [12], TMEM14A was described to stabilize mitochondrial membrane potential and thereby inhibit apoptosis in a yeast system [13]. However, their exact biological function is still unknown. Since animal mitochondria are the site for FA-degradation via ß-oxidation, a role for TMEM14 proteins in FA/lipid homeostasis, energy metabolism or disease (e.g., apoptosis) in vertebrates might be possible.
Levels and subcellular distribution of free FAs and polar lipids in Arabidopsis FAX1 mutants mainly correlated with a FA-export function, by which FAX1 influences cellular FA and acyl lipid homeostasis (for overview, see [1]). Most likely because of their toxicity and high metabolic fluxes for primary metabolites, changes in free FAs were not very pronounced. However, very-long–chain FAs (C20), which are elongated outside plastids and thus require previous export of C16–18 FAs, were significantly reduced in fax1 knockouts (Table 3). According to acyl-ACP synthesis rates and specificity of thioesterases in Arabidopsis chloroplasts, oleic acid (C18:1) is the major free FA exported from chloroplasts, followed by palmitic acid (C16:0) and only very little amounts of stearic acid (C18:0; compare [1,33,34]). FAX1 in yeast assays performed best for FA 16:0 (determined specificity range: 16:0 > 18:1 ~ 18:0 ~ 18:3) and thus, most likely, mainly is involved in the plastid export of free palmitic acid but also can transport oleic acid, which at the stromal side of the plastid IE is provided at highest substrate concentrations. The fact that in yeast, FAX1 was also able to transport α-linolenic acid (C18:3), which in planta is retained inside plastids, indicates that the protein does not discriminate between different degrees of unsaturation, but in general prefers C16 over C18 FAs. In chloroplast IE membranes, FAX1 most likely functions in a passive, carrier-like mode, driven by concentration gradients of free FA substrates (see above). Interestingly, accumulation of export-directed C16–18 FAs in flower tissue of fax1 knockouts (+0.24 > +0.04 > +0.01 mol% for 16:0 > 18:0 > 18:1), reflect the substrate specificity range of FAX1 in yeast (compare S3A Table and S4 Fig.). Furthermore, non-exported FA 18:3 significantly decreased (0.14 mol%; S3A Table) in flowers of fax1 knockouts, thereby maybe pointing to stronger fluxes of FAs into plastid-intrinsic pathways (e.g., synthesis of oxilipin hormones), due to a block in FA export via FAX1.
Besides changes in free FAs, 65% of the differential lipid patterns depicted in Table 3 underline the hypothesis of plastid FA-export via FAX1, best documented by the strong reciprocal changes in TAG oils. Here for almost 90% of all significantly distributed TAGs (compare S4 Table), the pattern in both FAX1 mutants and tissues perfectly matched to a FA-export function of FAX1. The distribution of 34:x glycolipids (MGDG, DGDG, SQDG), which increased in fax1 knockouts but decreased in FAX1 over-expressors, also corresponded to our theory. In this case, we can, however, not exclude a contribution of ER-made species, since the diacylglycerol backbone for the “34”-glycolipids can originate both, from prokaryotic (from plastids) and eukaryotic (from the ER) phospholipid precursors, respectively. Yet, Arabidopsis is a so-called 16:3 plant, which for galactolipids prefers the prokaryotic pathway with high levels of 16:3 acyl chains. In contrast, ER-derived “34” DAG-backbones contain 16:0 saturated acyl moieties (compare [1,34]). Thus, we can assume that MGDG 34:x and DGDG 34:x with more than four desaturated C-bonds are completely synthesized in plastids. For the strong MGDG 34:x reductions in FAX1ox leaves and flowers (2.8 and 0.5 mol%, Table 3) and the increase in fax1 flowers (+0.7 mol%) indeed 34:4, 34:5 and 34:6 are the major contributing species (see S2 Table, S3 Table) and therefore support our hypothesis. Most likely at least the abundant forms of phosphatidyl-glycerol (PG 34:3, 34:4) are exclusively made inside plastids as well (see [1]), and thus the pronounced overall increase of PG in fax1 leaf tissue (+3.2 mol%) also mainly is due to a block of FA-export via FAX1.
Our assumption that FAX1 mediates plastid FA-export is further confirmed by a large decrease of PC-levels in fax1 knockout tissues (up to 8.8 mol%) and a strong increase of PC in FAX1ox leaves (+3.0 mol%). However, also, considerable contrasting evidence is found for three ER-made lipids in flower tissue (i.e., +0.9–1.0 mol% MGDG 36:x, DGDG 36:x in fax1; -5.6 mol% PC in FAX1ox). The latter findings that only apply to lipid species synthesized in the cytosol/ER of flowers might be explained by the inhomogeneity of mature flowers, consisting of leaf, stalk, pollen, ovary, and seed/silique tissues, and/or by a preferential flow of FA-building blocks for lipids into TAG oils during seed development. Furthermore, for the bilayer-forming DGDG 36:x, a plastid export is described to act as surrogate lipid for the lack of PC at, e.g., phosphate-limited growth conditions (see [1] and references therein). Thus, the observed increase of DGDG 36:x species in fax1 knockout mutants might compensate for the strong decrease of PC in the same tissues (compare Table 3).
In summary, levels and subcellular distribution of free FAs and polar lipids in Arabidopsis FAX1 mutants mainly support a plastid FA-export function of FAX1. In addition, we can, however, not exclude plastid FA-export via different mechanisms or a bypass by other plastid FAX proteins (see below). Due to this potential functional redundancy of plastid FAX proteins, mutation of FAX1 alone does not affect all lipid species present in plants. Effects in leaf tissue, in particular of fax1 knockouts are somewhat more straightforward and stronger than in FAX1 over-expressing lines. The latter is not unexpected for mutation of a protein involved in transport, which is highly expressed in leaf tissues (see below and S5 Fig.). However, the impact of FAX1 mutation on TAG-oil levels might be of future biotechnological importance. Interestingly, already the enhanced FA-transport by FAX1 was able to significantly increase TAG contents in leaves and flower tissues. Furthermore, our finding is in line with higher TAG when FA-loading to the ER in seeds is improved by over-expression of the ABC transporter ABCA9 [10]. Thus, coupling of the bottlenecks FA-transport (e.g., FAX1, ABCA9) with those of FA-synthesis and acyl-editing processes and a seed-specific expression might boost plant oil production in future approaches.
Transcripts of At-FAX1 are present in all developmental stages and peak in leaf tissues (cotyledons, rosette, caulinary, senescent leaves, and flower sepals) as well as in early pollen development (S5A–C Fig.). Consequentially, the strongest phenotype of fax1 knockout mutants can be observed during growth (e.g., reduced rosette leaf size and biomass) and in particular in pollen grains, leading to almost complete male sterility due to the absence of pollen cell walls and impaired pollen release by anthers. For FAX1, we propose a function in FA-export from plastids of tapetum cells in anthers, which in fax1 knockouts leads to the strongly impaired assembly of exine layers and pollen coat, most likely because of the lack of FA-precursors for sporopollenin and/or tryphine synthesis (for a detailed description, see S2 Text).
Since FAX1 in Arabidopsis belongs to a family of seven proteins, the plastid-predicted FAX2, 3, and 4, whose expression is regulated throughout plant development as well (S5 Fig.), most likely can bypass the loss of FAX1 function in all tissues and organs, leading to the rather mild overall phenotype of fax1 knockouts. Especially in seed tissue (S5D, E Fig. and S6 Fig.), FAX2, 3, and 4 most likely play a more prominent role than FAX1. Indeed transcripts for FAX2 and FAX3—i.e., the highest plastid FAX genes in seed development and germination—showed to be significantly up-regulated (1.13- and 1.24-fold) in fax1 knockout flowers. Please note that with a relative signal of 1585 in wild type, FAX2 and FAX3 are strongly expressed in flower tissue we used for microarray analysis (among highest 9% of all genes on the chip; compare microarray dataset E-MTAB-3090 at www.ebi.ac.uk/arrayexpress).
The function of FAX1 in FA-delivery for pollen cell wall as well as for cuticular wax assembly is further underlined by differential gene expression in FAX1 mutants (see S1 Text, S5 Table, S6 Table), and by the occurrence of phenotypes similar to fax1 when biosynthesis pathways for FA/lipid-derived precursor material are mutated in Arabidopsis. These include a plastid-intrinsic fatty-acyl-ACP reductase (AlcFAR2/MS2), involved in primary fatty alcohol synthesis for anther cuticle and pollen sporopollenin formation [35]; as well as cytochrome P450 enzymes (CYP703A2, CYP704B1; [36,37]) that hydroxylate FAs, and the ACS ACOS5 [38] that activates FAs for sporopollenin synthesis in the cytosol of anther cells. Furthermore, several long-chain ACS (LACS1, 2, 4) are necessary to activate long-chain and very-long–chain FAs for building of cutin and wax as well as pollen exine layers [39,40]. In addition, several ABC transporters in the plasma membrane are required for deposition of surface lipids, displaying fax1-like phenotypes upon mutation: ABCG26 for pollen exine formation from tapetum cells, as well as ABCG11, G12, and G13 in lipid export from epidermis cells for formation of cuticular wax layers (for overview, see [41]). As for FAX1, pathways for synthesis of precursors of pollen cell wall and cutin/wax components often overlap. In stems of fax1 knockouts, we further identified strong regulation of two genes involved in wax biosynthesis: AlcFAR3/CER4 and CYP96A15/MAH1 (S6 Table). Because the latter enzyme is catalyzing synthesis of wax ketone components, its differential expression is in line with the observed lack of C29 ketones in fax1 knockout stems.
Besides deranged acyl lipid homeostasis, obviously also carbohydrate and cellular/cell wall metabolism was affected in FAX1 mutants as reflected by the impact on plant biomass production and differential regulation of gene expression (see S1 Text, S7–S10 Figs.). In general, we assume that these effects are rather secondary and most likely might result from still unknown signaling events, due to changed FA/lipid homeostasis. Since fax1 knockout plants are short of energy-rich lipids, they most likely turn down anabolic carbohydrate metabolism required for polysaccharide synthesis, resulting in, e.g., reduced biomass and secondary cell walls. The opposite effect is observed in FAX1 over-expressors, in which an excess of lipids most likely leads to more biomass and the production of additional cell layers in stems. These observations clearly demonstrate regulation of energy metabolism and a close correlation between the availability of FAs/lipids and the utilization of carbohydrates in growth processes. This link is further underlined by the finding that the flow of carbon into oil can be promoted by activating plastid FA synthesis and repressing starch synthesis [42]. In this context FAX1, to our knowledge, is not only the first membrane protein identified that mediates FA-export from plastids, but FAX1 and its relatives represent key transport proteins and thus—together with enzymes of FA/lipid-synthesis and modification—might provide powerful future tools to modulate plant lipid and bioenergy production [43].
Experiments were performed on Arabidopsis thaliana ecotype Columbia (Col-0, Lehle Seeds; Round Rock, United States). The T-DNA insertion lines SAIL_66_B09 (fax1–1) and GABI_599E01 (fax1–2) were purchased from NASC (Nottingham Arabidopsis Stock Center, Nottingham, United Kingdom) and GABI-Kat (MPI for Plant Breeding Research, Köln, Germany), respectively. To generate complementation lines of fax1–2 and over-expressing At-FAX1 under the control of the 35S promoter, the coding sequence of At-FAX1 was subcloned into pH2GW7 [44]. At-FAX1/pH2GW7 was transformed into Agrobacterium tumefaciens GV3101, which was used to transfect heterozygous fax1–2 and Col-0 plants as described [45]. Arabidopsis seeds were sown on soil, vernalized at 4°C in the dark for two days, and grown in a 16 h light (22°C; 100 μmol photons ⋅ m–2 ⋅ s–1) and 8 h dark (18°C) cycle.
At-FAX1 cDNA was purchased as SSP pUNI51 clone U12755 [46]. The corresponding mRNA (NCBI reference sequence NM_15588) is predicted to be 1,030 bp long, including 180 bp 5´- and 169 bp 3´-untranslated regions (UTRs; S1A Fig.). For amplification of FAX1 from pea, RT-PCR was performed using pea seedling cDNA as template and oligonucleotide primers designed according to a pea EST contig sequence [47]. The corresponding mRNA molecule was 1,115 bp long, with 143 bp 5´UTR, 699 bp coding region, and 273 bp 3´UTR (GenBank accession no. KF981436). For primer sequences, see S7 Table; for amino acid sequences, see Fig. 1A.
To generate a fusion of GFP to the preprotein At-FAX1, the coding sequence was subcloned into the p2GWF7 plasmid [44]. p2GWF7 provides a fusion of GFP to the C-terminal end of the respective proteins, which are expressed under the control of the constitutive 35S promoter. Transformation and analysis of Arabidopsis mesophyll protoplasts was performed as described [45]. GFP fluorescence was detected at 672 to 750 nm and chlorophyll autofluorescence was monitored at 503 to 542 nm by confocal laser scanning microscopy (Leica TCS SP5/DM 6000B, argon laser, excitation wavelength of 488 nm).
Pea chloroplasts isolated from leaf tissue of 10-day-old pea plantlets were sub-fractioned into OE and IE membranes, stroma and thylakoids as described [48]. Chloroplast envelopes, total protein extracts, and microsomal membranes from Arabidopsis plants were prepared as specified in [45] and [20], respectively. FAX1 antisera were raised in rabbit (Pineda Antibody Service, Berlin, Germany) against N-terminal peptide sequences of At-FAX1 (17 aa) and Ps-FAX1 (18 aa), respectively (see Fig. 1A). Antisera for marker proteins were produced as described previously [45,49]. Appropriate amounts of organellar or total cellular proteins were separated by SDS-PAGE, transferred to PVDF membranes and subjected to immunoblot analysis using primary antisera in 1:250 to 1:5000 dilutions in TTBS buffer (100 mM Tris-HCl pH 7.5, 150 mM NaCl; 0.2% Tween-20; 0.1% BSA). Non-specific signals were blocked by 3% skim milk powder and 0.1% BSA. Secondary anti-rabbit IgG alkaline phosphatase antibodies (Sigma-Aldrich) were diluted 1:30,000. Blots were stained using the alkaline phosphatase reaction with 0.3 mg/ml nitroblue tetrazolium (NBT) and 0.16 mg/ml bromochloroindolyl phosphate (BCIP) in 100 mM Tris pH 9.5, 100 mM NaCl, 5 mM MgCl2.
Genomic DNA of the T-DNA insertion lines fax1–1 and fax1–2 was screened by PCR genotyping. To identify plants with T-DNA insertion in both At-FAX1 alleles (homozygous), combinations of gene-specific primers that flank the predicted insertion sites with each other and with T-DNA-specific left border (LB) primers (S7 Table) were used. Positions and orientations of T-DNA inserts and oligonucleotide primers in fax1–1 and fax1–2 are shown in S1A Fig. To verify insertion sites, PCR-genotyping products were sequenced. T1 generations of generated FAX1 over-expression and complementaion lines were selected by hygromycin (30 μg/ml). Stable insertion of 35S::FAX1 was controlled by PCR-genotyping in all subsequent generations. Therefore, a vector-specific primer in combination with a FAX1 cDNA specific primer was used (S7 Table). In the T2 generation, complementation lines were selected for homozygous alleles of the original T-DNA insertion in fax1–2 (see above), resulting in lines Co#7 and Co#54. For FAX1 over-expression in Col-0 background, we selected the lines ox#2 and ox#4 in the T2 generation.
For microscopic analysis we used 5-week-old plants and dissected anthers from mature flowers or cut 1–2 mm2 stem segments 1 cm above the bottom of the second internode of the primary inflorescence stalk. We analyzed four individual fax1–2 knockouts, two of each Co#7, Co#54 complementation lines, and five Col-0 wild-type plants for anthers/pollen grains, and pictured stem tissue of independent fax1–1, fax1–2 knockouts, three ox#2, four ox#4 over-expressors, as well as seven individual Col-0 wild-type plants, respectively. Tissue was fixed immediately after harvest with 2.5% (w/v) glutaraldehyde (4°C, at least 24 h) in 75 mM cacodylate buffer (2 mM MgCl2, pH 7.0), rinsed several times with fixative buffer, and subsequently post-fixed with 1% (w/v) osmium tetroxide for at least 2.5 h in fixative buffer at 20°C. After two washing steps in distilled water, samples were stained with 1% (w/v) uranyl acetate in 20% acetone, dehydrated with a graded acetone series and embedded in Spurr’s low viscosity epoxy resin [50]. For light microscopy, semithin-sections (1–2 μm) were cut with a glass knife (Pyramitome 11800, LKB). Ultrathin-sections (50–70 nm) for transmission electron microscopy were prepared with an ultramicrotome (EM UC6, Leica) and post-stained with aqueous lead citrate (100 mM, pH 13.0). Micrographs were taken at 80 kV with a 268 electron microscope (Fei Morgagni).
The second to fourth internode region of primary inflorescence stalks from 7-week-old plants was used for wax and cutin analyses. For each replicate, stem segments from three to four individual plants were pooled, and samples were provided from two independent harvests of each FAX1 mutant line and respective wild-type controls. Determination of wax and cutin coverage of stems was essentially carried out as described previously [51,52]. Wax was extracted in chloroform and C24 alkane was added as internal standard. For cutin analysis, exhaustively extracted stems (1:1; methanol:chloroform) were transesterified using methanolic HCl, and cutin monomers were extracted in hexane containing C32 alkane as internal standard. Gas chromatographic and mass spectrometric analysis was carried out after derivatization of extracted wax and cutin with pyridine and BSTFA.
For each independent harvest (2-times for fax1 knockout, 4-times for FAX1 over-expressing lines) cauline leaves and flowers (stage 10–15, according to [53]) were sampled from at least ten individual, 7-week-old plants and grinded in liquid nitrogen. To be able to work on tissue of identical sample pools (i.e., from 7-week-old plants) for wax/cutin analysis, FA/lipid determination, and transcript profiling, as well as because FAX1 is highly expressed in cauline leaves (see S5A Fig.), we chose the latter instead of old rosette leaves. Tissue powder of each harvest was portioned into three aliquots of 50 mg, which were used to determine polar lipid and free FA contents. For details on data analysis, see S1 Table. Lipids/FAs were extracted from six (fax1 k.o) to 12 (FAX1 over-expressors) biological replicates using 1 ml of a pre-cooled (−20°C) methanol:methyl-tert-butyl-ether (1:3) mixture, spiked with 0.1 μg/ml PC 34:0 (17:0, 17:0) as internal standard. The samples were incubated for 10 min at 4˚C, followed by another 10 min incubation in an ice-cooled ultrasonication bath. After adding 650 μl of UPLC grade water:methanol (3:1), the homogenate was vortexed and centrifuged for 5 min in a table top centrifuge. The addition of water:methanol leads to a phase separation producing an upper organic phase, containing the lipids, and a lower phase containing the polar and semi-polar metabolites. The upper organic phase was removed, dried in a speed-vac concentrator, and re-suspended in 500 μl buffer B (see below) and transferred to a glass vial. 2 μl of this sample were injected onto a C8 reversed phase column (100 mm × 2.1 mm × 1.7 μm particles BEH C8, Waters), using a Waters Acquity UPLC system. The two mobile phases were water (UPLC MS grade, BioSolve) with 1% 1 M NH4Ac, 0.1% acetic acid (buffer A), and acetonitrile:isopropanol (7:3, UPLC grade BioSolve) containing 1% 1 M NH4Ac, 0.1% acetic acid (buffer B). The gradient separation, which was performed at a flow rate of 400 μl/min, was 1 min 45% A, 3 min linear gradient from 45% A to 35% A, 8 min linear gradient from 25 to 11% A, 3 min linear gradient from 11% A to 1% A. After washing the column for 3 min with 1% A the buffer was set back to 45% A and the column was re-equilibrated for 4 min (22 min total run time). Mass spectra were acquired as described [23,24,54] using either an Orbitrap Exactive mass spectrometer (Thermo-Fisher) for fax1 knockout lines or a Waters Synapt G1 (Waters) for FAX1 over-expressors, and corresponding wild types, respectively. The spectra were recorded using altering full scan and all-ion fragmentation scan mode, covering a mass range from 100–1,500 m/z. The resolution was set to 10,000 with 10 scans per second. Spectra were recorded from min 0 to min 20 of the UPLC gradients. The analysis of the spectra (alignment, peal picking, normalization and peak integration) was performed with the software package CoMet 2.0 (Nonlinear Dynamics) according to the instructions of the vendor.
For growth assays in yeast, the coding sequence of the mature At-FAX1 protein was subcloned into the yeast expression plasmid pDR195 (XhoI/BamHI). Therefore, we fused the open reading frame of the predicted mature At-FAX1, starting with aa 34 of the preprotein, behind an “ATG” base triplet by PCR amplification. The yeast mutant strains fat1 (LS2020-YB332) and faa1/faa4 (LS1849-YB525) are specified in [26]. Both strains were transformed with mature At-FAX1/pDR195 and the vector control pDR195 as described [45]. If not denoted elsewhere, liquid cultures of the respective yeast cells were grown to exponential phase in synthetic defined medium (SD-ura), containing 0.1% (w/v) glucose, 0.7% (w/v) yeast nitrogen base without amino acids, and necessary auxotrophic amino acids without uracil. Subsequently, 2 μl drops of the cultures were spotted in different dilutions onto SD-ura plates (2% agar), supplemented with 3.6 mM α-linolenic acid (0.1%, w/v in ethanol), and 1% tergitol (to increase α-linolenic acid solubility). For control plates, an equal amount of the solvent ethanol was added instead of α-linolenic acid. Assays in the presence of cerulenin were performed according to [26,27] in SD-ura media supplemented with 2% (w/v) glucose, 0.5% Brij 58, 0.7% KH2PO4, 10μM cerulenin and either 100μM of palmitic, stearic, oleic or α-linolenic acid. Growth of yeast cells on solid media was documented between 2 to 6 days at 30°C. OD600 measurements were performed in identical liquid media, inoculated to a starting OD600 of 0.05 or 0.06/0.03 for cerulenin experiments with the respective yeast cells. Cultures were continuously shaken at 30°C and the OD at 600nm was determined at indicated time points.
Tissue from flowers (stage 10–15 according to [53], compare S3 Fig.) and from the second to fourth internode of primary inflorescent stalks for each harvest was pooled from more than ten individual, 7-week-old plants (identical sample pool for lipid analysis) and used for preparation of RNA samples by the Plant RNeasy Extraction kit (Qiagen). RNA (200 ng) of four or five independently harvested samples (n = 4–5) from wild type (Col-0 and WT2, segregated from heterozygous fax1–2), fax1 knockout (fax1–1 and fax1–2 lines) as well as FAX1 over-expressors (ox#2 and ox#4 lines) was processed and hybridized to Affymetrix GeneChip Arabidopsis ATH1 Genome Arrays using the Affymetrix 3´-IVT Express and Hybridisation Wash and Stain kits (Affymetrix, High Wycombe, UK) according to the manufacturer’s instructions. Raw signal intensity values (CEL files) were computed from the scanned array images using the Affymetrix GeneChip Command Console 3.0. For quality check and normalization, the raw intensity values were processed with Robin software [55] default settings as described [19]. Specifically, for background correction, the robust multiarray average normalization method [56] was performed across all arrays (between-array method). Statistical analysis of differential gene expression of mutant versus wild-type samples was carried out using the linear model-based approach developed by [57]. In total, we analyzed the following comparisons (see S7 Fig.): (A) flowers: fax1 knockout (n = 5) versus wild type (n = 5); (B) flowers: FAX1 over-expressors (n = 8, four times each ox#2, ox#4) versus wild type (n = 5); (C) stems: fax1 knockout (n = 4) versus wild type (n = 4). The obtained p values were corrected for multiple testing using the nestedF procedure, applying a significance threshold of 0.05 in combination with the Benjamini and Hochberg false-discovery rate control [58]. All microarray data are available in the ArrayExpress database (www.ebi.ac.uk/arrayexpress) under accession number E-MTAB-3090.
Structural models of At-FAX1 and At-FAX6 were generated by Phyre2 [59], based on alignments with the PDB entries for human TMEM14C (c2losA) and TMEM14A (c2lopA), respectively. Identity of At-FAX1 with its template TMEM14C was 21% and for At-FAX6 with TMEM14A 36%, while confidence of both models was 99.9%, thereby indicating a high confidence and accuracy of the core models. Structural alignments were created with PyMOL [60].
|
10.1371/journal.pgen.1000313 | Sexually Antagonistic “Zygotic Drive” of the Sex Chromosomes | Genomic conflict is perplexing because it causes the fitness of a species to decline rather than improve. Many diverse forms of genomic conflict have been identified, but this extant tally may be incomplete. Here, we show that the unusual characteristics of the sex chromosomes can, in principle, lead to a previously unappreciated form of sexual genomic conflict. The phenomenon occurs because there is selection in the heterogametic sex for sex-linked mutations that harm the sex of offspring that does not carry them, whenever there is competition among siblings. This harmful phenotype can be expressed as an antagonistic green-beard effect that is mediated by epigenetic parental effects, parental investment, and/or interactions among siblings. We call this form of genomic conflict sexually antagonistic “zygotic drive”, because it is functionally equivalent to meiotic drive, except that it operates during the zygotic and postzygotic stages of the life cycle rather than the meiotic and gametic stages. A combination of mathematical modeling and a survey of empirical studies is used to show that sexually antagonistic zygotic drive is feasible, likely to be widespread in nature, and that it can promote a genetic “arms race” between the homo- and heteromorphic sex chromosomes. This new category of genomic conflict has the potential to strongly influence other fundamental evolutionary processes, such as speciation and the degeneration of the Y and W sex chromosomes. It also fosters a new genetic hypothesis for the evolution of enigmatic fitness-reducing traits like the high frequency of spontaneous abortion, sterility, and homosexuality observed in humans.
| Our study describes a new form of sexual genomic conflict that operates through the process of antagonistic green-beard effects. Although past theoretical and empirical work indicated that green-beard effects rarely operate in nature, our new theory shows why this conclusion may have to be reevaluated. We integrate modeling analysis with extant empirical work to show that the unique properties of sex chromosomes can lead to a previously unappreciated form of sexual conflict (sexually antagonistic zygotic drive) that may be widespread in nature. It operates through harmful epigenetic parental effects, asymmetrical allocation of parental investment to sons and daughters, and asymmetrical interactions between brothers and sisters. Sexually antagonistic zygotic drive is functionally analogous to meiotic drive except that it operates due to competition among opposite-sex siblings rather than between competing gametes.
| Sex chromosomes are unusual compared to the autosomes for three reasons. First, when present in the heterogametic sex, the two types of sex chromosome are transmitted to opposite sex offspring. Second, it is common for recombination to be suppressed over a part or all of their length. Third, non-recombining sex chromosomes can evolve to become far more dimorphic than autosomes. It has long been recognized that these characteristics can contribute to genetic conflict in the context of meiotic drive, but other forms of potential sex-linked genetic conflict have received relatively little attention (reviewed in [1]). Here we evaluate the potential for the special characteristics of the sex chromosomes to contribute to a meiotic-drive like process – sexually antagonistic zygotic drive (hereafter, SA-zygotic drive) – that operates due to competition among opposite-sex siblings, rather than gamete types. The phenotypes that fuel this process are sexually antagonistic green-beard effects (hereafter SA-GrBd-effects) that only operate when there is competition among siblings.
A green-beard effect [2],[3] is a complex trait coded by a pleiotropic gene, or a collection of tightly linked genes, with three distinct characteristics (Figure 1): they cause the carrier to i) produce a distinguishing phenotype (tag), ii) differentiate among other individuals based on the presence or absence of the phenotype (tag-differentiation), and iii) augment the fitness of other individuals expressing the phenotype (tag-directed-aid). A green-beard effect is antagonistic when it reduces the competitive ability of individuals that do not express the tag, thereby increasing the fitness of individuals carrying the gene that codes for it. Because green-beard effects require complex and multifarious pleiotropy, they have previously been presumed to be rare in nature [2],[3].
However, documented examples of green-beard effects do exist (e.g., [4]–[6]). For example, in the red fire ant (Solenopsis invicta) egg-laying queens are heterozygotes for the a and b alleles at the Gp-9-locus. Homozygous queens are absent because the b allele is a recessive lethal and developing aa queens are killed by ab heterozygous workers (but not by aa homozygous workers) [4]. Queens with the ab genotype that were experimentally rubbed against aa queens were also killed by heterozygous workers. These data indicate that the b allele (or an allele at a tightly linked locus) displays an antagonistic green-beard phenotype because it enhances its own propagation by killing aa competitors (identified by their smell) that do not carry it. Green-beard effects may also feasibly operate in humans and other placental mammals (by influencing resource transfer between maternal and fetal tissue) in the context of self-recognizing gene products, e.g., homophilic cell adhesion molecules that have extracellular domains that recognize copies of themselves expressed on other cells [7]. What has not been appreciated previously, however, is that the special characteristics of sex chromosomes greatly facilitate the evolution of SA-GrBd-effects whenever there is competition among siblings.
For simplicity – but without loss of generality – we will assume male heterogamety. There are, however, some important biological differences between male (XY) and female (ZW) heterogamety, and when appropriate, we will point out how such differences may influence the course of evolution. Lastly, when we refer to the two types of sex chromosomes, we will be referring to the portion of these chromosomes that does not recombine in the heterogametic sex.
Sex chromosomes are predicted to evolve to code for SA-GrBd-effects, and the sexually antagonistic zygotic drive that they propel, for three reasons. First, all X- and Y-linked genes co-segregate during male meiosis like a single Mendelian gene that is highly pleiotropic. As a consequence, different genes on the same sex chromosome, rather than pleiotropy of a single gene, can code for the multifarious phenotypes required for green-beard effects to operate. A second feature promoting X- and Y-coded SA-GrBd-effects is the presence of the master sex-determining gene on one of these chromosomes. This linkage creates a perfect association between the presence or absence of a father's X and Y in his offspring and all sexually dimorphic phenotypes that are coded by any gene in the genome, i.e., within a family, all daughter-specific traits are effectively paternal X-tags and all son-specific traits are effectively Y-tags (Figure 2). The final feature contributing to sex chromosomes being hot-spots for SA-GrBd-effects is competition among siblings. In this case, any X- or Y-coded phenotype that differentially influences the competitive ability of the two sexes of offspring can cause a SA-GrBd-effect in three ways (Figure 3):
The same logic applies to maternal SA-GrBd-effects in the context of ZW sex determination, but the opportunity for the epigenetic modification by the mother of an offspring's gene expression may be more substantial owing to her multifarious influences on the developing egg (e.g., deposition of steroid hormones in the yolk and RNAs in the egg's cytoplasm).
The logic of SA-zygotic drive is an extension of the concepts of meiotic and gametic drive that operates postzygotically during ontogeny rather than prezygotically during meiosis and gametogenesis. As a consequence, many of the evolutionary principles developed for meiotic drive in classic papers by Sandler and Novitski (1957) [8], Hiraizumi et al. (1960) [9], Hamilton (1967)[10], Hartl (1975) [11], and others will also apply to SA-zygotic drive. However, we will show in this paper that the postzygotic operation of SA-zygotic drive (unlike the prezygotic process of meiotic drive) has a unique mode of operation that creates unprecedented, broad-scale opportunity for green-beard effects to evolve. These SA-GrBd-effects are predicted to be capable of causing a wide diversity of maladaptive phenotypes that are expressed in the diploid phase of the lifecycle.
Previous theoretical work from our laboratories has shown that linkage to the W and Z chromosomes in species with female heterogamety facilitates the evolution of selfish genetic elements that code for heritable maternal effects [12]. Here we focus predominantly on X- and Y-coded green-beard effects that evolve due to paternal epigenetic effects, parental investment (PI) by either heterogametic sex (XY or ZW), and sibling-sibling interactions (competitive sib-sib-interactions).
In the following sections we first evaluate the biological feasibility of the evolution of SA-zygotic drive of the sex chromosomes via SA-GrBd-effects, and how the autosomes would be expected to respond to such evolution. We focus especially on the feasibility of paternal epigenetic effects, because of the constraints imposed on their transmission between father and offspring via the sperm. We next develop a mathematical model of SA-zygotic drive due to coevolution between X and Y coded SA-GrBd-effects. Before discussing our collective findings, we describe how SA-zygotic drive can provides a new genetic hypothesis for the evolution of enigmatic traits, like high-frequencies of spontaneous abortion, sterility, and homosexuality, that reduce Darwinian fitness.
Consider the expression of the paternal X and Y chromosomes during spermatogenesis at a time when the developing gametes remain functionally diploid, i.e., before the primary spermatogonial cell has divided into haploid spermatids, and also while the four developing spermatids derived from each spermatagonial cell remain connected by cytoplasmic bridges that permit RNA, steroid hormones, proteins and other molecules to be exchanged (i.e., most of spermatogenesis; [13]). With sib-competition, any X-coded epigenetic modification that influences gene expression in sons, and thereby reduces their competitive ability, would be favored by genic selection. An X-linked mutation producing such a paternal epigenetic effect represents a SA-GrBd-effect between a father and his offspring because it differentially helps those offspring that carry the mutation. For example, consider an X-coded mutation that was expressed during spermatogenesis and that epigenetically modified the expression of an autosomal gene (in the zygote or developing embryo of the next generation) in a manner that disrupted a male-specific ontogenetic pathway (such as dosage compensation in Drosophila melanogaster) and thereby reduced the competitive ability of sons during sib-competition. In this case, the green-beard ‘tag’ is the presence or absence of the male-specific ontogenetic pathway, the ‘tag-differentiation’ is the epigenetic modification of the expression of a gene in a male-specific ontogenetic pathway that harms only (or disproportionately) sons, and the ‘tag-directed-aid’ is the resulting increased competitive ability of daughters competing with debilitated brothers. When there is sib-competition, an X-coded green-beard effect that aids (harms) one sex of offspring necessarily harms (aids) the other sex – and hence such green-beard effects are necessarily sexually antagonistic. The same logic applies to Y-coded paternal epigenetic effects that help sons by harming daughters. For example, consider a Y-coded epigenetic effect that caused mis-expression of any gene located on the paternally inherited X chromosome. This phenotype would debilitate only daughters and thereby increase the fitness of the Y chromosome when there is sib-competition. Although the Y chromosome in many species may currently contain relatively few structural genes [14], this would not have been true historically before degeneration of the Y occurred. Furthermore, a highly degenerated Y chromosome with respect to structural genes may retain substantial regulatory potential as recently shown for D. melanogaster [15].
With male heterogamety, sexually antagonistic epigenetic effects must operate through the sperm, which provides far more formidable barriers to expression of paternal effects compared to that of maternal effects through the egg [16]: sperm are much smaller than eggs, nearly all paternal cytoplasm is stripped away during spermatogenesis, paternal imprinting via histone modification is restricted due to protamines replacing paternal histones, and paternal imprinting via methylation is made difficult due to the nearly global demethylation of the paternal chromosomes after fertilization, as occurs in mammals. Nonetheless, a large body of extant evidence indicates that sexually antagonistic paternal (and maternal) effects can and do operate in nature, as described below.
Most research on paternal epigenetic effects in animals has focused on methylation-based imprinting in mammals. This process, however, is unlikely to contribute substantially to SA-GrBd-effects coded by the sex chromosomes because it operates through cis-acting imprinting control regions (ICRs, which are associated with relatively small proportion of genes) [17]. In contrast, the X and Y are selected to produce trans-acting gene products that epigenetically modify the expression of other parts of the genome in offspring that do not carry the coding sex chromosome.
In Text S1, we summarize extant studies to provide evidence that: i) epigenetic maternal and paternal effects have evolved many times that selectively kill offspring that do not carry them, ii) mutations that cause antagonistic parental effects that selectively harm only one sex of offspring are well documented, at least in D. melanogaster in the context of maternal effects, iii) the expression levels of hundreds of genes in D. melanogaster are influenced by both maternal and paternal effects, with no evidence that this phenomenon is caused by imprinting-based parent-of origin effects iv) trans-acting epigenetic paternal effects (that are not parent-of-origin effects, and that influence offspring that do not carry the coding gene) can be produced by RNAs produced during spermatogenesis and transferred to the zygote (as RNA or cDNA), and v) epigenetic maternal effects that influence the competitive ability of one sex of offspring over the other can be produced by varying steroid levels in the yolk. Collectively these studies provide evidence that X and Y-coded (and Z and W-coded) SA-GrBd-effects can feasibly evolve through both paternal and maternal effects.
Here, we briefly overview some examples of the material covered in Text S1. Antagonistic maternal effects are well documented. In mice (HSR, scat+, OmDDK) and beetles (Medea factors), there are polymorphic alleles in natural populations that produce maternal effects that kill all of the siblings in a brood that do not carry them (reviewed in [1]). In D. melanogaster, there are at least three established loci that can mutate to alleles that kill sons via a maternal effect (snl, sok-1, and sok-2) and three that similarly kill only daughters (l(2)mat, da, and Ne) [18]. In birds, a maternal effect (elevated yolk androgen concentrations in the barn swallow, Hirundo rustica) causes enhanced growth rate of sons but reduced growth rate of daughters [19]. Trans-generational epigenetic paternal effects are also well documented. In Caenorhabditis elegans, a pair of tightly linked genes (peel-1 and zeel-1) code for a paternal effect that kills offspring that do not carry them [20]. In mice, a trans-generational epigenetic paternal effect, coded by an allele at the Kit locus, has been demonstrated to be mediated by RNAs produced during spermatogenesis and transmitted to the egg [21]. Human sperm transfer over 4,000 different types of RNA transcripts to the egg, including at least 68 miRNAs [16]. These studies demonstrate that mutations causing the phenotypes needed for SA-zygotic drive to operate do in fact occur.
Past evolution of antagonistic X- and Y-coded SA-GrBd-effects should have selected for adaptations by the affected sex chromosome to suppress them, and by the autosomes to suppress them whenever they harm one sex of offspring more than they help the other sex. A candidate phenotype for such suppression is the enigmatic early-inactivation of sex chromosomes (but not the autosomes) during the process of spermatogenesis. This is a well documented phenomenon in organisms as diverse as fruit flies, worms and mammals, but its adaptive significance is poorly understood [22],[23]. All chromosomes are inactivated during the latter stages of spermatogenesis when the sperm's DNA becomes highly condensed. However, the X and Y chromosomes are inactivated far in advance of the autosomes, during the early stages of spermatogenesis [24],[25]. Although the selective factors that led to the evolution of the early-inactivation of the sex chromosomes are unknown, the phenomenon is consistent with what would be expected if X and Y-coded SA-GrBd-effects have been important historically. If early-inactivation of the X and Y reduced the production of RNAs coded by these chromosomes during spermatogenesis, this would interfere with RNA-based epigenetic modification of genes in the developing sperm as well as the embryo (see Text S1). It may also protect these chromosomes from SA-GrBd-effects coded by the other sex chromosome by restricting access of gene products that modify chromatin structure (e.g., acetylation of histones). Early inactivation, however, does not completely preclude X and Y-coded SA-GrBd-effects from occurring. Recent studies indicate that ∼10% of genes on the X remain active throughout spermatogenesis in mice, and that some early inactivated X-linked genes regain activity during the latter stages of spermatogenesis [25]. Lastly, although early inactivation of the sex chromosomes might feasibly have evolved as a defense against SA-zygotic drive, meiotic drive of the sex chromosomes [26] and sex-linked sexually antagonistic alleles [23] would also select for this phenotype.
In sum, there is manifest evidence that sex chromosomes have the potential to evolve to code for SA-GrBd-effects that are mediated by parental epigenetic effects. Although the potential for such effects is greater through the egg in the case of female heterogamety, there is also substantial evidence that epigenetic paternal effects through the sperm also may be an important source of SA-GrBd-effects (Text S1). Antagonistic X and Y-coded SA-GrBd-effects may have been especially prominent during the initial stages of sex chromosome evolution, before early-inactivation of the sex chromosomes during spermatogenesis had evolved.
Parental investment (PI) in offspring can be elicited by specific signals from the offspring, such as vocalizations, begging behavior, or markings such as those associated with the gaping mouth of soliciting offspring [27]. Consider an X-linked mutation that causes a father to i) respond to a daughter-specific trait in a manner that increased PI, or ii) respond to a son-specific trait that in a manner that reduced PI. Such a mutation would be favored by genic selection because it would increase the probability of its own propagation even if the net fitness of the father declined owing to the reduction in the fitness of his sons [2],[3]. The same logic applies to a Y-linked mutation that increased PI allocated to sons at the expense of daughters. The potential for such sex-specific allocation of PI is illustrated by the barn swallow (Hirundo rustica), in which the begging vocalizations are distinct between sons and daughters [28], and the American kestrel (Falco sparverius), in which the male and female nestlings have markedly different plumage [29]. In red deer, females permit sons to suckle longer and more frequently compared to daughters [30], and such sex-specific discrepancies in parental investment are well documented across a wide diversity of taxa [31]. Both solicitation displays by offspring and response to them by parents have been shown to have measurable heritability across a wide diversity of taxa, and solicitation displays are known to be influenced by maternal effects [32]. Collectively these observations indicate that there is substantial evolutionary scope for sex chromosome-coded genes to evolve that cause parents to preferentially invest in one sex of offspring at the expense of the other sex, and hence to code for SA-GrBd-effects.
The logic for sex-linked SA-GrBd-effects that are mediated by competitive sib-sib-interactions is similar to that described above for parental investment (PI). The Y is selected to promote the competitive ability of brothers, the paternal X is selected to promote the competitive ability of sisters, and the maternal X and autosomes are selected to promote the survival of the brood as a whole. In other words, these chromosomes are selected in offspring in the same way that they are selected in their parents. There is a large body of empirical evidence indicating that siblings interact differently with each other in response to the sex of the interacting partners (e.g., [33],[34], so the requisite phenotypic variation is well established for the evolution of sex-linked SA-GrBd-effects that are mediated by competitive sib-sib-interactions. Evidence that SA-GrBd-effects have actually evolved would be established by showing that there are Y-linked genes that cause males to augment the survival of brothers at the expense of sisters, and vice versa for X-linked genes.
To illustrate how easily SA-GrBd-effects could evolve via competitive sib-sib-interactions consider facultative siblicide (i.e., siblings are killed by other siblings in some, but not all, broods), which occurs in many species of birds, and some mammals [33],[35]. If an X-linked gene caused its bearer to be less stimulated to kill a sister compared to a brother (because sister-specific traits were less stimulating in inducing siblicide compared to brother-specific traits), an antagonistic green-beard effect would be manifest. As another example, cannibalism is common in a wide diversity of species during juvenile development [36],[37]. If an X-lined gene caused females to be less likely to cannibalize their sisters and/or more likely to cannibalize their brothers, such a gene would necessarily produce a SA-GrBd-effect. The same logic applies to Y-coded genes that favor brothers over sisters. More generally, any gene located on the sex chromosomes that caused a sibling to be more, or less, stimulated to be aggressive or altruistic in response to sex-specific traits of competing siblings can feasibly lead to a SA-green-beard effect.
The accumulation of X- and Y-coded SA-GrBd-effects will sometimes lead to selection pressure on the autosomes to evolve counter-measures that rescue the affected sex from the antagonistic paternal effects. If an X- or Y-coded paternal effect increases the fitness of one sex of offspring more than it harms the other sex, then the autosomes receive a net benefit and they are not selected to block the antagonistic paternal effect. Selection to block Y- and X-coded antagonistic paternal effects will occur, however, whenever they reduce the average fitness of a brood (across both sexes), and hence reduce the fitness of the autosomes. However, unlike the strong selection on the X and Y to produce, and protect themselves from, sexually antagonistic paternal effects, selection on the autosomes to block them is relatively weak. To illustrate why, consider a new Y-linked mutation coding for a paternal effect that reduces the vigor of daughters and thereby increased the juvenile competitive ability of sons. Let the fitness gain to sons (or the Y) be a positive increment (sson) and the fitness loss to daughters (or the X) be a negative increment (sdaughter). The fitness effect on the autosomes is the average of sson and sdaughter. Since one s-value is positive and the other negative, they tend to be counterbalancing, so that selection on the autosomes to block harmful paternal effects is closer to zero than selection on either the X or the Y to produce them. Hence selection on the autosomes to block antagonistic paternal effects coded by the sex chromosomes is absent, when they increase the average fitness of a brood, or relatively weak, unless they were to lead to a strong, population-wide imbalance in the sex ratio (see [38] for constraints on selection in response to a biased sex ratio). Nonetheless, there is a large number of autosomal loci that may be capable of mutating to modifiers that shut down SA-zygotic drive. As a consequence, more extreme forms of SA-zygotic drive (that reduce net brood fitness) may be eventually silenced by counter-evolution on the autosomes, or to operate episodically when new forms of SA-zygotic drive evolve that are resistant to extant autosomal modifiers (see for example [39] and references in [1], chapter 3). The same logic applies to sex-linked SA-GrBd-effects mediated by PI and competitive sib-sib-interactions.
If a SA-GrBd-effect evolved that was coded by the Y and that favored sons at the expense of daughters, there would be counter-selection on the X to ameliorate this effect, and vice versa if a SA-GrBd-effect evolved that was coded by the X favoring daughters. Such selection and counter-selection could potentially lead to a genetic arms race (Figure 4) with the autosomes being selected to block X- and Y-coded antagonistic paternal effects only when the net fitness of the brood was reduced. Here we explore the fate of mutations located on the X and Y chromosome that code for i) paternal investment (PI) that is skewed toward the sex of offspring that carries them, ii) epigenetic paternal effects that interfere with the ontogeny of the sex of offspring that do not carry them (and thereby reduce their competitive ability during sibling competition), and iii) competitive sib-sib-interactions that reduce the competitive ability of the sex that does not carry them (by helping same sex siblings or harming opposite sex siblings).
We specifically model coevolution between the X and Y. Because the X and Y do not recombine with each other, we model them as alleles at a simple Mendelian locus that determines sex (XY is male, XX is female) and pleiotropically influences epigenetic parental effects, paternal PI, or competitive sib-sib-interactions. This simplification ignores the recombination that is possible between X chromosomes in females that will lead to reduced Hill-Robertson interference on the X compared to the Y. As a consequence, our model will somewhat underestimate the rate of adaptive evolution of the X. Our model also ignores any counter-evolution by the autosomes, but this simplification should not change our qualitative conclusions owing to the expected weaker selection on the autosomes (see above section).
We start by formulating a model of differential paternal investment in sons and daughters which we then study numerically. At the end of this section, we show that a similar approach can be used, and similar conclusions apply, for epigenetic parental effects and competitive sib-sib-interactions that harm the sex that does not carry them.
SA-zygotic drive provides a previously unexplored genetic model for the evolution of traits, such as sterility and homosexuality, which reduce Darwinian fitness, but yet can attain appreciable frequency in natural populations. We illustrate the heuristic potential of the concept of SA-zygotic drive by applying this genetic model to the unusual distribution of female homosexuality in human pedigrees (Figure 6, drawn from the data presented in Table 6 of [41]). We do not claim that this phenotype represents an established example of SA-zygotic drive, only that SA-zygotic drive provides a new functional form of hypothesis that can be tested to account for this – and other enigmatic – phenotypes that presently have no other genetic explanation.
Relative to a proband (i.e., a focal homosexual female), female homosexuality was observed at rates elevated above the background level on the paternal but not the maternal side of the family, and here only among the daughters of the fathers' brothers. A proband's sisters also had elevated rates of homosexuality. There was also some indication that probands' daughters may have had elevated levels of homosexuality, but the number of daughters assayed was small, and their elevated rate of homosexuality was not statistically significant when high stringency in identifying homosexual probands was applied.
The major pattern of female homosexuality in the pedigrees was that its occurrence was elevated only in relatives (sisters and paternal female cousins) whose fathers shared the same Y chromosome, and many of the same X-linked alleles. The observation that paternal aunts did not show elevated rates of homosexuality indicates that it was the X/Y combination of the father, rather than the Y alone, that was associated with an increased probability of female homosexuality. The weaker evidence for elevated rates of homosexuality in probands' daughters is also consistent with an epigenetic effect of the sex chromosomes since paternal epigenetic effects are know to sometimes carry-over to more than one generation (e.g., see description of the Kit-locus in Text S1).
The association of female homosexuality with only the patriline is consistent with the operation of SA-zygotic drive, yet we are aware of no previously available genetic model that predicts this association [42]. Male homosexuality has been found to be associated with the matriline, at least in some ethnic groups (e.g., [43], but see [44]) and more recent evidence indicates that it may be caused, in part, by sexually antagonistic alleles [45]. SA-zygotic drive provides a testable hypothesis for the association of female homosexuality with a different form of genomic conflict: SA-GrBd-effects.
We see no rationale for why the Y would directly be selected to cause female homosexuality. Nonetheless, the Y is selected to epigenetically disrupt daughter-specific developmental pathways that influence their vigor. These effects could feasibly influence female sexual development outside the context of vigor through pleiotropy and lead to female homosexuality, despite there being no direct selection for this specific phenotype.
SA-zygotic drive is also predicted to influence other enigmatic fitness-reducing traits that are controlled by sex-specific processes, like the high levels in humans of both sterility (e.g., ∼10% of couples are infertile, with males accounting for 30–50% of this value [46]) and spontaneous abortion (e.g., ∼70% of human conceptions spontaneously abort, [47], most of which are not due to aneuploidy [48]). The logic in these cases is identical to that described above for female homosexuality, but in this case the disrupted sex-specific developmental pathways lead to sterility and inviability of embryos rather than homosexuality. These examples illustrate how SA-zygotic drive provides a new theoretical framework that can be used to construct a more complete set of alternative genetic hypotheses when evaluating the evolution of traits that reduce Darwinian fitness.
Transmission asymmetries are the biological foundation for many forms of genetic conflict. For example, the mitochondria – and cytoplasmic endosymbionts like Wolbachia – are typically propagated across multiple generations only through the female line of descent (matriline). Transmission of these genomes through sons (patriline) is therefore an evolutionary dead-end, as is transmission through pollen in plants. In response to this transmission asymmetry between sons and daughters, the cytoplasmically transmitted genomes of some species have evolved to kill sons or eliminate pollen production (reviewed in [1],[49],[50]). The killing of male offspring by cytoplasmically transmitted genomes is most strongly favored by natural selection when there is sib-competition because removing sons from a brood increases the availability of resources for their sisters – thereby improving the propagation of the matriline. Here we have shown that the same logic can be extended to the asymmetrical transmission of sex chromosomes to sons and daughters – leading to the hypothesis of SA-zygotic drive.
The operation of SA-zygotic drive via epigenetic parental effects has two prerequisites: i) sibling competition and ii) parental-effect mutations that harm only one sex of offspring. The first prerequisite is well established in a wide diversity of taxa (reviewed in [33]). The second prerequisite is well established in D. melanogaster and birds, at least for the context of maternal effects (e.g., [18],[19], see Text S1), and the recent finding of autosomal zygotic drive in C. elegans [20] (see Text S1) makes it clear that the requisite genetic variation is feasible via paternal effects as well. This extant empirical information, when coupled with our modeling analysis, indicates that SA-zygotic drive via epigenetic parental effects almost certainly occurs in nature, and that antagonistic green-beard effects may be more evolutionarily important than indicated by their rare demonstration in other contexts from past studies (e.g., [4]–[6]). What remains to be established is its evolutionary scope.
SA-zygotic drive via PI and sex-specific competitive sib-sib-interactions is, in principle, simpler to evolve because it does not require trans-generational epigenetic effects. There is clear evidence that sex can strongly influence both PI and competitive sib-sib-interactions (as described in detail above), so the phenotypic traits needed to fuel SA-zygotic drive are clearly in place. Nonetheless, the operation of SA-zygotic drive via sexually antagonistic competitive sib-sib-interactions and PI remains to be explored empirically, and we hope that our study will foster the relevant research.
As described in the introduction, SA-zygotic drive via SA-GrBd-effects is an extension of the logic behind meiotic drive that acts at the diploid zygote and postzygotic stages. The evolutionary scope for SA-zygotic drive may, however, far surpass that of meiotic drive, and also that of autosomal-zygotic drive (e.g., [51]) and gestational drive [7]. In male meiotic drive, selfish elements accumulate because they kill or debilitate competitor sperm that do not carry them. In female meiotic drive, driving elements accumulate when they are less prone to being transported to polar bodies because the cell's molecular motors differentiate between the centromeres of the two homologous chromosomes. Because the dimorphism between sperm carrying different chromosomes (or between the centromeres of homologs in oocytes) is relatively small, there is restricted opportunity for meiotic drive elements to distinguish between them. The small effect that a sperm's haploid genome can have on its structure and function is illustrated by D. melanogaster in which sperm carrying <1% of the genome (only a single “dot” chromosome 4) are fully functional [52]. Similarly, there is relatively little dimorphism between zygotes and embryos that do and do not carry a genetic element, such as a Medea factor [51], that causes autosomal-zygotic drive, or between fetuses expressing different self-recognizing alleles hypothesized to mediate gestational drive [7]. In sharp contrast, there are many sexual dimorphisms (and the ontogenetic pathways that produce them) that distinguish male and female offspring. These numerous dimorphic phenotypes are expected to substantially increase the evolutionary scope for sex-linked, SA-zygotic drive to operate, since any one of them, irrespective of the genes coding for them, represents a phenotypic “tag” for a SA-GrBd-effect. In addition, sex-specific PI and sib-sib interactions, which are well documented in nature (see above), as well as epigenetic modification of any sex-specific phenotype (also well established in nature, see Text S1), can readily produce both “tag differentiation” and “tag-directed aid” whenever these phenotypes are coded by the sex chromosomes and there is competition among siblings. Therefore, SA-zygotic drive has the potential to be a far more pervasive process than meiotic drive, gestational drive, and autosomal-zygotic drive.
The accumulation of Z- and Y-linked mutations that reduce the competitive ability of daughters, or W- and X-linked mutations that reduce the competitive ability of sons, would be expected to create counter-selection on the opposite sex chromosome (and sometimes the autosomes) to rescue the affected sex from harm, and thereby potentially lead to a genetic arms race. If such an arms race occurred, it would contribute to i) rapid genetic divergence between allopatric lineages – thereby potentially contributing to the evolution of postzygotic reproductive isolation during the process of speciation, ii) the decay of the nonrecombining sex chromosome via genetic hitchhiking, and iii) the evolution of elevated levels of sterility, embryo inviability, and homesexuality that exceed what would be expected by mutation-selection balance.
Each time a new SA-GrBd-effect mutation is recruited to the nonrecombining W or Y chromosome, one or more mildly deleterious mutations can accumulate on this chromosome due to genetic hitchhiking (hitchhiking-decay, [53]–[55]). If there is a substantial pool of SA-GrBd-effect mutations that can potentially accumulate on nascent W or Y chromosomes, then coevolution between the W or Y and the rest of the genome could be a powerful process driving their decay.
Antagonistic coevolution between X and Y-coded SA-GrBd-effects, and sometimes including their autosomal suppressors, would be expected to cause otherwise conserved genes to evolve rapidly. The consequent genetic divergence between allopatric populations could be a potent factor leading to Dobzhansky-Muller incompatibilities [56],[57]. In accordance, recent evidence indicates the sex chromosomes are coding hotspots for Dobzhansky-Muller incompatibilities in Drosophila [58]. SA-zygotic drive also provides an unexplored genetic route to the evolution high frequencies of fitness-reducing traits like sterility and homosexuality due to its predicted disruption to sex-specific ontogenetic pathways, as described above.
If SA-zygotic drive can so readily evolve, then why has it not already been widely reported, as has meiotic drive? One explanation is that early inactivation of the sex chromosomes during gametogenesis has largely shut down SA-zygotic drive in most species with ancient X and Y sex chromosomes, which included most multicellular model organisms. However, this same logic would apply to sex-linked meiotic drive, which has been observed in model organisms like Drosophila. Another explanation is that SA-zygotic drive has been misidentified as meiotic drive in non-model organisms that have not been analyzed genetically. A more satisfying explanation, however, is that most SA-zygotic drive may not have the strong effects that would lead to easily noticeable phenotypes, such as strongly distorted brood sex ratios. Antagonistic mutations that code for parental effects that kill offspring that do not carry them (e.g., Medea in Tribolium and peel-1/zeel-1 in C. elegans, which have only recently been discovered) may metaphorically represent the tip of an iceberg of a larger number of potential SA-GrBd-effects that have smaller effects, and therefore would not be detected unless specifically looked for with large sampling effort.
Two lines of evidence suggest that SA-zygotic drive would typically not produce an easily observable lethal phenotype. First, most sperm-mediated trans-generational epigenetic effects (other than methylation-based imprinting, which is not expect to fuel SA-zygotic drive, as described earlier) that have been studied to date do not fully silence their target genes (i.e., they act more like rheostats than on/off switches; reviewed in [59]). Second, even if the target gene of an epigenetic modification were silenced, the vast majority of loss of function mutations are not homozygous-, hemizygous, nor heterozygous-lethal (e.g., as established in Drosophila; [60]–[62]), although it is common for non-lethal mutations in Drosophila (and lethal mutations in the heterozygous state) to reduce the juvenile competitive ability of their carriers (reviewed in [63]). For these reasons, trans-generational epigenetic effects that cause reduced competitive ability (but not unconditional lethality) of the sex of offspring that does not carry them, are expected to play the predominant role in coding for SA-zygotic drive.
Is SA-zygotic drive expected to be a common, but overlooked, phenomenon? We have provided what we think is convincing evidence that SA-zygotic drive, fueled by SA-GrBd-effects, is a plausible evolutionary process because the requisite phenotypes for its operation are known to occur. It is a more difficult matter, however, to predict how commonly this phenomenon is likely to be manifest in nature. Large, easily observed SA-GrBd-effects (that harm one sex more than they help the other, e.g., son- or daughter-killers), would select for suppressors on the autosomes. In this case, the numerical excess of autosomal compared to sex-linked genes should lead to autosomal silencing of this form of SA-zygotic drive, or at least make it episodic. However, less extreme forms of SA-zygotic drive are not predicted to be opposed by the autosomes, as described above, so this – more difficult to discern – form of SA-zygotic drive is predicted to be most common. In this case the prevalence of SA-zygotic drive will depend only on the mutation rate to alleles coding for small SA-GrBd-effects – a parameter that is presently unknown.
Throughout this manuscript we have emphasized harm, rather than altruism, as the phenotype mediating SA-zygotic drive. We have done this because we have assumed that there is competition among siblings for limiting resources. In this case, any phenotype that aids one sex of offspring in a family will make this sex more competitive, and thereby harm the opposite sex. Thus, helping one sex in a brood will necessarily harm the other sex. We also have focused predominantly on the sex chromosomes themselves. However, in some cases the mitochondria and other cytoplasmic genomes will co-segregate with a sex chromosome (e.g., the W sex chromosome in species with female heterogamety co-segregates with all cytoplasmically transmitted genomes). In this case, SA-zygotic drive also may be influenced by phenotypes coded by the cytoplasmic genomes that co-segregating with the sex chromosomes.
Our theory of SA-zygotic drive can be used to generate testable predictions. The major – and counterintuitive – prediction concerning SA-zygotic drive is that a father's Y chromosome will be observed to sometimes strongly influence the fitness of his daughters and his X will similarly influence his sons. In the case of female heterogamety, analogous predictions apply to the W and Z chromosomes. The empirical work described above (e.g., the Kittm1alf mutation in mice [21] and the sex-specific maternal effect mutations in D. melanogaster [18] and birds e.g., [19] proves that these types of effects can feasibly evolve (see Text S1). It has also been established in inbred strains of mice that a father's Y chromosome can influence the behavior [64] and immune function [65] of his daughters. Our theoretical study provides a motivation for researchers to screen in future studies for an influence of the X and Y (and W and Z) on the sex of offspring that does not carry them.
A second prediction is that heritable paternal effects on offspring fitness should be found to be more common, and larger in magnitude, in species with male heterogamety, and within this group this pattern should be strengthened as the degree of monandry and sib-sib interactions increase. The absence of strong paternal effects in species lacking male parental care is commonly assumed in studies of quantitative genetics. Our theoretical work, however, predicts that this assumption will sometimes be violated due to polymorphism (sex linked or autosomal) influencing the expression of paternal SA-GrBd-effects.
A taxonomic prediction is that SA-zygotic drive should be especially prevalent in birds. This taxon has unusually high levels of monogamy (within a breeding season and despite low levels of extra-pair fertilizations, [66], an absence of inactivation of the W and Z sex chromosomes during oogenesis [23], and high levels of parental care and sib-sib interactions. Birds also have female heterogamety which facilitates parental epigenetic effects through the mother's large contribution to the embryo of RNAs and steroid hormones. The combination of these characteristics makes birds an ideal taxon to test for the existence of SA-zygotic drive.
The main prediction concerning competitive sib-sib-interactions is that, in species with sex chromosomes, same-sex sibling interactions should be more altruistic and less aggressive compared to between-sex interactions (excluding species with other factors magnifying same-sex sib competition, such as those with local mate competition or early dispersion of only one sex of offspring). A similar prediction has been made earlier by several other researchers (reviewed in [67]) based on the idea that X and Z sex chromosomes segregate the same way that haploid genomes do in species with haplodiploid sex determination. In haplodipoid species, full sisters are more closely related to each other (R = proportion of shared polymorphic alleles = ¾) than to brothers (R = 1/4), and more closely related than bothers are to each other (R = ½). As a consequence, sister-sister interactions are predicted to be the more cooperative. Assuming that the heteromorphic sex chromosome (Y or W) is too degenerate to code substantially for cooperation, the X and Z have the same relationship in brothers and sisters as whole genomes do in haplodiploids, and hence X and Z-linked genes are predicted to evolve to make members of the homogametic sex to be more cooperative with each other. There is some support for this prediction based on taxonomic comparisons. For example, long-term cooperative groups are more common among brothers in birds and sisters in mammals [68]. However, we have found no relevant information (pro or con) in the literature concerning the more specific prediction of SA-zygotic drive that during sib-competition opposite-sex individuals will be more competitive with each other compared to same-sex individuals. We suspect, however, that this information may have been collected incidentally in many studies of animal behavior – but unreported. Our study should provide an impetus to publish such comparisons.
The main prediction concerning PI is that, all else being equal, asymmetry in its allocation to sons and daughters should be higher, and sometimes more variable, in the heterogametic compared to the homogametic parent. The ‘all else being equal’ qualifier is important here because in taxa like birds males may vary in PI more than females owing to varying uncertainty in paternity. We have been unable to find any studies reporting this metric (so we have found neither positive nor negative evidence), but again we suspect that it may have been collected incidentally but unreported in past studies of animal behavior.
Lastly, when there is sib-competition, sexual dimorphism of offspring is predicted to be reduced in species with sex chromosomes, and within this group, lower yet when there is PI from the heterogametic parent. To illustrate the rationale for this prediction, suppose that an X-coded paternal effect evolved that caused fathers to increased PI in response to a daughter-specific trait, or reduce PI in response to a son-specific trait. Sons would be selected to converge in phenotype with their sisters, leading to the evolution of reduced sexual dimorphism during the period of sib-competition.
Nonrecombining sex chromosomes create an unappreciated opportunity for the evolution of zygotic drive via sexually antagonistic green-beard effects whenever there is competition among siblings. The evolutionary scope for SA-zygotic drive is predicted to exceed that of meiotic, gestational, and autosomal-zygotic drive because all sexually dimorphic traits can acts as “tags” for sexually antagonistic green-beard effects. These sexually antagonistic phenotypes can, in principle, lead to an arms race between the two types of sex chromosomes (sometimes also including the autosomes, which can slow, and temporarily or permanently halt, the process) that can i) accelerate the degeneration of the heteromorphic sex chromosome, ii) cause genes that would otherwise be highly conserved to diverge among allopatric lineages and thereby leading to the evolution of Dobzhansky-Muller incompatibilities during speciation, and iii) lead to the disruption of sex-specific ontogenetic pathways that can lead to increased levels of expression of traits, like homosexuality and sterility, that lower Darwinian fitness. We need to stress in closing, however, that we have only established the potential for SA-zygotic drive to operate in nature and it will remain a feasible but unproven possibility until suitable empirical testing has been undertaken.
For simplicity we assume that each mating results in two offspring. Let the parameter bson characterize the bias in paternal investment toward sons in families with one daughter and one son, with bson = 0, bson>0, and bson<0 implying equal investment in both offspring, higher investment in the son and higher investment in the daughter, respectively. Let x and y be the (additive) effects of X- and Y-linked genes in the father on the bias of his paternal investment. More specifically, we let bson = y−x and bdaughter = x−y = −bson, so that X-linked genes favored by selection (that increase x) cause the father to invest more in his daughter while Y-linked genes favored by selection (that increase y) cause him to invest more in his son. We assume that the fitness of a brother and a sister in a brother-sister brood are w(bson) and w(bdaughter) = w(−bson), respectively, where w(.) is a symmetric function changing from 0 to 1 as bson changes from −∞ to +∞ with w(0) = 0.5 and w(bson)+w(bdaughter) = 1 (see below). Interpreting fitness as the amount of a resource available, the latter two equalities imply that the overall amount of resource is fixed (at 1) and that with no bias (i.e. if bson = bdaughter = 0), both sex of offspring get an equal share (equal to 0.5). The symmetry of this relationship is motivated by the idea that an extra unit of PI given to one sex of offspring is taken away from the other sex of offspring, and this implicitly assumes that the benefit of an extra unit of PI is equal to the cost of losing a unit of PI. Finally, we assume that fitness of each offspring in the families with the same-sex of offspring is equal to 0.5. Under these conditions, the average fitness of sons and daughters of fathers with effects (x, y) are(1)The average fitness of sons and daughters given by eq. 1 are both limited to the interval [0.25, 0.75].
The evolutionary dynamics in this model were analyzed by using stochastic, individual-based simulations that allowed for the effects of random genetic drift, mutation, and selection. Generations were discrete and non-overlapping and the population size was fixed at N males and N females. Individuals entered the mating pool with probabilities proportional to wson and wdaughter for males and females, respectively, and mating was random within the mating pool. The number of matings (and families produced) per individual of each sex was a binomial random variable. Mutation occurred in both parents with probability μ per chromosome per generation and changed effects x or y by a random value taken from a normal distribution with a mean of zero and a standard deviation of one. The fitness function w(.) was specified as:
where α>0 is a parameter measuring the strength of selection (larger values of α imply stronger selection; see Figure 7). We assumed that initially there was no genetic variation and the x and y effects of all individuals were set to zero. We varied the mutation rate μ and the strength of selection α while the number of individuals of each sex was always set at N = 1000. For each parameter combination, we did 20 runs each for 10000 generations. Overall, the dynamics are expected to be very similar to those observed in models of sexual conflict over mating rate [69]–[71].
|
10.1371/journal.pgen.1000461 | Genetic Evidence That the Non-Homologous End-Joining Repair Pathway Is Involved in LINE Retrotransposition | Long interspersed elements (LINEs) are transposable elements that proliferate within eukaryotic genomes, having a large impact on eukaryotic genome evolution. LINEs mobilize via a process called retrotransposition. Although the role of the LINE-encoded protein(s) in retrotransposition has been extensively investigated, the participation of host-encoded factors in retrotransposition remains unclear. To address this issue, we examined retrotransposition frequencies of two structurally different LINEs—zebrafish ZfL2-2 and human L1—in knockout chicken DT40 cell lines deficient in genes involved in the non-homologous end-joining (NHEJ) repair of DNA and in human HeLa cells treated with a drug that inhibits NHEJ. Deficiencies of NHEJ proteins decreased retrotransposition frequencies of both LINEs in these cells, suggesting that NHEJ is involved in LINE retrotransposition. More precise characterization of ZfL2-2 insertions in DT40 cells permitted us to consider the possibility of dual roles for NHEJ in LINE retrotransposition, namely to ensure efficient integration of LINEs and to restrict their full-length formation.
| Long interspersed elements (LINEs) are transposable elements that mobilize and amplify their own copies within eukaryotic genomes. Although LINEs had been considered as “junk” DNA, recent studies have suggested that the LINE-induced alterations of host chromosomes are a major driving force for eukaryotic genome evolution. LINEs mobilize via a mechanism called retrotransposition, in which transcribed LINE RNA is reverse transcribed into DNA that is then integrated into the host chromosome. Although the role of LINE-encoded proteins in retrotransposition has been revealed, the participation of host-encoded proteins has not been well investigated. Here, using knockout chicken DT40 cell lines, we present genetic evidence that the host-encoded proteins involved in repair of DNA double-strand breaks participate in LINE retrotransposition. More precise characterization of LINE insertions in DT40 cells suggested dual roles for these host DNA repair proteins in LINE retrotransposition; one function is required for efficient integration of LINEs and the other restricts their full-length formation.
| Long interspersed elements (LINEs) and short interspersed elements (SINEs) are transposable elements widely distributed in eukaryotic genomes [1],[2]; as such, they substantially affect genome complexity and evolution [3],[4]. These elements mobilize and amplify their own sequences by a mechanism called retrotransposition. LINEs are 4–7 kbp in length and typically encode two open reading frames (ORFs), ORF1 and ORF2, both of which are essential for LINE retrotransposition [5],[6]. During retrotransposition, LINEs are first transcribed into messenger RNA (mRNA) from which the LINE-encoded proteins are translated (Figure S1A). Next, the LINE mRNA and proteins form a complex [7],[8] and move to target sites on a host chromosome where the LINE-encoded endonuclease (EN) nicks a strand on the DNA duplex. The LINE-encoded reverse transcriptase (RT) then reverse transcribes the LINE mRNA using the 3′ hydroxyl group generated by the nick as a primer; this reaction is called target-primed reverse transcription [5],[9],[10]. Thereafter, the newly synthesized LINE is integrated into the host chromosome, at which time sequence alterations are generated at the target site. The position of the second strand cleavage is considered to define which kind of target site alteration is generated (Figure S1B) [11]. In the model, second-strand cleavage downstream of the initial first-strand nick generates target site duplication (TSD), cleavage at the same site generates blunt end joining (BEJ), and cleavage upstream generates target site truncation (TST). However, the precise mechanism of the integration remains unclear (Figure S1A). A DNA double-strand break (DSB) would necessarily need to be generated at the target site to integrate the newly synthesized LINE element. In fact, overexpression of human LINE L1 in mammalian cultured cells induces DSBs in the host chromosomal DNA [12]. Accumulating evidence has revealed that several host-encoded DNA repair proteins are involved in the mobility reactions of retrotransposons, such as yeast LTR retrotransposons and bacterial group II introns [for review, 13]. However, the roles of host factors in LINE retrotransposition remain unclear. Only a few genetic studies have identified host proteins that are involved in LINE retrotransposition: the ataxia-telangiectasia mutated (ATM) protein—a protein kinase involved in cellular responses to DSBs—is suggested to participate in L1 retrotransposition [12], and the ERCC1/XPF endonuclease—which functions in nucleotide excision repair—is involved in limiting L1 retrotransposition [14]. It is conceivable that the LINE retrotransposition reactions involve other host factors, such as proteins of the non-homologous end-joining (NHEJ) pathway, that predominate in DSB repair in vertebrate cells [15].
The core components involved in vertebrate NHEJ are the Ku70 and Ku80 heterodimer (Ku70/80), the catalytic subunit of DNA protein kinase (DNA-PKcs), DNA ligase IV (LigIV), and Xrcc4. Initially, Ku70/80 binds to the broken DNA ends. DNA-PKcs is recruited to the ends by Ku70/80, with which it maintains the broken ends in proximity and provides a platform for the recruitment of other enzymes [16]. The kinase activity of DNA-PKcs—which is activated upon recruitment to the broken ends—is considered to enhance the DSB signal via phosphorylation of many downstream targets, although physiological targets of the phosphorylation remain obscure [for example, 17]. LigIV, which forms a tight complex with Xrcc4, is responsible for ligation of the broken DNA ends [18],[19]. There are other proteins implicated in NHEJ. During NHEJ, a pair of broken ends that are incompatible for ligation is processed into compatible ends by a nuclease(s), such as Artemis, and/or a polymerase(s), although their significance in NHEJ is less clear. Interestingly, broken ends are still repaired by NHEJ in cells deficient in the core NHEJ components such as Ku proteins or LigIV, suggesting that NHEJ can be achieved by at least two distinct pathways [20],[21]. To distinguish these two processes, the NHEJ pathway that depends on the core components is denoted ‘classical’, and the other pathway, which can occur without the core components, is called ‘alternative’. In contrast to the classical NHEJ, the enzymes responsible for the alternative NHEJ remain uncharacterized.
Recently, we established a new system to detect LINE retrotransposition in the chicken B lymphocyte cell line, DT40, using two different kinds of LINEs, zebrafish ZfL2-2 and human L1 [22] (Figure S2). Here, we applied this system to Ku70−/−, Artemis−/−, and LigIV−/− DT40 cell lines to determine the effect of these knockouts on the retrotransposition frequencies (RFs) of ZfL2-2 and L1. We then characterized ZfL2-2 insertions retrotransposed in the chromosomal DNA of DT40 cells to obtain evidence for the involvement of NHEJ factors in the LINE integration reaction. In addition, we examined the possible involvement of DNA-PKcs in LINE integration in human HeLa cells using NU7026, an inhibitor of DNA-PKcs activity.
To investigate whether host factors participating in NHEJ are involved in LINE retrotransposition, we examined RFs of two types of LINEs that have different structural characteristics—zebrafish ZfL2-2 and human L1—using wild-type (WT) and five knockout DT40 cell lines. The knockout DT40 cell lines were deficient in the genes encoding Ku70, Artemis, LigIV, SHIP1, or Rad18; the first three cell lines are related to NHEJ, and the others are not (Figure 1, Tables S1, S2). Because the intrinsic colony-forming capacities varied among these cell lines, we compensated for this aspect by including the plating efficiency in the RF calculation (see Materials and Methods). The RF of ZfL2-2 decreased by about 2- to 8-fold relative to the WT DT40 in all NHEJ-deficient cell lines examined here (Figure 1B; Ku70−/−, Art−/− and LigIV−/−). On the other hand, knockout of the Rad18 or SHIP1 gene, neither of which is related to the NHEJ pathway, did not affect the RF (Figure 1B; Rad18−/−and SHIP1−/−). These results suggest that the NHEJ pathway plays a role in ZfL2-2 retrotransposition in these chicken cells. Similar retrotransposition results were obtained using L1, although the decrease in the L1 RF in Ku70−/− and LigIV−/− was smaller than that for ZfL2-2 (Figure 1C; see also Table S1, S2).
To confirm that the RF decrease in the Ku70-defective cells was caused by Ku70 disruption, ZfL2-2 retrotransposition was assessed in three DT40 cell lines, WT, Ku70−/− and LigIV−/−, with transient expression of a cloned chicken Ku70 gene (Figure 1D, Table S3). Transcription of the cloned and/or endogenous Ku70 genes in each sample was verified by RT-PCR. Expression of the cloned Ku70 in WT and LigIV−/− cells did not significantly alter the ZfL2-2 RF. In contrast, exogenous Ku70 expression in Ku70−/− cells dramatically increased the ZfL2-2 RF to a level comparable to WT cells. These results indicate that the decrease of ZfL2-2 RF in Ku70−/− cells was indeed caused by Ku70 disruption.
The NHEJ-defective DT40 cell lines are sensitive to intense ionizing radiation [23],[24], indicating that the cells cannot efficiently repair radiation-induced DSBs, causing cell death. If the expression of ZfL2-2 or L1 in DT40 cells induces DSBs in chromosomal DNA as in the case of the L1 expression in HeLa cells [12], the NHEJ-defective DT40 cells may be more sensitive to such LINE-induced DSBs than WT cells. If this is the case, it is possible that the decrease of ZfL2-2 and L1 RF observed in the NHEJ-defective DT40 cells only reflects cell death caused by the LINE-induced DSBs, which cannot be compensated for by the plating efficiency in our assay (see Figure S3). To examine this possibility, we monitored the viability of WT and mutant DT40 cells transfected with the LINE expression vector. As shown in Figure S4, when two different fluorescence protein expression vectors (enhanced green fluorescence protein (EGFP) and DsRed-Express) were mixed and co-electroporated into DT40 cells, most transfected (fluorescence-positive) cells (>80%) express both of the two fluorescent proteins, and the amounts of proteins expressed from the co-transfected vectors were roughly proportional to each other (Figure S4D and S4E). Hence, to trace the LINE-expressing cells, an EGFP expression vector was electroporated together with the LINE expression vector into DT40 cells, and the EGFP expression and its intensity were monitored as shown in Figures S5, S6, S7, S8, S9, S10, S11, and S12. This EGFP monitoring was conducted from 3 to 8 days after electroporation, during which cell division occurred at least eight times (data not shown). EGFP expression observed on the eighth day was minimal, showing the detection limit. As shown in Figure 2, the relative ratio of the amount of EGFP-expressing cells at each time point to that of the third day was similar between WT and Ku70−/− DT40 cells up to the end of the monitoring. In addition, the time course of the relative ratio of the geometric mean and the median of the EGFP intensity was similar between the WT and Ku70−/− DT40 cells (Figure 2). Moreover, the time course of the values did not change when a point mutation that abolishes LINE EN activity was introduced in the ZfL2-2 and L1 elements. Similar results were obtained from Artemis−/− and LigIV−/− cell lines (Figure S13, S14). These results indicate that LINE EN expression does not influence the viability of WT and NHEJ-defective cell lines. Thus, the decrease in LINE RF in the NHEJ-defective cell lines is likely to be related to the involvement of NHEJ in LINE retrotransposition in DT40 cells.
LINE retrotransposition typically depends on the LINE's own EN [5],[25],[26]. In contrast, a previous study reported that a fraction of human L1 retrotransposition in Chinese hamster ovary (CHO) cells was not dependent on L1 EN [27],[28]. This L1 EN-independent retrotransposition was enhanced in mutant CHO cells defective in a gene involved in NHEJ [27]. This report prompted us to consider that ZfL2-2 and L1 might atypically retrotranspose in NHEJ-defective DT40 cells through an EN-independent manner. To examine this possibility, we examined RFs of the EN-defective ZfL2-2 and L1 elements in the NHEJ-defective DT40 cell lines (Table S1, S2). In both NHEJ-defective and WT cell lines, no G418-resistant colonies were formed with these EN mutants, indicating that mobilization of ZfL2-2 and L1 in the DT40 cell lines examined here is dependent on their own ENs. Although we have not resolved the reason why the dependence of LINE retrotransposition on EN differs in chicken and hamster cells, this may reflect the differences in DNA repair that exist between these cells as discussed by Morrish et al. [27].
To determine in which step of the retrotransposition reaction each NHEJ factor is involved, we determined and analyzed the 5′ and 3′ junction sequences of 102 ZfL2-2 inserts in chromosomal DNA of WT, Ku70−/−, Artemis−/− and LigIV−/− DT40 cells (26, 25, 24 and 27 insertions, respectively; Table S4). We previously showed that ∼40% of ZfL2-2 elements in the zebrafish genome had extra nucleotides at the 5′ junction, whereas ∼50% had microhomologies [29]. At the 3′ junction, on the other hand, ∼80% of these elements had microhomologies [29]. Similar tendencies were observed at both junctions of ZfL2-2 insertions in DT40 cells, and these tendencies were not altered by NHEJ defects (Table S5). Also, the length distribution of the 5′ and 3′ microhomologies did not differ between the WT and NHEJ-deficient DT40 cells (Figure S15). However, the ZfL2-2 insertions in Ku70−/− and Artemis−/− cells were significantly longer than those in WT cells (Figure 3A and 3B; P = 0.008 and 0.036, respectively). In particular, full-length elements were recovered only from NHEJ-deficient cells (Figure 3A, 3C). Indeed, the fraction of full-length insertions differed significantly between WT and Ku70−/− cells and between WT and Artemis−/− cells (P = 0.010 and 0.046, respectively). These results indicate that Ku70 and Artemis inhibit the generation of longer inserts in DT40 cells, suggesting that these NHEJ factors, at least in part, participate in LINE 5′ truncation.
We classified the target site alterations of the ZfL2-2 insertions in DT40 cells into five categories: long TST (L-TST, >20 bp), short TST (S-TST, ≤20 bp), BEJ, short TSD (S-TSD, ≤20 bp) and long TSD (L-TSD, >20 bp) (Figure 3D, Table S5). Consistent with our previous data regarding ZfL2-2 elements present in the zebrafish genome [29], a large fraction of ZfL2-2 insertions (20 of 25, 80%) in WT DT40 had a TSD, and the rest of them had a TST (Figure 3D). Most insertions (24 of 25) in the WT cells had short target site alterations (≤20 bp), and only one had a L-TSD (1228 bp), indicating that long target site alterations are relatively rare in WT cells. On the other hand, L-TST (343–50187 bp) insertions were frequently observed in Ku70−/− cells (Figure 3D; 5 of 25), suggesting that Ku70 prevents the generation of L-TST.
We next focused on insertions with short target site alterations (Figure 3E). Insertions with S-TSD predominated in all cell lines. Still, 5 of 25 insertions (20%) in WT cells and 3 of 20 insertions (15%) in Ku70−/− cells had an S-TST. In contrast, only one of 23 insertions (4%) in Artemis−/− cells had an S-TST, and no S-TSTs were observed in LigIV−/− cells. The difference in occurrence of S-TSTs between the WT and LigIV−/− cells is statistically significant (Figure 3E; P = 0.016). These results indicate that LigIV (and possibly Artemis) plays an important role in generating S-TSTs.
To examine whether NHEJ is also involved in LINE retrotransposition in cells other than chicken DT40, we performed the retrotransposition assay in human HeLa cells (Figure 4). No knockout HeLa cell line is available, but DNA-PKcs kinase activity can be specifically inhibited by NU7026 [30]. We first confirmed that NU7026 kills HeLa cells in a dose-dependent manner only in the presence of a DSB inducer, etoposide [31]. HeLa cells treated with NU7026 became more sensitive to etoposide, indicating that the NHEJ repair capacity is suppressed by NU7026 (Figure 4A). The RFs of both ZfL2-2 and L1 decreased with increasing concentrations of NU7026, suggesting that the NHEJ pathway is also involved in LINE retrotransposition in HeLa cells (Figure 4B, Table S6, S7). Consistent with the results using DT40 cells, ZfL2-2 retrotransposition was more sensitive than L1 retrotransposition to NU7026.
Host repair systems are likely to be involved in the later stages of LINE retrotransposition [12],[13],[29],[32],[33]. For example, bioinformatic studies have suggested that the ‘alternative’ NHEJ pathway is involved in LINE retrotransposition [33] (see below). There is, however, no bioinformatic evidence for such involvement of the ‘classical’ NHEJ or experimental evidence for a role in LINE retrotransposition of any host repair system except for the ATM kinase [12]. Here, we studied the effects of defects in ‘classical’ NHEJ on ZfL2-2 retrotransposition and found that such defects considerably decrease the ZfL2-2 RF, suggesting that a large fraction of ZfL2-2 insertion events in DT40 cells utilizes these classical NHEJ factors. In addition, the characterization of ZfL2-2 insertions revealed that disruption of the genes encoding NHEJ components extended the length of inserted ZfL2-2 elements, allowing more full-length insertions (Ku70−/− and Artemis−/−); frequently generated L-TSTs (Ku70−/−); and diminished the generation of S-TSTs (LigIV−/−). These results suggest that NHEJ proteins are involved in the 5′ joining of ZfL2-2 insertions during retrotransposition, as detailed below.
During retrotransposition (Figure S1A), the ZfL2-2 RNA-protein complex chooses a target site, at which the ZfL2-2 EN nicks the first strand of the host DNA. The ZfL2-2 RT then initiates reverse transcription of the ZfL2-2 RNA from the nick. Most ZfL2-2 elements in DT40 cells as well as those in the zebrafish genome have a certain length of truncation at the 5′ end (5′ truncation), which is a characteristic of a typical LINE element. The mechanism by which the 5′ truncation is generated is, however, unclear. Our data provide a possible mechanism for the 5′ truncation. The Ku70 defect produced longer insertions (Figure 3A, 3B), implying that the Ku70/80 complex can obstruct the progression of the ZfL2-2 RT. For instance, transient dissociation of the RT from the template RNA could allow Ku70/80 to associate with the end of the newly synthesized ZfL2-2 DNA (Figure 5) because Ku70/80 is able to interact with a single-to-double-strand transition of DNA [34]. The Ku70/80 association may interfere with further reverse transcription and initiate a joining reaction between the premature ZfL2-2 cDNA and upstream target DNA, resulting in a 5′ truncation. Because deficiencies of Artemis and LigIV—which act downstream of Ku70 in NHEJ—also caused longer insertions (Figure 3A, 3B), the progression of the NHEJ pathway might be related to the switching of reaction modes from reverse transcription to 5′ joining.
Ku70/80 protects DNA ends from exonucleolytic degradation [35]. Consistently, Ku70−/− cells frequently produced ZfL2-2 insertions containing long chromosomal DNA deletions (Figure 3D; L-TST, 343–50187 bp). This suggests that Ku70/80 is associated with the end of the upstream target DNA as well as the end of the ZfL2-2 element during integration, and protects the chromosomal DNA from degradation (Figure 5). In the case of TST generation, genomic information is altered not only by inserting the ZfL2-2 sequence but also by deleting the pre-existing sequence. Thus, Ku70/80 may also serve as a barrier against the loss of genomic information caused by ZfL2-2 retrotransposition.
The variability of target site alteration has been accounted for by the difference in the position of the second strand cleavage [11] (Figure S1B); however, it remains unclear what other factor or factors are involved in this variation. We found that LigIV−/− cells did not produce S-TSTs, whereas ∼20% of insertions in WT cells had an S-TST (P = 0.016), indicating that S-TST generation is dependent on LigIV activity. Moreover, the S-TST frequency was also decreased in Artemis−/− cells. Therefore, the 5′ overhang (generated by the second-strand cleavage upstream of the first nick; see Figure S1B) at the chromosomal end may be processed predominantly by Artemis and then ligated to the ZfL2-2 5′ end by LigIV (Figure 5). Hence, our results suggest that these NHEJ factors contribute to variation among target site alterations.
Taken together, our data suggest the possibility that NHEJ proteins, originally recruited for the repair of chromosomal breaks generated by the ZfL2-2 EN, are necessarily utilized for ZfL2-2 integration. Deficiencies of NHEJ proteins remarkably decreased the ZfL2-2 RF, indicating that NHEJ proteins are required for efficient retrotransposition. On the other hand, these NHEJ factors restricted the generation of full-length ZfL2-2 copies by diverting initiated retrotransposition reactions toward the generation of truncated ZfL2-2 copies. The restriction of full-length copies that have the potential to undergo subsequent retrotransposition limits the amplification of ZfL2-2 copies in the next generation. Interestingly, Deininger's group showed that many more DSBs than retrotransposition events are generated by L1 EN expression in HeLa cells [12], suggesting that a considerable fraction of L1-induced DSBs are repaired without L1 insertions. Because DSBs are predominantly repaired by NHEJ in vertebrate cells, it is plausible that these L1-induced DSBs are fixed by NHEJ. Thus, the NHEJ pathway probably limits retrotransposition at two different phases: 1) inhibition of LINE cDNA integration itself by immediate repair of DSBs, resulting in direct limitation of retrotransposition and 2) production of truncated insertions, resulting in limitation of retrotransposition in the next generation. Because more active L1 elements produce longer insertions [36], the rapid or efficient progression of reverse transcription may counteract both of these NHEJ limitations. Thus, rapid cDNA synthesis prior to the operation of the NHEJ pathway may be vital for successful LINE amplification. Taken together, these observations indicate that there is opposition between DNA repair and LINE retrotransposition. Similarly, a retrotransposon conflict hypothesis has been proposed by Sawyer and Malik, in which NHEJ proteins are proposed to be hijacked for mobilization of Ty LTR retrotransposons or recruited to defend against them [37].
Our data show that disruption of the NHEJ pathway in DT40 cells did not completely suppress ZfL2-2 retrotransposition (Figure 1). Therefore, a pathway(s) other than classical NHEJ may exist to connect the ZfL2-2 integrants and the end of the target DNA at the 5′ junction. As proposed by Zingler et al. [33], one possibility is the ‘alternative’ NHEJ pathway, which joins two DNA ends via microhomology in a manner independent of NHEJ factors such as Ku70/80 and LigIV. However, the fraction of insertions having 5′ microhomology was not elevated in Ku70−/− or LigIV−/− cells (Table S5), and thus this putative mechanism cannot fully account for the observed residual Ku70- or LigIV-independent retrotransposition activity. Rather, the template-jump model [33], [38]–[43] seems more likely for the 5′ joining process, although the DNA ligase responsible remains unidentified. Indeed, a large proportion of extra nucleotides found at the 5′ junction of ZfL2-2 insertions in DT40 cells appears to be synthesized by the template jump reaction (unpublished data).
Cell death caused by LINE EN expression was not detected in chicken DT40 cells in our experimental system (Figure 2, S13 and S14), although it causes considerable cell death in human HeLa cells [12] (data not shown). Neither DT40 nor HeLa cells have detectable levels of p53, a tumor suppressor that induces apoptosis or cell cycle arrest against DNA damage [44],[45]. Thus, although a lack of p53 appears to confer tolerance to chromosomal instability in DT40 cells [44], it cannot explain the observed differential sensitivity to LINE EN between DT40 and HeLa cells. We thus speculate that this differential sensitivity may reflect the presence of an intrinsic LINE retrotransposition mechanism in each cell line. The fact that LINE retrotransposition in chicken DT40 and hamster CHO cells is differentially dependent on EN supports this idea. Hence, comparative analysis of the mechanism of LINE retrotransposition in different cell lines and organisms is indispensable for understanding the generality and specificity of the LINE amplification mechanism.
Deficiencies of NHEJ proteins in DT40 and HeLa cells also decreased the L1 RF, suggesting that NHEJ factors participate in retrotransposition of human L1 as well as zebrafish ZfL2-2 in these vertebrate cells. The degree of decrease in the L1 RF was, however, smaller than that for ZfL2-2 (Figure 1, 4), indicating that NHEJ is not much involved in retrotransposition of human L1. This implies that each LINE has its own dependency on NHEJ and probably other repair system(s); in other words, that the mechanism of LINE retrotransposition is considerably distinct between each LINE in the light of the participation of host repair systems. A major structural difference between these LINEs is the absence (ZfL2-2) or presence (L1) of ORF1p. Because L1 ORF1p has been suggested to be involved in the 5′ joining [46], ORF1p might make L1 more independent of the host NHEJ system. Our study indicates that the factors of classical NHEJ are involved in the repair of breaks generated by LINEs during retrotransposition. Our results also indicate that the NHEJ pathway is not the only mechanism by which such breaks can be repaired. Elucidation of the entire ensemble of host factors involved in LINE mobilization will help us understand the interaction between hosts and molecular parasites during evolution.
pBZ2-5 expresses the WT zebrafish LINE ZfL2-2 containing the neomycin resistance gene that is disrupted by an intron in the antisense orientation (mneoI) [25]. p131.11 expresses the mneoI-marked ZfL2-2 element containing a point mutation (E72A) in the EN sequence [22]. pAZ2-2, which expresses the WT ZfL2-2 element marked by mneoI400/ColE1 [11], was constructed as follows. The mneoI400/ColE1 cassette was amplified from pCEP4/L1.3mneoI400/ColE1 [11] by PCR using primers Neo-NotF-1 (5′-TGTGTGTGGCGGCCGCGCACAAACGACCCAACACCC-3′) and Neo-BamR-1 (5′-CACACGGATCCGCTGCAGCATAGCCTCAGG-3′). The PCR fragment of mneoI400/ColE1 was digested with NotI and BamHI. Using the mneoI400/ColE1 fragment, the NotI and BamHI fragment of pBB4 [25], which contains mneoI, was replaced, resulting in pBB5-9. The 3′ tail of ZfL2-2 was amplified from pBZ2-5 by PCR using primers Z2-3′F1 (5′-ATATGGATCCTGAAACTTGCCTTTAGTACTTATTCATTGTTGC-3′) and Z2-3′R1 (5′-ATATGGATCCTTTACATTTACATTTACATTTAGTCATTTAGCAGACGC-3′). The PCR fragment of the 3′ tail was digested with BamHI and inserted in the BamHI site of pBB5-9, resulting in pAZ2-2. pJM102/L1.3 expresses the WT L1 (L1.3) containing the marker mneoI [27]. pJM102/L1.3 H230A expresses the mneoI-marked L1 (L1.3) containing a point mutation (H230A) in the EN sequence [27]. pEGFP-FLAG-1 expresses an enhanced green fluorescence protein (EGFP) [22]. Chicken Ku70/pAneo was constructed by cloning the chicken Ku70 gene into the expression vector pAneo [23]. The expression vectors were all purified using the QIAfilter Plasmid Midi or Mega kit (Qiagen).
WT DT40 and its SHIP1−/− and IP3R−/− derivatives were purchased from RIKEN Bioresource Center (cell numbers RCB1464, 1465, and 1467). Ku70−/−, DNA ligase IV−/− and Rad18−/− DT40 cell lines were established previously [23],[24],[47]. The Artemis−/− DT40 cell line was kindly provided by Dr. Minoru Takata [48]. These DT40 cells were cultured in RPMI medium 1640 (Invitrogen) supplemented with 10% fetal bovine serum, 1% chicken serum, 20 U/ml penicillin, 20 µg/ml streptomycin, and 10 µM β-mercaptoethanol, in a humidified atmosphere with 5% CO2 at 37°C or at the temperatures indicated.
The retrotransposition assay procedure in DT40 cells has been described [22]. Briefly, DT40 cells were cotransfected with pEGFP-FLAG-1 (15 µg) and one of the LINE expression vectors (15 µg), pBZ2-2, p131.11, pJM102/L1.3, or pJM102/L1.3 H230A [22],[27]. Transfection was carried out by electroporation at 250 V and 960 µF for ZfL2-2 expression vectors, and at 200 V and 960 µF for L1 expression vectors using the GENE Pulser (Bio-Rad). After the transfected cells were incubated at 33°C for 3 days, the number of EGFP-positive and EGFP-negative cells were counted by flow cytometry to measure the transfection efficiency. To detect retrotransposition, the electroporated cells (∼1×106 cells per dish) were plated in soft agarose medium containing G418 (1.6 mg/ml). In parallel, to determine the plating efficiency, the electroporated cells (200 cells per dish) were plated in soft agarose medium without G418. After an 11-day incubation at 37°C, visible colonies were counted. Plating efficiency was calculated as the number of visible colonies on the plate (without G418) as a percentage of the 200 cells plated. RF was calculated as: RF = G/(E×P/100), where G represents the number of G418-resistant colonies, E represents the number of EGFP-positive cells, and P represents the plating efficiency.
DT40 cells (WT, Ku70−/− or LigIV−/−) were cotransfected with pEGFPFLAG-1 (10 µg), pBZ2-2 (10 µg), and one of two expression vectors (10 µg), chicken Ku70/pAneo or pAneo. Transfection was carried out by electroporation at 250 V and 960 µF using a GENE Pulser. After transfection, cells were processed by the same procedure described above, and the RF was calculated.
To detect transcription of Ku70 from the exogenous and/or endogenous gene, cells transfected with the three plasmid DNAs were harvested 3 days after transfection. Total RNA was extracted from the cells using the RNeasy Mini kit (Qiagen). The column-based preparation was repeated to avoid any DNA contamination. RT-PCR was performed using the total RNA (1 µg) as the template with primers cKu70R1 (5′-CAGAGACAGTGAGCTTGCCC-3′) and cKu70F2 (5′-CGCTGGATATGCTGGAACCA-3′). As a control, transcription of the chicken β-actin gene was detected by RT-PCR using primers cActinF1 (5′-GGTCAGGTCATCACCATTGG-3′) and cActinR1 (5′-TGCATCCTGTCAGCAATGCC-3′).
DT40 cells were cotransfected with pEGFP-FLAG-1 (15 µg) and one of the LINE expression vectors (15 µg), pBZ2-2, p131.11, pJM102/L1.3, or pJM102/L1.3 H230A [22],[27]. Transfection was carried out by electroporation at 200 V and 950 µF using the GENE Pulser Xcell (Bio-Rad). After electroporation, the cells were incubated at 33°C for 3 days. Then the cells were subcultured at 37°C. The percentage of EGFP-positive cells and their EGFP fluorescence intensity were monitored by flow cytometry at intervals of 12 h from 3 to 8 days after electroporation. Ten thousand cells were counted for each measurement by flow cytometry.
Circular DNA containing ZfL2-2 insertions was isolated using the procedure developed by Gilbert et al. [11]. Briefly, DT40 cell clones derived from each G418-resistant colony produced by pAZ2-2 were cultured separately until the total number of cells reached ∼1×107 per clone. Genomic DNA was isolated from each clone using the GenElute mammalian genomic DNA miniprep kit (Sigma). Genomic DNA (∼20 µg per clone) was digested with 75 U of HindIII for 6 h at 37°C. The digested DNA (∼20 µg) was then self-ligated overnight by T4 DNA ligase (350 U) in 500 µl solution at 16°C. Ninety percent of the circular DNA was incorporated in E. coli DH10BT1R (Invitrogen) by electroporation with the GENE Pulser Xcell (Bio-Rad) under conditions of 2,500 V, 25 µF and 100 Ω, and the electroporated cells were plated on kanamycin-containing (70 µg/ml) plates. Circular DNA containing a mneoI400/ColE1-marked ZfL2-2 insertion (with its flanking chicken genomic DNA) was isolated from the kanamycin-resistant cells. The 5′ and 3′ junctions of each isolated ZfL2-2 insertion were sequenced using the appropriate primers. Sequences flanking each ZfL2-2 insertion were used as probes in BLAT searches to identify the preintegration site in the chicken genome database (http://genome.ucsc.edu; the May 2006 chicken (Gallus gallus) v2.1 draft assembly).
Exponentially growing HeLa-RC cells [25] were exposed to increasing concentrations of NU7026 with or without etoposide (1 µM) for 2 h. After treatment, the cells were trypsinized and reseeded into new 100-mm dishes at densities of 350 or 3,500 cells/dish and grown in fresh medium containing no drug. After 10 days, colonies were fixed with 100% ethanol and stained with 2% Giemsa solution. The survival rate was calculated as the number of colonies as a percentage of the reseeded cells.
HeLa-RC cells (2×105 cells/well) were seeded in 6-well dishes [25]. NU7026 of the indicated concentration was added to the medium 1 day after seeding. One hour after the addition of NU7026, the cells were transfected with 1 µg plasmid DNA (pBZ2-5 or pJM102/L1.3). The cells containing the plasmid were selected with hygromycin (200 µg/ml) for 6 days. NU7026 treatment was continued during the hygromycin selection. The hygromycin-resistant (HygR) cells were trypsinized and reseeded into new 100-mm dishes (at the density of 100,000 cells/dish for pBZ2-5 and 5,000 cells/dish for pJM102/L1.3) and grown in medium with 400 µg/ml G418. In parallel, 10,000 HygR cells for pBZ2-5 or 2,000 HygR cells for pJM102/L1.3 were also reseeded in a 100-mm dish and grown in medium without G418 to measure the plating efficiency. After a 12-day incubation, cell colonies were fixed by 100% ethanol and stained with 2% Giemsa solution. The plating efficiency was calculated as the number of visible colonies on the plate (without G418) as a percentage of the number of cells plated. RF was calculated as the number of G418-resistant colonies per HygR cell, compensating for the plating efficiency.
|
10.1371/journal.pgen.1006761 | Identification of breast cancer associated variants that modulate transcription factor binding | Genome-wide association studies (GWAS) have discovered thousands loci associated with disease risk and quantitative traits, yet most of the variants responsible for risk remain uncharacterized. The majority of GWAS-identified loci are enriched for non-coding single-nucleotide polymorphisms (SNPs) and defining the molecular mechanism of risk is challenging. Many non-coding causal SNPs are hypothesized to alter transcription factor (TF) binding sites as the mechanism by which they affect organismal phenotypes. We employed an integrative genomics approach to identify candidate TF binding motifs that confer breast cancer-specific phenotypes identified by GWAS. We performed de novo motif analysis of regulatory elements, analyzed evolutionary conservation of identified motifs, and assayed TF footprinting data to identify sequence elements that recruit TFs and maintain chromatin landscape in breast cancer-relevant tissue and cell lines. We identified candidate causal SNPs that are predicted to alter TF binding within breast cancer-relevant regulatory regions that are in strong linkage disequilibrium with significantly associated GWAS SNPs. We confirm that the TFs bind with predicted allele-specific preferences using CTCF ChIP-seq data. We used The Cancer Genome Atlas breast cancer patient data to identify ANKLE1 and ZNF404 as the target genes of candidate TF binding site SNPs in the 19p13.11 and 19q13.31 GWAS-identified loci. These SNPs are associated with the expression of ZNF404 and ANKLE1 in breast tissue. This integrative analysis pipeline is a general framework to identify candidate causal variants within regulatory regions and TF binding sites that confer phenotypic variation and disease risk.
| The promise of effective personalized medicine is dependent upon the ability to identify genetic variants in the population that influence disease risk and then use this information to accurately predict the likelihood of disease incidence for individual patients. High-risk individuals may be entered into clinical trails, pre-clinical intervention strategies, or increased frequency of screening to detect early disease onset. However, the contribution of any one genetic variant to increase disease susceptibility is typically small, with many potential causal variants in the genomic region associated with risk. Therefore, it is important to understand the biological mechanisms by which the variants within a genetic region influence disease susceptibility by refining the set of all variants to those that are highly plausible to be causal. Herein, we describe a method to integrate molecular genomics data with genetic epidemiological data to inform on the underlying molecular mechanisms that influence breast cancer risk. This approach identifies the important transcription factors that directly regulate gene expression to modulate disease susceptibility.
| Genome-wide association studies (GWAS) have identified more than 90 genomic loci and common genetic variants associated with breast cancer [1–5]. The single nucleotide polymorphisms (SNPs) associated with breast cancer have been shown to be enriched in DNA regulatory regions [6, 7], with few residing in coding regions of genes. The mechanisms by which most of these variants contribute to breast cancer biology remain unknown [8–11]. The effects of putative causal non-coding SNPs are challenging to interpret as they may alter transcription factor (TF) binding sites [12], lncRNA structure [13], splicing [14], transcription start or termination signals, or DNA shape [15]. Non-coding SNPs that alter TF binding sites are the most easily interpreted because they have the potential to modulate gene expression to mediate their effects on disease risk [16]. Therefore, it is possible to identify putative causal SNPs by focusing on those that alter TF binding sites in breast tissue.
TF dysregulation is a hallmark of many cancers [17, 18]. Genes encoding TFs in tumor cells are often amplified, deleted, rearranged via chromosomal translocation, or subjected to point mutations that result in a gain- or loss-of-function [18]. For example, transcriptional amplification of c-Myc reduces rate-limiting constraints for tumor cell growth and proliferation; high c-Myc expression correlates with tumor aggression and poor clinical outcome [19]. Estrogen receptor (ER) is a TF that regulates cell proliferation, which is the defining feature of luminal breast cancers [20]. Identifying the full set of TFs that function within a cell type remains a challenge.
Enzymatic accessibility assays identify open chromatin in the genome, which is an indirect measure of regulatory element activity and TF binding events [21–25]. TFs that directly or indirectly recruit cofactors, such as histone modifiers and nucleosome remodelers, recognize sequence motifs that are enriched in regions of open chromatin characterized by active histone marks and enzymatic hypersensitivity peaks [26, 27]. These cofactors maintain the chromatin structure at regulatory elements. Therefore, one strategy for identifying the functional TFs in a cell is to query the sequence underlying accessible regions for over-represented motifs [28–31]. Alternatively, depletion of signal within hypersensitive regions (footprints) [30, 31] coupled to motif analysis can be used to infer TF binding. The reliance on hypersensitivity footprinting to define a near-comprehensive set of TF motifs is limiting because many TFs do not have footprints for biological reasons [31] and the enzymes exhibit sequence-specificity that can be misinterpreted as footprint signatures [31–33]. These methods strictly identify motifs that are over-represented within regions of open chromatin, but families of TFs often contain paralogous DNA binding domains and thus recognize indistinguishable sequence motifs. One can directly measure the expression of TFs to identify putative functional TFs among related TFs [34]; however, expression of a TF is an imperfect proxy for TF function. For example, many nuclear receptors are not transcriptionally functional in the absence of ligand regardless of expression levels. Here we propose an approach that integrates open chromatin genomic data and gene expression data to identify candidate TFs that are functional in breast cancer-relevant cells.
Expression quantitative trait loci (eQTL) analysis identifies genetic variants that correlate with gene expression differences in a population. eQTL analysis complements genetic association data by predicting causal genes whose expression differences dictate organismal phenotypes [35, 36]. The gene that is responsible for a trait may be located relatively far from the GWAS associated SNPs, as the causal SNPs may modulate TF binding and TFs can act distally to regulate gene transcription. In these cases, preferential binding of a TF to one allele causes differential regulation of gene expression to confer the phenotype [37]. Several studies have provided evidence of causal relationships for gene expression mediating the association between GWAS SNPs and traits [16, 38]. Phenotype-associated SNPs are enriched for eQTLs, suggesting that eQTL analysis can enhance discovery of causal genes associated with complex phenotypes [39, 40].
The goal of this study is to gain mechanistic insight into how breast cancer susceptibility alleles confer disease risk. First we identify sequence motifs in breast-relevant tissue and cell lines that are enriched within enzymatic hypersensitive sites and support their functional role with evolutionary conservation analysis and gene expression data. Next we identify breast cancer risk alleles that are predicted to modulate TF binding and we perform eQTL analysis to find candidate causal risk genes.
Regulatory regions are bound by a milieu of protein complexes and regulatory nucleic acids, such as transcription factors, histone modifiers, nucleosome remodelers, and lncRNA. Sequence-specific TFs are directly or indirectly responsible for the recruitment of downstream transcriptional modifiers in a sequence-dependent manner. Therefore, we sought to identify all TFs that bind within open chromatin in breast cancer-relevant cells and tissues. We performed ATAC-seq in a mammary epithelial cell line (MCF10A) to complement publicly available DNase-seq data from Encyclopedia of DNA Elements (ENCODE) [41] and the Roadmap Epigenomics project [42]. We quantified open chromatin and identified regulatory elements genome-wide in five breast cancer-relevant cell lines and breast tissue: MCF7 cells, MCF10A cells, T47D cells, cultured human mammary epithelial cells (HMEC), and primary breast variant HMECs (vHMEC). Many regions of chromatin accessibility are shared between the cell lines, although the degree of accessibility for a region can vary between cell types (Fig 1). Enzymatic accessibility coverage (peaks) [28, 29] or depletions of signal in hypersensitive region (footprints) [30, 31] coupled to motif analysis are routinely used to identify TFs that maintain open chromatin structure.
We performed iterative rounds of de novo motif analysis using the sequence underlying enzymatic hypersensitivity peaks to identify overrepresented TF recognition sites in each data set (S1 File). All but one of the motifs we identified were previously characterized and described in databases [43–47]. These motifs found in hypersensitive peaks are, on average, evolutionarily conserved (Fig 2). Additionally, we identified a potential TF recognition sequence that has no known cognate TF binding partner; we refer to this sequence element as an orphan motif (Fig 2B). This orphan motif is evolutionarily conserved, as measured by phastCons [48] and phyloP [49] scores, in hypersensitivity peaks. Hypersensitivity footprints result from protection of the DNA by a bound TF [50]; however, approximately half of all TF-bound motifs do not exhibit composite footprints [31, 51] because the TF dissociates during the nuclei isolation procedure. We corrected the DNase data for intrinsic sequence bias [33], but we do not observe a composite footprint for this orphan motif (Fig 3B). However, the hypersensitivity pattern surrounding the motif is not uniform—the region downstream of this orphan motif is more hypersensitive than upstream. This directional pattern of enzyme accessibility is common with many TFs [29], including CTCF (Fig 3A). We hypothesize that this orphan motif is functional and directs chromatin accessibility by serving as a recognition site for an uncharacterized TF.
In addition to this orphan motif, we identified hundreds of position-specific weight matrices (PSWM) from our exhaustive analyses, many were redundant between cell lines and found in multiple rounds of motif analysis in the same cells. To reduce the complexity of these data, we mapped the comprehensive set of motifs found in MCF7, MCF10A, T47D, HMEC and vHMEC into distinct non-redundant motif families. This operation resulted in identification of 37 sequence motif classes across the five breast cancer-relevant cell lines and tissue (S1 Fig). Many TFs share paralogous DNA binding domains and these TFs often recognize the same sequence motifs. We defined the full set of TFs that recognize each motif by using known TF/sequence interaction data from ChIP-seq [43, 45, 52], protein binding microarrays [53, 54], and SELEX [46, 47] data. We identified 23 TF families in at least two cell lines/breast tissue (Table 1). Fourteen TF families were uniquely found in one cell line or tissue (Table 1). These 37 TF families represent the binding motifs for 235 distinct TFs (Table 1). However, these TFs are not all expressed in breast tissue. We identified the TFs that are most likely candidates for maintaining open chromatin and the gene regulatory expression profiles in breast-relevant cells by examining the relative expression of all of the TFs in each family using TCGA expression data (Fig 4 and S2 Fig). For example, ESR1, ESR2, and PPAR-γ contain paralogous DNA binding domains and they recognize indistinguishable sequence elements (S1 Fig). We find that ESR1 is the most highly expressed TF that recognizes the motif (Fig 4). This result is consistent with the biological role of ESR1 in the etiology of breast cancer and breast biology compared to ESR2 and PPAR-γ. Similarly, FOXA1 is the most well-charactered TF within the Forkhead Box family of TFs in terms of estrogen signaling and interplay with ER [55]. As expected, we find that FOXA1 is the most highly expressed TF in the family of 22 Forkhead Box TFs (Fig 4). We find that many of these highly expressed TFs correlate with breast cancer survival time in a subtype-specific manner (S3 Fig). Silencing of IRF7 pathways in breast cancer cells promotes breast cancer metastasis, and high expression of the IRF7-regulated genes with breast cancer is associated with prolonged survival [56]. Similarly, we find that high expression of IRF7 is correlated (P = 0.029) with positive breast cancer patient outcome in Luminal A subtype (S3 Fig). We find that high expression of BATF (P = 0.0035) and TP73 (P = 0.0077) is correlated with breast cancer patient survival in HER2+ and Basal-like subtypes, respectively (S3 Fig). Taken together, these data support the notion that TF expression levels may serve as biomarkers of patient outcome.
A major goal of this study was to identify a set of plausible causal SNPs that modulate TF binding from a list of SNPs associated with breast cancer in GWAS-defined loci. We identified a total of 463 SNPs in strong LD (r2 ≥ 0.8) with the most associated breast cancer GWAS SNPs defined in 93 distinct genomic loci; these SNPs are predicted to affect the binding of TFs belonging to at least 30 TF families. Six examples of candidate causal breast cancer-associated SNPs are shown in Table 2. Transcription factors from the following TF families are predicted to have their binding affected: CTCF, GABPA, RUNX, GRHL2, USF1, ZBTB33, and ZNF143. For example, SNP rs11540855 (3′ UTR of ABHD8 on chromosome 19p13.11) is within a DNase-defined regulatory element in human mammary epithelial cells and should affect the binding of CTCF (Fig 5). rs11540855 is in strong LD with rs8170 (r2 = 0.98), which is associated with breast cancer risk [58–60]. Likewise, rs3760982 (1.1kb 5′ of KCNN4 on chromosome 19q13.31) is associated with breast cancer susceptibility [3] and we find that its A allele is predicted to enhance RUNX binding (Fig 5). More examples of candidate causal breast cancer-associated SNPs disrupting TF binding sites within breast cancer GWAS loci are shown in the supporting information (S4 Fig).
Two SNPs (rs4414128 and rs8103622) are predicted to strongly affect CTCF binding (Fig 6 and S5 Fig); therefore, we tested our predictions of how the alleles would affect CTCF binding by analyzing ENCODE ChIP-seq data. The SNPs rs4414128 and rs8103622 showed allele-specific binding that was consistent with the direction predicted based on tolerated degeneracy from the consensus binding site (Fig 6 and S5 Fig). For rs4414128 (Fig 6), 34 of 44 cell types/replicates show the expected C preference with a range between 52% and 75%. Eight cell lines/replicates show modest allelic imbalance favoring the T allele (31–49% of the reads spanning the SNP); two experiments are balanced. Importantly, both replicates of human mammary epithelial cells (HMEC) show an allelic imbalance favoring C in 58% and 75% of the reads. Across 37 cell types (and replicates) that are heterozygous and normal karyotype, rs8103622 (S5 Fig) shows allele-specific preference of C as expected in 34 cell types/replicates with the range between 53% and 88%. The other three instances exhibit allelic balance, with allele frequencies between 48% and 52%.
We identified rs11540855 as the most significant (P-value: 2.75E-10) eQTL SNP in the GWAS locus that is in strong LD (r2 = 0.90) with the most associated breast cancer GWAS SNP rs8170 [58] (Fig 7C). The G/G genotype at rs11540855 is predicted to compromise CTCF binding (Fig 5) and G/G individuals have, on average, higher expression levels of ANKLE1 (Fig 7A and 7B). The most significantly associated GWAS SNP in this locus (rs8170-T) increases the risk of ER-negative breast cancer with an odds ratio of 1.10 [60] and the T allele is in LD with the G allele of rs11540855. Therefore, higher expression of ANKLE1 (Fig 7A and 7B) is associated with increased breast cancer risk. We also identified rs3760982 as one of the most associated eQTL SNPs (P-value: 3.16E-07) and rs3760982 is correlated with ZNF404 expression (Fig 8C). The A/A genotype at rs3760982 is predicted to increase RUNX binding (Fig 5) and is correlated with higher expression of ZNF404 in breast cancer tumor samples (Fig 8A) and breast tissue (Fig 8B). The rs3760982-A allele is associated with an increased risk (odds ratio of 1.06) of breast cancer [3], thus higher expression of ZNF404 correlates with increased breast cancer risk. Therefore, we prioritized SNPs within GWAS loci that are predicted to affect transcription factor binding and module expression of ANKLE1 and ZNF404 to confer breast cancer risk.
GWAS have discovered more than 90 genetic loci and common genetic variants associated with breast cancer susceptibility [1–5], and the majority of SNPs in these loci are enriched in non-coding regions. Non-coding genetic variants can contribute to complex traits and diseases through many molecular mechanisms [12–14, 61, 62] including having an effect on TF binding affinities, which can result in differential gene expression. Herein, we describe an integrative genomics methodology to identify a near-comprehensive set of TFs that are actively maintaining open chromatin in a cell type. We identify which GWAS-relevant SNPs are predicted to modulate TF binding intensity and use ChIP-seq data to confirm our predictions. Lastly, we use eQTL data to identify the likely target genes of SNPs that affect TF binding affinity. Taken together, this approach can identify likely causal SNPs associated with breast cancer risk.
We found that rs3760982 variants are predicted to modulate RUNX binding (Fig 5) and we confirm previous work showing that rs3760982 is an eQTL for ZNF404 (Fig 8) [3, 63]. The A allele of rs3760982 conforms more stringently to the RUNX consensus sequence and this allele is predicted to enhance RUNX binding; allele-specific ChIP-seq in breast tissue, would test whether RUNX family TFs prefer binding the A allele in vivo. The RUNX family of TFs are canonical transcriptional activators [64–67], so we hypothesize that increased RUNX binding is a mechanism by which the ZNF404 is regulated. Testing this hypothesis would necessitate specific gene editing of rs3760982 and subsequent measurements of ZNF404 expression. While CRISPR-mediated [68] deletions of genetic elements is routine, precise changes of specific alleles remains a challenge. Deletion of the rs3760982 variant by CRISPR, followed by measuring ZNF404 expression would confirm or refute the role of rs3760982 variants in ZNF404 expression. One could also test allele-specific expression of alleles within transcription units that are phased with rs3760982 variants. We propose using precision global run-on sequencing [69] to measure nascent RNA expression to capture informative intronic SNPs. While genomic approaches are a means to develop novel hypotheses, the advent of genetic editing approaches permits hypothesis testing to define mechanisms by which genes and genetic variants contribute to disease risk.
We found that rs11540855 is an eQTL for ANKLE1 (Fig 7) and rs11540855 variants are predicted to affect CTCF binding (Fig 5). The rs11540855 SNP is in high LD (r2 = 0.90) with the breast cancer-associated GWAS SNP rs8170, which was first found as a modifier of breast cancer risk in BRCA1 mutation carriers [58]. Subsequently, this SNP was found to be associated with breast cancer susceptibility in ER-negative breast cancer [58–60]. ANKLE1 is an evolutionarily conserved non-membrane-bound LEM protein that harbors endonuclease activity, but its cellular functions remain uncharacterized [70, 71]. Future work will need to determine the allele-specificity of CTCF binding at rs11540855 and test the role that this site has upon ANKLE1 expression. These approaches will be able to define the relationship between TF binding and gene expression, but it is challenging to develop a physiologically relevant model of breast cancer risk that incorporates human genetic variation. GWAS-identified SNPs are common and typically confer relatively small differences in risk. Further, the cumulative affects of differential gene expression over the lifetime of an individual cannot be easily recapitulated in a controlled environment.
Our research identified a previously uncharacterized DNA sequence motif that is enriched in open chromatin, evolutionarily conserved, and is associated with directional hypersensitivity profiles (Figs 2 and 3). We hypothesize that this orphan motif is the recognition site for a previously uncharacterized transcription factor. Future work, such as DNA affinity chromatography [72], will be needed to identify this candidate TF.
Genomic approaches are ideally suited to address fundamental biological questions in a relatively unbiased manner. Integrative genomic approaches and analyses can clarify the null-hypothesis and permit the development of novel hypotheses that were previously inconceivable. These approaches serve as a first-step in understanding the biology of breast cancer risk and targeted experimental follow-up is necessary to define the mechanistic roles of genes and genetic variants in breast cancer susceptibility and disease progression.
We cultured MCF10A cells in Dulbecco’s modified Eagle’s medium (Invitrogen) with 5% horse serum (Invitrogen), 1% penicillin/streptomycin (Invitrogen), 20 ng/ml EGF (Peprotech), 0.5 μg/ml hydrocortisone (Sigma), 100 ng/ml cholera toxin (Sigma) and 10 μg/ml insulin (Sigma) in a humidified incubator at 37°C with 5% CO2. The ATAC-seq library was prepared as previously described [73] with several modifications: 1) IGEPAL CA-630 was omitted from the lysis buffer; 2) we performed two additional wash steps with lysis buffer; and 3) we performed PCR-clean up using AMPure XP beads to select DNA <600 bp. The MCF10A ATAC-seq data were deposited in the Gene Expression Omnibus (GEO) database, with accession number GSE89013. We mapped reads to the hg38 human reference genome using Bowtie2 [74] and merged replicate aligned files. We used the merged data for all subsequent analysis; refer to S1 File for ATAC-seq data analysis details.
We performed iterative rounds of de novo motif analysis using a 120-base pair window centered on the summit of hypersensitivity as defined by ATAC-seq or DNase-seq (S1 File). In each cell type we found hundreds of over-represented position specific weight matrices (PSWMs). We identified all instances of each PSWM within breast-specific regulatory elements, while accounting for the possibility that the reference genome contains variants that will conform more or less strictly to the PSWM. This approach allowed us to identify potential binding sites that contain SNPs, even if the reference allele does not match the queried PSWM.
Although we identified hundreds of distinct PSWMs, many PSWMs are similar to one another and are likely to represent redundant specificity of a single TF or TF family. To consolidate similar PSWMs into known TF families, we systematically classified several public PSWM repositories [43–47] into families with distinct features. PSWMs were first divided into clusters based on connectivity; connectivity between motif nodes was measured by negative log10 E-value as calculated by TOMTOM [75]. An edge was inferred between two motif nodes if their similarity exceeded a negative log10 E-value of 10. We defined a motif cluster as a connected set of nodes; connectivity is defined by the existence of a path between every pair of nodes. A fast greedy modularity algorithm [76] further divided each motif cluster into families.
We downloaded a curated list of breast cancer associated SNPs from the GWAS catalog [77]. The SNP that exhibits the most statistically significant association with breast cancer in any locus may not be causal due to linkage disequilibrium (LD) and the sampling variation that interrogated the specific SNP. To better define the list of likely causal variants for each locus, we identified all SNPs satisfying the following three criteria: 1) SNPs that are in strong LD (r2 ≥ 0.8) with the most significant reported GWAS SNP; 2) SNPs that are located within putative TF binding sites identified by hypersensitivity assays; and 3) SNPs that are within critically important positions that affect TF binding affinity (Information Content ≥ 0.5).
We analyzed ENCODE CTCF ChIP-seq data for allele-specific preference of SNPs that are predicted to modulate CTCF binding affinity. All CTCF ChIP-seq data are provided within S1 File. We analyzed the highest intensity CTCF sites to ensure that sequencing reads would span the query SNP. We exclusively queried normal karyotype cell lines that had SNPs that were heterozygous within each locus to reduce the chances that copy number variations (i.e., aneuploidy) of alleles would skew our analyses.
To identify putative causal genes whose expression may be affected by polymorphisms, we performed eQTL analysis using TCGA breast cancer data [57] with fastQTL [78]. We imputed the genotypes from TCGA SNP6 arrays that were hybridized with DNA extracted from the blood of patients with breast cancer. We retrieved genotype data from dbGaP (phs000178.v9.p8) and imputed genotypes using the Michigan Imputation Server [79] with the following parameters: 1000G Phase 1 v3 Shapeit2 Reference Panel, ShapeIT Phasing, Mixed Population, and Quality Control/Imputation Mode. Following imputation, we removed SNPs with the following features: Hardy-Weinberg Equilibrium p < 1 × 10−6 and minor allele frequency (MAF) < 5%. We used UCSC-curated TCGA RNA-seq data [57] from breast cancer patient solid tumor samples as the gene expression data to identify eQTLs. TCGA clinical data were incorporated as the covariates for eQTL analysis such as sample RNA concentration, RIN value, sex, and ethnicity. We performed Principal Component Analysis (PCA) on the quantitative variables from clinical data and used the first three principal components as covariates. We retained other qualitative variables as categorical covariates.
|
10.1371/journal.pgen.1001365 | An Evolutionary Genomic Approach to Identify Genes Involved in Human Birth Timing | Coordination of fetal maturation with birth timing is essential for mammalian reproduction. In humans, preterm birth is a disorder of profound global health significance. The signals initiating parturition in humans have remained elusive, due to divergence in physiological mechanisms between humans and model organisms typically studied. Because of relatively large human head size and narrow birth canal cross-sectional area compared to other primates, we hypothesized that genes involved in parturition would display accelerated evolution along the human and/or higher primate phylogenetic lineages to decrease the length of gestation and promote delivery of a smaller fetus that transits the birth canal more readily. Further, we tested whether current variation in such accelerated genes contributes to preterm birth risk. Evidence from allometric scaling of gestational age suggests human gestation has been shortened relative to other primates. Consistent with our hypothesis, many genes involved in reproduction show human acceleration in their coding or adjacent noncoding regions. We screened >8,400 SNPs in 150 human accelerated genes in 165 Finnish preterm and 163 control mothers for association with preterm birth. In this cohort, the most significant association was in FSHR, and 8 of the 10 most significant SNPs were in this gene. Further evidence for association of a linkage disequilibrium block of SNPs in FSHR, rs11686474, rs11680730, rs12473870, and rs1247381 was found in African Americans. By considering human acceleration, we identified a novel gene that may be associated with preterm birth, FSHR. We anticipate other human accelerated genes will similarly be associated with preterm birth risk and elucidate essential pathways for human parturition.
| The control of birth timing in humans is the greatest unresolved question in reproductive biology, and preterm birth is the most important medical issue in maternal and child health. To begin to address this critical problem, we test the hypothesis that genes accelerated in their rate of evolution in humans, as compared with other primates and mammals, are involved in birth timing. We first show that human gestational length has been altered relative to other non-human primates and mammals. Using allometric scaling, we demonstrate that human gestation is shorter than predicted based upon gestational length in other mammalian species. Next, we show that genes with rate acceleration in humans—in coding or regulatory regions—are plausible candidates to be involved in birth timing. Finally, we find that polymorphisms in the human accelerated gene (FSHR), not before implicated in the timing for birth, may alter risk for human preterm birth. Our understanding of pathways for birth timing in humans is limited, yet its elucidation remains one of the most important issues in biology and medicine. The evolutionary genetic approach that we apply should be applicable to many human disorders and assist other investigators studying preterm birth.
| Despite the important public health consequences of preterm birth [1], [2], determinants of human parturition remain largely uncharacterized. While some important physiological antecedents of labor have been identified in model organisms, such as progesterone withdrawal in rodents, such signals do not seem to precede human labor. Because humans are born developmentally less mature than other mammals [3], [4], birth timing mechanisms may differ between humans and model organisms that have been typically studied [5].
Evidence suggests that parturition has changed along the human lineage in response to other uniquely human adaptations. The dramatic increase in brain size, along with the human pelvis becoming narrower to facilitate bipedalism, places unique constraints on birth in humans compared even with evolutionarily close relatives such as Neanderthals and chimpanzees [6], [7]. Given the historically high mortality rate associated with pregnancy, these human traits may generate selective pressure to initiate parturition at a relatively earlier time in gestation compared to non-human primates to avoid cephalopelvic disproportion and arrested labor by delivery of a relatively smaller, less mature fetus. High rates of human versus non-human primate divergence in human pregnancy-related genes, such as genes in the reproduction Gene Ontology (GO) category [8], [9] as well as GO categories related to fetal development, including transcription factors [10], nuclear hormone receptors [10], transcriptional regulation [11] and development [9], support the notion that human gestation length has been altered to accommodate features unique to human pregnancy.
Genetic influences on birth timing in humans appear to be substantial, based on family and twin studies [12], [13], [14]. However, association studies using candidates selected from suspected pathways have not detected robust susceptibility variants for preterm birth. Genome-wide association studies (GWAS) are promising but will require large numbers of well-characterized subjects in order to overcome the challenge of multiple statistical comparisons. Here, we test the hypothesis that the set of genes accelerated on the human lineage will include genes that play important roles in regulating parturition and harbor variants that influence preterm birth risk. We identified and analyzed genes showing marked divergence between humans and other mammals, defined by relative nucleotide substitution rates in coding and highly conserved noncoding regions, for association with preterm birth. We find that genes with evidence of rate acceleration in humans may provide an informative group of candidates, and demonstrate that the human accelerated gene, follicle-stimulating hormone receptor (FSHR), may alter risk for preterm birth.
Because of large human head size and narrow birth canal cross-section compared to other primates [6], we hypothesized that genes involved in parturition have evolved rapidly along the human phylogenetic lineage to decrease the length of gestation and alleviate the complications arising from these constraints. We performed a comparative analysis of life history traits in mammals to further evaluate whether the relative gestational period in humans has decreased compared to other primates and mammals. Data acquired by Sacher and Staffeldt [15] and reanalyzed by us show that both adult and neonatal higher primates (simians) have higher brain to body weight ratios compared to other mammals (Figure 1A, 1B and Table S1 for list of species). The difference in brain/body size ratios in higher primates relative to other mammals makes it possible to ask whether gestation in higher primates is linked to brain size or body size. Higher primates and other mammals have equivalent gestational periods with respect to brain weight (Figure 1C). In contrast, the gestational period in higher primates is longer relative to the length of gestation in mammals with equivalent neonatal body weights (Figure 1D). This suggests that the length of gestation is expected to change with brain size but not body size.
Humans have evolved the highest adult brain to body weight ratio of any mammal [16]. In contrast to the evolution of brain/body ratios in higher primates, where both adult and neonatal ratios are increased relative to other mammals, the increase in the brain/body ratio in humans relative to other primates is present in adults but not neonates (Figure 1B). The simplest explanation is that human adult brain/body ratios have changed independently of neonatal ratios. However, the ratio of brain/body weight is highest at birth and declines until adulthood. Thus, an alternative explanation is that both adult and neonatal brain/body ratios have increased in humans, as in other higher primates, but that a concurrent decrease in the length of gestation lowered the neonatal brain/body ratio. This second possibility is supported by the relative immaturity of human neonates compared to other primates [3], [4] and that the length of human gestation, relative to either neonatal brain or body weight, is shorter than most other higher primates (Figure 1C, 1D).
To examine the evolution of gestation length relative to neonatal brain and body weight in primates we inferred the evolution of these characters across a phylogenetic tree. For both gestation-neonatal body ratio (Figure 2A) and gestation-neonatal brain ratio (Figure 2B) there is a consistent trend of a relatively shorter length of gestation on branches leading to humans. Of note, humans have the lowest gestation-neonatal body ratio (Figure 2A) or gestation-neonatal brain ratio (Figure 2B) of all the 20 primates evaluated. The gestation-neonatal brain ratio for humans is 69% that of gorilla and 45% that of chimpanzee. The gestation-neonatal body ratio of human is 49% that of gorilla and 50% that of chimpanzee.
In light of this evidence for human adaptation for birth timing, we examined whether genes involved in parturition would display accelerated protein evolution along the human lineage measured by an increased rate of amino acid altering to synonymous nucleotide substitutions (dN/dS; Figure S1). We found that, of 120 suggested candidate genes for preterm birth that were included in the ENSEMBL database, 7 showed statistically significant increased rate acceleration (i.e. increased dN/dS; p<0.05) along the human lineage in comparison to the other lineages. Table 1 shows these 7 genes plus 2 other genes significantly accelerated along the human-chimpanzee ancestor lineage (complete analysis of dN/dS provided in Dataset S1). Of these, common variants of PGR [17] and MMP8 [18] have previously been found to contribute to preterm birth risk. Using criterion agnostic to possible involvement with preterm birth, and measuring genome-wide changes, we identified 175 genes either accelerated along the human (40 genes) or on the human and human-chimpanzee ancestor lineages combined (135 genes) at a 5% false discovery rate (FDR) [19] from this analysis of protein-coding sequences.
Motivated by this evidence of protein coding region evolution for genes involved in parturition and that acceleration has also been found to act on noncoding regions, we developed a method to identify human accelerated noncoding sequences [11], [20]. We identified a total of 401 elements significant along the human lineage and 2,103 elements significant along the human and human-chimpanzee ancestor lineages at a 5% FDR. To choose candidate genes, we calculated gene-wise p-values for each gene locus by assigning each conserved element to its nearest RefSeq gene [21] and a Fisher's combined p-value across the locus. This resulted in identification of a total of 279 candidate genes (complete analysis of human accelerated non-coding regions provided in Dataset S2). 150 of the genes identified as human accelerated in the protein-coding sequence and highly conserved noncoding elements screens, selected based on expression and functional information suggesting potential roles in parturition, were analyzed for association with preterm birth (Table S2).
Because recent data suggests that heritability of risk of preterm birth acts largely through the maternal genome [14], and the Finnish have low environmental risk and high genetic homogeneity compared to other populations, we genotyped Finnish (165 case, 163 control) mothers for 8,490 SNPs in the gene regions of our prioritized list of 150 human accelerated genes. The most significant finding was rs6741370 (p = 8.1×10−5) in the follicle-stimulating hormone (FSH) receptor gene (FSHR). 91 SNPs were significant at the p<0.01 level by allelic tests (Table S3). However, no SNPs were significant after correcting for 5,377 independent tests, considering relationships among markers, by the Bonferroni method (p<9.3×10−6). Of note, 8 of the 10 most statistically significant SNPs were located in FSHR. We identified FSHR as human accelerated in the noncoding analysis, with 40 changes in 4,218 bp of 17 conserved elements (human lineage p = 5.4×10−5, Dataset S2). Moreover, FSHR was revealed as rapidly evolving in a study of noncoding conserved elements by Prabhakar and colleagues [20], which otherwise had limited overlap with our gene list (see Methods). FSHR also harbors SNPs with extreme iHS values in the Yoruban population, reflecting extended haplotype homozygosity and suggesting a recent selective sweep [23]. Bird and colleagues [24] identified a region less than 1 megabase downstream of the FSHR gene boundaries as rapidly evolving in their study, further supporting human acceleration of the locus. Finally, because of being paralogous with other G-protein coupled receptors, such as the luteinizing hormone receptor, FSHR was excluded from our genome-wide coding region analysis. Therefore, we separately analyzed FSHR coding region acceleration along the human lineage. We found that the human-specific dN/dS was 1.41 which was significantly accelerated (p = 0.0045) in comparison to a constrained model for other primates and mammals using a 5 way multi-Z alignment in HYPHY where dN/dS was 0.174 over the entire tree (human, chimpanzee, rhesus, dog, mouse). The human-specific dN/dS for FSHR greater than 1 provides evidence for recent positive selection in addition to rate acceleration in humans. This information, together with the known importance of variation in human FSHR in subfertility [25], [26], a risk factor for preterm delivery independent of the use of assisted reproductive technologies [27], [28], and evidence suggesting its expression in uterus and cervix [29], [30], [31], motivated its specific study.
11 SNPs in FSHR showing potential association in the screening analysis (p<0.1) were genotyped in European American (147 preterm, 157 control), African American (79 cases, 171 controls) and Hispanic (Mexican) American (73 preterm, 292 control) mothers (Table 2 and Table S4). Several SNPs exhibited suggestive association (p<0.01) with preterm birth risk. Three SNPs in the African American mothers, rs11686474, rs11680730 and rs12473815, were significant after correcting for multiple testing (OR 1.63–1.82 (95% CI 1.11–1.21), 10 independent tests; p≤0.005). The allele frequency for this high linkage disequilibrium block differs considerably between HapMap CEU and YRI populations. To determine whether this association reflects a functional effect of local variation and not an artifact of population stratification with greater African ancestry in the case population relative to controls, we analyzed a limited set of ancestry informative markers using STRUCTURE. We found a small number of individuals (10, 3 cases and 7 controls) in our African American cohort that grouped more closely with the HapMap CEU cluster than the HapMap YRI cluster, though the relative distribution of these between cases and controls did not statistically differ from the relative sizes of the group. We performed a logistic analysis including the quantitative measure of CEU clustering as a covariate. The CEU cluster value was not significant in the model (p = 0.77), and adjusting for this in the regression model had little effect on statistical significance (e.g., unadjusted allelic p-value for rs12473815 = 0.0032, adjusted p = 0.0047). While we do not find evidence that population substructure confounds the association study in our African American cohort, we acknowledge that further study exploiting a larger number of subjects along with more dense ancestry markers will be needed for definitive conclusions to be drawn regarding association in this population. We did not find a statistically significant association in our European American or Hispanic cohorts for this LD block in FSHR, though risk trends for the minor allele (OR 1.08–1.38) were in the same direction as the Finnish and African American populations. This finding may reflect the limited sample size analyzed, or a specific role for variants in this LD block in the genetically isolated, homogeneous Finnish population and ancestrally distinct African American population.
In FSHR, these 4 SNPs in high LD lie within intron 2 of FSHR (Figure 3) and show little LD with variants outside of this intron, based on available information from the International HapMap Project database [32]. Variants in this intron may tag yet uncharacterized variants in coding regions or nearby regulatory sequences. Alternatively, an intronic variant in FSHR may affect risk directly by altering functional sequences contained within the intron, such as microRNA binding sites, splice regulatory sites or transcription regulation sites. For instance, a variant in a splice enhancer site may change splicing patterns in favor of transcripts that promote preterm birth risk, as several alternatively spliced FSHR isoforms have been observed with altered function [33]. Further suggesting functional importance of this LD block, rs12473870 is significantly associated (p<0.0001) with altered expression of CCNJ, FURIN, DDR1, TBCD10A, and NAGA in quantitative trait databases for YRI populations (http://scan.bsd.uchicago.edu/newinterface/about.html). Risk-promoting variation in this gene may contribute to birth timing, rather than size at birth, based on additional tests examining gestational age or birth-weight Z-score as a quantitative trait, rather than preterm birth affection status (Table S5). Hence, FSHR may represent a novel gene involved in birth timing and preterm birth risk.
FSHR encodes the follicle-stimulating hormone (FSH) receptor. FSH is secreted from the pituitary and, in females, acts primarily on receptors in the ovaries to stimulate follicle development and synthesis of estrogens. Investigators also have observed FSHR protein and mRNA expression in nongonadal tissues, including uterus and cervix [29], [30], [31]. In these tissues, FSHR may mediate uterine relaxation, as suggested by FSH's ability to modify electrical signaling in the myometrium, independent of estrogen and progesterone [29]. Padmanabhan and colleagues [34] noted a progressive rise in bioactive serum FSH levels during pregnancy. Because high levels of FSH are known to downregulate FSHR expression [35], increasing levels of FSH may lead to gradual desensitization to the hormone and resultant increase in contractility as term approaches. Additionally, evidence from the FSHR haploinsufficient mouse [36] suggests that FSHR levels affect the relative abundance of progesterone receptor isoforms A (PR-A) and B (PR-B). Increased PR-A: PR-B ratios, occurring in human pregnancy normally near term and observed in FSHR haploinsufficient mice in non-pregnant states, are correlated with increased myometrium contractility. Hence, dysregulation of FSHR may contribute to early uterine contractility and promote preterm birth.
Aspects of our approach pose limitations on interpretation of this work. First, we assigned conserved elements to the nearest RefSeq gene to calculate gene-wise p-values; however, conserved elements may not be associated with the nearest gene per se, potentially affecting the accuracy of the estimate gene-wise p-values. Additionally, because we use adjacent genes to estimate expected synonymous and nonsynonymous rates for a given locus, human accelerated genes that are located physically nearby other genes undergoing human acceleration, such as gene families with multiple members in the same region, may miss detection. The variability in number of probes represented on the Affymetrix Genome-Wide Human SNP Array 6.0 within the gene regions of the 150 human accelerated genes tested poses another limitation. Although the coverage is adequate for most human accelerated genes, there are some genes with too few probes tested to support or refute their potential association with preterm birth; as a result, this study may have failed to detect association between preterm birth and human accelerated genes underrepresented on this genotyping array. Lastly, while precedence exists for intronic variants affecting protein structure and function [37], [38], additional study is needed to prove whether any of the SNPs associated with preterm birth in this work have a functional effect.
We find that human gestational length has been altered relative to other non-human primates and mammals. Using allometric scaling, we demonstrate that human gestation is shorter than predicted based upon gestational length in other mammalian species. By using comparative genomics to identify genes with an accelerated rate of change in humans, we identified a gene that shows evidence of association with preterm birth that otherwise would not have been revealed by current models of parturition physiology [39]. Moreover, our approach exploits a filter for relevant genes based upon rate of evolution in humans to more efficiently utilize currently available datasets for preterm birth, which are probably underpowered to detect variants of effect sizes reported in GWAS of other complex traits. Our approach represents an alternative method for a priori gene discovery in which fewer comparisons are made than in GWAS, thus potentially retaining more power to detect effect sizes typical for common variants. We provide evidence that FSHR, identified by these means, may alter risk for preterm birth. We anticipate that other human accelerated genes will similarly be associated with preterm birth risk and elucidate the essential pathways for human parturition.
Data acquired by Sacher and Staffeldt [15] was used to examine the relationships among brain size, body size and gestation length among mammalian species. Specifically, we compared logarithm-transformed values for these traits between human, primate and non-primate mammals, using linear regression implemented in R [40]. Additionally, we used allometric data from this paper and the primate phylogeny delineated by Purvis [41] to trace the evolution of gestation-neonatal body size ratio, and gestation-neonatal brain size ratio, using Mesquite [42]. Given a phylogenic tree, the Mesquite method uses parsimony to reconstruct the ancestral states by assuming a squared change for a continuous character from state x to state y is (x–y)2.
We obtained a set of 10,639 human gene predictions from the ENSEMBL database with one-to-one orthologs in the chimpanzee, macaque, mouse, rat, dog, and cow genomes (Release 46) [43]. We limited our analysis to only those proteins where the human, chimpanzee, macaque, and at least 75% of the mammalian genomes were present (Text S1). The list of 120 possible candidate genes for preterm birth assessed for dN/dS included those in the Institute of Medicine report [39], SPEED (pregnancy), GeneCards (parturition), and progesterone/prostaglandin metabolic pathways.
We obtained a set of highly conserved elements from UCSC Genome Browser [44] and tested 443,061 noncoding sequences with a conservation score > = 400. From the 17-way MultiZ alignments that are publicly available (downloaded March, 2007) [45], we extracted the human, chimpanzee, macaque, mouse, rat, dog and cow sequences (Text S1).
We used the phylogeny ((Human, Chimpanzee), Macaque), ((Mouse, Rat), (Dog, Cow))). The evolutionary models were implemented in the HYPHY package [46] and we used the Q-value software [19] to establish statistical thresholds to achieve 5% false discovery rates (p-value distributions and pi_0 values in Figure S2).
Previous studies of both coding [9], [46] and noncoding [11], [21] sequences identify regions evolving under positive selection by a rate of evolution faster than a neutral rate. However, we felt that this criterion is too restrictive since some genes may have an increased rate of evolution along the human lineage relative to other mammals, but not increased above the neutral rate. To include genes with a significantly increased rate in humans compared to other mammals for testing in a population association study, we identify genes as human accelerated by testing whether omega along the human (or human+human-chimpanzee ancestor) lineage is significantly higher than omega along the non-human lineages (or non-human+non-human-chimpanzee ancestor). Here, omega is dN/dS-adj or dNC/dNC-adj, where dNC is the noncoding rate and dS-adj and dNC-adj are the adjacent synonymous rates from the 10 upstream and 10 downstream genes and the adjacent noncoding rates from 25 kb of conserved noncoding sequences, respectively. Thus, we test whether the data is more likely under a model with 1 omega value or 2 omega values (Figure S1). The coding sequence model used the MG94×HKY85 [47] model of codon evolution. The noncoding sequences model used an HKY85 model. For both tests, the alternative model has one additional degree of freedom and the significance of the change in likelihood was determined using chi-squared statistics. Both models use adjacent coding or conserved noncoding sequences to estimate the expectation for a given sequence that accounts for variable mutation rates across the genome and lineage-specific differences in effective population size, by allowing for branch-specific differences in selective constraint. Our list of human accelerated coding region gene list showed low overlap with previous studies that required for dN/dS>1 in their analyses (6% with Clark et al. [9], 0% Nielson et al. [48]) and more overlap with Arbiza et al. [49] (26%) which considered rate acceleration on the human lineage by methods more similar to ours than those used by [9], [48] (Figure S3). For human accelerated conserved noncoding elements in humans, 22% of the elements we identified were in common with Prabhakar et al. [20]. Considering unique genes associated with human accelerated conserved noncoding elements in humans, 11% of our genes also were identified by Prabhakar et al. [20], and 4% identified by Pollard et al. [11]. Similar to our study, 4% of unique genes in the Prabhakar study overlapped with those identified by Pollard et al. (Figure S4).
We calculated gene-wise p-values for each gene locus by assigning each conserved element to its nearest RefSeq gene [21] and a Fisher's combined p-value across the locus. Chi-squared analysis was used to determine the statistical significance of observed and expected genes with p<0.05 in suggested preterm birth candidate and overall human gene lists.
To minimize the number of tests we would perform and thereby retain more power to detect small effects, we selected a subset of genes likely to be involved in parturition, based on expression and functional information, to use as candidate genes. Duplicated genes from a list developed by Bailey and colleagues [50] identified as pregnancy, fetal, placental or hormone-related genes were also included as candidates. A total of 150 of genes were used as candidate genes in subsequent analysis (Table S2).
Mothers of preterm or term infants were enrolled for genetic analysis by methods approved by Institutional Review Boards/Ethics Committees at each participating institution. Informed consent was obtained for all participants. Mothers with preterm birth were included if the birth was spontaneous (non-iatrogenic), singleton, had no obvious precipitating stimulus (trauma, infection, drug use), and was less the 37 weeks (Yale University; New York University) or 36 weeks (University of Helsinki; University of Oulu; Centennial Hospital, Nashville, TN) of completed gestation. DNA from blood or saliva was prepared by standard methods. Race/ethnicity was assigned by self-report. For the African American cohort, no differences in allele frequency were found in the distribution of 24 ancestry informative markers selected across the genome comparing cases and controls (all p>0.05 performing Chi square analysis between cases and controls; data not shown). All specimens were linked with demographic and medical data abstracted from maternal/neonatal records.
Initial genotyping of the Finnish cohort was performed using the Affymetrix Genome-Wide Human SNP Array 6.0. Genotypes were called from cell intensity data by the birdseed v2 algorithm, implemented in Affymetrix Genotyping Console 3.0. We selected SNPs represented on the array within the gene regions of candidate genes for analysis. SNPs examined in replication cohorts were genotyped using the Sequenom iPLEX massARRAY technology (Sequenom, San Diego, CA).
Data cleaning and analysis was performed with Whole-genome Association Study Pipeline (WASP) [51] and PLINK [52]. We excluded individuals in the Affymetrix Genome-Wide Human SNP Array 6.0 analysis based on genotyping quality (<95% call rate) and possible cryptic relatedness, and SNPs based on the following criteria: not in Hardy-Weinberg Equilibrium in controls (p<0.001 chi-squared test), <95% genotype call rate, minor allele frequency (MAF) <0.05, duplicate probes. Our primary analysis considered preterm birth affection status (i.e. delivery <36 weeks) as a binary trait, comparing allele and genotype frequencies between case and control groups by chi-squared test. We also examined gestational age and birth-weight Z-score as quantitative traits, standardized to normal distributions (μ = 0, σ = 1) using a Wald test to compare the mean phenotype between different allele or genotype classes. We corrected for multiple testing using the simpleM method [53], which estimates the number of independent tests, given the LD relationships among SNPs, used to adjust the significance level. Genetic ancestry in the African American population was inferred using STRUCTURE 2.3.1 [54] and the available ancestry informative markers that had been genotyped. Assuming K = 4 with the admixture function on and allowing 10,000 iterations and 10,000 burn-in cycles, genetic ancestry was determined for study samples using unrelated individuals from Hapmap Phase 3 (112 CEU, 113 YRI, and 48 ASW) as learning populations for STRUCTURE.
|
10.1371/journal.ppat.1006157 | The Rice Dynamin-Related Protein OsDRP1E Negatively Regulates Programmed Cell Death by Controlling the Release of Cytochrome c from Mitochondria | Programmed cell death (PCD) mediated by mitochondrial processes has emerged as an important mechanism for plant development and responses to abiotic and biotic stresses. However, the role of translocation of cytochrome c from the mitochondria to the cytosol during PCD remains unclear. Here, we demonstrate that the rice dynamin-related protein 1E (OsDRP1E) negatively regulates PCD by controlling mitochondrial structure and cytochrome c release. We used a map-based cloning strategy to isolate OsDRP1E from the lesion mimic mutant dj-lm and confirmed that the E409V mutation in OsDRP1E causes spontaneous cell death in rice. Pathogen inoculation showed that dj-lm significantly enhances resistance to fungal and bacterial pathogens. Functional analysis of the E409V mutation showed that the mutant protein impairs OsDRP1E self-association and formation of a higher-order complex; this in turn reduces the GTPase activity of OsDRP1E. Furthermore, confocal microscopy showed that the E409V mutation impairs localization of OsDRP1E to the mitochondria. The E409V mutation significantly affects the morphogenesis of cristae in mitochondria and causes the abnormal release of cytochrome c from mitochondria into cytoplasm. Taken together, our results demonstrate that the mitochondria-localized protein OsDRP1E functions as a negative regulator of cytochrome c release and PCD in plants.
| Plants have developed a hypersensitive response (HR) that shows rapid programed cell death (PCD) around the infection site, which in turn limits pathogen invasion and restricts the spread of pathogens. Although many studies reported the characterization of PCD in different pathosystems in the last decade, the molecular mechanisms on how PCD is initiated and how it regulates host resistance are still unclear. Lesion mimic mutants exhibit spontaneous HR-like cell death without pathogen invasion and are ideal genetic materials for dissecting the PCD pathway. In this study, we characterized the lesion mimic gene OsDRP1E that negatively regulates plant PCD through the control of cytochrome c release from mitochondria. Our results suggest that the E409V point mutation in the dynamin-related protein OsDRP1E affects the morphogenesis of mitochondrial cristae that leads to the cytochrome c release into cytoplasm. This study provides new insights into the function of dynamin-related proteins in plant immunity.
| Programmed cell death (PCD) occurs in animals and plants, and the primary forms of PCD in mammals include apoptosis, autophagy, and necrosis [1]. In plants, PCD occurs during normal growth, development, and responses to biotic and abiotic stresses [2]. In plant disease resistance, PCD in the hypersensitive reaction (HR) is accompanied by the accumulation of reactive oxygen species (ROS) [3] and is triggered by the activation of plant resistance proteins after recognition of their corresponding effectors from the pathogen. The HR characteristically involves spontaneous PCD around the infection sites, which limits pathogen invasion and restricts the spread of pathogens [4].
Some mutant plants exhibit spontaneous HR-like cell death even without pathogen invasion. Based on their cell death phenotypes, these mutants were designated lesion mimic or spotted leaf mutants [5]. A number of lesion mimic and spotted leaf mutants have been described in many plant species, including maize [6], Arabidopsis thaliana [7], barley [8], and rice [9,10]. To date, more than 60 lesion mimic genes have been identified in plants [5]. These genes encode proteins that play various regulatory roles in different pathways, such as sphingolipid and fatty acid biosynthesis [11], chloroplast activity and photosynthesis [12], transcriptional regulation [13], signal perception at the plasma membrane [14], Ca2+ signal transduction [15], and ubiquitination-mediated protein degradation [10]. Therefore, various pathways regulate the complex process of PCD in plants.
The mitochondrion, the bioenergy hub of the cell, plays central roles in biochemical pathways for energy production, signal transduction, and cellular metabolism [16,17]. In addition, mitochondria play a major role in the regulation of apoptosis in animals [18]. Cytochrome c plays an important part in this process, serving as one of the first markers of the molecular events preceding apoptosis [19]. During apoptosis, cytochrome c, the sole water-soluble component of the electron transfer chain, is released from the intermembrane space of the mitochondria into the cytosol [20,21]. Cytosolic cytochrome c binds to Apaf-1 to promote the assembly of apoptosomes and recruits procaspase-9 to these complexes, which subsequently initiates an apoptotic protease cascade [22]. Several proteins, including BH3, Bim, and tBid, are involved in the conformational changes to PCD-related proteins such as Bax and Bak, allowing them to form oligomers on the mitochondria. Oligomerized Bax and Bak trigger apoptosis by causing permeabilization of the mitochondrial outer membrane and activation of OMA1 [17].
Several lines of evidence demonstrated that mitochondria also participate in plant PCD [23,24]. The induction of PCD in Arabidopsis cell cultures by ceramide, protoporphyrin IX, or the avirulence factor AvrRpt2 leads to morphological changes in the mitochondria, as well as the release of cytochrome c [25]. A work in Arabidopsis also detected changes in the dynamics and morphology of mitochondria during the onset of cell death [26]. These findings suggest that mitochondria play a role in regulating PCD in plants, but it is still not clear how mitochondrial proteins regulate plant PCD [27]. In recent years, several studies have observed the release of cytochrome c from the mitochondria into the cytosol before plant cell death following toxin protein and elicitor treatments [28–30]. However, the proteins that regulate cytochrome c release during plant PCD are currently unknown.
Dynamin-related proteins (DRPs) and dynamin-like proteins belong to the structurally conserved yet functionally divergent dynamin superfamily. These proteins are commonly found in prokaryotic and eukaryotic organisms including mammals, plants, fungi and bacteria [31,32]. In general, classical dynamin family proteins have five distinct domains: the N-terminal GTPase domain, which binds to guanosine triphosphate (GTP) and hydrolyzes GTP to guanosine diphosphate; the “middle” domain, which is involved in the formation of homo-polymers based on self-interaction; the pleckstrin homology domain, which is related to lipid binding; the GTPase-effector domain, which interacts with the GTPase domain and regulates GTPase activity; and the C-terminal proline-rich domain, which participates in protein–dynamin interactions [31,33,34]. The multi-domain DRPs self-assemble into complex higher-order rings and helices and trigger the fusion or fission of organelles. Studies in Arabidopsis have shown that DRPs play various roles in different pathways: clathrin-dependent endocytosis (DRP1 subfamily) [35], induction of cell death (AtDRP1E during powdery mildew infection) [36], pinching of the clathrin-coated vesicles (DRP2A) [37], vesicular trafficking through the perception of PAMP-triggered immunity (PTI) signaling (DRP2B) [38], and regulated fission of mitochondria and peroxisomes (DRP3) [39] and chloroplasts (DRP5) [40]. However, the roles of rice DRPs remain poorly understood. To date, OsDRP2B is the only DRP that has been shown to regulate cellulose biosynthesis in rice [41,42].
In this study, we characterized a spontaneous lesion mimic mutant, designated dj-lm (dongjin-lesion mimic), which was found among plants of the japonica rice (Oryza sativa) cultivar Dongjin (DJ) grown in our greenhouse. Using map-based cloning, we cloned the mutated gene and found that the cell death phenotype of dj-lm resulted from a point mutation of the rice dynamin-related gene OsDRP1E. This point mutation abolished the self-interaction of OsDRP1Es in yeast, disrupted high-order complex formation in planta and reduced the protein’s intrinsic GTPase activity in vitro. Our results show that the E409 residue is required for the localization of OsDRP1E to the mitochondria, and the point mutation affects the morphology of the mitochondrial cristae and the release of cytochrome c into the cytoplasm, which leads to PCD in rice plants.
Under green house or field conditions, the leaves of dj-lm mutants showed small, dark brown lesions by 30 d to 45 d after germination (Fig 1A). The lesions increased in both quantity and size with the maximum abundance reached at around 2.5 month. Then the cell death lesions gradually covered the entire leaf area, aggravating from the tip to the whole leaf (S1 Fig). After Trypan blue staining, the dj-lm leaves exhibited numerous dark blue spots (Fig 1B), indicating the occurrence of extensive cell death. When we analyzed H2O2 accumulation using 3,3′-diaminobenzidine (DAB) staining, many brownish spots appeared around the lesion sites on dj-lm leaves, whereas almost no brown spots were detected on wild-type DJ leaves (Fig 1C). At the heading stage, dj-lm plants also exhibited a typical senescence phenotype, with withering leaves (Fig 1D). In addition to the cell death and senescence phenotypes, major agronomic traits including plant height, seed setting rate, tiller number, flag leaf angle, 1000-grain weight, and panicle length were affected in the dj-lm plants (S1 Table).
To determine whether the mutation in dj-lm led to enhanced resistance to pathogens, we first inoculated six-week-old DJ and dj-lm plants when the dj-lm plants displayed lesions with the compatible Magnaporthe oryzae isolate RO1-1 using the punch inoculation method. The lesions on dj-lm leaves were approximately one-quarter the size of those on DJ leaves (Fig 1E and 1F). Moreover, the relative fungal biomass on dj-lm was approximately 13% of that on DJ (Fig 1G). Additionally, we tested the disease response of six-week-old rice plants to the bacterial blight pathogen Xanthomonas oryzae pv. oryzae (Xoo) and found that the disease lesions on dj-lm were approximately one-sixth of the length of those on DJ after infection with the Xoo isolate PXO99 (Fig 1H and 1I). These results clearly demonstrate that dj-lm plants have significantly increased broad-spectrum resistance against both M. oryzae and Xoo.
ROS generation occurs as an early event in plant cell death [43]. In this study, we used luminol-based chemiluminescence to detect ROS generation in leaf disks from six-week-old plants as reported previously [44]. In the water control as the mock treatment, ROS levels in dj-lm were approximately twice those of DJ (Fig 2A), which was consistent with the DAB staining results (Fig 1C). Following chitin treatment, the luminol count in dj-lm reached its highest values, roughly 4- to 4.5-fold more than that in DJ, at approximately 10 minutes after chitin application (Fig 2A). However, we could not detect a difference in ROS burst between DJ and dj-lm in plants challenged with flg22 (Fig 2B).
To determine whether the transcription of defense-related genes, such as senescence-associated genes, cell death-related genes and pathogenesis-related (PR) genes, was affected in the mutant, we analyzed the expression of these genes from six-week-old DJ and dj-lm plants using quantitative RT-PCR. Consistent with the enhanced disease resistance of dj-lm, the transcriptional levels of the senescence-associated gene OsI85, the cell death-related genes OskS4 and OsNAC4 and the PR genes PR1a, PR5, and AOS2 were significantly higher in the dj-lm plants than in DJ (Fig 2C).
To isolate the mutant gene that controls the cell death phenotype, we employed a map-based cloning strategy. For the genetic analysis, we crossed dj-lm with wild-type DJ and the indica cultivar 9311. The F1 progenies from the DJ × dj-lm and 9311 × dj-lm crosses did not have any lesions on their leaves, but the F2 populations displayed segregation of the wild-type and lesion mimic phenotypes. The segregation ratio was approximately 3:1 (x2<x20.05 = 3.84, P>0.05) in both populations, suggesting that the phenotype of dj-lm is controlled by a single recessive gene (S2 Table). A total of 3,400 F2 recessive individuals from the 9311 × dj-lm cross were used for DNA marker and phenotype segregation analysis. The phenotypes and genotypes of recombinant individuals were further confirmed in the F3 generation. For the initial mapping, 184 of the 920 pairs of SSR markers from Gramene (http://www.gramene.org) were well distributed on the 12 rice chromosomes and showed polymorphisms between dj-lm and 9311. Linkage analysis with the molecular marker and lesion phenotype data in the 9311 × dj-lm F2 mapping populations delimited the DJ-LM candidate gene to a 101-kb genomic region between the InDel marker ZQ14 and the telomere on the long arm of chromosome 9 (Fig 3A). There are 16 putative open reading frames (ORFs) annotated in this genomic region according to the RGAP website (http://rice.plantbiology.msu.edu/) (Fig 3B). Since ten of these ORFs were annotated as retrotransposons, we focused on the six remaining genes (S3 Table). Because we did not detect any difference in the transcript levels of these genes between DJ and dj-lm (S2 Fig), we further sequenced a 35-kb genomic region spanning these genes and discovered only one A-to-T nucleotide substitution (Fig 3C). This single-nucleotide polymorphism corresponds to the 1226th nucleotide of the ORF within the LOC_Os09g39960 locus (on the 12th exon), resulting in an amino acid change from E to V at the 409th residue of the annotated protein, OsDRP1E, with a molecular weight of 70 kDa (Fig 3D).
To confirm whether this mutation is responsible for the lesion mimic phenotype, we cloned an 11-kb genomic fragment of LOC_Os09g39960, including a 2,937 bp upstream promoter region, a 4,806 bp coding region, and a 3,047 bp downstream region from DJ into the binary vector pCAMBIA1300. The resulting construct, designated 1300-OsDPR1E, was introduced into dj-lm calli through Agrobacterium-mediated transformation. We generated 36 individually transformed T0 lines and grew them in the greenhouse. None of these plants exhibited lesions, unlike the lesion mimic control plants transformed with the empty vector (EV) (Fig 3E, left panel). To verify that the recovery of the wild-type phenotype was indeed due to the reintroduction of 1300-OsDPR1E into the mutant background, we sequenced the mutated OsDRP1E region in the two complemented lines, 1300-OsDRP1E-2 and -7 (Fig 3E, right panel). The sequencing analysis confirmed that the introgression of the wild type OsDRP1E into the mutant background exists in the two complemented lines (Fig 3E, right panel). In addition, we evaluated the disease resistance phenotype and ROS burst in the wild-type DJ, an empty-vector transformed line and two complemented lines. The complemented lines showed the same disease phenotype against M. oryzae and Xoo and ROS burst after chitin treatment as the wild-type DJ (S3 Fig). These results clearly demonstrate that the E409V point mutation causes the lesion mimic phenotype in dj-lm.
Bioinformatics analysis showed that the DJ-LM gene encodes a dynamin-related protein, OsDRP1E, comprising three conserved domains: the N-terminal GTPase domain, the dynamin central region, and the dynamin GTPase effector domain (Fig 3D), as determined based on the annotations at NCBI (http://www.ncbi.nlm.nih.gov/cdd). Transcription analysis using RT-PCR revealed that OsDRP1E was universally expressed in all rice tissues tested, with relatively high expression in roots and leaves (S4A Fig). As the rice plants grew older, the transcription of OsDRP1E generally decreased, but the differences among the investigated growth stages (week 4 to week 14) were not significant (S4B Fig). In addition, the expression level of OsDRP1E was not affected by inoculation with the compatible M. oryzae isolate RO1-1 or the incompatible rice isolate RB22 (S5 Fig).
Phylogenetic analysis of the DRPs from different eukaryotic organisms, including human, yeast, Arabidopsis, and rice, revealed that OsDRP1E belongs to the plant DRP1 subgroup (S6A Fig). Moreover, amino acid sequence alignment of the DRPs from various origins revealed high sequence similarity in the DRP central domain, and it demonstrates that E409 in OsDRP1E is one of the most highly conserved amino acid residues in the proteins analyzed (S6B Fig). These results suggest that the residue E409 in OsDRP1E and other DRPs is structurally and functionally important.
DRPs can form higher-order complexes through self-interaction assembly and the formation of higher-order complexes is a prerequisite for their roles in various cellular processes, such as endocytosis and mitochondrial division [45]. Recent structural studies have revealed that the dynamin central domain plays a vital role in the self-assembly of DRPs [46–48]. We reasoned that the E409V mutation might affect the self-interaction of OsDRP1E based on the observations that the E409 site is located at the self-interaction region and is a highly conserved amino acid residue. To test this hypothesis, we first analyzed the self-interaction of OsDRP1E in the yeast two-hybrid system. As shown in Fig 4A, strong self-interaction was detected in wild-type OsDRP1E, but the mutant OsDRP1E (hereafter referred to as E409V) failed to self-interact in yeast. We then performed native PAGE to examine whether the E409V mutation affects the ability of OsDRP1E to form a higher-order complex. The fusion proteins OsDRP1E-GFP and E409V-GFP were transiently expressed in N. benthamiana through agro-infiltration. The expression levels of OsDRP1E-GFP and E409V-GFP in planta remained similar when analyzed by SDS-PAGE followed by immunoblot detection of GFP (Fig 4B, bottom panel). By contrast, the results of blue native PAGE (BN-PAGE) followed by immunoblot analysis to detect GFP showed that high molecular weight complexes formed from wild-type OsDRP1E in planta, while only dimers or tetramers formed from the E409V mutant protein (Fig 4B, upper panel). These results indicate that the E409 residue of OsDRP1E is required for its self-interaction to form higher-order protein complexes.
DRPs belong to a group of large GTPases with molecular weights above 70 kDa. In contrast to small GTPases, dimerization or higher-order assembly of DPRs promotes the activity of large GTPases and is required for their biological function [47,49]. To determine whether the E409V mutation affects the GTPase activity of OsDRP1E, we examined the in vitro GTPase activity of the purified maltose-binding protein (MBP) fusions MBP-OsDRP1E and MBP-E409V (S7A Fig) using a GTPase colorimetric assay. As shown in Fig 5A, no obvious color change was observed in control reactions with the purified MBP protein and H2O. Although both MBP-OsDRP1E and MBP-E409V displayed catalytic activity towards the substrate GTP, the former had much stronger activity (Fig 5A). The difference in the activity of these two proteins was confirmed in dosage and time-course assays. Clearly, the levels of phosphates released by MBP-OsDRP1E were significantly higher in both a time-dependent (Fig 5B, S7B Fig) and dosage-dependent manner compared to the E409V mutant (Fig 5C).
Online subcellular localization prediction analysis using the program Euk-mPLoc2.0 specific for plant protein [50] (http://www.csbio.sjtu.edu.cn/bioinf/plant-multi/) and RSLpred specific for rice protein [51] (http://www.imtech.res.in/raghava/rslpred/) showed that OsDRP1E is a mitochondrial protein. To determine the subcellular localization of OsDRP1E, we agro-infiltrated OsDRP1E-YFP and E409V-YFP constructs (S8A Fig) into N. benthamiana leaves. Immunoblot analysis using anti-YFP antibody showed similar protein levels of OsDRP1E-YFP and E409V-YFP in agro-infiltrated N. benthamiana leaves (S8B Fig). Confocal microscopy showed spotty, bright fluorescent signals when OsDRP1E-YFP was expressed in N. benthamiana, while the signal from E409V-YFP was distributed evenly in the cytoplasm, resembling the signal of the YFP control (S9A Fig). To exclude the possibility that the C-terminal fusion of YFP might affect the subcellular localization of OsDRP1E, we also investigated the fluorescence patterns of N-terminal YFP-fused OsDRP1E and E409V in planta. As expected, the E409V point mutation abolished the bright speckled signals (S9B Fig).
To verify the subcellular localization of OsDRP1E, we transfected rice protoplasts with the GFP-tagged construct and stained the transfected protoplasts with MitoTracker CMXRos (a mitochondria-specific dye). As shown in Fig 6A, the green fluorescent signals from OsDRP1E-GFP exactly over-lapped the red signals from MitoTracker CMXRos in rice protoplasts. Also, uniform GFP signals were detected in whole cells transfected with E409V-GFP (Fig 6A). We then co-expressed GFP-tagged OsDRP1E and DsRED-tagged COX4, which is a mitochondrial marker protein, in N. benthamiana, and confirmed that the loss of mitochondrial localization was due to the E409V mutation. Similar differences in GFP signals were observed in the N. benthamiana cell that co-expressing OsDRP1E-GFP and E409V-GFP (Fig 6B). Taken together, these results demonstrate that OsDRP1E localizes to the mitochondria and that the E409 residue is essential for the mitochondria-specific localization of OsDRP1E.
The discovery that the E409V point mutation abolished OsDRP1E retention in the mitochondria, together with the finding that the mutated protein showed lower GTPase activity, prompted us to investigate whether mitochondrial morphology was affected by the functional loss of OsDRP1E. We observed the ultrastructure of the mitochondria in mesophyll cells (S10 Fig) from four or eight-week-old DJ and dj-lm plant leaves by transmission electron microscopy. The dj-lm and DJ plants had similar overall number and shapes of mitochondria. However, swelling cristae with vesicle-like structures and reduced intermembrane content were observed in the dj-lm mesophyll cells collected from the first and second leaves in four and eight-week-old plants. The ratio of vesicle-like to normal cristae was approximately from 9–14% in DJ compared to 75–79% in dj-lm in different growth periods (Fig 7), and the vesicle-like structure of cristae is similar to those in previous reports [52,53].
Studies in animals have shown that the release of cytochrome c into the cytoplasm induces caspase activity, ultimately leading to PCD [54]. To explore the association of the cytochrome c levels and the cell death in rice plants, we isolated subcellular fractions from the plants and determined the cytochrome c levels in the cytoplasm and mitochondria using immunoblotting analysis with anti-cytochrome c antibody (Fig 8). We found that the increase of the cytochrome c levels in the cytoplasm as plants aged from four-week old to eight-week-old. Interestingly, the cytosol cytochrome c levels were higher in dj-lm than in DJ no matter the mutant plants exhibited cell death lesions or not and reached its maxim values in the dj-lm plants after lesion appeared (Fig 8). On the contrary, the mitochondrial cytochrome c levels in dj-lm remained similar as those in DJ when plants were eight-week-old. To further confirm that the OsDRP1E is indeed required for the cytochrome c releasing, we compared the cytochrome c level from four-week-old and eight-week-old plants of wild type DJ, empty vector transgenic control (EV) and complemented lines (C2 and C7) (S11 Fig). Similar levels of cytochrome c were observed in the DJ and the complemented lines whereas the empty vector control transgenic line (EV) displayed higher protein level of the cytochrome c in 4-week or 8-week old plants. Taken together, these results demonstrate that the E409V point mutation in OsDRP1E affects the morphology of the mitochondrial cristae and leads to increased release of cytochrome c into the cytoplasm, which may represent the main trigger for the development of the lesion mimic phenotype in dj-lm plants.
In this study, we found that the dj-lm mutant displayed a spontaneous cell death phenotype and enhanced resistance to rice blast and bacterial blight pathogens. Using map-based cloning and a genetic complementation approach, we demonstrated that the dynamin-related protein OsDRP1E is a negative regulator of cell death and that the E409V point mutation in OsDRP1E leads to a lesion mimic phenotype. In the past two decades, dynamin and DRPs have been extensively studied in animal and yeast systems. Mutations in DRPs such as OPA1 [55], Mitofusin2 [56], Atlastin [57], and Drp1 [58] have been identified as the causes of many genetic disorders in humans. For example, a lethal mutation (A395D) located at the central domain of Drp1 causes neonatal death in humans [58]. As Drp1 participates in mitochondrial and peroxisomal fission, a nonfunctional mutation of Drp1 leads to the formation of elongated mitochondria [58]. Likewise, mutations at the other conserved residues, G350 and G362, which are also located in the central domain of Drp1, impair the retention of Drp1 on the mitochondria and lead to the production of elongated mitochondria [45]. From a structural point of view, these conserved amino acids, including E409 of OsDRP1E as well as G350, G362, and A395 of Drp1, map to Interface 3, as recently revealed by resolving the crystal structure of the Dynamin-3 tetramer [59]. Interface 3, together with Interface 1, are required for the assembly of tetramers from Dynamin-3 dimers. A series of single-site mutations within Interface 3 of Dynamin-3 yields dimeric proteins, causing deficient liposome binding and reduced GTPase activity. Similarly, our assays of the self-association of E409V-GFP in planta showed that the majority of the mutant proteins are dimerized, with fewer tetramers compared to the higher-order complexes formed by wild-type OsDRP1E-GFP. These results suggest that the E409V mutation in OsDRP1E might weaken the polar interactions of the negatively charged amino acid, thereby hampering the formation of higher-order complexes in planta. Because the formation of the higher-order complexes promotes the hydrolysis activity of the GTPase domain of dynamin, the E409V mutation affects the GTPase activity of OsDRP1E, as demonstrated in the present study.
In mammals, abnormal mitochondrial cristae are associated with the apoptosis process [60], which in turn causes the release of cytochrome c from the mitochondria into the cytoplasm, followed by the induction of caspase-like activity in the cell, ultimately leading to cell death [61]. Mitochondrial outer membrane permeabilization (MOMP) is a crucial event during apoptosis that leads to the release of cytochrome c. Studies in HeLa cells have demonstrated that Drp1 plays important roles in regulating MOMP and the morphology of mitochondria [61]. In plants, abnormal mitochondrial cristae trigger changes in MOMP and the release of cytochrome c from the mitochondria in early embryonic cells [62]. A previous study revealed the presence of cytochrome c in the cytosolic compartment obtained by subcellular protein fractionation followed by western blot analysis [63]. These events also occur during PCD in plants under abiotic stress [64], and they may disrupt ATP biosynthesis, which is dependent on the normal structure of mitochondria [65]. In the current study, we detected abnormal distribution of cytochrome c between the mitochondria and cytoplasm in dj-lm plants, thus establishing a link between the functional inactivation of OsDRP1E caused by the E409V mutation and mitochondria-mediated cell death, presumably via the release of cytochrome c into the cytoplasm. We hypothesis that vesicle-like cristae with increased spaces between the membranes may allow cristae-localized cytochrome c to flow freely into the inter membrane space and subsequently to cytosol to trigger cell death activation pathway. However, the exact mechanism of cell death caused by the dysfunction of OsDRP1E requires further investigation. Nevertheless, plants harbor a known PCD pathway controlled by Type I meta-caspases such as AtMC1 and AtMC2 [66]. Therefore, we speculate that a PCD pathway exists in plants that is mediated by the mitochondria through a cytochrome c-caspase-like activation pathway (as in animals) and that cytochrome c might act epistatically on the meta-caspases. To determine whether cytochrome c induces cell death in plants, we infiltrated different concentrations of cytochrome c into N. benthamiana leaves and observed phenotypes at different time points after infiltration. We did not see any obvious cell death phenotype in the treated leaves. We speculate that exogenous cytochrome c may be not able to cross the plasma membrane or be inhibited in an inexplicit mechanism in plant cells to activate the caspase-like pathway as in animal cells. Therefore, more research is needed to investigate the function of cytochrome c in plant PCD.
In Arabidopsis, DRP3A and DRP3B are the closest homologs to human Drp1 and are functionally redundant during mitochondrial fission. The null mutants drp3a and drp3b-1 have mitochondria that are slightly longer than those of wild-type plants, while drp3a/drp3b-1 double mutant has mitochondria that form an extremely elongated, interconnected network structure [67]. However, in this study, we did not obtain direct evidence that OsDRP1E is responsible for mitochondrial fission, as no elongated mitochondria were found in dj-lm. This result might be due to the functional redundancy between OsDRP1E and OsDRP1C or OsDRP1D, as phylogenetic analysis revealed that these three DRPs belong to a close clade. Nevertheless, the observation of an abnormally high percentage of bubble-like structures of mitochondrial cristae in dj-lm plants, together with the loss of mitochondrial localization of the mutant protein OsDRP1E-E409V, strongly suggest that OsDRP1E participates in the maintenance of mitochondrial membrane structures.
The functionally divergent DRPs can target various organelles in plants [36,66]. These studies helped reveal the functional divergence of this group of multifaceted proteins. Using confocal microscopy, we demonstrated that OsDRP1E is localized to mitochondria based on the following observations. First, OsDRP1E tagged with YFP or GFP at either the N- or C-terminus produced similar speckled patterns in N. benthamiana, while YFP- and GFP-tagged E409V mutant proteins did not produce these specific, speckled patterns. Second, when we used the mitochondria-specific dye MitoTracker or the marker protein COX4 as an indicator, the florescent signals from OsDRP1E-GFP co-localized with these mitochondrial markers. The observation that E409V leads to targeting of the protein to the cytosol might be due to the inability of OsDRP1E to form polymers at the mitochondrial membrane. Finally, the differences in localization patterns between OsDRP1E and E409V matched the functional consequences of the mutation, as revealed by the changes in mitochondrial morphology observed by transmission electron microscopy. However, it is currently unclear whether the targeting of OsDRP1E to the mitochondria occurs via recruitment by unknown adapter proteins or through direct association with the mitochondrial membrane. Therefore, identifying and characterizing interacting proteins of OsDRP1E will provide new insights into the OsDRP1E-mediated regulation of PCD in rice.
Apart from physical barriers, ROS burst is the first layer of defense in plant PTI signaling [68]. Mitochondria play an important role in mediating the balance of ROS levels in plant cells. Because OsDRP1E may participate in the maintenance of mitochondrial membrane structures, mutation of the gene might cause structure changes and elevated ROS levels. Indeed, we detected higher ROS levels in dj-lm plants than in wild type, even in the absence of any treatment, suggesting the existence of basal-level activation of the defense pathway in dj-lm. Interestingly, the ROS levels were significantly higher in dj-lm plants after chitin treatment, while there was no difference between the ROS bursts detected in DJ and dj-lm after flg22 treatment. These results suggest that OsDRP1E-induced ROS generation is limited to the chitin-signaling pathway.
Most lesion mimic mutants display enhanced disease resistance [23,69] and significant up regulation of defense-related genes such as: PR1a and PR5, marker genes associated with defense-related responses in rice [70,71]; OsKS4 and AOS2, encoding important biosynthetic enzymes in the phytoalexin and jasmonic acid biosynthesis pathways, respectively [72,73]; OsNAC4, encoding a protein that participates in the induction of HR cell death and may regulate the transcription of multiple genes, including OsHSP90 and IREN [74], and Osl85, a senescence-associated gene that functions in fatty acid metabolism [75]. In summary, the higher transcript levels of these defense-related genes and senescence-associated genes correlate well with the enhanced resistance to rice blast and bacterial blight pathogens and the senescence phenotypes observed in dj-lm plants.
Rice cultivars DJ (Oryza sativa ssp. japonica) and 9311 (Oryza sativa ssp. indica) were used in this study. Rice plants were cultured in a growth chamber at 26/22°C under a 14 h light/10 h dark cycle or in a paddy field on our experimental farm in June through October. Agronomic traits of rice plants grown in the paddy field were measured, including plant height, seed setting rate, tiller number, flag leaf angle, 1000-grain weight, and panicle length.
Leaves from dj-lm plants containing lesions and leaves from DJ at the same growth stage (eight-week-old) were submerged in lactic acid-phenol-Trypan blue solution (0.25% Trypan blue, 25% lactic acid, 23% water-saturated phenol, and 25% glycerol) for staining, as previously described [69]. Briefly, the leaf samples were incubated in a boiling water bath for 10 min, cooled to room temperature and incubated in the Trypan blue staining solution supplemented with chloral hydrate (0.25%) for 48 h.
DAB staining was used to detect H2O2 accumulation in the leaves as described previously [76]. Briefly, the leaves of eight-week-old rice plants were submerged in DAB solution (0.1%, pH3.8) at 26°C for 8 h in the light. After draining off the DAB solution, the leaves were boiled for 10 min in a water bath containing 95% ethanol for destaining, followed by incubation in 95% ethanol at room temperature.
The leaves were subjected to punch inoculation to measure the rice blast resistance of DJ and dj-lm plants using M. oryzae isolate RO1-1, as previously described [44]. Briefly, six to eight week-old leaves were lightly wounded using a mouse ear puncher, and 7 μl of spore suspension (5×105 spore· ml-1) was added to the wound site, which was then sealed in a small chamber with transparent tape. The inoculated plants were incubated in the dark for 24 h in a growth room at 28°C with 100% relative humidity, and then moved to a growth chamber at 26/22°C under a 14 h light/10 h dark cycle with 80% relative humidity. Disease symptoms and fungal biomass in the infected leaves were surveyed 7 d after inoculation. The fungal biomass in the infected leaf tissue was quantified using the method was described in a previous study [44]. Briefly, the infected rice tissue about 3 × 1cm was cut for DNA extraction using the CTAB method. After RNase A treatment, DNA- based qPCR was performed using Bio-Rad iQ2 PCR system (Bio-Rad). The threshold cycle value (CT) of M. oryzae Pot2 gene against the CT of rice Os-Ubq gene was used to calculate the relative fungal biomass in rice leaves. The CT of Os-Ubq was subtracted from the CT of Pot2, and then, using the equation ECT (Os-UBQ)–CT (Mo-Pot2) that represents the ratio of (Mo-Pot2/Os-Ubq) to calculate the relative fungal biomass, in which the amplification efficiency, E, is 2 for the primer pairs designed for the respective genes.
The leaf-clipping method was used to measure the bacterial blight resistance of DJ and dj-lm plants using the isolate PXO-99 in greenhouse-grown plants as described previously [77]. Briefly, tips of the top-two fully expanded leaves of eight-week-old DJ and dj-lm which showed lesion mimics were cut with scissors and inoculated with Xoo isolate PXO99 solution (OD595 = 0.5). The inoculated plants were moved to greenhouse at 28°C, 12/12 h light/dark photoperiod. The lesion length was measured at 14 d after inoculation.
Leaf disks were excised from the fully expanded leaves (the second or third leaf from the top) of six to eight-week-old plants using an ear-hole puncher and floated on sterile distilled water overnight. Three leaf disks were placed in a 1.5 ml microcentrifuge tube containing 100 μl of luminol (Bio-Rad Immun-star horseradish peroxidase substrate 170–5040), 1μl of horseradish peroxidase (Jackson Immuno Research) and 100 nM flg22 or 8 nM hexa-N-acetyl-chitohexaose, with sterile distilled water for the control. The tube was immediately placed in a Glomax 20/20 luminometer (Promega) and the luminescence was recorded at 15 s intervals for 30 min.
Total RNA was extracted with Trizol reagent (Invitrogen) according to the manufacturer’s protocol. After DNaseI treatment, 2 μg of RNA was added to a 20 μl reaction system to synthesize first-strand cDNA using the Reverse Transcription System (Promega) according to the manufacturer’s instructions. Using 1.0 μl of 1:10 diluted cDNA as template, PCR was performed in a 20 μl reaction volume with Bio-Rod SYBRII Super-Mix buffer on a Bio-Rad iQ2 PCR system (Bio-Rad). The rice actin gene was used as the internal control. Gene-specific primers for PCR are listed in S4 and S5 Tables.
Genetic analysis was performed using 133 individuals from the F2 population of the dj-lm × DJ cross and 126 individuals from the F2 population of the dj-lm × 9311 cross. F2 recessive individuals from the dj-lm × 9311 cross were used for DNA marker and phenotype segregation analyses. The phenotype and genotype of each recombinant individual was confirmed in the F3 generation. For the initial mapping, SSR markers from Gramene (http://www.gramene.org) were used for linkage analysis. For fine-mapping of the candidate dj-lm mutant gene, InDel markers were developed based on the sequence differences between the japonica variety NPB (http://rgp.dna.affrc.go.jp/) and the indica variety 9311 (http://rise2.genomics.org.cn/page/rice/index.jsp). The primers used for fine mapping are listed in S4 Table. The PCR products were separated by electrophoresis in 8% polyacrylamide gels or 3% agarose gels depending on the amplicon size. For complementation tests, the wild-type DJ-LM genomic DNA fragment was cloned into binary vector pCAMBIA1300. This derivative construct or the empty vector was mobilized into Agrobacterium stain EHA105 by electroporation and used to transform the dj-lm mutant. The transformants were grown in a growth chamber for phenotypic and genotypic investigations.
Alignment of the DRP amino acid sequences was performed using CLUSTAL W with DRP amino acid sequences obtained from NCBI (blast.ncbi.nlm.nih.gov/Blast.cgi).The phylogenetic trees were constructed by the neighbor joining method [78] using MEGA 6.06 software.
The ProQuest yeast two-hybrid system (Invitrogen) was used to screen the OsDRP1E-interacting proteins according to the product manual. The coding sequence of OsDRP1E and OsDRP1E (E409V) were cloned into bait vector pDBleu and prey vector pPC86, respectively. The bait and prey vectors were co-transformed into yeast strain MAV203 and the transformants were selected on synthetic dextrose medium without Leu and Trp (SD-Leu-Trp). The single transformed yeast was subjected to 10-fold serial dilutions and plated on SD-Leu-Trp-His medium including 0 or 40 mM 3-amino-1, 2, 4-triazde (3AT, Sigma-Aldrich). Three independent experiments were performed, and positive clones on the SD-Leu-Trp-His plates were stained with 2.5 mM X-gal to detect β-galactosidase activity.
The fusion constructs of OsDRP1E and E409V with GFP or YFP were transformed into Agrobacterium strain EHA105 via electroporation. Six-week-old N. benthamiana leaves were infiltrated with EHA105 transformants containing the appropriate constructs as described previously [44]. After infiltration for 48 h, the leaf samples were collected for confocal microscopy and immunoblot analysis. Confocal microscopy was performed using a Zeiss LSM710 confocal laser-scanning microscope. For immunoblot analysis, 100 mg fresh N. benthamiana leaf samples were finely ground in liquid nitrogen and combined with 100 μl 2 × loading buffer (10% glycerol, 50 mM Tris-Cl [pH6.8], 2% β-mercaptoethanol, 0.02% bromophenol blue, 2% SDS). After boiling for 5 min in a water bath and centrifugation for 5 min at 13,000 rpm at room temperature, 15 μl of the supernatant was loaded onto an SDS-PAGE gel for immunoblot analysis using anti-GFP and anti-YFP antibody (1:5000 anti-GFP/anti-YFP dilution, Roche).
Blue native PAGE was performed as described previously [64]. Briefly, 100 mg samples of fresh of N. benthamiana leaves that transiently expressed the OsDRP1E-GFP or E409V-GFP fusion proteins were ground in liquid nitrogen using a mortar and pestle. A Native PAGE Sample Prep Kit (Invitrogen) was used to isolate the native tobacco proteins. The ground samples were combined with 400 μl 1 × Tris-buffered saline buffer (2% Triton X-100), vortexed, and incubated on ice for 30 min. The homogenates were centrifuged twice at 17,000 g, 4°C, 20 min per centrifugation. Then, 25 μl of the supernatant was transferred to a new tube, combined with 3 μl of 5% Coomassie Brilliant Blue G250 and separated in a 4 to 16% native PAGE gel (Invitrogen) according to the manual. Immunoblot analysis was performed using anti-GFP antibody (1:5000). Chemiluminescence was detected using an Image Quant LAS 4000.
Fusion constructs of MBP-OsDRP1E and MBP-E409Vwere transformed into E. coli BL21 (DE3) for protein expression. BL21 was grown at 28°C to OD600 of 0.6, and IPTG was added to a final concentration of 2 mM, followed by incubation for 6–8 h at 28°C. The cell pellets were harvested by centrifugation for 15 min at 5000 g, 4°C and resuspended in 1/10 volume bacterial culture in 0.5 M Tris-HCl buffer (pH8.0) containing protease inhibitor cocktail (Roche). After sonication, the lysates were centrifuged for 10 min at 15,000 g, 4°C. Protein purification was performed using an Amylose resin (NEB, E8021) column according to the product manual. GTPase activity was determined using a GTPase Assay Kit (Innova Biosciences, 602–0120). The enzyme activity was determined based on the amount of phosphate released during GTP hydrolysis, which was calculated according to a phosphate standard curve prepared using the 0.1 mM phosphate stock included in the GTPase Assay Kit.
Rice protoplast isolation and transfection were performed as described previously [79]. Briefly, 1 μg of plasmid was transfected into rice protoplasts using the polyethylene glycol 4000 (PEG4000)-mediated transfection method. MitoTracker staining was performed according to the product manual. Briefly, the protoplasts were incubated in 200 nM MitoTracker CMXRos (Invitrogen) in W5 buffer (154 mM NaCl, 125 mM CaCl2, 5 mM KCl, 2 mM MES, pH5.6) for 30 min at room temperature and washed three times in W5 buffer. Fluorescence images were taken under a Zeiss LSM710 confocal laser-scanning microscope at 559 nm excitation and 560 nm emission.
Four-week-old and eight-week-old rice leaves (same region in the first and second leaves from the top and three biological leaf samples were taken) were cut into 1 mm2 sections and submerged in 2.5% glutaraldehyde in sodium phosphate buffer (pH7.2) for 4 h at 4°C. The samples were prepared as previously described [24]. The images were observed under a transmission electron microscope (H-7650B, Hitachi LTD). To quantify the abnormal mitochondria in the mutant and WT plants, we counted approximate thirty mitochondria in each sample. A mitochondrion is considered abnormal when the ratio of vesicle-like to normal cristae is more than 50% and a mitochondrion is considered normal when the ratio of vesicle-like to normal cristae is less than 50%.
The mitochondria were isolated using the plant mitochondrial extraction kit (Biohao Biotechnology Co. #P0045) according to the product manual. Briefly, 200 mg samples of fresh rice leaves were ground in liquid nitrogen using a mortar and pestle. The ground samples were combined with 1,000 μl of cold lysis buffer (0.5%β-mercaptoethanol), and vortexed. The homogenates were centrifuged for 10 min at 1,000 g, 4°C. The supernatant, containing cytoplasmic and mitochondrial proteins, was transferred to a new tube and centrifuged for 10 min at 16,000 g at 4°C. The supernatant was collected as the cytoplasmic protein fraction, and the pellet was washed in 500 μl washing buffer and was centrifuged for 5 min at 1,000 g at 4°C. The supernatant from the washing buffer was centrifuged for 10 min at 16,000 g at 4°C. The mitochondrial protein fraction pellet was dissolved using 100 μl store buffer. The cytoplasmic and mitochondrial protein fractions were mixed with the same volume of 2 × loading buffer, respectively, and incubated in a boiling water bath for 5 min. The mixed samples were loaded onto a SDS-PAGE gel for immunoblot analysis using the following antibodies at the appropriate dilutions: 1:8000 anti-cytochrome c, 1:6000 anti-VDAC and 1:6000 anti-HSP90 (Agrisera). Chemiluminescence was detected using an Image Quant LAS 4000 and the software Image J was used to measure the relative protein levels.
The OsDRP1E/E409V-GFP, OsDRP1E/E409V-YFP and YFP-OsDRP1E/E409V fusion constructs were generated for the subcellular localization and BN-PAGE experiments. The full-length cDNAs of OsDRP1E and OsDRP1E (E409V) containing an ORF without the stop codon were amplified with primers 1-F/R including the SmaI and KpnI restriction sites, and the PCR product was inserted into pYBA-1132 (-GFP), pYBA-1155 (-YFP), and pYBA-1135 (YFP-), respectively, after double digestion with SmaI and KpnI. For the yeast two-hybrid interaction assay, the coding sequence of OsDRP1E and OsDRP1E (E409V), including the SmaI and SpeI restriction sites, were amplified with primers 1-F/primer2-R. The PCR product was inserted into bait vector pDBleu or prey vector pPC86, respectively, after double digestion with SmaI and SpeI. For the GTPase activity assay, the coding sequence of OsDRP1E and OsDRP1E (E409V), including the BamHI and SalI restriction sites, were amplified with primer3-F/R, and the PCR product was inserted into vector pMalC2 after double digestion with BamHI and SalI. The fusion plasmid was transformed into E. coli strain DE3 for the OsDRP1E protein GTPase activity assay. For the complementation test, the 5’ terminal portion and 3’ terminal portion of the OsDRP1E genomic fragments were amplified using primers BQ5-2R-kpnI/BQ4-1F and BQ4-3F/BQ3-1F-SalI, respectively. The fragments with the correct sequences were sub cloned (via two steps) into pCAMBIA1300 using a combination of KpnI/SalI and KpnI digestion. The primer information is listed in S6 Table.
Sequence data from this work can be found in the Rice Genome Annotation Project or GenBank database under the following accession numbers and GI numbers: OsDRP1A (AK065908), OsDRP1B (AK072230), OsDRP1C (AK061703), OsDRP1D (AK073186), OsDRP1E (AK069270), OsDRP2A (AK102187), OsDRP2C (AK069134), OsDRP3A (AK073965), OsDRP3B (AK105435), OsDRP3C (AK111167) from rice; AtDRP1A (NP_851120), AtDRP1B (NP_191735), AtDRP1C (NP_172936), AtDRP1D (NP_850420), AtDRP1E (NP_567094) from Arabidopsis; Drp1 (NP_036193) from H. sapiens; Dlp2 (AAF51235) from D. melanogaster; Drp1 (AAL56621) from Caenorhabditis elegans; Dlp (Q09748) from Schizosaccharomyces pombe; Dnm (AAA99998) from Saccharomyces cerevisiae. PYBA-1132 (KF876796); pYBA-1135 (KF876799); pYBA-1155 (KF876807).
|
10.1371/journal.ppat.1003207 | Cooperativity Between CD8+ T Cells, Non-Neutralizing Antibodies, and Alveolar Macrophages Is Important for Heterosubtypic Influenza Virus Immunity | Seasonal epidemics of influenza virus result in ∼36,000 deaths annually in the United States. Current vaccines against influenza virus elicit an antibody response specific for the envelope glycoproteins. However, high mutation rates result in the emergence of new viral serotypes, which elude neutralization by preexisting antibodies. T lymphocytes have been reported to be capable of mediating heterosubtypic protection through recognition of internal, more conserved, influenza virus proteins. Here, we demonstrate using a recombinant influenza virus expressing the LCMV GP33-41 epitope that influenza virus-specific CD8+ T cells and virus-specific non-neutralizing antibodies each are relatively ineffective at conferring heterosubtypic protective immunity alone. However, when combined virus-specific CD8 T cells and non-neutralizing antibodies cooperatively elicit robust protective immunity. This synergistic improvement in protective immunity is dependent, at least in part, on alveolar macrophages and/or other lung phagocytes. Overall, our studies suggest that an influenza vaccine capable of eliciting both CD8+ T cells and antibodies specific for highly conserved influenza proteins may be able to provide heterosubtypic protection in humans, and act as the basis for a potential “universal” vaccine.
| Influenza virus continues to pose a significant risk to global health and is responsible for thousands of deaths each year in the United States. This threat is largely due to the ability of the influenza virus to undergo rapid changes, allowing it to escape from immune responses elicited by previous infections or vaccinations. Certain internal determinants of the influenza virus are largely conserved across different viral strains and represent attractive targets for potential “universal” influenza vaccines. Here, we demonstrated that cross-subtype protection against the influenza virus could be obtained through simultaneous priming of multiple arms of the immune response against conserved elements of the influenza virus. These results suggest a novel strategy that could potentially form a primary component of a universal influenza vaccine capable of providing long-lasting protection.
| Influenza virus remains a significant threat to global health, and results in 200,000 hospitalizations and 3,000–49,000 deaths each year in the United States [1]–[3]. The ability of influenza virus to rapidly mutate and/or undergo reassortment, allows the virus to evade protective immunity obtained from previous infections or vaccinations [4]. Annual influenza vaccines induce an antibody response specific for the highly variable surface glycoproteins of influenza: neuraminidase (NA) and hemagglutinin (HA). These seasonal vaccines typically take months to produce and rely on the accurate prediction of the influenza serotypes that will be circulating in the next flu season [5]. Thus, if the prediction is not accurate or a pandemic strain emerges, current vaccines offer little protection. Much research has therefore, focused on the development of a “universal” vaccine that will target the conserved, internal regions of the influenza virus, and confer protection against multiple influenza virus serotypes.
Significant research in the influenza field has focused on the design of vaccines capable of eliciting influenza virus-specific CD8+ T lymphocytes [6]–[8]. Since, CD8+ T cells are able to recognize internal, conserved regions of the influenza virus, these cells may be able provide cross-subtype, or heterosubtypic, protection against the influenza virus [9]–[13]. A large body of work supports the potential viability of a CD8+ T-cell based vaccine [14]–[17]. In mice, vaccination with internal proteins such as influenza nucleoprotein (NP), leads to higher frequencies of NP-specific CD8+ T cells prior to infection, and lower viral titers after challenge with H1N1 and H3N2 strains of influenza [18]–[22]. Furthermore, influenza virus-specific memory T cells are detected in the peripheral blood of healthy adolescents and adults, and there is some evidence for heterosubtypic immunity in humans that has been proposed to be due to T cells [23]–[25].
However, several groups have reported that the number of influenza virus-specific memory CD8+ T cells in the lung airways of mice declines over time corresponding with a loss of heterosubtypic protection [26]–[29]. While there is some conflicting data over whether heterosubtypic protection wanes, if true this gradual loss of the virus-specific CD8+ T cell population represents a serious concern in the generation of CD8+ T cell based vaccines [30]. Interestingly, recent work suggests that non-neutralizing antibodies targeting the internal proteins of influenza, specifically NP, can provide some protection against the influenza virus infection through a mechanism involving Fc receptors [31]–[34]. Unlike neutralizing antibodies, which are able to prevent viral entry or exit, non-neutralizing antibodies typically target antigens that reside inside virions and/or infected cells. Despite recent progress, it is still not clear what role T cells play in non neutralizing antibody-mediated heterosubtypic protection elicited in an immune competent host. Additionally the mechanisms by which non-neutralizing antibodies can provide protection against the influenza virus remain elusive.
Several groups have also noted the potential role of CD4+ T cells in providing protection against influenza virus [6], [35], [36]. These reports indicated that CD4+ T cells can form a lung-resident population following influenza virus infection where they can serve a protective role in mediating enhanced viral clearance and survival following lethal challenge through a variety of mechanisms including IFNγ secretion [35], [37]. Recent studies using human volunteers infected with influenza virus also point to a key role for pre-existing CD4 T cell responses in limiting the severity of influenza virus infection and disease [38]. Intriguingly, influenza-specific memory CD4+ T cells have also been reported to synergize with naïve B cells and CD8+ T cells to provide protection against influenza viral infection [36]. Whether virus-specific CD8 T cells also exhibit such cooperativity in protective immunity is unclear.
In this study, we demonstrate that, in most settings, influenza virus-specific CD8+ T cells alone are insufficient to provide optimal protection against influenza virus. However, when virus-specific non-neutralizing antibodies are present together with virus-specific CD8+ T cells, complete protection is achieved against a lethal influenza virus challenge. Moreover, this cooperative protection is dependent, at least in part, on the presence of alveolar macrophages (AM) or other respiratory tract phagocytes, suggesting that non-neutralizing antibodies are able to eliminate influenza virus-infected cells through antibody-dependent cell-mediated cytotoxicity (ADCC) and/or phagocytosis. We demonstrate a novel mechanism by which antibodies and CD8+ T cells targeted against the conserved regions of the influenza virus act in concert to provide heterosubtypic protection. Our results complement recent work on the synergy between memory CD4+ T cells and naïve B and CD8 T cells [36] and suggest that elicitating multiple arms of the adaptive immune response may represent a potent mechanism by which heterosubtypic protection against the influenza virus can be achieved.
It has been reported that CD8+ T cell activity correlates with reduced influenza virus shedding following rechallenge [13]. Since, CD8+ T cell epitopes are often located in the internal, conserved regions of the influenza virus, the generation of influenza virus-specific CD8+ T cells may provide protective immunity against heterosubtypic influenza strains. Thus, we tested whether influenza virus-specific CD8+ T cells could mediate protective immunity using a recombinant viral approach to identify and track responses. We used influenza viruses in which the GP33-41 epitope from lymphocytic choriomeningitis virus (LCMV), was inserted into the NA stalk region of the H3N2 influenza X31 (X31-GP33) and H1N1 influenza PR8 (PR8-GP33) viruses [39], [40]. Influenza viruses expressing the GP33 epitope have been shown to induce a robust GP33 response in mice [39], [41]. Mice were primed with either LCMV Armstrong or X31-GP33 intranasally (i.n.) and rechallenged, along with a control group of naïve mice, with PR8-GP33 30 days later. The antibodies generated against the surface glycoproteins of the H3N2 X31-GP33 virus do not neutralize the H1N1 PR8-GP33 challenge virus [42]–[44]. The GP33-specific CD8+ T cell population elicited from primary viral challenge, however, should be capable of responding to the secondary infection, allowing the role of CD8+ T cells in protection against the influenza virus to be investigated.
The level of protection conferred upon secondary challenge was determined using three assays. Morbidity was assessed by weight loss. Pulse oximetry was also used to evaluate lung function. Finally, real time quantitative (qRT-PCR) was used to detect viral RNA and determine viral load at several time points following infection. X31-GP33 primed mice were completely protected from influenza rechallenge by all 3 measures (Fig. 1A). These mice experienced almost no impairment of lung function or loss in weight, and had low viral load at all time points measured. In contrast, the LCMV Armstrong immunized mice, despite generating a robust GP33-specific CD8+ T cell response, exhibited little if any protection from the PR8-GP33 rechallenge. Apart from a slight delay in compromised lung function, the LCMV Armstrong immune mice were indistinguishable from the naïve group and experienced a 25% decline in weight and lung function by day 9 post rechallenge. To determine if the difference in protection between the X31-GP33 and LCMV Armstrong immune groups was due to the X31-GP33 immune mice having a larger influenza virus-specific CD8+ T cell response, we quantified the total immune response in these mice 6 days after rechallenge in the lung, bronchoalveolar lavage (BAL), and spleen. Responses were analyzed using intracellular staining to evaluate the number of cells in these mice able to produce interferon gamma (IFNγ) in response to stimulation with overlapping peptide pools for the influenza virus proteins HA, NA, non-structural protein 1 (NS1), NS2, polymerase acidic (PA), polymerase basic (PB), NP, as well as the LCMV GP33 peptide. We found that the LCMV Armstrong immune mice had a similar or slightly larger antiviral CD8+ T cell response directed against the recombinant influenza virus following rechallenge in both the lung and BAL despite lack of protection (Fig. 1B). Similar results were obtained using mice immunized intraperitoneally (i.p.) with LCMV Armstrong (Fig. S1A, S1B, S1C), though LCMV i.n. immunized mice had slightly enhanced viral control compared to the LCMV i.p. primed mice (Fig. S2). Overall, while the route of priming may have some impact, these results indicate that the magnitude of the virus-specific CD8+ T cell response alone might not be a major determinate of protection against influenza viral challenge and suggest that other factors could be responsible for protection against influenza virus in X31-GP33 immune mice.
We first sought to evaluate whether the agent used to prime the mice could have an impact on protection in our system. The specific priming agent used has been reported to confer differences in protection against influenza virus in several vaccine studies [45], [46]. Thus, mice were primed with LCMV Armstrong, Listeria-GP33 (LM-GP33), or Vaccinia-GP33. Each group of mice had a similar GP33-specific CD8+ T cell population despite being immunized with different bacterial or viral agents. We found that regardless of the priming agent used, all groups experienced severe weight loss, decline in lung function, and high viral load following rechallenge with PR8-GP33 (data not shown). Thus, in this setting, the priming agent did not have an obvious direct correlation with whether or not protection was achieved.
We next examined whether the epitope against which the CD8+ T cells were primed impacted protection. Different epitopes have been shown to elicit varying levels of protection to the influenza virus [47]. We therefore tested whether priming with another CD8 epitope shared between X31-GP33 and PR8-GP33, the dominant Db-restricted NP366-374 (NP366) epitope from influenza virus nucleoprotein, could elicit better protection than the GP33 response. Mice were primed with recombinant viruses (or bacteria) expressing GP33, NP366, or a non-influenza determinant (the LCMV nucleoprotein), and challenged 30 days later. None of these approaches achieved substantial protection against PR8-GP33 rechallenge, as each group had substantial weight loss, high viral load and reduced lung function (Fig. 2A). Thus, at least for the determinants examined, the specific epitope was not a major factor in the lack of protection observed in this model system.
Next, we used a prime-boost strategy to test whether the more robust immune response induced upon boosting was superior in providing protection compared to non-boosted memory CD8+ T cells. Mice were immunized with LM-GP33 and then boosted with LCMV Armstrong 30 days after initial priming. These mice were then rechallenged with PR8-GP33 either 8 days or 30 days after the boost. We also immunized a group of mice with LCMV Armstrong and rechallenged with PR8-GP33 8 days later. The mice rechallenged with influenza virus 8 days after initial priming displayed delayed morbidity with the initiation of weight loss on ∼day 6 post rechallenge rather than at ∼day 2–3 in naïve mice (Fig. 2B). This transient delay in weight loss suggested that GP33-specific effector CD8+ T cell response present at 8 days after acute LCMV infection was capable of providing some initial protection following influenza virus infection. This group, however, still lost significant weight by day 9 post rechallenge and had reduced lung function as well as high viral load. It is worth noting that the dose of LCMV Armstrong used here has been demonstrated in our lab and others to be cleared by day 8 [48]. The prime boost group that was rechallenged 30 days after the “boost” did not experience this transient delay in weight loss and showed kinetics of morbidity similar to the LCMV Armstrong immune groups described above in terms of magnitude of weight loss and decline of lung function suggesting that the greater magnitude of GP33-specific CD8+ T cell response in this group was insufficient to mediate protection. In contrast, the prime-boost group that was rechallenged 8 days after the “boost” showed no signs of influenza-related pathology in terms of weight loss and lung function. Despite this lack of morbidity these mice still had very high viral loads until day 9 post rechallenge (Fig. 2B). While priming (and the prime-boost regimen) was able to induce a robust population of GP33-specific response capable of producing IFNγ and tumor necrosis factor (TNFα) in response to stimulation (Fig. S3), this response was still insufficient to mediate viral clearance. Thus, although this prime boost group is similar to the X31-GP33 group in terms of weight loss and lung function following rechallenge, viral control was relatively poor.
The protection from morbidity in the mice challenged 8 days after boosting was not due simply to elevated bystander inflammation as mice subjected to two other prime-boost strategies that lacked CD8+ T cells specific for influenza virus showed rapid weight loss (Fig. 2C). Furthermore, the lack of weight loss achieved by the virus-specific prime-boost strategy was recapitulated when LM-GP33 was substituted with VV-GP33 (Fig. 2C), suggesting that the lack of weight loss in these mice is not dependent on the identity of the priming agent, but on the rapid initiation of an influenza virus-specific CD8+ T cell response. These results are in line with reports that “boosted” memory CD8+ T cells are better than primary memory CD8+ T cells in controlling some acute infections [49], [50]. However, while pathology was reduced, this immune response was not sufficient to efficiently control viral load.
We next investigated whether we could achieve enhanced CD8+ T cell-mediated protection using adoptive transfer strategies analogous to adoptive transfer approaches used for influenza virus-specific CD4+ T cells [35], [36]. Ly5.1+ mice were immunized with LM-GP33 and 30 days later boosted with LCMV Armstrong. On day 8 following the boost, CD8+ T cells were isolated and 0.8×106, or 1.6×106 GP33+ CD8+ T cells were adoptively transferred to Ly5.2+ naïve mice. The recipient mice were rechallenged with PR8-GP33 the next day. Little to no protection was observed as measured by weight loss and viral load compared to a PBS-treated control group (Fig. 3A–B), despite high numbers of GP33 specific CD8+ T cells present in the lungs of these mice 6 days after rechallenge (Fig. 3C).
To determine whether the observed lack of protection was due to an insufficient number of in vivo primed adoptively transferred GP33-specific CD8+ T cells, we primed P14 TCR-transgenic CD8+ T cells (specific for LCMV GP33-41) using an in vitro approach that allowed the generation of large numbers of activated GP33-specific CD8+ T cells. We then adoptively transferred 2×106, 10×106 or 20×106 GP33-specific CD8+ T cells into naïve mice and challenged these mice with influenza virus. Mice given 20×106 or 10×106 GP33-specific CD8+ T cells were almost completely protected from influenza-related morbidity and experienced virtually no decline in weight or lung function (Fig. 3D). Additionally, even the group given 2×106 GP33-specific CD8+ T cells exhibited improved protection with only a 15% decline in weight and a 20% reduction in lung function. Despite the greatly reduced morbidity in these mice, the viral loads measured by qRT-PCR for viral RNA were almost indistinguishable from the PBS-treated control group. This lack of difference in viral control was confirmed using an assay for infectious virus to ensure that qRT-PCR-based approach was accurately reflecting replicating virus rather than residual viral debris or viral RNA independent of replicating virus (Fig. S4). Thus, similar to the prime-boost group described above (Fig. 2C), the immune response elicited by the adoptively transferred, in-vitro generated effector CD8+ T cells was insufficient to reduce viral load despite the improvement in measures of morbidity.
Overall, these results indicate that both in vivo and in vitro generated GP33-specific CD8+ T cells alone were insufficient to provide optimal protection against a pathogenic influenza virus challenge. While CD8+ T cells in large enough numbers are able to provide some protection as measured by weight loss and lung function, they are unable to significantly reduce viral load. Moreover, our results suggest that the mechanism of cross-subtype protection in X31-GP33 immune mice is unlikely to be exclusively CD8+ T cell-dependent.
To evaluate the mechanism of cross-subtype protection in X31-GP33 primed mice, we next examined the dependence of protection in this setting on T cells (Fig. 4). CD8+ and/or CD4+ T cells were depleted from X31-GP33 immune mice prior to PR8-GP33 rechallenge with depletion being verified as >97% in the lungs. In all cases, depletion of CD8+ and/or CD4+ T cells did not increase the severity of weight loss. There was a slight but non-significant decrease in oxygen saturation following the depletion of CD8+ T cells, which was amplified when both CD4+ and CD8+ T cells were depleted (Fig. 4A). (Note, that CD4+ and CD8+ T cells were depleted simultaneously using anti-Thy1.2, which may also deplete double negative T cells, natural killer (NK) cells and innate lymphoid cells). Interestingly, the depletion of CD8+ T cells resulted in a considerable increase in viral load while CD4+ T cell depletion caused no significant difference in viral load. The viral load of the CD8+ T cell depleted group was still lower than that found in naïve mice challenged with PR8-GP33 (see Fig. 1A), although this difference was non-significant. Furthermore, the viral load was identical between the group in which only CD8+ T cells were depleted and the one in which both CD8+ and CD4+ T cells were depleted (Fig. 4A). This result agrees with previous reports that CD4+ T cells play only a minor role in modulating influenza viral titers, and that depletion of CD4+ T cells does little to alter the course of viral infection [51], [52]. While it is known that memory CD4+ T cells can cooperate with naïve B or CD8+ T cells in the context of influenza infection [36], it is unknown whether a similar cooperativity occurs with influenza virus-specific CD8+ T cells. It is interesting that CD8+ T cells in this setting were needed for control of virus while in the previous experiments (Fig. 2B, 3D) CD8+ T cells seemed to be able to control weight loss but not viral load. This difference may be due to differences in depletion versus immunization or adoptive transfer approaches or other mechanisms such as changes in immunopathology because of larger numbers of CD8 T cells suppressing other responses (e.g. CD4 T cells).
X31-GP33 immune mice are largely protected against symptoms of influenza virus infection in the absence of CD8+ and CD4+ T cells, suggesting other possible mechanisms contributing to protection. One possibility is that B cells have a role through the action of non-neutralizing antibodies specific for determinants shared between the X31 and PR8 influenza strains. To test this notion, we immunized µMT mice, which lack B cells [53], with X31-GP33. When challenged 30 days later these mice demonstrated no protection against PR8-GP33 despite a very similar influenza virus-specific CD8 T cell response compared to B6 mice (Fig. 4B and C). These data suggested that B cells are essential for heterosubtypic protection. One concern is that the immunological response may be altered in µMT mice due to the total lack of B cells. Therefore, we examined MD4 transgenic mice that have normal numbers of B cells, but have a transgenic B cell receptor specific for hen egg lysozyme [54] and are therefore unable to generate an influenza virus-specific antibody response. Similar to the µMT mice, however, MD4 mice immunized with X31-GP33 were also not protected and experienced severe weight loss upon rechallenge with PR8-GP33 (Fig. 4B). X31-GP33 immune MD4 mice also had reduced lung function and high viral load (Fig. 4B). To test whether the lack of non-neutralizing antibodies might underlie the defect in X31-GP33 immune MD4 mice, we transferred serum collected from X31-GP33 immunized B6 mice (referred to as X31 serum) into X31-GP33 primed MD4 mice and rechallenged with PR8-GP33. Protective immunity, as measured by all three parameters was improved (Fig. 5). Collectively, these data suggested that an influenza virus-specific B cell response was essential for X31-GP33 based heterosubtypic protection. As cross-neutralizing antibodies are not induced between the X31 and PR8 influenza strains [42]–[44], non-neutralizing antibodies are likely contributing to protection. One possible interpretation of the data presented thus far is that both CD8+ T cells and non-neutralizing antibodies might be necessary for optimal protection.
To evaluate whether non-neutralizing antibodies in conjugation with influenza virus-specific CD8+ T cells can elicit robust heterosubtypic protection in B6 mice, we transferred serum from X31-GP33 immune mice into LCMV Armstrong immune mice. When given X31-GP33 serum, LCMV immune mice containing GP33-specific memory CD8+ T cells displayed significantly reduced weight loss and viral load compared to the LCMV Armstrong immune group that had received PBS or serum from naïve mice (Fig. 6A). LCMV immune mice that received X31-GP33 serum also maintained nearly 100% blood oxygen saturation following PR8-GP33 challenge. The protection achieved by transfer of X31-GP33 serum to LCMV Armstrong-immune mice in terms of weight loss and lung function was nearly equivalent to that achieved with transfer of serum from PR8-GP33 immune mice containing neutralizing antibodies, although PR8 serum resulted in more effective control of viral replication (Fig. 6A). To determine whether the protective factor in the serum was indeed antibodies, we administered serum that had been depleted of IgG and IgA to LCMV Armstrong immune mice and challenged with PR8-GP33 [55]. These mice exhibit no evidence of protection indicating that X31-GP33 mediated protection is antibody-dependent (Fig. 6A). While it is possible that there may be more total IgG in the transferred X31 serum compared to naïve serum, there was no obvious correlation between the total IgG levels and protection in these experiments (data not shown). Our results indicate that optimal heterosubtypic protection against influenza is elicited only when both GP33-specific CD8+ T cells and non-neutralizing antibodies are present.
While many previous studies have demonstrated that antibodies induced by X31 do not neutralize PR8 and vice versa [42]–[44], it was possible that a new neutralizing determinant might have been formed due to the insertion of the GP33 sequence into the NA stalk. Thus, we transferred X31-GP33 serum into naïve mice one day prior to challenge with PR8-GP33. These mice displayed no signs of protection and experienced severe weight loss and decline in lung function (Fig. 6B). The naïve group given PR8 serum was completely protected from challenge with PR8-GP33 due to the presence of neutralizing antibodies (Fig. 6B). This result strongly suggested that the antibodies found in X31-GP33 serum were non-neutralizing. Additionally the lack of protection found in the naïve group given X31-GP33 serum indicates that both antigen-specific CD8+ T cells and non-neutralizing antibodies were needed for protection.
To further examine the cooperativity between non-neutralizing antibodies and virus-specific CD8+ T cells in heterosubtypic protection we transferred X31-GP33 serum, naïve serum, or PBS into LCMV Armstrong immune mice. We then rechallenged these mice with PR8-WT instead of PR8-GP33. In this setting the LCMV Armstrong primed mice given X31-GP33 serum, as well as the groups given naïve serum or PBS were not protected from PR8-WT rechallenge by any measure (Fig. 6C). Thus, the protective immunity in this setting was dependent on both non-neutralizing antibodies and recognition of viral determinants by primed CD8+ T cells.
To explore whether this cooperativity-based protection could be achieved using natural influenza-virus derived epitopes we immunized mice with VVNP366 to induce an influenza virus-specific T cell response. We then waited 30 days and transferred X31 or naïve serum into these mice one day prior to challenge with the H1N1 swine influenza virus strain SW/33. SW/33 is not genetically engineered, but, like PR8-GP33, is pathogenic in mice. Similar to LCMV Armstrong immune mice, VV366 immune mice given X31 serum were protected against viral challenge in terms of both weight and lung function, with these mice also having a trend to lower viral load compared to mice given naïve serum (Fig. 6D). This finding strongly indicates that the cooperativity-based protection is not simply an artifact of out recombinant influenza virus system, but rather a likely physiologically relevant mechanism of protection against influenza virus challenge.
The means by which non-neutralizing antibodies operate in cooperative protection is unclear. Among the possible mechanisms are: antibody-dependent cell-mediated cytotoxicity (ADCC), Fc receptor (FcR) mediated phagocytosis, and the complement pathway. To distinguish between these possibilities we used mice either lacking FcRγ or interleukin-15 (IL-15). FcRγ -/- mice are deficient in the gamma chain subunit of the FcgRI, FcgRIII and FceRI receptors resulting in functionally impaired macrophages, neutrophils, mast cells, basophils and Natural Killer (NK) cells. IL-15 -/- mice, on the other hand, are deficient in NK cells, but not these other cell types allowing the role of NK cell-mediated ADCC in heterosubtypic protection to be tested. Wild type, FcRγ -/-, and IL15 -/- mice were primed with LCMV Armstrong, rested 30 days, and then given either X31-GP33 serum or naïve serum 1 day prior to rechallenge. The LCMV immune IL-15-/- mice given X31-GP33 serum were protected upon PR8-GP33 challenge, although the decrease in viral load in these mice was only a trend. These results might reflect the moderate defect in CD8 T cell memory in these mice [56]–[58], although influenza virus-specific T cell memory in the respiratory tract appears independent of IL-15 [59]. In contrast, the FcRγ -/- mice were not protected against infection as measured by any parameters tested regardless of whether the mice were given X31-GP33 or naïve serum (Fig. 7A). Together, these results suggested that non-neutralizing antibody-based protection was FcRγ-dependent, but that NK cells were non-essential.
To further examine cooperative heterosubtypic immunity we first immunized B6 mice with X31-GP33. After 30 days we treated these mice i.n. with clodronate-loaded liposomes to deplete alveolar macrophages (AM) (and possibly other airway-resident phagocytes), cobra venom factor to deplete complement, or anti-NK 1.1 to deplete NK cells. Another group was given empty liposomes as a control. Following PR8-GP33 rechallenge, the only group left unprotected was the clodronate treated group in which AM were depleted. These mice experienced severe morbidity and high viral load despite having an unimpaired CD8+ T cell response (Fig. S5A). All other groups remained healthy and controlled the infection (Fig. 7B). This result suggested that heterosubtypic immunity mediated by non-neutralizing antibodies and CD8+ T cells was, at least in part, dependent on cells depleted by clodronate liposomes including AM. It is important to note that while we found clodronate liposome treatment to be non-toxic to uninfected mice and to result in ∼70% depletion of alveolar macrophages in the BAL fluid three days following a single clodronate treatment (Fig. S5B), it is possible that depletion of other airway populations such as dendritic cell or inflammatory macrophages could occur. The unimpaired CD8+ response seen in clodronate liposome treated mice however, suggests that any clodronate depletion of dendritic cells in the airway was insufficient to significantly impact presentation of antigen to CD8+ T cells. Furthermore, preliminary studies using adoptive transfer of alveolar macrophages obtained from the BAL of naïve mice into LCMV Armstrong immune FcRγ -/- mice suggested that reintroducing alveolar macrophages could partially rescue weight loss in half the mice when given in conjunction with X31 serum (Fig. S6). While further studies are necessary, these data are consistent with the notion that AM are involved in cooperative heterosubtypic protection.
To further evaluate the role of AM and NK cells in cooperative heterosubtypic protection, we depleted AM or NK cells in LCMV Armstrong immune mice as described above, and administered X31-GP33 serum to these mice one day prior to rechallenge. We found that only the group treated with clodronate liposomes exhibited severe weight loss and a decline in lung function (Fig. 7C). The group depleted of NK cells did display a trend towards higher viral loads then the group given only X31-GP33 serum, but this difference was not significant and this group still experienced almost no weight loss or decline in lung function, further suggesting that NK cells (or perhaps other NK1.1+ cells) only play a minor role in the mechanism of non-neutralizing antibody-based protection. Overall, this finding strongly indicates that the mechanism of non-neutralizing antibody-based protection is dependent on cells depleted by clodronate liposomes, including AM, likely through FcR-dependent AM phagocytosis or ADCC of influenza virus-infected cells.
The aim of universal influenza vaccination approaches is to provide long-lasting protection against a wide range of viral serotypes. Creating a universal vaccine by inducing CD8+ T cells specific for conserved internal proteins of influenza virus has received considerable attention, but remains an unrealized goal. In this study, we demonstrate that influenza virus-specific CD8+ T cells can cooperate with non-neutralizing antibodies to provide efficient cross-subtype influenza virus-specific protection. While non-neutralizing antibodies against M2e or other conserved determinants have recently been examined, our data indicate a previously unappreciated role for cooperativity between non-neutralizing antibodies and CD8+ T cell responses in the induction of optimal protection from serologically distinct influenza virus strains. This mechanism represents a novel approach by which a universal influenza vaccine could be developed.
Currently there are several promising universal influenza vaccine candidates in development. Among these candidates are broadly neutralizing antibodies, which are able to target the conserved stem region of the influenza virus [60]–[64]. These broadly neutralizing antibodies have been found to be cross reactive among different H1 or H3 influenza subtypes and are likely to represent a major advance in generating more effective influenza virus vaccines. However, current antibodies specific for the H1 stem are largely effective only against heterologous H1 and H5 viruses, and antibodies against the H3 stem are only effective against H3 viruses [60]. Interestingly, neutralizing antibodies were sometimes induced following vaccination with a pandemic H1N1 vaccine, but were of too low magnitude to induce robust heterosubtypic protection [64]. Until neutralizing antibodies can be generated against an antigen conserved between many different influenza subtypes, humans will remain vulnerable to the threat of a pandemic from a novel influenza strain such as H7N7, H9N2, etc. [65]. Another promising potential universal influenza vaccine targets the ectodomain of matrix protein 2 (M2e) [66]. The M2e sequence is conserved across influenza virus subtypes, and humoral anti-M2e immunity has been shown to protect against influenza virus challenge in mice [67], [68]. However, M2e-based protection does not prevent or resolve infection and is of a lower potency that HA-specific antibodies, making an M2e-dependent therapy more likely to act as a safety net in the case of the emergence of pandemic influenza strains rather than a replacement for current vaccines [69]. One concern associated with both broadly neutralizing antibodies and M2e based vaccines is that widespread use of these vaccines will introduce immune pressure promoting the evolution of antigenic escape viruses [66]. There have already been reports of escape viruses being generated in response to M2e antibodies, with one study finding that virus mutants with antigenic changes in M2e emerged in 65% of virus-infected mice treated with anti-M2e, although some level of protection remained despite these mutations [69]. An interesting avenue of future research will be to determine if cooperativity between T cells and non-neutralizing antibody can be used to boost the protection elicited through these vaccination strategies. Virus-specific CD8+ T cells do not seem to be generated by HA-stalk or M2e immunization strategies [70], so a vaccine in which CD8+ T cells can be elicited specific for conserved influenza virus determinants, combined with approaches to generate HA-stalk or M2e targeting antibodies, may offer improved protection. Furthermore, the overlapping protection provided by virus-specific CD8+ T cells should help reduce the possibility of an escape virus emerging.
Recent reports have implicated NP-specific IgG in heterosubtypic immunity to influenza virus [28], [29]. This previous work found that NP protein was detectable in the BAL and nasal washes of influenza virus-infected mice, thereby allowing the NP antigen to interact with NP-specific antibodies and form complexes to stimulate antiviral immune responses. These studies also demonstrated that 5 daily antibody injections starting 3 days prior to infection were required to reduce viral load in naïve mice, using a 0.25 lethal dose 50% (LD50) influenza rechallenge. Our findings extend this work in determining that cooperativity between influenza virus-specific CD8+ T cells and antibodies is important for heterosubtypic protection in immune competent mice. NP-specific antibodies are likely to be a primary component of the X31-GP33 serum used in our work. The cooperativity we demonstrated may mean that excessively high amounts of non-neutralizing antibodies might not be required if virus-specific CD8+ T cells are also present. Since previous influenza virus infections or vaccinations have likely induced anti-NP antibodies and influenza virus-specific memory CD8+ T cells in most adults, it is interesting that more heterologous protection does not seem to exist in humans. One reason may be that sufficiently high titers of anti-NP antibodies are not present in adults to mediate cooperative protection. Indeed one report indicates that trivalent inactivated influenza virus vaccine (TIV) only rarely and modestly boosted existing levels of anti-NP IgG [32]. Alternatively, perhaps such cooperative immunity is, in fact, one reason for the relatively low mortality in healthy adults for most strains of influenza virus. Our results suggest that it will be interesting to test whether this cooperative immunity might wane with increasing age, a theory supported by several reports [26]–[29]. It is possible that the lack of CD8 T cell boosting by the yearly vaccine allows CD8 T cell memory to decline over time even in healthy young adults. Future studies will be necessary to test some of these ideas in humans.
The mechanism by which non-neutralizing antibodies operate in the setting of heterosubtypic immunity remains poorly understood. Unlike neutralizing antibodies, non-neutralizing antibodies do not prevent viral entry into host cells and must therefore employ a different means of action to reduce viral load. FcRs have been reported to be important mediators of this process, but the specific cell types directly involved in reducing influenza viral load and pathology by this mechanism are unclear. Alveolar macrophages have been shown to play a critical role in influenza virus protection, likely through ADCC or phagocytosis [71], [72]. Some reports have also suggested that NK-cells may be involved in influenza virus protection through ADCC [59], although several recent studies have found this to be unlikely [73], [74]. The complement pathway could also have an important role due to the ability of complement to bind and lyse infected cells or enveloped virus in the presence of antibodies. Complement has been demonstrated to be able to neutralize the influenza virus in the presence of natural antibodies [75], and complement component C3 may be important in T cell priming and migration to the lungs [76]. Indeed, C3 deficiency in humans correlates with recurrent infections of the upper and lower respiratory tract [77]. In the current studies we found that heterosubtypic protection was dependent, at least in part, on alveolar macrophages. Intranasal administration of clodronate liposomes have been shown to selectively deplete alveolar macrophages while leaving the interstitial macrophage population as well as other cell types in the lungs intact [71], [78]–[80]. However, the possibility of off target effects of the clodronate liposome approach cannot be fully excluded. The unimpaired CD8+ T cell response found in clodronate-treated mice (Fig. S3A) however, suggested that depletion of dendritic cells was unlikely to be responsible for the lack of protection in this setting. Thus, alveolar macrophages are likely a major cell type impacted by this treatment and are expected to act to help reduce viral load through recognition of the Fc region of non-neutralizing antibodies. This recognition can lead to ADCC and/or antibody-dependent cell-mediated phagocytosis directed against infected cells bound by non-neutralizing antibodies.
The reduced effectiveness of seasonal influenza vaccines and greater infection-related morbidity and mortality in the elderly is thought to be due to alterations in both the innate and adaptive immune response that occur with age [3], [81]–[84]. Among the alterations reported in the elderly that could influence immunity to infection are changes in macrophages, NK cells, neutrophils, pathogen recognition via Toll-like receptors, innate cell cytokine production [85], [86], as well as decreased numbers, proliferation and signaling of B and T cells [87]–[91]. Interestingly, there have also been reports of changes in FcRs that occur with age, which could lead to defects in FcR-dependent effector functions [92]. While it has been shown that CD8+ T cell and neutralizing antibody-based protection obtained at a young age is still protective many years later [93]–[95], less is known about non-neutralizing antibody-dependent protection. Since this type of protection relies on FcR-dependent effector mechanisms to clear infected cells it will be important to determine if the heterosubtypic protection observed in young mice is also seen in aged groups of animals. A major goal of “universal” influenza vaccines is to elicit cross-subtype influenza virus protection in both young and aged populations. Hence, age-related defects in the immune system are a critical issue that must be addressed in future studies to determine if cooperative protection is an effective strategy in eliciting heterosubtypic influenza protection.
Collectively, we have shown that influenza virus-specific CD8+ T cells in cooperation with non-neutralizing antibodies are able to provide optimal protection against a lethal influenza virus rechallenge. This protection is only exhibited when both influenza virus-specific CD8+ T cells and non-neutralizing antibodies are present. Furthermore, non-neutralizing antibodies likely contribute to influenza virus clearance, possibly through a mechanism involving alveolar macrophages. It should be pointed out that while we have focused largely on CD8+ T cells, it is possible that cooperative protection will occur for CD4+ T cells and non-neutralizing antibodies. Indeed, there is good evidence that CD4+ T cells can contribute to protective immunity to influenza virus [35], [96], [97] and cooperate with naïve B and CD8+ T cells [36]. It will be important to address this issue in the future. This work provides novel insights into cross-subtype influenza virus protection and could have implications for the development of a universal influenza vaccine.
This study was carried out in accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. Protocols were approved by the Institutional Animal Care and Use (IACUC) committees of the Wistar Institute, (animal welfare assurance number A3432-01) or University of Pennsylvania (animal welfare assurance number A3079-01). The Wistar and University of Pennsylvania Animal Care and Use Programs are fully accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International (AAALAC).
C57BL/6 and Ly5.1+ mice were purchased from the National Cancer Institute (Frederick, MD) or Jackson Laboratories (Bar Harbor, ME). Age- and sex matched IL-15-/- mice were obtained from Taconic (Germantown, NY), FcRγ knockout mice (FcRγ KO; strain name B6.129P2-Fcer1gtm1RavN12) were purchased from Taconic Farms, Inc. (Hudson, NY), and B cell-deficient B6.129S2-IghtmICgn/J (µMT) mice and anti-HEL B-cell receptor (BCR)-transgenic C57BL/6-TgN (IghelMD4) mice (referred to as MD4) were obtained from the Jackson Laboratory. For primary or secondary infections, mice were inoculated using the following pathogens, doses and routes: with LCMV Armstrong (LCMV Arm; 2×105 PFU i.p. or 5×104 PFU i.n); recombinant X31 influenza virus expressing the LCMV GP33 epitope (X31-GP33; 1.6×105 TCID50 i.n.); vaccinia virus (VV) expressing the LCMV GP33 epitope (VVGP33), VV expressing the influenza virus NP366 epitope (VVNP366), and VV expressing LCMV Nucleoprotein (VV-NP (LCMV)) all used at 3×105 PFU i.n.; Listeria (LM) expressing the GP33 epitope (LM-GP33; i.v.); Vesicular stomatitis virus expressing the ovalbumin (OVA) epitope (VSV-OVA; 2×106 PFU i.v.); LCMV Armstrong V35A which lacks the GP33 epitope (LCMV-V35A; 2×105 PFU i.p.). For rechallenge experiments, mice were given either recombinant PR8 influenza virus expressing the LCMV GP33 epitope (PR8-GP33; 3 LD50 i.n.); wild type PR8 influenza virus (PR8-WT; 3 LD50 i.n.); wild type swine influenza virus (SW/33; 3 LD50 i.n). For both strains of PR8 1LD50 = ∼250 TCID50. Prior to i.n. infections, mice were anesthetized by i.p. injection of ketamine hydrochloride and xylazine (Phoenix Scientific, San Marcos, CA) in 0.2 ml Life Technologies HBSS (Invitrogen, Carlsbad, CA). In some experiments, mice were anesthetized with 2.5% Avertin (0.2–0.35 ml) i.p. Recombinant influenza virus strains containing the LCMV GP33–41 epitope inserted in the neuraminidase stalk region were obtained from Dr. Richard J. Webby (St. Jude Children's Research Hospital, Memphis, TN) and have been previously described [35], [36]. These viruses were propagated in eggs, and stored at −80°C. The replication and pathogenicity of these recombinant X31 and PR8 strains were not substantially different from their nonrecombinant counterparts (data not shown). Viral titers were determined by plaque assay on Vero cell monolayers (for LCMV and VV) or on Madin-Darby canine kidney cell monolayers (for X31-GP33 and PR8- GP33) as previously described [98]. For all experiments, naïve C57BL/6 mice were also infected with influenza virus to allow comparison of weight loss, viral load, and pulse oximetry between different experiments. The concentration of infectious virus in lungs was determined by titration of homogenized tissues in Madin-Darby canine kidney cell (MDCK) microcultures as described previously [6]. Lung titers are expressed as dilution of lung extract at which 50% of the MDCK cultures revealed virus growth (TCID50/ml).
For all adoptive transfer experiments, congenic mice differing in Ly5 (Ly5.1 versus Ly5.2) were used. For adoptive transfers CD8+ T cells were purified (90% purity) using magnetic beads (CD8+ T cell isolation kit, MACS beads; Miltenyi Biotec, Auburn, CA).
The MouseOx Pulse-oximeter (Starr Life Sciences, Oakmont PA) was used to measure blood oxygen saturation (SpO2) in PR8-GP33 infected mice during the course of infection. A depilatory agent (Nair, Church & Dwight Co.) was applied to the neck of anesthetized mice 1 day prior to influenza infection to remove hair and delay future hair growth. For readings, the oximeter clip was placed on the neck and percent SpO2 was measured each second over several minutes, data shown is the average of SpO2 readings recorded over 3–5 minutes per mouse.
Alveolar macrophages were isolated and transferred as previously described [71]. Briefly AM were isolated from BAL with PBS-EDTA from C57BL/6 mice. 10 mice were sacrificed as donors for every recipient mouse. A 23-gauge cannula was inserted into the trachea, and cells were collected by washing the airway lumen with 3×0.5 ml PBS-EDTA. The obtained BAL fluid was centrifuged, and cells were washed twice with PBS, counted, and resuspended in PBS. 4.5×105 cells were then transferred i.n. into recipient mice at a volume of 50 ul/mouse.
NK1.1 cells were depleted in vivo by i.p. injection (0.2 mg/injection) of rat mAb PK136. CD8+ T cells were depleted by i.p. injection of rat mAb 53.6, CD4+ T cells were depleted using rat mAb GK1.5 and both CD4+ and CD8+ T cells were depleted simultaneously using anti-Thy1.2 (0.2 mg/injection; clone 30H12, isotype Rat IgG2b obtained from BioXCell). All antibody treatments were givens days -3, -1, 2, and 5 post PR8-GP33 rechallenge. Depletion was confirmed by flow cytometric analysis on day 6 post rechallenge in the lungs. All in vivo mouse antibodies were purchased from Bio X Cell (West Lebanon, NH).
Alveolar macrophages were depleted using the liposome-mediated macrophage depletion technique based on the intracellular delivery of the drug dichloromethylene diphosphonate (clodronate). Preparation of clodronate-liposomes and applications of the technique was done as previously described [79]. Alveolar macrophages were depleted by i.n. administration of 50 ul of clodronate-liposomes on days -3, -1, and 2 post PR8-GP33 rechallenge.
Viral quantitative real-time RT-PCR was performed essentially as previously described [26]. Briefly, total RNA was purified from lungs of PR8-GP33 infected mice using the RNeasy Mini Kit (Qiagen, Valencia, CA). Reverse transcriptions were primed with random primers and performed using the High Capacity cDNA Reverse Transcription Kit from Applied Biosystems (Foster City, CA). Real-time quantitative PCR (qRT-PCR) was performed on cDNA using TaqMan Universal PCR Master Mix (Applied Biosystems) and probes and primers specific to the influenza PA protein with all samples analyzed in triplicate. Reactions were run on a real-time PCR system (ABI7500; Applied Biosystems). Amount of influenza viral RNA per sample was then calculated using known standards. The total amount of virus per lung was then calculated using the mass of the lung portion taken for viral RNA determination in relation to the total lung mass. The TCID50 of each sample was determined by calculating the volume of virus per lung (using the viral RNA determination of the PR8-GP33 stock) and then calculating the total TCID50 in the lungs using the known TCID50 per unit volume of the viral stock. The limit of detection was determined by performing qRT-PCR on lung samples from uninfected mice and represented by a dashed line.
PA sense: CGGTCCAAATTCCTGCTGAT. PA antisense: CATTGGGTTCCTTCCATACA. PA probe: 6FAMCCAAGTCATGAAGGAGAGGGAATACCGCTTAMRA
Lymphocytes were isolated from tissues as previously described [99]. Briefly, mice were euthanized and the hepatic vein cut. Lungs were perfused by injection of PBS into the hepatic artery or the right heart ventricle. Lungs were cut into pieces and incubated in 0.2 mg/ml collagenase D (Roche Diagnostic, Indianapolis, IN) at 37°C for 35 min. Spleens and lymph nodes were homogenized using a cell strainer. In all tissues, red blood cells (RBCs) were lysed using ACK lysing buffer (Quality Biologicals, Gaithersburg, MD), and lymphocytes were washed and counted.
Serum was collected from naïve and day 30+ X31-GP33, or PR8-GP33 infected mice. Serum samples from individual mice were pooled and 1 ml of pooled serum/mouse was injected i.p. into mice on day -1 prior to PR8-GP33 rechallenge. In some instances to verify that the antibodies present in the serum were responsible for any protective effects, serum was depleted of IgG and IgA using Protein A and G SpinTrap (GE Healthcare, Pittsburgh, PA) according to manufacturer's instructions.
Effector CD8 T cells were generated in vitro by peptide-stimulation of TCR-transgenic splenocytes (obtained from a P14 transgenic mouse) specific for the LCMV glycoprotein peptide (P14 mice specific for GP33-41). Briefly, spleen cells were incubated with 5 µM GP33 peptide for two hours. The peptide was washed off, media replaced and the cells were cultured for 48 hrs in 24-well plate, and maintained afterwards in 75T culture flasks in IL-2 - supplemented media for 5 days. The media was changed every 48 hours. A daily sample from the culture was examined by flow cytometry for the expression level of the activation markers, CD44 and CD25. On day 5, the cells were harvested, washed in PBS, counted and resuspended in PBS for adoptive transfer.
Lymphocytes were stained using standard techniques and analyzed by flow cytometry. Virus-specific CD8 T cells were quantified using MHC class I peptide tetramer staining. MHC class I peptide tetramers were made and used as described [98]. Antibodies to CD8 and CD44 were purchased from eBioscience (San Diego, CA). Staining and analysis were performed as previously described [92]. Function was investigated by intracellular cytokine staining following antigen stimulation (IFNγ, TNFα, IL-2, CD40L). Briefly, 1×106 splenocytes were cultured in the absence or presence of the indicated peptide (0.2 mg/ml) and brefeldin A for 5 h at 37°C. Influenza virus pooled peptides were used to evaluate the influenza virus-specific CD8+ T cell responses. This pool contains 147 overlapping peptides from influenza virus NP and M proteins, and we also included the GP33 peptide in this pool. For later experiments the overall influenza-specific CD8+ T cell response was evaluated via intracellular cytokine staining following stimulation with peptides from the influenza proteins HA, NA, NS1, NS2, PA, PB, NP, as well as the LCMV epitope GP33. Following staining for surface antigens as described above, cells were stained for intracellular cytokines using the Cytofix/Cytoperm kit (BD Biosciences). Samples were collected using an LSRII flow cytometer (BD Biosciences).
Results represent the mean ± SEM unless indicated otherwise. Statistical significance was determined by paired or unpaired Student's t test. Statistical analyses were performed using Prism GraphPad software v5.0. (*, p<0.05; **, p<0.01; ***, p<0.001).
Neuraminidase-956530; Interferon gamma-15978; Tumor necrosis factor-21926; Hemagglutinin- 956529; Nucleoprotein (Influenza)-956531; Nucleoprotein (LCMV)-956592; Non-structural protein 1–956533; Non-structural protein 1–956532; Polymerase acidic-956535; Fc receptor-109615; Glycoprotein (LCMV)-956590; Matrix protein 2–956528; Interleukin 15–16168. All accession ID numbers are recorded from the Entrez Gene database.
|
10.1371/journal.ppat.1003491 | Irf8-Regulated Genomic Responses Drive Pathological Inflammation during Cerebral Malaria | Interferon Regulatory Factor 8 (IRF8) is required for development, maturation and expression of anti-microbial defenses of myeloid cells. BXH2 mice harbor a severely hypomorphic allele at Irf8 (Irf8R294C) that causes susceptibility to infection with intracellular pathogens including Mycobacterium tuberculosis. We report that BXH2 are completely resistant to the development of cerebral malaria (ECM) following Plasmodium berghei ANKA infection. Comparative transcriptional profiling of brain RNA as well as chromatin immunoprecipitation and high-throughput sequencing (ChIP-seq) was used to identify IRF8-regulated genes whose expression is associated with pathological acute neuroinflammation. Genes increased by infection were strongly enriched for IRF8 binding sites, suggesting that IRF8 acts as a transcriptional activator in inflammatory programs. These lists were enriched for myeloid-specific pathways, including interferon responses, antigen presentation and Th1 polarizing cytokines. We show that inactivation of several of these downstream target genes (including the Irf8 transcription partner Irf1) confers protection against ECM. ECM-resistance in Irf8 and Irf1 mutants is associated with impaired myeloid and lymphoid cells function, including production of IL12p40 and IFNγ. We note strong overlap between genes bound and regulated by IRF8 during ECM and genes regulated in the lungs of M. tuberculosis infected mice. This IRF8-dependent network contains several genes recently identified as risk factors in acute and chronic human inflammatory conditions. We report a common core of IRF8-bound genes forming a critical inflammatory host-response network.
| Cerebral malaria is a severe and often lethal complication from infection with Plasmodium falciparum which is driven in part by pathological host inflammatory response to parasitized red cells′ adherence in the brain microvasculature. However, the pathways that initiate and amplify this pathological neuroinflammation are not well understood. As susceptibility to cerebral malaria is variable and has been shown to be partially heritable, we have studied this from a genetic perspective using a mouse model of infection with P. berghei which induces experimental cerebral malaria (ECM). Here we show that mice bearing mutations in the myeloid transcription factor IRF8 and its heterodimerization partner IRF1 are completely resistant to ECM. We have identified the genes and associated networks that are activated by IRF8 during ECM. Loss-of-function mutations of several IRF8 targets are also shown to be protective. Parallel analysis of lungs infected with Mycobacterium tuberculosis show that IRF8-associated core pathways are also engaged during tuberculosis where they play a protective role. This contrast illustrates the balancing act required by the immune system to respond to pathogens and highlights a lynchpin role for IRF8 in both. Finally, several genes in these networks have been individually associated with chronic or acute inflammatory conditions in humans.
| IRF8 is a member of the Interferon Regulatory Factor (IRF) family of transcription factors that plays a central role in interferon signaling, response to infection and maturation of dendritic cells (DCs) and other myeloid lineages [1], [2]. IRF8 regulates elements of constitutive gene expression in both myeloid and lymphoid cells and can also activate or suppress pathogen responsive transcription programs following exposure to type I or type II interferons, lipopolysaccharides, and a range of additional microbial products [1], [2]. Heterodimerization of IRF8 with members of the IRF (IRF1, IRF4) or ETS (PU.1) families leads to DNA binding and transcriptional regulation of target genes containing ISRE (GAAAnnGAAA) and EICE-type (GGAAnnGAAA) canonical motifs in their promoters [3]–[5].
During hematopoiesis, IRF8 promotes differentiation of myeloid progenitors towards the mononuclear phagocyte lineages (monocytes, macrophages, DCs) by acting as an antagonist of the polymorphonuclear granulocyte pathway [6]–[9]. This is accomplished through positive regulation of pro-apoptotic signals (Cdkn2b, Nf1, Bax), and negative regulation of pro-survival signals (Bcl2, Bcl-XL) in CD11b+ myeloid precursors [10]–[12]. Mice bearing either a targeted null allele [7] or a severely hypomorphic mutation (BXH2, Irf8R294C) [9] show either complete absence of all CD11c+CD8α+ dendritic cells subsets (Irf8−/−) while BXH2 mice (Irf8R294C) retain some plasmacytoid DCs (B220+/CD11clow) [13]–[15]. Furthermore, both mutants display a chronic myeloid leukemia-like phenotype dominated by expansion of Gr1+/CD11b+ immature myeloid cells [6]–[9]. Additionally, IRF8 is required for B lymphocyte lineage specification, commitment, and differentiation, including expression of certain biochemical pathways that play a key role in the specialized functions of these antigen-presenting cells (APCs) [2], [5], [16].
During infection, IRF8 activates antimicrobial defenses in myeloid cells, propagates pro-inflammatory signals and is required to amplify early immune responses by these cells. IRF8 activity is essential to express Il12p40, Il12p35 and Il18 in response to IFNγ [6], [17]–[21] and is therefore required for APC-mediated Th1 polarization of early immune responses [1], [2]. Irf8-deficient mice, then, display defective Th1 responses (absence of antigen specific and IFNγ producing CD4+ T cells) [22], enhanced Th17 responses [23], and are susceptible to in vivo infection with many intracellular pathogens [18], [24]–[26] including tuberculosis [22] and blood-stage malaria [27]. Irf8-deficient macrophages are also extremely susceptible to ex vivo infection by intracellular pathogens [28]–[30]. Studies using genome-wide transcript profiling, chromatin immunoprecipitation [3]–[5] and individual gene targets [1] show that IRF8 regulates multiple aspects of antimicrobial defenses in mononuclear phagocytes. These include antigen recognition and processing, phagosome maturation, production of lysosomal enzymes and other cytoplasmic microbicidal pathways.
IRF8 mutations in humans cause pathologies remarkably similar to those observed in Irf8 mutant mice, affecting the myeloid compartment in general and DCs in particular [31]. We have shown that homozygosity for a DNA-binding incompetent and transcriptionally inactive human IRF8 mutant variant (IRF8K108E) is associated with severe recurrent perinatal bacterial and fungal infections, with absence of blood monocytes and DCs, and a lack of IL-12 and IFNγ production following in vitro stimulation of blood cells [31]. We also reported a milder autosomal dominant form of IRF8-deficiency (IRF8T80A) in two patients suffering from Mendelian susceptibility to mycobacterial disease (MSMD) with recurrent episodes of mycobacterial infections following perinatal vaccination with M. bovis BCG. These patients showed selective depletion of the CD11c+ CD1c+ DC subset and impaired production of IL-12 by circulating peripheral blood cells. The IRF8T80A variant displays negative dominance and can suppress the trans-activation potential of wild type IRF8 for known transcriptional targets such as NOS2 and IL-12 [31]. Interestingly, T cells raised in the absence of an intact DC compartment in patient IRF8K108E show impaired function (production of IFNγ and other cytokines in response to non-specific stimuli) [31]. Finally, recent results from genome-wide association studies (GWAS) have pointed to a role for IRF8 in the complex genetic etiology of several human inflammatory diseases. Strong and independently replicated associations have been detected between polymorphic variants within or near IRF8 gene for systemic lupus erythematosus [32], ulcerative colitis [33], Crohn's disease [34], [35], and multiple sclerosis (MS) [36]–[38]. Details of one study in MS patients showed that the IRF8 susceptibility allele (rs17445836) was associated with higher expression of both IRF8 mRNA and downstream IFNβ-responsive targets [36].
We have used an experimental model of murine cerebral malaria (ECM) induced by infection with Plasmodium berghei ANKA (PbA) to investigate the role of Irf8 in pathological inflammation. In this model, adherence of PbA-infected erythrocytes to brain microvasculature leads to acute and rapidly fatal neuroinflammation. Symptoms such as tremors, ataxia and seizures appear between d5 and d8 in susceptible mice, progressing to morbidity and death within hours. Irf8-deficient BXH2 mice (Irf8R294C) did not develop any neurological symptoms and were found to be completely resistant to PbA-induced ECM. Comparative transcript profiling of PbA-infected wild-type C57BL/6 (B6) and BXH2 mice, together with IRF8 chromatin immunoprecipitation coupled to high-throughput DNA sequencing (ChIP-seq) have identified a list of key Irf8 targets whose expression is associated with acute ECM-associated neuroinflammation. This list has substantial overlap with genes activated in mouse lungs following infection with M. tuberculosis (Mtb), suggesting a shared core inflammatory response to infection that is protective against Mtb infection but deleterious in ECM. These studies identify IRF8 as a key regulator of acute neuroinflammation during ECM and a major inflammatory mediator.
BXH2 is a recombinant inbred mouse strain derived from B6 and C3H/HeJ (C3H) parents which displays a myeloid defect in the form of immature myeloid hyperplasia and susceptibility to multiple infections. We have previously used high resolution linkage analysis, positional cloning and candidate gene sequencing to demonstrate that myeloid hyperplasia and susceptibility to infections are caused by a severely deleterious hypomorphic allele at Irf8 (Irf8R294C) that spontaneously arose during the breeding of this strain [9], [12]. To assess the contribution of Irf8 to pathological inflammation, we infected BXH2 mice with PbA parasites, the murine agent of experimental cerebral malaria (ECM). Parasite replication in the blood, appearance of neurological symptoms and overall survival were recorded over 18 days (Figure 1). While all B6 mice developed ECM and succumbed by d9, BXH2 mice were completely resistant to the ECM phase, succumbing later to hyperanemia caused by uncontrolled blood-stage replication of the parasite (Figures 1A and 1B). [BXH2×B6]F1 mice showed significant resistance to PbA-induced ECM when compared to susceptible B6 parental control (p<0.0001), with approximately 50% of the animals surviving past d9 (Figure 1A). Additional phenotyping of a small group of segregating [BXH2×B6]F2 mice (n = 24) identified ECM-resistance only in mice either homozygote or heterozygote for the Irf8R294C allele, confirming that the protective effect we observed is due to the Irf8R294C mutation with minimal or no contribution of the mixed B6/C3H genetic background of BXH2 (Figure 1C). These data show that a) partial or complete loss of IRF8 function protects mice against lethality in an ECM-associated neuroinflammation model, and b) that the ECM-protective effect of the Irf8R294C mutation is inherited in a co-dominant fashion. Resistance to ECM in BXH2 mice was not associated with decreased parasite burden, as B6, BXH2 and [BXH2×B6]F1 mice showed similar levels of circulating blood parasitemia at d5, d7 and d9 post-infection (p>0.1)(Figure 1B). However, as the infection progressed, some BXH2 mice developed extremely high levels of blood parasitemia between d12-d21 in sharp contrast to any surviving controls or [BXH2×B6]F1s. This high parasitemia, rather than cerebral inflammation, was responsible for the observed mortality.
Lethal ECM in PbA-infected mice is associated with endothelial dysfunction, including loss of integrity of the blood brain barrier (BBB) [39]–[41]. Using an Evans Blue dye extravasation assay (Figure 1D), PbA-infected B6 mice display obvious BBB permeability by d7, indicated by the blue color, while infected BXH2 mice retained BBB integrity both early (d7) and late (d16) during infection. The BBB integrity was also assessed by flow cytometry cellular profiling of perfused brains from day 6 infected mice. In B6 brains, ECM was associated with presence of both CD4+ and CD8+ lymphocytes, as well as CD11b+/Ly6C+ granulocytes and monocytes (Figure 1E). Interestingly and despite the immature myeloid hyperplasia in peripheral myeloid and lymphoid organs, infiltration of T cells and myeloid cells in the brain was not seen in PbA-infected BXH2, nor was it seen in [BXH2×B6]F1 (Figure 1E). Together, these results indicate that ECM susceptibility in B6 is associated with infiltration of myeloid and lymphoid cells at the site of pathology. Such infiltration is not seen in ECM-resistant mice that are either homozygous (BXH2) or heterozygous ([BXH2×B6]F1) for Irf8R294C.
These results demonstrate that although IRF8 dysfunction is protective against ECM, functional IRF8 is required to control blood stage replication of PbA late in infection. Partial IRF8 activity in [BXH2×B6]F1 is sufficient to protect against high blood-stage replication.
To gain further insight into the genes, proteins and pathways that play a role during pathological neuroinflammation, and identify those whose expression is regulated by Irf8, we used transcript profiling of BXH2 and B6 brains either prior to or during PbA-infection. Principal components analysis (PCA) clustered the samples along two axes: component 1, which explained 39.4% of the variance and was associated with infection status (infection component), and component 2, associated with mouse strain (genetic component), which explained 24.4% of the variance (Figure 2A). PCA also indicates that PbA infection had a much stronger impact on transcriptional profiles in B6 mice than in BXH2, with the B6 d7 infected samples forming a remote out-group. In contrast, the BXH2 d7 cluster was only moderately shifted by infection and remained much closer to the BXH2 d0 group (compared to B6 d0 vs. B6 d7) indicating more modest response to infection by BXH2.
We used paired t-tests to assess the infection-induced transcriptional responses in both B6 and BXH2 mice as a way to extract gene lists relevant to ECM susceptibility. As suggested by the PCA, B6 response to infection was robust, with 292 genes showing statistically significant differences in expression (Figure 2B, d0 vs. d7; fold change ≥2, padj<0.05). On the other hand, response to infection in BXH2 was more modest with 81 genes reaching statistical significance. More than half of the genes (n = 48) regulated by infection in BXH2 were common to the B6 set and may indicate Irf8-independent regulatory mechanisms. This analysis also identified a set of 117 genes that show significantly different levels of expression in B6 and BXH2 mice prior to infection. Only ∼10% of these genes (n = 16) were further significantly modulated by PbA infection (Figure 2B). Importantly, this analysis further identified a key subset of 231 genes that were specifically regulated in B6, but not BXH2, mice during infection, therefore associated with pathology.
We also performed a two-factor ANOVA test accounting for both differences in basal level of gene expression in the brain (B6 vs. BXH2 at day 0), and infection-induced transcriptional response to PbA (Figure 2C). This identified a total of 107 genes (123 probes; fold change ≥2, padj<0.05) that were strongly regulated by infection in an Irf8-dependent fashion (padj-interaction <0.05) (Figure 2C, Table S1). Euclidean hierarchical clustering of this gene list identified three major categories of transcripts. Group 1 genes (n = 15) expression levels were increased by infection in both strains (more pronounced in B6 than BXH2), group 2 genes were increased by infection in B6 mice but not significantly induced in BXH2 (n = 62) and group 3 genes (n = 30) expression was reduced in infected mice (stronger repression in B6 compared to BXH2). Using the online Database for Annotation and Integrated Discovery (DAVID) tool to examine the list of genes regulated by infection in B6 mouse brains indicated substantial enrichment for immune response (4.4-fold enrichment above Illumina WG-6 v2.0 chip background, padj = 3.3×10−12), antigen processing and presentation (11.0-fold enrichment, padj = 1.4×10−9), defense response (3.8-fold enrichment, padj = 1.7×10−8), chemotaxis (5.2-fold enrichment, padj = 3.4×10−3) and inflammatory response (3.6-fold enrichment, padj = 3.7×10−3). Increased genes on these lists include potent pro-inflammatory chemokines that recruit myeloid and lymphoid cells to the site of infection and/or tissue injury such as Cxcl9, Cxcl10, Ccl4, and Ccl12; myeloid cell receptors associated with phagocytosis of microbes (Fcgr3, Fcgr4) and maturation of phagosomes (small GTPases Igtp, Irgm1, Gbp2, Gbp3); IRF8's hetero-dimerization partner (Irf1), and early type I interferon response (Oasl2, Ifit3). Genes under these categories were expressed more highly in B6 than in BXH2, consistent with the notion that resistance to ECM-associated neuroinflammation in BXH2 is linked to reduced Irf8-dependent inflammatory and innate immune responses, with a strong involvement of the myeloid compartment.
Organizing the list of genes associated with ECM pathology (B6 d7/d0 pairwise) using Ingenuity Pathway Analysis revealed enrichment for several significant biological networks. The top-scoring network included down-regulation of the hematological response and increase of intercellular signaling (Figure S1A). Several immune cell chemoattractants (Ccl4, Ccl7, Cxcl9) are featured in this network, suggesting recruitment of monocytes, NK cells and T-cells to the site of brain inflammation. Very little modulation of either of these pathways is seen in BXH2 during infection, with most components not making the 2 fold change threshold (Figure S1B). The second network highlights significant increase of interferon-responsive genes and inflammatory mediators, with type 1 interferon, Irf7 and Stat1 regulation occupying highly connected nodes (Figure S1C). Most of these inflammatory processes are increased to a lesser degree in BXH2 mice (Figure S1D and details in Table S1). The third network focuses on antigen processing and presentation with immunoglobulins and Fc receptors increased at the center of the B6 infection-responsive image (Figure S1E). Similar to what was seen in network 1, significant increase of these genes was largely missing from the BXH2 response (Figure S1F).
Since total brain RNA was used in our studies, genes differentially regulated in response to infection in an Irf8-dependent fashion may represent direct transcriptional targets of IRF8 or may be secondary targets corresponding to markers of cell populations differentially recruited to the site of infection in B6 and BXH2 mice. To distinguish between these possibilities, we mapped genome-wide IRF8 binding sites using ChIP-seq. The resulting sequence reads were mapped to mouse mm9 genome assembly and IRF8 binding peaks were identified using MACS peaks finding algorithm [42]. In order to validate ChIP-seq results, we confirmed IRF8 recruitment on known target sites [3] by independent ChIP-qPCR experiments (Figure 3A). IRF8-bound genes were identified as those containing a binding peak within a 20 kb window from their transcriptional start site (TSS). The list of IRF8-bound genes by ChIP-seq was intersected with the list of genes regulated by PbA in a strain, infection and Irf8-associated fashion from the two-factor ANOVA analysis (Figure 2C, Table S1). This intersection revealed a strong enrichment of IRF8 binding sites in genes increased during infection, with IRF8 binding sites detected in 85% of Group 1 genes (13/15) and 50% of Group 2 genes (31/62) (Figure 2C, Table S1). Differentially down-regulated genes did not show any enrichment with only 13% (4/30) of Group 3 genes containing IRF8 peaks, lower than background peak association (21% of all genes represented on the Illumina array; Figure 3B). These results strongly suggest that during neuroinflammation, IRF8 functions as a direct transcriptional activator of multiple genes coding for key pro-inflammatory pathways.
To identify IRF8-bound genes associated with ECM neuropathology, we queried the list of all genes whose expression is regulated by infection in Irf8-competent B6 mice for the presence of IRF8 binding sites (pair-wise comparison of B6 d7/d0). This analysis (Figure 3B) showed very strong enrichment for IRF8 binding sites in the vicinity of genes increased by infection, with 74% of increased genes (92/125; p<0.0001, Fisher's Exact test) bearing one or more IRF8 binding sites within 20 kb of the TSS (Figure 3C for ChIP-seq profile examples and complete list in Table S2). Genes decreased in response to infection did not show enrichment above background (Figure 3B and Table S2). Using the AmiGO gene ontology annotation tool [43], this list (Figure 2D and see details in Table S3) was found to contain numerous genes involved in inflammatory and innate immune responses, a finding most pronounced in the subset (74%) of genes with IRF8 binding sites. This included inflammatory cytokines and chemokines involved in chemotaxis of myeloid and lymphoid cell types to the sites of infection (Ccl4, Ccl5, Ccl7, Ccl12, Cxcl9, Cxcl10), early innate immune recognition and responses (Nlrc5, Ifi205), response to viral infections (Oasl2, Mx2, Oas1g), type I interferon responsive genes and pathways (Ifit2, Ifit3, Isg15, Rsad2), antigen capture (C1q, C4b, Fcerg1), phagosome maturation (Irgm1, Irgm2, Igtp, Gbp2, Gbp3), antigen processing (Tap1, Tap2) and Class I and Class II MHC-dependent antigen presentation in myeloid cells (B2m, H2-Ab1, H2-D, H2-K, H2-L, H2-Q, H2-T22 ). Other IRF family members implicated in early response to antigenic stimuli or danger signals (Irf1, Irf7, Irf9) were also induced (Table S3). These IRF8-regulated pro-inflammatory pathways appear linked primarily to the myeloid cellular compartment.
In both humans [31] and mice [22], mutations in IRF8 cause susceptibility to mycobacterial infections. Thus, we compared the list of genes regulated by infection in B6 brains during PbA cerebral malaria (d7/d0) with the list of genes contributing to the protective response in the lungs of M. tuberculosis infected B6 mice (d30/d0) [22] (Table S2). B6 and BXH2 mice have opposite phenotypes in the two disease models, with B6 being susceptible to ECM and resistant to M. tuberculosis, while BXH2 succumb rapidly to pulmonary tuberculosis [22]. Strikingly, out of the 123 genes increased more than 2-fold during ECM, 66 followed the same trend during M. tuberculosis infection (p<0.0001, Fisher's Exact Test). There was minimal overlap in the down-regulated genes (21 M. tuberculosis-regulated genes overlapping with the 170 PbA-regulated, including 6 in the opposing direction). The overwhelming majority (80%) of genes increased during both infections contained at least one IRF8 binding site, again highlighting IRF8 as having a central role during inflammation and host response to infections (Table S2).
To visualize the core networks engaged in this common IRF8-regulated host response, we analyzed the 53 genes containing IRF8 binding sites, and which were increased during both ECM and in pulmonary tuberculosis (Figure 4 and see Table S2 for details). Several networks with a clear focus on type 1 and type 2 interferon pathways characteristic of a pro-inflammatory innate immune response were detected. All these networks are clearly less activated in PbA infected BXH2 mice. One obvious network (Figure 4A) includes three Irfs (Irf1, Irf7, Irf9) and other hallmarks of type 1 interferon and MHC class 1 response. A second network (Figure 4B) centers on IFNγ/STAT1 signaling, with interferon-inducible GTPases, pro-inflammatory cytokines and chemokines forming the downstream response. A third, classically myeloid network (Figure 4C) features Fc-γ receptors, immunoglobulins and MHC class 2 molecules, as well as T-cell activating IL-12, CCL4 and CCL5. Across all networks, the most highly connected genes were Irf1, Irf7, Ifng, and Stat1, consistent with a key role for IRF8 in regulating pro-inflammatory innate immune responses in myeloid cells.
To validate the role of several identified IRF8 targets and associated pathways in pathological neuroinflammation, we phenotyped targeted knock-out mice for several of these loci. These included infection-regulated genes bearing IRF8 binding sites (Irf1, Ifit1, Isg15, Nlrc4, Il12p40 and Irgm1; examples of IRF8 ChIP-seq profiles are shown in Figure 3C), and genes known to play key roles in early innate immune response (Ifng, Jak3, Stat1, Il12p40). Results from these experiments (Figure 5) show that Ifng−/−, Jak3−/− and Stat1−/− mutant mice were completely resistant to ECM following PbA infection, highlighting key roles for these molecules in the progression or amplification of the pathological inflammatory response and supporting their central positions in the network models (Figures S1C, 4A and 4B) [44], [45]. Loss of Irf1 [46], IL12p40 [47] and Irgm1 delayed appearance of neurological symptoms and prolonged survival of PbA-infected mice but did not ultimately confer complete protection (Figure 5). These results validate that certain of IRF8's transcriptional targets during PbA infection in lymphoid and myeloid cells play critical downstream roles in pathology. On the other hand, Ifit1−/−, Isg15−/− and Nlrc4−/− mutant mice remained susceptible to PbA-induced ECM, suggesting that although these proteins may play important roles in neuroinflammation, their inactivation is not sufficient to induce protection.
To clarify the immunological basis of ECM resistance in BXH2, and in [BXH2×B6]F1, cellular immunophenotyping of spleen cells was carried out, both at steady state (d0) and 6 days following infection with PbA, immediately preceding the appearance of neurological symptoms (Figure 6–7). In uninfected spleens, BXH2 were characterized by a dramatic expansion of CD11b+/Ly6C+ immature myeloid cells comprised of monocyte-like F4/80+ (Figure 6A Day 0 - Gate R1 and Figure S2A Gate R1) and granulocyte-like Ly6G+ cells (Figure 6A Day 0 – Gate R2 and Figure S2A Gate R2), which resulted in reduced proportions of CD4+ and CD8+ T lymphocytes (Figure 7A Day 0). In contrast, lymphoid and myeloid cell populations are highly similar in spleens of ECM-susceptible B6 and ECM-resistant [BXH2×B6]F1 mice prior to PbA infection (Figure 6A and 7A). During P. berghei infection, we noted that ECM-resistance in BXH2 and [BXH2×B6]F1 compared to B6 controls was associated with reduced numbers of splenic myeloid DCs (Fig. 6D; Ly6C−CD11b+CD11c+MHCII+), which was concomitant to reduced but not absent serum levels of IL12p40 (Figure 6E) and reduced production of IL12p40 by spleen cells ex vivo 6 days post-infection (Figure 6F). We also noted that during infection, the ECM-protective effect of the Irf8 mutation is phenotypically expressed in the lymphoid compartment of BXH2 and [BXH2×B6]F1. Indeed, at day 6 post-infection ECM-resistant BXH2 and F1 mice did not show the significant increase in CD8+ T cells that was detected in ECM-sensitive B6 mice (2× increase; Figure 7C and S3), while CD4+ T cell numbers were increased in all groups by the same factor. Together, these changes were accompanied by a significantly reduced production of IFNγ (6 days post-infection) by splenocytes ex vivo measured under non-stimulated conditions (4–5×decrease), or in response to PMA/ionomycin, or to IL12p70 treatment (Figure 7E). We also noted a 16 and 3 fold reduction in the level of serum IFNγ in infected ECM-resistant BXH2 and F1 mice compared to B6 controls, respectively (Figure 7D). These studies suggest that heterozygosity and homozygosity for a loss of function allele at Irf8 (Irf8R294C) impacts the activity of both myeloid cells and T cells during P. berghei infection (resulting in anergy of CD8+ T cells).
Irf1 cooperates with Irf8 to regulate gene expression [1], and Irf1−/− mice are resistant to ECM (Figure 5). Therefore, we also conducted immunophenotyping studies of Irf1−/− mice before (d0) and during P. berghei infection. As previously described [48], Irf1−/− mice display defective lymphocyte maturation, with normal CD4+ and greatly reduced CD8+ T cells numbers (Figure 7A–C). Following P. berghei infection, Irf1−/− mice exhibit slight reduction in IL12p40 serum level (p = 0.0585) and a significant decrease in IL12p40 production by splenocytes (Figure 6E–F). Altered IL12p40 production and reduced CD8+ T cell numbers (Figure 7A, C) is associated with severe reduction in IFNγ production, both in serum and splenocytes cultures supernatants (Figure 7D–E). Irf1−/− splenocytes produce much less IFNγ than ECM-susceptible B6 mice in response to external secondary trigger such as PMA/ionomycin (Figure 7E). Taken together, the combined analysis of Irf1 and Irf8 mutant mice associates reduced pro-inflammatory cytokines production (IL12p40 and IFNγ) to resistance to ECM.
The demonstrated role of IRF8 in the ontogeny of the myeloid lineage, its known role in defense against infectious pathogens and the growing body of evidence from GWAS in humans linking IRF8 variants to chronic inflammatory conditions such as multiple sclerosis [36], systemic lupus erythematosus [32] and inflammatory bowel disease [34], [35], [49] prompted us to investigate a possible role for IRF8 in acute pathological inflammatory reactions. To do so, we used a mouse model of acute cerebral malaria encephalitis caused by infection with PbA, which involves lethal neuroinflammation associated with recruitment of inflammatory mononuclear and polymorphonuclear leukocytes, and loss of integrity of the blood brain barrier (BBB). We found that the severe loss of Irf8 function in BXH2 mice completely protects against this pathology, preventing the development of neurological symptoms and prolonging survival post-infection. Interestingly, the protective effect was inherited in a co-dominant fashion as 50% of Irf8R294C/+ F1 heterozygotes survived through the cerebral phase when infected with PbA (Figure 1A–B). These finding establish that Irf8 is critical to the development of acute lethal neuroinflammation and further implicate Irf8 as a major regulator of this pathological response. Moreover, results from Irf8R294C/+ F1 heterozygotes indicate that Irf8 regulates key pro-inflammatory cells and pathways in a gene dosage dependent fashion.
In addition to its established role in ontogeny and function of myeloid cells, Irf8 is required for certain aspects of B lymphocyte development and T lymphocyte function including Th1 and Th17 responses [5], [23], [50]. Cellular immunophenotyping was conducted in naïve and PbA-infected tissues to identify the cell population(s) associated with and most likely responsible for pathological inflammation, whose absence is associated with ECM-resistance in BXH2 and in [BXH2×B6]F1 mice. These studies showed that a) ECM-resistance is independent of the immature myeloid hyperplasia (CD11b+/Ly6C+) characteristic of BXH2, as this trait is absent from resistant [BXH2×B6]F1 (Figure 6A), and b) is linked to brain infiltration of both myeloid (CD11b+/Ly6C+) and lymphoid cells (TCRβ+/CD4+ and TCRβ+/CD8+ T cells) in B6 mice (Figure 1E), and c) is concomitant with reduced production of IL12p40 by myeloid cells and impaired production of IFNγ by T cells following infection with P. berghei.
To identify the gene dosage dependent pathways that are activated by Irf8 during neuroinflammation, we compared brain transcript profiles from PbA-infected B6 and BXH2 mice and extracted a list of genes that are induced by infection in an Irf8-dependent and independent fashion (two-factor ANOVA and pairwise analyses) (Figure 2, Table S1). In parallel, we carried out ChIP-seq experiments to map genome-wide IRF8 binding sites. We compared the position of these binding sites to the gene lists generated by transcript profiling and identified both IRF8-bound genes (Table S2), and IRF8-bound genes regulated in an Irf8-allele specific fashion (Tables S1). There was substantial overlap between these gene sets, which were similarly dominated by markers of antigen-presenting cells (APC) including antigen processing and presentation, production of type I interferon, pro-inflammatory cytokines/chemokines and others. These analyses confirm that Irf8 plays a prominent role in the unique functions of APCs including antigen capture and microbial phagocytosis (C1q, C4b, Fcgr4, Fcgr1), cytoplasmic inflammasome platforms (Nlrc5, Ifi205), phagosome maturation and recruitment of small GTPases (Irgm1, Irgm2, Igtp, Gbp2, Gbp3), endoplasmic reticulum membrane associated antigen transport (Tap1, Tap2), and both Class I and Class II MHC-dependent antigen presentation in APCs (B2m, H2-A, D, K, L, Q, T molecules). These gene lists also featured a number of inflammatory cytokines and chemokines involved in chemotaxis of myeloid and lymphoid cell types to the sites of infection (Ccl4, Ccl5, Ccl7, Ccl12, Cxcl9, Cxcl10). Network analysis of the IRF8 targets bound and activated in response to PbA infection confirms that Irf8 directly regulates several myeloid-specific, pro-inflammatory pathways that are ultimately responsible for pathological inflammation (Figure S1).
These findings are compatible with a simple functional model where myeloid cells (including APCs) and lymphoid cells are rapidly recruited in large numbers to the site of PbA infection and associated tissue injury - namely capillaries of the blood brain barrier. This initiates a robust IRF8-dependent pro-inflammatory cascade. Local amplification of this response by recruited cells leads to excessive production of immunopathological soluble mediators such as IFNγ and TNFα by T lymphocytes and induces other transcription factors including Stat1 and other IRF family members (Irf1, Irf7, Irf9). Damage to the BBB causes further infiltration of pro-inflammatory cytokine-secreting cells (Figure 1E), and ultimately appearance of lethal neurological symptoms in susceptible mice. Absence of Irf8 blunts this pathological response and allows mutant BXH2 mice to avoid developing neuroinflammation during ECM, thus surviving the critical acute phase. Results from immunophenotyping studies (Figures 6–7) indicate that the protective effect of Irf8 (and Irf1) mutations is associated with simultaneously diminished pro-inflammatory responses by both the myeloid and the lymphoid compartments; this is most evident for the blunted production of IFNγ in ECM-resistant BXH2, Irf1−/− or BXH2×B6]F1 mice, when compared to ECM-susceptible B6 (serum level, and by CD4+ and CD8+ T cells with or without stimulation with PMA/ionomycin).
At steady-state, BXH2 mice show a severe depletion of myeloid DC subsets but display expansion of immature myeloid progenitors, including granulocytes-like cells (Cd11b+/Ly6C+) in peripheral tissues [12]–[15]. Literature in support of [51], [52] or arguing against [53] a role for granulocytes/neutrophils in the pathogenesis of ECM has been published. However, myelocytic hyperplasia in BXH2 is not responsible for ECM resistance, since [BXH2×B6]F1 mice are also resistant and do not display this myeloid expansion (Figures 1A and 6). Although the depletion of DCs, along with a concomitant reduction in IL-12 production and antigen-specific T-cell priming in BXH2 is likely to account for a significant component of ECM-resistance, we propose that even in the context of normal myeloid cell numbers, reduced IRF8-dependent transcriptional activation of APC-specific and T-cell specific pathways is sufficient to significantly blunt inflammatory response and protect against acute pathological inflammation. This is based on several observations. First, IRF8 behaves primarily as a transcriptional activator, not a repressor, in myeloid cells as can be seen by the enrichment of IRF8 binding sites in increased genes only (Figure 3B). Table S1 also highlights that for each gene regulated in a strain and infection specific way, Irf8-competent mice invariably show a higher magnitude fold change than BXH2, and the majority of these genes are increased (Group 1 and 2) rather than decreased (Group 3). Second, Irf8R294C/+ F1 heterozygotes show normal numbers of myeloid cells (DCs, macrophages) and lymphoid cells, but still exhibit significant resistance to PbA induced ECM (Figures 1). Thirdly, the inactivation of several direct transcriptional targets of IRF8 (identified in our study as bound and regulated by IRF8 during PbA infection) including the phagosome-associated small GTPase Irgm1 (Figure 5), the pro-inflammatory cytokines Il12p40 (Figure 5 and [47]), Cxcl9 and Cxcl10 [54], the Cd11b receptor Icam1 [55], [56], and the transcriptional activator and IRF8 dimerization partner Irf1 [57] have been shown to increase resistance to ECM in mouse knockout mutants (Figure 5 and Refs [58], [59]). Finally, inactivation of additional IRF8 targets, detected herein by ChIP-seq, have previously been shown to protect against PbA-induced ECM, including Ifng [60], Jak3 [45], Cd8, Cd14, Cd40, Hc, Fcgr2, Lta and Ltbr [61]. Together, these results highlight the role of IRF8 in regulating pro-inflammatory pathways in myeloid and lymphoid cells during ECM-associated neuroinflammation.
Although IRF8-dependent activation of pro-inflammatory pathways in myeloid cells has detrimental and pathological consequences during PbA infection, it clearly plays a protective role in other infections including pulmonary tuberculosis. Indeed, using the same analysis and stringent statistical parameters, we noted a strong overlap between the list of genes increased in brains of B6 mice in response to PbA infection and up-regulated in lungs 30 days following aerosol infection with M. tuberculosis (Table S2). Of the 123 genes increased in PbA-infected brains, more than half (n = 66) were also up-regulated in M. tuberculosis-infected lungs, and nearly three quarters (90/123) of the PbA-regulated genes harbored an IRF8 binding site. Amongst the 66 genes up-regulated during both M. tuberculosis infection and during ECM, a striking 80% (n = 53) display one or more IRF8 binding sites. Furthermore, inactivating mutations in several of these common genes including Irf8 [22], Irf1 [62], and Irgm1 [63] cause susceptibility to pulmonary tuberculosis, while conveying some degree of protection against ECM. Mutations in additional ECM increased genes such as Tap1 and B2m also cause susceptibility to tuberculosis [64], while their direct effects on ECM susceptibility have yet to be tested. We propose that this set of 53 genes represents the core Irf8-dependent pro-inflammatory response pathways that play key roles in protection against TB, and pathological inflammation associated with ECM. Network analysis of these direct Irf8 targets highlights central nodes specific to myeloid cells in general, and dendritic cells in particular, including pro-inflammatory cytokines, type 1 interferon response and their transcriptional regulators.
Inactivation of Irf8 or certain of its transcriptional targets leads to complete protection against ECM. On the other hand, reduced Irf8 function causes partial protection (in Irf8R294C/+ F1 heterozygotes). Such gene-dosage dependent effects raise the possibility that even small changes in expression or activity of IRF8 may have phenotypic consequences, with increased Irf8 expression possibly associated with enhanced and/or chronic inflammation. Results from GWAS have pointed to IRF8 as one of the genetic factors implicated in the complex genetic architecture of several human inflammatory conditions. For example, SNPs ∼60 kb downstream of IRF8 are associated with increased risk for both ulcerative colitis [33] and Crohn's disease (lead SNP: rs16940202) [35], [49]. In agreement with our hypothesis, the same set of SNPs exhibit a remarkably similar association pattern with IRF8 expression levels in colon and rectum (Gros and Georges, unpublished). The findings are consistent with the notion that one or more regulatory variants increase IBD risk by enhancing intestinal IRF8 expression. In addition, IRF8 is a key risk factor in multiple sclerosis (MS), and its association with this disease has been validated in multiple GWAS and meta-analyses [36], [38]. In MS, disease risk is associated with an expression SNP (rs17445836) mapping 61 kb downstream of IRF8 which is associated with both disease susceptibility and higher peripheral blood mononuclear cell IRF8 mRNA levels [36]. Finally, a SNP near IRF8 was found associated with systemic lupus erythematosus [65], a disease where production of type I interferon is central to pathogenesis. These results not only support a role for IRF8 in human chronic inflammatory conditions but further suggest that, in agreement with our results in mice, even modest changes in expression or activity of Irf8 in the context of persistent microbial or autoimmune stimulus, may lead to chronic or pathological inflammation. Furthering this proposal, we note that several IRF8 targets regulated during neuroinflammation in PbA-infected mice have also been detected as genetic risk factors in GWAS of human chronic inflammatory conditions, including the MHC (type 1 diabetes, rheumatoid arthritis, lupus, MS, psoriasis), CCL7 (IBD), IRF1 (IBD), IRF7 (Lupus) and ICAM1 (IBD) (Table S3). This highlights the role of IRF8 and its regulated pathways in pathological inflammation in humans.
The mouse model of acute neuroinflammation induced by PbA infection has proven valuable to identify novel genes, proteins and pathways involved in pathological inflammatory conditions. This model may help prioritize genes identified in human GWAS for therapeutic development, including assessing activity of novel anti-inflammatory drug candidates for use in common human inflammatory conditions.
All mice were kept under specific pathogen free conditions and handled according to the guidelines and regulations of the Canadian Council on Animal Care. Mice experimentation protocol was approved by the McGill Facility Animal Care Committee (protocol number: 5287).
C57BL/6J (B6), BXH2, Il12p40−/−, Irf1−/−, and Isg15−/− mutant mice were obtained from The Jackson Laboratory (Bar Harbor, ME). Stat1−/− mutant mice were purchased from Taconic Farms (Germantown, NY). Ifng−/− deficient mice were obtained from Dr. M. M. Stevenson (Montreal General Hospital Research Institute), Ifit1−/− mutant mice were obtained from Dr. M. Diamond (Washington University School of Medicine, St-Louis), Irgm1−/− from Dr. J. D. MacMicking (Yale, New Haven, CT) and Nlrc4−/− from Millenium Pharmaceuticals, Inc. and Dr. R. A. Flavell (Yale, New Haven, CT). [BXH2×B6]F2 mice were generated by inter-crossing [BXH2×B6]F1 mice.
P. berghei ANKA (PbA) was obtained from the Malaria Reference and Research Reagent Resource Centre (MR4), and was stored frozen at −80°C. Prior to experimental infections, PbA was passaged in B6 mice until peripheral blood parasitemia levels reached 3–5%, at which point animals were euthanized by CO2 inhalation, exsanguinated and an infectious stock was prepared. All experimental infections were done via intraperitoneal (i.p.) injection of 106 parasitized red blood cells (pRBC). Blood parasitemia was monitored during infection by microscopic examination of thin-blood smears stained with Diff-Quick (Dade Behring, Newark, DE, USA). The appearance of neurological symptoms (shivering, tremors, ruffled fur, seizures, paralysis) associated with cerebral malaria (ECM) was monitored closely, and affected animals were immediately sacrificed as previously described [39]. Survival curves were compared using Kaplan-Meier statistics.
To monitor the integrity of the blood brain barrier during experimental ECM, groups of control and PbA-infected B6 and BXH2 mice were injected intraperitoneally with 0.2 ml of 1% Evan's Blue dye (E2129; Sigma-Aldrich, Oakville, ON, Canada) in sterile phosphate-buffered saline (PBS) on d7 and d16 (BXH2 only) post-infection (n = 3 mice/condition). The dye was allowed to circulate for 1 h, then the mice were sacrificed by CO2 inhalation, perfused with PBS and the brains were dissected and photographed.
B6, BXH2, [BXH2×B6]F1 and Irf1−/− mice were infected with PbA and sacrificed at day 6 post-infection. Whole blood was collected by cardiac puncture and serum separated by centrifugation. Levels of circulating IL12p40 and IFNγ were measured using a commercially available ELISA kit, according to manufacturer's instructions (Biolegend and eBioscience, respectively). Spleens from PbA-infected mice were collected, single-cell suspensions prepared and re-suspended in complete RPMI. 4×106 cells were plated in 48-well tissue culture plates and were stimulated with PMA/Ionomycin (eBioscience) or IL12p70 (Biolegend) for 48 hours. Culture supernatants were collected and assayed for IL12p40 and IFNγ production.
Naïve or PbA infected B6, BXH2, and [BXH2×B6]F1 mice were sacrificed, exsanguinated and perfused with PBS containing 2 mM EDTA. Spleens were collected, single-cell suspensions prepared and resuspended in FACS buffer (PBS, 2% FBS, 0.02% NaN3). Infiltrating brain leukocytes were enriched by Percoll gradient centrifugation, as previously described [66]. Briefly, brains were gently disrupted using a dounce homogenizer, and then separated over discontinuous 70/30% Percoll gradients. Cells from the interface were collected, washed and resuspended in FACS buffer. Cells were surface stained for 30 minutes in the dark at 4°C with the following cocktails: APC-eFluor780 anti-CD45 (30-F11), APC anti-TCRβ (H57-597), PE anti-CD4 (GK1.5), FITC anti-CD8a (53-6.7) to stain lymphoid cells, or APC-eFluor780 anti-CD45 (30-F11), PE-Cy7 anti-CD11b (M1/70), PE anti-Ly6C (HK1.4), BV421 anti-Ly6G (1A8, Biolegend), eFluor660 anti-F4/80 (BM8), FITC anti-MHCII (M5/114.15.2), and PerCP-Cy5.5 anti-CD11c (N418) to stain myeloid cells. Acquisition was performed using an eight-color FACS Canto II flow cytometer (BD Biosciences) and data analyzed using FlowJo software (Tree Star). For the spleen, between 5×104 and 105 live cells were acquired; for the brain, a maximum number of cells was acquired, but did not exceed 5×103. Cell aggregates were gated out based on the forward scatter (FSC)-height versus FSC-area plot and the live cell gate established based on the side scatter (SSC)-area versus FSC-area. Leukocytes were gated as CD45+ cells in the spleen and as CD45hi cells in the brain; microglia were excluded based on CD45low staining. Data were expressed as the percentage of total CD45+ cells or absolute numbers according to total splenocyte numbers (Figure S2B). All antibodies are from eBioscience, unless otherwise stated.
Whole brains were dissected from B6 and BXH2 mice either prior to (d0) or d7 post infection (n = 3/condition). Total brain RNA was isolated using TRIzol reagent (Invitrogen, Burlington, Canada) according to the manufacturer's instructions, followed by further purification with RNeasy columns (Qiagen, Toronto, Canada) and hybridized to Illumina MouseWG-6 v2.0 microarrays at Genome Quebec Innovation Centre, Montreal, Canada. Unsupervised principal components analysis was done in R, using the lumi [67] package to transform with vst (variance stabilizing transformation) and to perform quartile normalization. For other analyses, microarray expression data was log2-transformed, median normalized and analyzed using GeneSifter (Geospiza) software. Groups were compared using either pairwise t-tests (≥2-fold cutoff, Benjamini-Hochberg corrected padj-values<0.05) or two-factor ANOVA (≥2-fold cutoff, Benjamini-Hochberg corrected padj-values<0.05) to identify genes whose expression is modulated in a strain-dependent, infection-dependent and/or interactive fashion. Lists of genes that were differentially expressed were clustered according to fold change using Multi Experiment Viewer [68]. Raw data can be accessed through the Gene Expression Omnibus (# pending).
The J774 mouse macrophage cell line was grown to 80% confluence in complete Dulbecco's modified Eagle's medium (DMEM). The cells plated in 150 mm tissue culture-grade Petri dishes (Corning Inc., Corning, NY) were treated with 400 U/ml IFNγ (Cell science, Canton, MA) and CpG DNA oligonucleotides (5′-TCCATGACGTTCCTGACGTT-3′) for 3 h. Chromatin immunoprecipitations were performed as previously described with few modifications [69]. Briefly, treated cells were crosslinked for 10 min at 20°C with 1% formaldehyde in culture medium. Crosslink was stopped with ice-cold PBS containing 0.125 M glycine for 5 min. Nuclei were prepared and chromatin was sonicated with a Branson Digital Sonifier (Branson Ultrasonics, Danbury, CT) to an average size of 250 bp. Sonicated chromatin was incubated overnight on a rotating platform at 4°C with a mixture of 20 µl Protein A and 20 µl Protein G Dynabeads (Invitrogen, Carlsbad, CA) pre-bound with 6 µg of normal goat IgG (sc-2028) or IRF8 (sc-6058×) antibodies (Santa Cruz Biotechnologies, Santa Cruz, CA). Immune complexes were washed sequentially for 2 min at room temperature with 1 ml of the following buffers: Wash B (1% Triton X-100, 0.1% SDS, 150 mM NaCl, 2 mM EDTA, 20 mM Tris-HCl pH 8), Wash C (1% Triton X-100, 0.1% SDS, 500 mM NaCl, 2 mM EDTA, 20 mM Tris-HCl pH 8), Wash D (1% NP-40, 250 mM LiCl, 1 mM EDTA, 10 mM Tris-HCl pH 8), and TEN buffer (50 mM NaCl, 10 mM Tris-HCl pH 8, 1 mM EDTA). After decrosslinking, the DNA was purified with QIAquick PCR purification columns following manufacturer's procedure (Qiagen, Mississauga, Ca). IRF8 ChIP efficiency relative to the IgG control was assessed by qPCR using the Perfecta SYBR green PCR kit (Quanta Bioscience, Gaithersburg, MD) for known IRF8 binding sites [3] (oligonucleotide sequences are available upon requests).
A total of 8 independent ChIPs were pooled for each condition (IRF8 and IgG). Libraries and flow cells were prepared following Illumina's recommendation (Illumina, San Diego, CA), with a size selection step targeting fragments between 250 and 500 bp. The ChIP libraries were sequenced on Illumina HiSeq 2000 sequencer. The sequencing yielded 86 and 79 million 50 bp sequence reads for IgG control and IRF8 samples, respectively. The reads were mapped to the mouse mm9 genome assembly using Bowtie with the following parameters: -t –solexa1.3-qual –sam –best mm9 [70]. The mapping efficiency was 91.7% for IgG and 91.9% for IRF8 samples. To identify IRF8 binding peaks, we used the MACS 1.4.1 peak finder with the following parameters: –bw 250 –mfold 7,30 –pvalue 1e-5 -g mm [42]. This analysis yielded 11216 genomic regions bound by IRF8 with p-values under the threshold of 10−5. The genes identified as affected by PbA infection in the expression profiling experiment were queried for the presence IRF8 binding peaks in a 20 kb interval around the gene transcription start site (TSS). This analysis was also performed for all the genes represented on the Illumina mouse WG-6 v2.0 array used in the microarray experiments, to assess the background association of IRF8 peaks with surrounding genes (Figure 3B).
The list of all genes differentially regulated in B6 mice during infection (B6 d7/d0 pairwise) was uploaded into Ingenuity Pathway Analysis and networks were generated based on known direct or indirect interactions from published reports and the IPA databases. Seventeen networks were constructed and the three most significant were re-drawn in Adobe Illustrator CS4 14.0.0 (Adobe Systems Inc.). IRF8 binding sites from ChIP-seq data were cross-referenced and genes were colored according to their pairwise fold change during infection for each strain (Figure S1).
A second set of networks was generated according the same procedure using the list of 53 genes detailed in Table S1 possessing at least one IRF8 binding site and up-regulated by both PbA and Mtb infection (Figure 4).
|
10.1371/journal.pgen.1004273 | Folliculin Regulates Ampk-Dependent Autophagy and Metabolic Stress Survival | Dysregulation of AMPK signaling has been implicated in many human diseases, which emphasizes the importance of characterizing AMPK regulators. The tumor suppressor FLCN, responsible for the Birt-Hogg Dubé renal neoplasia syndrome (BHD), is an AMPK-binding partner but the genetic and functional links between FLCN and AMPK have not been established. Strikingly, the majority of naturally occurring FLCN mutations predisposing to BHD are predicted to produce truncated proteins unable to bind AMPK, pointing to the critical role of this interaction in the tumor suppression mechanism. Here, we demonstrate that FLCN is an evolutionarily conserved negative regulator of AMPK. Using Caenorhabditis elegans and mammalian cells, we show that loss of FLCN results in constitutive activation of AMPK which induces autophagy, inhibits apoptosis, improves cellular bioenergetics, and confers resistance to energy-depleting stresses including oxidative stress, heat, anoxia, and serum deprivation. We further show that AMPK activation conferred by FLCN loss is independent of the cellular energy state suggesting that FLCN controls the AMPK energy sensing ability. Together, our data suggest that FLCN is an evolutionarily conserved regulator of AMPK signaling that may act as a tumor suppressor by negatively regulating AMPK function.
| The FLCN gene is responsible for the hereditary human tumor disease called Birt-Hogg-Dube syndrome (BHD). Patients that inherit an inactivating mutation in the FLCN gene develop lung collapse as well as tumors in the kidney, colon, and skin. It is not clear yet what the exact function of this protein is in the cell or an organism. In this study, we used a simple model organism (the round worm C. elegans) to study the function of FLCN. We found that it is involved in the regulation of energy metabolism in the cell. FLCN normally binds and blocks the action of another protein (AMPK), which is involved in the maintenance of energy levels. When energy levels fall, AMPK is activated and drives a recycling pathway called autophagy, where cellular components are recycled producing energy. In the absence of FLCN in worms and mammalian cells, like in tumors of BHD patients, AMPK and autophagy are chronically activated leading to an increased energy level, which makes the cells/organism very resistant to many stresses that would normally kill them, which in the end could lead to progression of tumorigenesis.
| Birt-Hogg-Dubé syndrome (BHD) is an autosomal dominant neoplasia disorder that was originally described by Hornstein and Knickenberg in 1975 and by Birt, Hogg, and Dubé in 1977 as a disorder associated with colon polyps and fibrofolliculomas of the skin [1], [2]. Toro et al. recognized in 1999 that BHD patients were also predisposed to develop kidney cancer mostly of the onococytic, chromophobe, or mixed subtype [3]. However, later studies showed a predisposition for all subtypes of kidney cancer including clear cell and papillary subtypes [4]. In addition, BHD confers an increased risk of pulmonary cysts, spontaneous pneumothorax, and cysts of the kidney, pancreas, and liver [3], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14].
The gene responsible for BHD, FLCN, was mapped to chromosome 17p11.2 by linkage analysis [15], [16] and identified in 2002 by positional cloning [14]. FLCN encodes a novel cytoplasmic 64kDa protein FLCN, which is expressed in most epithelial tissues [17]. BHD patients carry a loss of function germline mutation in one FLCN allele and acquire a second hit somatic mutation or loss of heterozygosity (LOH) in the remaining wild-type copy in their renal tumors [18], [19]. In addition, strains of rats, mice, and dogs with a germline mutation in the Flcn gene developed spontaneous kidney tumors with a loss of function in the second allele pointing to a tumor suppressor function of FLCN [20], [21], [22], [23]. However, homozygous deletion of Flcn resulted in embryonic lethality in these species [24], [25]. Finally, ablation or restoration of FLCN in human cancer cells revealed tumor suppressor function in xenograft and soft agar assays [24], [26].
Though the FLCN protein presents no significant homology to any known protein, it is highly conserved from unicellular organisms (yeast) through mammalian species (rodents, dog, humans). Moreover, two 130 kDa folliculin-interacting proteins, FNIP1 and FNIP2 have been identified [27], [28], [29] and implicated in some of the FLCN phenotypes in B-cell and stem cell differentiation, and the regulation of apoptosis upon DNA damage [30], [31], [32], [33]. Several studies identified both FLCN and FNIP1/2 as AMPK (5′AMP-activated protein kinase) binding proteins [28], [34], [35], [36]. However, no clear role for FLCN/FNIP1/2 in AMPK function has been described, since both inhibition and stimulation of AMPK have been reported upon loss of function of these genes [32], [37]. Strikingly, the majority of naturally occurring FLCN mutations predisposing to BHD were predicted to generate truncated proteins unable to bind AMPK pointing to an essential role of this interaction in the tumor suppressor function. Since we and others have observed that FLCN regulates cellular metabolism [37], [38], [39], we hypothesized that FLCN may regulate cellular energy metabolism through its interaction with AMPK.
AMPK is an evolutionarily conserved master regulator of energy metabolism [40], [41], [42]. When energy levels drop, AMP or ADP bind to the γ regulatory subunit of AMPK and induce an allosteric conformational change [43], [44]. This change leads to the activation of AMPK through phosphorylation of a critical threonine residue (Thr172) in the catalytic subunit and inhibition of its dephosphorylation. When animals and cells encounter stressful environmental conditions leading to lower energy levels, activated AMPK phosphorylates downstream metabolic targets to generate ATP and maintain bioenergetics [40], [41], [42]. For instance, AMPK activates autophagy, a lysosome-dependent degradation process that recycles cytosolic components to generate new cellular components and produce energy [45]. Recently, AMPK was shown to activate autophagy via binding and phosphorylation of the autophagy initiation kinase ULK1, Beclin 1, and Vps34 [46], [47], [48].
Since studies in mammalian cells have led to unclear roles for FLCN in AMPK function, we decided to study the genetic relationship between FLCN and AMPK in the model organism C. elegans. FLCN and AMPK are conserved in C. elegans and loss-of-function mutants are viable. AMPK activation promotes lifespan extension in C. elegans [49], [50], [51], [52] and increases resistance to oxidative and other stresses [51], [52], [53], [54], [55], [56], [57].
Here we show that FLCN controls a distinct evolutionarily conserved energy stress pathway by acting as a negative regulator of AMPK function. Loss of FLCN function led to constitutive AMPK activation, which increased autophagy, resulting in inhibition of apoptosis, higher bioenergetics, thereby enhancing survival to several metabolic stresses. Specifically, we find that the chronic activation of autophagy upon loss of FLCN modifies cellular metabolism, providing an energetic advantage that is sufficient to survive metabolic stresses such as oxidative stress, heat, and anoxia. We confirmed these C. elegans results in FLCN- deficient mouse embryo fibroblasts (MEFs) and human cancer cells demonstrating strong conservation of this pathway throughout evolution. Our results demonstrate that FLCN inhibits AMPK and autophagy functions, which may lead to inhibition of tumorigenesis.
The FLCN gene product is highly conserved from C. elegans to humans with 28% identity and 50% similarity (Figure 1A). To determine the function of FLCN and whether it genetically interacts with AMPK in C. elegans, we used a strain carrying the flcn-1(ok975) mutation. The ok975 mutation is an 817 bp insertion-deletion, predicted to truncate the protein at residue 141 resulting in a null or loss-of-function allele (Figure 1B). In accordance, the C. elegans FLCN-1 polyclonal antibody that we developed recognized a gene product at the predicted size in N2 wild-type but not in flcn-1(ok975) animals (Figure 1C). Importantly, we did not detect obvious developmental or morphological defects in flcn-1(ok975) animals compared to wild-type.
The C. elegans AMPK ortholog (aak-2; α2 catalytic subunit) modulates longevity and tolerance to stresses including oxidative stress, heat, anoxia, and dietary restriction [49], [50], [51], [52], [54], [56]. Since we did not observe a difference in lifespan between wild-type and flcn-1(ok975) animals (Figure 1D and Table S1), we investigated the function of FLCN-1 in stress response by treatment of animals with mild and acute oxidative stress. The flcn-1(ok975) mutation conferred an increased resistance to 4 mM and 100 mM paraquat (PQ), a superoxide inducer [58], which could be rescued in two different transgenic lines expressing FLCN-1 (Figures 1E and 1F and Tables S2 and S3). In addition, treatment with flcn-1 RNAi increased the resistance of wild-type animals to low and high concentrations of PQ but did not further increase the resistance of the flcn-1(ok975) mutant animals, supporting that the ok975 mutation is a loss-of-function allele (Figures 1G and 1H). A similar resistance phenotype was observed upon H2O2 treatment and was rescued with the two transgenic lines expressing flcn-1 (Figures 1I and S1A). To exclude the possibility that the changes in stress resistance are not due to effects on lifespan of the animals, assays performed on 4 mM PQ were accompanied with lifespan controls (Table S2). In conclusion, these results demonstrate that loss of FLCN increases resistance to oxidative stress in C. elegans.
Since FLCN binds to AMPK in mammalian cells, we aimed to determine whether flcn-1 and aak-2 interact genetically in C. elegans. Similarly to published results [50], [51], [52], [54], aak-2(ok524) mutant animals were more sensitive to PQ stress compared to wild-type (Figures 2A and 2B and Tables S2 and S3). Strikingly, flcn-1(ok975); aak-2(ok524) double mutant animals displayed reduced survival upon treatment with 4 mM PQ (Figure 2A and Table S2) and 100 mM PQ (Figure 2B and Table S3), similarly to aak-2(ok524) single mutants, indicating that aak-2(ok524) is required for the flcn-1(ok975) phenotype. The C. elegans AMPKα1 homolog (AAK-1) was previously shown to be dispensable for oxidative stress resistance [54]. Accordingly, the aak-1(tm1944) mutation did not abolish the increased survival of flcn-1(ok975) mutants to PQ (Figure 2C and Table S3). To further test whether the increased survival of flcn-1(ok975) mutants was also dependent on PAR-4, the C. elegans ortholog of LKB1 and major upstream kinase of AMPK [40], [54], we measured survival to PQ upon par-4 loss. Interestingly, par-4(it57), a strong loss of function allele, only partially suppressed the flcn-1(ok975) survival phenotype, leading to a significant increase in the resistance of flcn-1(ok975); par-4(it57) animals to PQ compared to par-4(it57) animals alone, suggesting that additional inputs might activate AAK-2 to mediate survival (Figure 2D and Table S3).
Based on these results, we anticipated that loss of flcn-1 might lead to a constitutive activation of AAK-2 since the observed flcn-1 loss of function phenotype is similar to the reported AAK-2 overexpression in terms of oxidative stress resistance [49], [52]. Although we did not observe an increased abundance of phospho-AAK-2 at residue Thr234 (corresponding to Thr172 in human AMPKα) in flcn-1(ok975) animals compared to wild-type, our data demonstrate a significant increase in phospho-AAK-2 levels in flcn-1(ok975); par-4(it57) double mutants compared to par-4(it57) animals (Figure 2E and 2F). This is consistent with the observation that par-4 is not fully required for the stress resistance phenotype (Figure 2D and Table S3). To exclude the possibility that the increased phosphorylation of AAK-2 is due to a flcn-1-dependent increase in total AAK levels, we measured the mRNA expression of aak-1 and aak-2 in wild-type and flcn-1(ok975) animals and did not observe a significant difference (Figure S1B). Interestingly, the residues flanking the mammalian AMPKα (Thr172) are conserved in AAK-2 (Thr234), but are different in AAK-1 and therefore, AAK-1 is unlikely to be detected by the antibody used (Figures 2E, lane 5). The increased phosphorylation of AAK-2 only in par-4(it57) mutants can be explained by the fact that PAR-4 is the major kinase that phosphorylates AAK-2 in certain tissues of the animal or cellular sub-compartments, which would mask the FLCN-dependent phosphorylation signal on AAK-2 induced by other upstream kinases [54]. Taken together, these results imply that flcn-1 negatively regulates aak-2 in C. elegans, and that loss of flcn-1 confers an aak-2-dependent resistance to oxidative stress.
The insulin/IGF-1-like (DAF-2)/FOXO3a (DAF-16) and target of rapamycin (TOR) signaling pathways are known to control lifespan and stress response in C. elegans and other organisms and have been linked to AMPK signaling [50], [52], [59], [60], [61]. While daf-2(e1370) mutants exhibited an increased survival to PQ compared to wild-type animals, the flcn-1 mutation further increased the resistance of daf-2(e1370) animals (Figure S2A and Table S3). Consistently, daf-16(mu86) slightly reduced but did not suppress the resistance of the flcn-1(ok975) animals to PQ (Figure S2B and Table S3). Moreover, we found that the PQ resistance of the flcn-1(ok975) animals treated with TOR (let-363) RNAi was significantly higher than wild-type animals fed with the same RNAi (Table S3).
Transcriptional upregulation of ROS detoxification enzymes prior to stress could explain the increased survival of flcn-1(ok975) [62]. We did not observe a significant increase in the gene expression of superoxide dismutases (sod-1, sod-2, sod-3, sod-4, and sod-5) or catalases (ctl-1, ctl-2, and ctl-3), in flcn-1(ok975) mutants when compared to wild-type (Figure S2C). Furthermore, we quantified the oxidative damage to protein and DNA. Levels of protein carbonylation and DNA damage were equal in both wild-type and flcn-1(ok975) animals under basal conditions and were similarly induced after PQ treatment (Figures S2D and S2E). Taken together, these findings suggest that the increased survival of flcn-1(ok975) mutant to PQ may not be dependent on classical oxidative stress resistance mechanisms. A likely mechanism of survival upon loss of flcn-1 on PQ might involve inhibition of apoptosis.
Autophagy has been shown to mediate resistance to oxidative stress across evolution without a clear mechanistic explanation [63]. To investigate whether the increased oxidative stress resistance of flcn-1(ok975) mutants was due to autophagy, we measured autophagy levels using several methods. Using electron microscopy, we noticed the frequent appearance of autophagic vacuoles (Figure 3B) in flcn-1(ok975) mutants at the basal level compared to wild-type animals which increased under PQ treatment (Figures 3A and 3C). To confirm this result, we used a reporter strain that carries the integrated transgene expressing GFP::LGG-1 (LC3 ortholog). LC3 localizes to pre-autophagosomal and autophagosomal membranes, and GFP-positive puncta are thought to represent autophagosomal structures in this strain [64], [65], [66]. To exclude effects on the transgene expression, we determined the mRNA levels of LGG-1 in wild-type and flcn-1(ok975) animals and in the GFP::LGG-1 and flcn-1; GFP:: LGG-1 transgenic lines. In both cases, the transcript levels of LGG-1 were not significantly different between wild-type and flcn-1(ok975) animals demonstrating equal expression (Figures S3A and S3B). Importantly, we observed a significant increase in the number of GFP-LGG-1 positive puncta in flcn-1(ok975) mutants compared to wild-type animals under basal conditions (Figure 3D). Consistently, treatment of GFP::LGG-1 animals with flcn-1 RNAi increased the number of GFP-LGG-1 puncta (Figure 3D). Previous studies in yeast, C. elegans, and mammalian cells have demonstrated that LGG-1-II (or LC3-II) is degraded inside the autolysosomes, and that the GFP fragment is resistant to degradation and accumulates when autophagy is induced [67], [68], [69], [70]. Therefore, we performed western blot analysis on wild-type and flcn-1 protein extracts to assess the level and cleavage of GFP-LGG-1. Importantly, western blot analysis showed that both cleaved LGG-1-II and released GFP were increased in flcn-1(ok975) mutants, indicating higher autophagic activity (Figure 3E).
AMPK has recently been shown to directly induce autophagy in mammals via phosphorylation of autophagy proteins including ULK-1, VPS-34 and BEC-1 [46], [47], [48]. Moreover, loss of aak-2 reduced autophagy in daf-2 mutant animals, while aak-2 overexpression induced autophagy [46]. Based on these results, we questioned whether the increased autophagy in flcn-1(ok975) animals depends on aak-2. Importantly, RNAi treatment against aak-2 significantly reduced the number of puncta in flcn-1(ok975) mutants, demonstrating an aak-2-dependent mechanism (Figure 3D).
Inhibition of autophagy genes in C. elegans reduced survival to certain stresses such as anoxia, starvation and pathogens [66], [71], [72]. However, the requirement of autophagy genes in resistance to oxidative stress was not previously reported. We aimed to determine whether the increased survival of flcn-1(ok975) animals to oxidative stress was dependent on autophagy. Strikingly, inhibition of the essential authophagy genes atg-7 and bec-1 using RNAi markedly abolished the resistance of flcn-1(ok975) to PQ (Figures 3F–3I and Tables S2 and S3). Taken together, these results demonstrate that loss of flcn-1 induces autophagy, which is required for flcn-1-mediated stress resistance.
Autophagy is a process that generates catabolic substrates for mitochondrial ATP production and allows cellular macromolecules to be recycled. Since we did not observe a difference in oxidative damage between wild-type and flcn-1(ok975) mutant, and since PQ is known to severely decrease ATP levels by inhibiting oxidative phosphorylation [73], [74], [75], we wondered if flcn-1 is mediating an increased resistance to energy stress by employing autophagy as a source of energy. To test this hypothesis, we measured ATP levels prior and after 13 hours of 10 mM PQ treatment. Strikingly, we found that flcn-1 mutant animals have higher levels of ATP before PQ treatment compared to wild-type (Figure 4A). As expected, PQ treatment decreased the ATP levels in both wild-type and flcn-1, yet ATP levels in flcn-1(ok975) nematodes remained higher than wild-type. Importantly, flcn-1(ok975) mutants treated with PQ exhibited equal amounts of ATP when compared to the non-treated wild-type animals (Figure 4A). To test if the increased energy in flcn-1(ok975) mutants is dependent on autophagy, we treated the wild-type and flcn-1 nematodes with atg-7 RNAi and measured ATP levels. Strikingly, downregulation of autophagy completely suppressed the increased ATP levels in flcn-1(ok975) mutants in presence or absence of PQ (Figure 4A).
To further confirm that loss of flcn-1 confers resistance to low energy levels, we measured the resistance of wild-type and flcn-1(ok975) nematodes to heat stress and anoxia, both of which are known to result in a strong depletion of energy [76]. Accordingly, when exposed to 35°C, the mean survival of flcn-1(ok975) animals was significantly higher compared to wild-type (Figure 4B and Table S4). In addition, the recovery rates after a 26 hours anoxic injury were faster in flcn-1(ok975) compared to wild-type (Figure 4C and Table S5). In conclusion, our data describe a novel mechanism for AAK-2-dependent resistance to oxidative stress, which depends on maintenance of energy homeostasis via autophagy.
The interplay between autophagy and apoptosis determines the decision between life and death which is very important for the genetic integrity of the cell [77]. The activation of autophagy has been shown to protect against cell death in C. elegans and mammals [72], [77], [78], [79]. To see whether flcn-1 controls apoptosis in animals, we determined the number of apoptotic cell corpses in the gonad arms of wild-type and flcn-1 animals upon PQ treatment. As expected, we found that PQ treatment significantly increased the number of apoptotic corpses in wild-type animals (Figure S4A). However, the increase in flcn-1 was much lower suggesting that loss of flcn-1 protects against cell death. To determine whether the decreased cell death in flcn-1 nematodes depends on the activation of autophagy, we pretreated the wild-type and flcn-1(ok975) nematodes with atg-7 RNAi and then measured the number of apoptotic corpses upon PQ treatment. Importantly, the inhibition of the autophagy gene atg-7 increased the number of apoptotic corpses, up to the same level, in both wild-type and flcn-1 suggesting that the FLCN-1-dependent activation of autophagy protects against cell death (Figure S4A).
Importantly, the apoptotic pathway is conserved from animals to mammals. When cells are destined to die, the BH3 only protein EGL-1 binds and inhibits the BCL-2 homolog CED-9, which activates the caspase CED-3 and leads to death [80]. Therefore, we treated wild-type and flcn-1(ok975) animals with egl-1 and ced-3 RNAi and assessed their survival to 100 mM PQ. Importantly, downregulation of egl-1 or ced-3 using RNAi increased the resistance of wild-type animals which was not observed in flcn-1(ok975 animals suggesting that the inhibition of the apoptotic pathway leads to the increased resistance (Figures S4B–S4D and Table S3). Consistently, treatment of wild-type and flcn-1(ok975) animals with ced-9 RNAi abolished the survival of flcn-1 animals (Figure S4 and Table S3). Importantly treatment of aak-2(ok524) animals with ced-3 RNAi did not increase their resistance to PQ suggesting that this phenotype depends on AAK-2 (Table S3). The increased stress response by inhibition of apoptosis has been recently reported [81]. Although it is not clear whether the apoptosis inhibition in the gonad arms delays organismal death in C. elegans or whether apoptotic genes acquire non-apoptotic functions, our data suggest an AMPK-dependent involvement of the apoptotic pathway in the increased survival of flcn-1 animals to PQ stress (Table S3).
To test whether the FLCN functions that we identified in C. elegans are evolutionarily conserved, we used wild-type (Flcn+/+) and knockout (Flcn−/−) MEFs. First, we examined the cellular resistance to serum starvation (-FBS), which reduces energy levels and induces oxidative stress in a physiological manner (Figure 5A) [58]. Flcn−/− MEFs were unaffected by serum starvation, as demonstrated by a significant maintenance of cell survival after 4 days of serum starvation, which eliminated almost 80% of wild-type cells. Rescue of wild-type FLCN expression (resc.) reverted this protective phenotype (Figure 5A). Consistent with these data, Flcn−/− MEFs were more resistant to 2 mM H2O2 treatment compared to wild-type (Figure S5A).
Moreover, in accordance with the C. elegans results, phospho-AMPK levels were increased upon loss of Flcn in MEFs, which could be rescued by expression of wild-type Flcn, suggesting that FLCN acts a negative upstream regulator of AMPK (Figures 5B and S2B). Additionally, down regulation of Flcn by shRNA in MEFs lacking AMPKα (Ampk−/−) did not increase resistance to serum starvation suggesting that the increased resistance to starvation-induced stress depends on AMPK (Figures 5C and S5B).
Next, we asked whether the AMPK activation in Flcn−/− MEFs could be further activated. Importantly, phosphorylation levels of AMPK and its target ACC were maximal in Flcn−/− cells and did not further increase upon serum starvation (Figure S5C). Similarly, treatment with AICAR (5-amino-1-β-D-ribofuranosyl-imidazole-4-carboxamide), an AMP analogue, increased AMPK activation in wild-type as marked by elevated pACC levels but not in Flcn−/−MEFs (Figure S5C). These results demonstrate that loss of FLCN leads to maximal AMPK activation, which appears uncoupled from its energy sensing function.
Moreover, we wondered whether loss of FLCN also increases autophagy in MEFs similarly to the results obtained in C. elegans. Importantly, Flcn−/− MEFs displayed an increased number of autophagosomes at the basal level and under serum starvation conditions when compared to wild-type cells as determined by the GFP-LC3 reporter (Figures 5D and 5E). To validate these results, we used a GFP-mCherry/LC3 reporter [82]. Upon physiological pH in newly formed autophagosomes or when autophagy is impaired, both GFP and mCherry colocalize in puncta whereas upon lysosomal fusion and acidification, the GFP signal is lost and only mCherry is detected. As expected, the number of mCherry puncta was increased in Flcn−/− MEFs, pointing to a normal lysosomal acidification and completion of autophagy (Figure S6A). In addition, chloroquine (CQ) treatment, which inhibits the acidification of autolysosomes, further increased the number of autophagosomes in Flcn−/− MEFs suggesting that the autophagy process is not impaired (Figures 5E and S6A). We also measured the rate of conversion of LC3I to LC3II by western blot analysis. The ratio of LC3II to LC3I was increased in Flcn−/− MEFs at the basal level and was reverted by FLCN rescue (Figure S6B). In agreement with the heightened autophagy, we observed an increase in the activating AMPK-dependent phosphorylation site at the autophagy initiating kinase ULK1 (Figure S5C). To determine whether the increased resistance to serum starvation was due to increased autophagy as we observed in C. elegans, we inhibited autophagy using CQ or Bec-1 shRNA. Inhibition of autophagy strongly suppressed the survival advantage upon serum starvation in Flcn−/− MEFs (Figures 5F and S6C). Finally, we measured apoptosis in response to serum starvation and inhibition of autophagy in MEFs (Figure S7D). Apoptosis was strongly increased in wild-type MEFs and suppressed in Flcn−/− MEFs upon serum starvation. Furthermore, inhibition of autophagy using CQ abolished the suppression of apoptosis in Flcn−/− MEFs. In conclusion, these results correspond well with the data we obtained in C. elegans.
Next, we determined whether the chronic activation of autophagy is leading to an energy surplus, which is required for the stress resistance phenotype conferred by loss of FLCN. Similarly to what we found in C. elegans, Flcn−/− MEFs displayed increased ATP levels under basal conditions (Figure 5G). Serum starvation decreased ATP levels in wild-type MEFs, while Flcn−/− MEFs maintained ATP levels at wild-type levels (Figure 5G). Importantly, inhibition of autophagy with CQ abolished the increase in ATP in Flcn−/− MEFs (Figure 5G). The ATP, ADP, and AMP levels as well as phospho-creatine, a short term energy reserve, were all increased at the basal level in Flcn−/− cells as compared to wild-type, suggesting a general increase in cellular metabolism and energy storage (Figures S7A and S7C). Importantly, serum starvation decreased energy levels in Flcn−/− cells down to wild-type basal levels (Figures S7A and S7C).
Normally, AMPK is activated upon energy deprivation (increased ADP/ATP and AMP/ATP ratios), which is inconsistent with the increased basal ATP levels conferred by loss of FLCN [40], [42]. To understand this discrepancy, we determined the adenylate energy charge of wild-type and Flcn−/− MEFs. The adenylate energy charge, which is expressed by the ratio [ATP] + 0.5[ADP]/[ATP]+[ADP]+[AMP], was proposed as a convenient indicator of the cellular energy status [83]. Under normal conditions, ATP levels were increased in Flcn−/− MEFs but the total energy charge was comparable to wild-type (Figure S7B). However, serum starvation significantly reduced the energy charge in wild-type MEFs but not in Flcn−/− MEFs, suggesting that Flcn−/− cells derive their energy from an intracellular source of ATP [83] (Figure S7B).
To determine whether loss of FLCN in human cancer cells also conferred an advantage in energy homeostasis, we used the follicular thyroid carcinoma cells FTC-133 lacking FLCN expression, which we rescued for FLCN using stable transfection. First, we confirmed the findings that we obtained with wild-type and Flcn−/− MEFs. As expected, loss of FLCN conferred an increased phosphorylation of AMPK as well as a higher LC3 cleavage demonstrating that AMPK activation and autophagy in FLCN-deficient FTC cells is elevated compared to rescued cells (Figure S8A). Next, we aimed to determine whether the increased autophagy promoted by loss of FLCN heightens ATP levels at normal conditions in FTC cells. Similarly to what we observed in MEFs, loss of FLCN requires autophagy to increase ATP levels (Figure S8B). To determine whether autophagy contributes to anchorage-independent growth conferred by loss of FLCN, we performed soft agar assays in the presence or absence of autophagy inhibition using atg7 shRNAs. As expected, loss FLCN significantly increased the number of colonies growing in soft agar in an autophagy-dependent manner (Figures S8C and S8D). Taken together, these data demonstrate that loss of FLCN leads to an autophagy-dependent increase in ATP levels enabling FLCN-deficient animals/cells to resist metabolic stresses, which could constitute a tumor suppression mechanism.
Maintenance of cellular bioenergetics and management of oxidative stress are essential for life. Here we highlight the discovery of an evolutionarily conserved signal transduction pathway mediated by the tumor suppressor FLCN and AMPK that is essential for resistance to metabolic stress. Loss of FLCN in C. elegans and mammalian cells leads to constitutive activation of AAK/AMPK, which in turn increases autophagy. Chronic activation of autophagy leads to increased ATP production and confers resistance to energy depleting stresses by inhibition of apoptosis. Together, our data identify FLCN as a key regulator of stress resistance and metabolism through negative regulation of AMPK.
Several questions arise from these results. First, how is AAK-2 being activated in flcn-1(ok975) animals upon loss of PAR-4? Several upstream AMPK kinases other than LKB1 have been identified in mammalian cells, have been shown to affect AMPK activity [40]. Although these kinases have not been linked to AAK-2 in C. elegans, our data showed a significant basal phosphorylation of AAK-2 in par-4(it57) mutant animals. This is in agreement with a recently published study showing that starvation and mitochondrial poisons increased phospho-AAK-2 levels in par-4(it57) mutant animals, and that the starvation-induced aak-2 phenotypes were partially dependent on PAR-4 [84]. The fact that the AMPK signaling pathway is evolutionarily conserved suggests that AMPK upstream kinases other than PAR-4 are likely to exist in C. elegans. For instance, Pak1/Camkk-beta was first identified in yeast as Snf-1/AMPK-activating kinase and was proven later to act upstream of AMPK in mammalian systems [85], [86], [87]. Very recently, CAMKII overexpression was shown to increase lifespan in C. elegans, although the link to AAK-2 was not investigated [88].Together, our data demonstrate that FLCN-1-dependent regulation of AAK-2 mediates an important novel pathway for stress resistance. Interestingly, this pathway is distinct from previously described AAK-2-mediated oxidative stress resistance mechanisms that involve ROS detoxification [51], [52], [89].
Another unexpected finding is that loss of flcn-1 did not modulate C. elegans longevity under normal growth conditions. The observed increased AAK-2 activation upon loss of flcn-1 is masked by PAR-4-dependent phosphorylation of AAK-2. General overexpression of AMPK extends lifespan and increase stress resistance [46], [49], [50], [51], [52], [53], [56], [90]. Our data suggest that the signaling cascade downstream of FLCN-1/AAK-2 is different from the AAK-2 responses that modulate longevity. It is possible that the PAR-4-dependent activation of AAK-2 extends lifespan and increases stress resistance, while the AAK-2 activation by other upstream kinases only increases resistance to stress. Another possibility would be that a tissue-specific or sub-cellular AAK-2 activation might lead to different outcomes.
Importantly, our data indicate that loss of FLCN-1 extends lifespan only upon treatment with high concentrations (100 uM) of the DNA synthesis inhibitor 5-fluoro-2′-deoxyuridine (FUDR) (Figure S9F and Table S1), a phenotype that has been recently reported by Gharbi et al. [91]. This drug is frequently used in C. elegans aging studies to prevent eggs from hatching and has been recently reported to “artifactually” affect lifespan in mitochondrial C. elegans mutants and modulate metabolism in the daf-2 mutant strain [92], [93], [94]. It is not clear why flcn-1(ok975) animals exhibit an extension of lifespan only upon treatment with FUDR and what would be the potential mechanism of FUDR action. FUDR may act as metabolic stressor especially that high FUDR concentrations above 100 uM seem to be required to observe the effect on lifespan in flcn-1(ok975) animals (Figure S9). Interestingly, lower concentrations of FUDR (5–10 uM) that are also sufficient to inhibit germ line proliferation had no effect on lifespan.
Here we show that the enhanced resistance to oxidative stress in the absence of FLCN-1 does not result from a decrease in oxidative damage or an increased transcriptional upregulation of ROS-detoxifying enzymes [51], [52], [89]. Instead, we show that loss of FLCN-1 activates AAK-2 thereby inducing autophagy. Accordingly, downregulation of unc-51, the ortholog of the autophagy kinase ULK1, was shown to suppress the increased number of positive GFP::LGG-1 foci upon overexpression of the kinase domain of AAK-2 in C. elegans [46]. More evidence was gathered in mammalian systems to support the direct activation of autophagy by AMPK [46], [47], [48]. In addition, we show that autophagy is required for the increased ATP in flcn-1(ok975) animals and Flcn−/− MEFs suggesting that the chronic activation of autophagy in the absence of FLCN recycles building blocks to produce ATP promoting stress resistance.
When energy levels drop in the cell, AMP or ADP bind to the γ regulatory subunit of AMPK and induce an allosteric conformational change [43], [44], which leads to AMPK activation through phosphorylation of Thr172 in the catalytic subunit via inhibition of dephosphorylation activities. It is striking that loss of FLCN induces AMPK and autophagy in flcn-1(ok975) mutant animals and Flcn−/− MEFs, which exhibit high energy levels. These observations suggest that FLCN might be involved in the control of the energy sensing ability of AMPK. The increased activation of AMPK despite high energy levels has been recently reported upon inhibition of two other inhibitors of AMPK activity [95], [96].
The roles of AMPK and autophagy in cancer are puzzling [97], [98], [99], [100]. Both AMPK activation and autophagy have been shown to acquire anti- and pro-tumor functions [97], [98], [99], [100]. Our results imply that the AMPK-dependent activation of autophagy might be essential for FLCN-deficient tumor cells to acquire an energetic advantage and drive tumorigenesis. In analogy with our results, autophagy was recently shown to be required for tumor growth in many cancer models [101], [102], [103]. A similar role for VHL, another renal tumor suppressor, in the regulation of autophagic events in renal cell carcinomas has been recently described [104]. In fact, the inhibition of autophagy by MiR-204 suppressed the tumor growth in VHL-deficient cells [104]. Moreover, the LC3B/ATG5-dependent autophagy was shown to be required for the development of VHL-deficient renal cell carcinomas in nude mice [104]. We suggest that the AMPK-dependent autophagy activation upon loss of FLCN promotes the survival of transformed cells, which normally undergo severe metabolic stresses as caused by hypoxia and lack of blood vessels [105]. In agreement with our results, four groups have recently reported that AMPK activation drives tumorigenesis via metabolic stress adaptation of different tumors [99], [106], [107], [108]. In fact, AMPK is the best-characterized target of the tumor suppressor LKB1 and it was shown recently that loss of LKB-1/AMPK had a positive effect on tumor initiation but a negative effect on tumor progression/dissemination [98], [109].
The fact that FLCN negatively regulates AMPK strongly implicates that it exerts physiological functions other than being a tumor suppressor. In the past ten years, FLCN was reported to be involved in the regulation of apoptosis, rRNA synthesis, TGF-β signaling, B-cell and stem cell differentiation, ciliogenesis, mitochondrial biogenesis, TOR signaling, epithelial polarization, and cytokinesis without the elucidation of the molecular mechanism [24], [25], [26], [31], [33], [35], [39], [110], [111], [112], [113]. While this report was under review, two reports have shown that loss of folliculin leads to mTOR inhibition and that it is involved in nutrient sensing via acting as a GTPase activating enzyme for the RAG GTPases [114], [115]. It is possible that FLCN acts in two complexes. It binds to RAGs under starvation conditions leading to mTOR inhibition, while in normal conditions FLCN would bind AMPK, inhibiting its activity. However, it is not clear how the reported inhibition of mTOR activity upon loss of FLCN could lead to tumorigenesis, since mTOR was shown to be hyper-activated in tumors of BHD patients and mice devoid of FLCN [25].
In conclusion, we used the model organism C. elegans to decipher a genetic pathway, which is regulated by FLCN. We show that FLCN negatively regulates the activity of AMPK, which leads to increased autophagy, energy and survival to metabolic stress. Moreover, we confirmed conservation of this pathway in mammalian cells and suggest that chronic AMPK activation upon loss of FLCN potentiates tumorigenesis via increased autophagy leading to metabolic stress resistance and inhibition of apoptosis, which are two hallmarks of cancer cells [116].
The FLCN-1 nematode polyclonal antibody was generated in rabbits with a purified GST-FLCN-1 recombinant protein by the McGill animal resource center services. Commercial antibodies and reagents used in this study are listed in Text S1.
C. elegans strains were obtained from the Caenorhabditis Genetics Center (see Table S6). The flcn-1(ok975) strain was outcrossed eight times to wild-type. Nematodes were maintained and synchronized using standard culture methods [117]. The RNAi feeding experiments were performed as described in [118], and bacteria transformed with empty vector were used as control. For all RNAi experiments, phenotypes were scored with the F1 generation except for aak-2 knockdown (F2). Lifespan assays were performed according to standard protocols [119].
Expression constructs were generated using the pPD95.77 vector. pRF4 rol-6(su1006) was used as a co-injector marker. Transgenic lines were generated by microinjection into the gonad of adult hermaphrodite using standard techniques. 2.8 kb endogenous promoter of flcn-1 was generated by PCR from wild-type genomic DNA (Forward primer 5′AAAACTGCAGCGTCTTCTCGTTTCACAGTAGTCA-3′ and reverse primer: 5′GCTCTAGATTGAATTCTGTAAAAACATGAATTTGA-3′) and cloned into pPD 95.77 at PstI and XbaI sites. flcn-1 cDNA was obtained from an RT–PCR reaction performed on wild-type animals RNA extracts using the following: forward primer 5′GCTCTAGAATGCAAGCAGTAATAGCACTTTGT-3′ and Reverse primer 5′CGGGATCCACGAGCAGTAGAGGTTTGAGACTG-3′. flcn-1 cDNA was subsequently cloned into pPD 95.77 (GFP expression plasmid-with flcn-1 endogenous promoter region) at XbaI and BamHI sites.
Primary MEFs were isolated from C57BL/6 E12.5 Flcn floxed mice. Ampk+/+ and Ampk−/− MEFs were described in [93]. Cell lines were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 100 U/ml penicillin and 100 µg/ml streptomycin (Invitrogen). For details on shRNA procedures, transfections, FLCN deletions and immortalization see supporting information.
Resistance to acute oxidative stress (100 mM PQ and H2O2) was determined as described in [52], [54]. Chronic oxidative stress was assessed on thirty post-fertile animals using 4 mM PQ and survival was measured daily. For heat stress, one-day adult animals were transferred to NGM plates and exposed to 35°C. Survival was measured at indicated time points until all animals died. Concerning anoxia stress, one-day old adult animals were transferred to NGM plates and left in a Bio-Bag Environmental Chamber Type A (Becton Dickinson Microbiology Systems) for 26 hours at 20°C. Recovery rates were scored at indicated time points. For MEFs, cells were seeded (2×104 cells) in 12-well plates and FBS-free media was added 24 hours after plating. Cell numbers were counted daily and survival rates were determined as the percent cell number compared to day 0.
For C. elegans, autophagy levels were assessed in hypodermal seam cells of L3 animals using the GFP::LGG-1 transgenic reporter strain DA2123 (See Table S6). For MEFs, wild-type and Flcn−/− cells were infected with the pMigR-1-LC3-GFP or the mcherry/GFP-LC3 constructs [82], seeded on coverslip (50000 cells in 6 well-plate), serum starved for 12 hours and fixed with 4% paraformaldehyde. Autophagic-GFP positive puncta were quantified in at least 200 cells. Pictures from nematodes and MEFs were acquired with a Zeiss fluorescence confocal microscope.
For C. elegans, synchronized young adults were collected and washed in M9 buffer. Pellets were treated with three freeze/thaw cycles and boiled for 15 min. ATP content in C. elegans was measured using an ATP determination kit (Invitrogen) and in MEFs using CellTiter-Glo Luminescent Cell Viability Assay (Promega). For C. elegans, levels were normalized to protein levels and in MEFs normalized to cell number.
Genomic DNAs from worm pellets were purified using Phenol/Chloroform extraction and treated with RNase A for 1 hour at 37°C. OxiSelect oxidative DNA damage ELISA assay was performed with 8 µg of DNA following manufacturer's instructions (Cell Biolabs). Protein oxidative damage was assessed using Oxyblot Protein Oxidation Detection Kit (Millipore).
Synchronized young adult nematodes were harvested and total RNA was extracted with Trizol, purified using the RNeasy kit (Qiagen). Quantitative real-time PCR (qRT-PCR) was performed using Express SYBR Green qPCR supermix (Invitrogen) and LightCycler480 system (Roche). Catalase and SOD transcripts were normalized to housekeeping genes cdc42, pmp-3, and Y45F10D.4 using Genorm [120]. AAK transcripts were normalized against cdc-42. For primer sequences see Table S7.
Cells and synchronized young adult nematodes were washed with ice-cold PBS and M9 respectively and lysed in the AMPK lysis buffer [121] supplemented with the complete protease and phosphatase inhibitors (Roche), 1 mM DTT, and benzamidine 1 µg/ml. Worm pellets were sonicated and cleared by centrifugation. Percent pAAK-2/pAMPK levels were quantified using ImageJ software and normalized for the AMPK levels.
Synchronized young adult nematodes were treated with 50 mM PQ/M9 or M9 alone for 2 hours, washed and plated on NGM plates allowing 30 minutes recovery. TEM immersion fixation and embedding was performed according to [122]. See Text S1.
FTC cells were trypsinized, counted and resuspended in complete DMEM/F12 media. Two layered soft agar assay were undertaken in six well plates. The bottom layer contains 0.6% agar in complete DMEM/F12 media. The second layer encompasses 0.3% agar mixed with 0.5 million cells. Plates were cultured at 37°C in 5% CO2.
For worms, apoptotic germ cell corpses were visualized using Acridine Orange (AO) as described in [123]. Worms were incubated for 2 hours in M9 with or without 50 mM PQ in OP50, supplemented with 2 µl/ml AO (stock of 10 mg/ml). Worms were then washed and transferred into light-protected recovery NGM plates for 45 min before visualization. In MEFs, apoptosis levels were determined using the Annexin V: PE Apoptosis Detection Kit I (BD Pharmingen) according to manufacturer's protocol. Fluorescence intensity corresponding to apoptosis levels was detected using FACSCalibur flow cytometer (excitation 488; emission 575/26; BD Biosciences).
Targeted metabolite analysis was performed on an Agilent 6430 triple quadrupole mass spectrometer equipped with a 1290 Infinity UPLC system (Agilent Technologies). Metabolites were separated using a 4.0 µm, 2.1×100 mm Cogent Diamond Hydride column (MicroSolv Technology Corporation). Quantification was accomplished using MassHunter Quantitative Analysis software (Agilent). See Text S1 for details.
Data are expressed as means ±SEM. Statistical analyses for all data were performed by unpaired student's t-test, ANOVA, using Excel (Microsoft, Albuquerque, NM, USA), SPSS (IBM, Armonk, NY, USA) and prism software (GraphPad). For lifespan curve comparisons we used the Log-rank Mantel Cox test using GraphPad from Prism Statistical significance is indicated in figures (* P<0.05, **P<0.01, ***P<0.001) or included in the supplementary tables.
|
10.1371/journal.ppat.1002414 | The Enteropathogenic E. coli (EPEC) Tir Effector Inhibits NF-κB Activity by Targeting TNFα Receptor-Associated Factors | Enteropathogenic Escherichia coli (EPEC) disease depends on the transfer of effector proteins into epithelia lining the human small intestine. EPEC E2348/69 has at least 20 effector genes of which six are located with the effector-delivery system genes on the Locus of Enterocyte Effacement (LEE) Pathogenicity Island. Our previous work implied that non-LEE-encoded (Nle) effectors possess functions that inhibit epithelial anti-microbial and inflammation-inducing responses by blocking NF-κB transcription factor activity. Indeed, screens by us and others have identified novel inhibitory mechanisms for NleC and NleH, with key co-operative functions for NleB1 and NleE1. Here, we demonstrate that the LEE-encoded Translocated-intimin receptor (Tir) effector has a potent and specific ability to inhibit NF-κB activation. Indeed, biochemical, imaging and immunoprecipitation studies reveal a novel inhibitory mechanism whereby Tir interaction with cytoplasm-located TNFα receptor-associated factor (TRAF) adaptor proteins induces their proteasomal-independent degradation. Infection studies support this Tir-TRAF relationship but reveal that Tir, like NleC and NleH, has a non-essential contribution in EPEC's NF-κB inhibitory capacity linked to Tir's activity being suppressed by undefined EPEC factors. Infections in a disease-relevant intestinal model confirm key NF-κB inhibitory roles for the NleB1/NleE1 effectors, with other studies providing insights on host targets. The work not only reveals a second Intimin-independent property for Tir and a novel EPEC effector-mediated NF-κB inhibitory mechanism but also lends itself to speculations on the evolution of EPEC's capacity to inhibit NF-κB function.
| Enteropathogenic Escherichia coli (EPEC) remain an important cause of infant diarrhoea and death in developing countries. Undoubtedly, a key pathogenic event relates to the bacteria's ability to inhibit the expression of anti-microbial and inflammation-inducing molecules regulated by the NF-κB transcription factor. This inhibitory process was previously linked to undefined EPEC effectors blocking NF-κB function. In this manuscript, we report that the most extensively studied EPEC effector, Tir, possesses a potent and specific ability to inhibit the cytokine, TNFα from activating NF-κB. Moreover, this finding leads to the discovery of a novel inhibitory mechanism relating to the induced degradation of the TNFα Receptor-Associated Factor (TRAF) adaptor protein. Additional work supports reported NF-κB inhibitory activities for multiple non-LEE-encoded effectors and key roles for two (identifying possible target proteins) leading to speculations on the evolution of EPEC's multi-effector strategy for inhibiting NF-κB activity.
| The EPEC disease process depends on a protein delivery system, encoded by the Locus of Enterocyte Effacement (LEE) Pathogenicity Island, that transfers effector proteins directly into the cytoplasm of infected epithelia [1]–[3]. This delivery apparatus is composed of a Type Three Secretion System (T3SS) and a filamentous extension - formed by the EPEC secreted/signalling protein A (EspA) tipped by EspB/EspD to form a pore in the host plasma membrane - generating a conduit for transferring effectors into host cells [1], [3]. The LEE region also encodes other factors, including the bacterial Intimin surface protein and six effectors: Translocated-Intimin receptor (Tir), Mitochondrial-associated protein (Map), EspF, EspG, EspH and EspZ (with EspB also exhibiting effector functions) [1], [3]. Prototypic EPEC (E2348/69) has at least fourteen additional non-LEE-encoded (Nle or Esp nomenclature) effector genes distributed on six horizontally-acquired mobile genetic elements [3], [4].
The EPEC disease process is characterised by a number of histo-pathological events including (i) initial non-intimate attachment to epithelial cells, (ii) bacteria sinking into the microvillus surface, (iii) intimate interaction with the host plasma membrane, (iv) nucleation of actin beneath intimately-adherent bacteria and (v) extensive loss/effacement of microvilli [1]–[3]. Studies with the Caco2 small intestinal model have provided insights on these events and revealed a plausible mechanism to explain the rapid onset of EPEC-induced watery diarrhoea [5]. Moreover, the use of such models have uncovered EPEC's ability to disrupt cell-cell interactions [5]–[7] subsequently verified by in vivo studies [8], [9]. While disruption of cell-cell interactions is an inflammatory event, human EPEC infections are normally associated with unexpectedly weak inflammation [2] thereby suggesting that the pathogen employs inhibitory mechanisms. Indeed, studies have revealed that EPEC inhibits Nuclear Factor κB (NF-κB) function - responsible for inducing the expression of anti-microbial and inflammation-related molecules - before barrier function is disrupted [10].
Host cell detection of foreign antigens (by Toll-like receptors; TLR) and cytokines (such as TNFα and IL1β by TNFR and IL1R, respectively) triggers a cascade of phosphorylation and ubiquitination events leading to IKK (Inhibitor of KappaB kinase) complex activation [11]–[13]. The activated IKK complex, composed of two kinases (IKKα, IKKβ and a regulatory subunit (NEMO/IKKγ), phosphorylates IκB (Inhibitor of κB) to induce its proteasomal-dependent degradation thereby releasing NF-κB for import into the nucleus to transcribe genes [14]. NF-κB function is regulated through many mechanisms including IκB re-synthesis, modification of NF-κB (or accessory factors) and altering NF-κB access to promoters [13]–[15]. In addition, NF-κB activity can be regulated at the level and/or function of signalling pathway components that includes kinases, phosphatases, ubiquitin ligases, de-ubiquitinases and adaptor proteins [13], [16]–[18]. TLR, IL1R and TNFR signalling to the IKK complex depends on TNFα Receptor-Associated Factor (TRAF) adaptor proteins and TGFβ-Activating Kinase 1 (TAK1) with pathway-specific components including kinases such as Receptor-Interacting Protein1 (RIP1) and adaptors such as Myeloid Differentiation primary response gene (88) (MyD88) [13], [18]–[20](see Fig. 1).
EPEC inhibition of NF-κB activity triggered by flagellin (recognised by TLR5) or cytokines (including TNFα and IL1β) in the Caco-2 small intestinal model depends on the pathogen possessing a functional effector-delivery system [10], [21]. Previous work has argued against a need for all LEE effectors, Intimin and a subset of Nle effectors (Orf3/EspG2, NleA, NleF and NleH) in this EPEC NF-κB inhibitory process thereby implicating other Nle effectors [10]. Indeed, screening programs by us and others have revealed novel NF-κB inhibitory activities for NleC and NleH with critical co-operative roles for NleE1 and NleB1 [22]–[29]. Here, we show that the LEE-encoded Translocated-Intimin receptor has a potent and specific Intimin-independent ability to inhibit NF-κB activation leading to the discovery of a novel inhibitory mechanism. Moreover, our work provides insights on the NleE1/NleB1 subversive process and on the possible evolution of EPEC's capacity to inhibit NF-κB activity.
EPEC requires a functional T3SS to inhibit antigens (such as flagellin) and cytokines (such as TNFα and IL1β) from activating NF-κB [10], [21]. To gain insight on the inhibitory mechanism, we examined EPEC's ability to interfere with NF-κB activity driven by plasmid-expression of TLR/IL1R and TNFR signalling pathway components (see Fig. 1). Thus, HeLa cells were co-transfected with plasmids expressing luciferase under the transcriptional control of NF-κB - via five repeats of the κB consensus promoter [30] - and specific signalling pathway components [30]–[32] prior to infecting with EPEC strains and quantifying luciferase cellular activity (see Materials and Methods). As EPEC can induce HeLa cells detachment [33], we employed the eae mutant (which lacks the Intimin surface protein) that, like wild type EPEC, inhibits NF-κB activity [10] but has little capacity to detach HeLa cells (Quitard et al., unpublished). The luciferase assay revealed basal NF-κB activity within uninfected HeLa cells as reported [23], with no significant change following infection with the eae or T3SS (effector-delivery defective) strains (Fig. 1A). Plasmid expression of TNFR pathway-related kinase (IKKα, IKKβ, TAK1 and RIP1) or adaptor (TRAF2) proteins increased cellular luciferase levels by 8 to 24 fold. Interestingly, the T3SS mutant inhibited increases driven by plasmid-expressed RIP-1 (Fig. 1A) supporting the presence of T3SS-independent inhibitory mechanisms in the HeLa model [28], [29]. By contrast, the eae mutant infection inhibited luciferase activity driven by plasmid-expression of all components except IKKα and IKKβ (Fig. 1A). Parallel studies on TLR/IL1R pathway components revealed another T3SS-independent inhibitory mechanism relating to MyD88 (Fig. 1B), with the eae mutant inhibiting signalling driven by plasmid-expressed TRAF6 and MyD88 (Fig. 1B). Western blot analyses verified the T3SS-dependent delivery of effectors (EspF and Tir; not shown) and plasmid-expression of examined host proteins, with the latter revealing unexpected infection-related increases in the cellular level of Flag-tagged TRAF proteins (Fig. 1C). Thus, consistent with previous studies [28], [29], EPEC infection of HeLa cells inhibits signalling to NF-κB by T3SS-independent and -dependent mechanisms. Moreover, the work implies that T3SS-dependent inhibitory mechanism(s) relates to effector(s) acting at or upstream of the IKKα/β complex.
To identify Nle effectors postulated to inhibit NF-κB activity [10], putative nle effector genes [3], [4] were cloned into mammalian expression vectors and co-transfected into HeLa cells with the NF-κB luciferase reporter vector for screening. Indeed, this approach identified a NF-κB inhibitory activity for NleC leading to its definition as a zinc metalloprotease that degrades NF-κB complexes [23] as supported by independent studies [22], [24], [25]. Interestingly, inclusion of an available Tir-expressing construct [34] in the screening program indicated that this LEE effector could prevent TNFα from activating NF-κB. To investigate this putative NF-κB inhibitory activity in more detail, the tir gene was sub-cloned into pEGFP-N1 to generate a Tir-eGFP fusion protein with pEGFP-N1 serving as a negative control. Fig. 2A reveals similar basal luciferase activity in cells transfected with pEGFP or ptir-EGFP plasmids, with TNFα leading to a significant increase in NF-κB reporter activity for pEGFP, but not ptir-EGFP transfected cells. The relevance of this finding to NF-κB function was illustrated by Western blot analysis where expression of Tir-eGFP, unlike eGFP, inhibited TNFα from inducing the phosphorylation-associated activation of IKKα/β kinases and the NF-κB component, p65 (Fig. 2B). Inhibition specificity was illustrated by unaltered total cellular levels of IKKα/β and p65 proteins (Fig. 2B). Furthermore, fluorescent microscopy examinations revealed p65 within the nucleus of ∼18% of eGFP or Tir-eGFP expressing cells, with TNFα treatment increasing this to ∼70% for eGFP expressing cells but only ∼35% for Tir-eGFP expressing cells (Fig. 2C). As IL8 secretion requires NF-κB activity [35], we examined the extra-cellular levels of this chemokine (see Materials and Methods). Consistent with the luciferase NF-κB reporter data (Fig. 2A), pEGFP and ptir-EGFP transfected cells released similar basal levels of IL8, with TNFα treatment increasing IL8 secretion levels from pEGFP but not ptir-EGFP transfected cells (Fig. 2D). Thus, expressing Tir-eGFP within HeLa cells specifically prevents TNFα from transducing signals that activate NF-κB in a manner linked to a blockage in the phosphorylation-associated activation of IKK components needed to release NF-κB for nuclear import. The absence of other EPEC factors in these experiments illustrate that this novel property of the Translocated-Intimin receptor (Tir) effectors occurs independently of Intimin.
To support the specific and Intimin-independent nature of the Tir inhibitory activity, studies evaluated Tir's ability to block TNFα-induced IL8 secretion following its delivery into HeLa cells by Yersinia pseudotuberculosis as previously described [34]. Importantly, the control Tir-negative Yersinia strain (which lacks most of its own T3SS-delivered effectors) failed to inhibit TNFα-induced IL8 secretion, whereas the Tir-expressing variant inhibited this process to a similar degree as EPEC-delivered effectors (Fig. 3). Western blot analyses verified Yersinia-delivery of Tir where it underwent partial host kinase-mediated modification, compared to EPEC-delivered Tir (not shown), as previously described [34]. Given that TNFα augmentation of IL8 secretion requires NF-κB activity [35], this work supports the premise that Tir (in the absence of other EPEC factors, including Intimin) possesses a potent and specific ability to prevent TNFR-induced signalling from activating NF-κB.
To gain insight on how Tir inhibits TNFα-induced NF-κB activation, HeLa cells were co-transfected with plasmids encoding (i) the NF-κB luciferase reporter protein, (ii) TNFR signalling pathway components and (iii) eGFP or Tir-eGFP proteins prior to assaying cellular luciferase levels. This work revealed that Tir-eGFP, but not eGFP, inhibited luciferase activity driven by plasmid-expression of TRAF2 and RIP1, but not TAK1, IKKβ (Fig. 4A) or IKKα (not shown). Fluorescence microscopy studies were undertaken to determine the cellular location of over-expressed signalling components and to assess if Tir-eGFP expression induced detectable changes. Staining for plasmid-expressed Tir and IKK kinase proteins revealed diffuse cytoplasmic signals in contrast to cytoplasmic aggregates/clusters for the TRAF2 (Fig. 4B and Fig. S1) and MyD88 (not shown) adaptor proteins. Cytoplasmic clustering of plasmid-expressed TRAF2 has been reported [36]. Imaging of co-transfected cells revealed similar, distinct and partially-overlapping signals for Tir/IKK, Tir/MyD88 and Tir/TRAF signals, respectively (Fig. 4B and Fig. S1). Intriguingly, Tir-eGFP expression was associated with a loss of TRAF clusters (Fig. S1B) as supported by quantification studies (Fig. 4C). Thus, Tir may inhibit plasmid-expressed TRAF2 from transmitting signalling to NF-κB by inducing the disaggregation and/or degradation of activation-associated clusters.
To examine predicted Tir-TRAF interactions, GFP-Trap beads were used to isolate eGFP and Tir-eGFP proteins from cells co-transfected with the Flag-tagged TRAF2 expressing plasmid. Fig. 5A reveals eGFP and Tir-eGFP within input cellular extracts and their isolation by the GFP-Trap beads. While the IKKα/β proteins were present in the input pool, they did not co-isolate with eGFP or Tir-eGFP (Fig. 5A). Probing for Flag-tagged TRAF2 revealed a prominent monomer-sized band with smaller amounts of a trimer-sized TRAF species in the input pool (Fig. 5A). TRAF2 function is linked to the formation of homo- or hetero-trimers [13], [37]. Intriguingly, the minor trimer-sized TRAF2 species preferentially isolated with Tir-eGFP, though some monomer was co-isolated (Fig. 5A). This work suggests that Tir interacts (either directly or indirectly) with the activation-associated multimeric form of TRAF adaptor proteins.
Examination of the input samples (Fig. 5A) suggested that Tir expression may decrease the cellular levels of multimeric (and perhaps monomeric) Flag-tagged TRAF2 protein (Fig. 5A). This premise was supported by demonstrating that co-expression of Tir-eGFP with Flag-tagged TRAF2 could lead to the complete loss of TRAF2 from cell extracts, while the actin loading control protein remained unchanged (Fig. 5B). As inflammatory signalling is commonly regulated by targeting components for proteasomal-dependent degradation [38], we used the proteasomal inhibitor MG132. Whilst MG132 inhibitory activity was confirmed, as per a parallel study [23], it failed to prevent the Tir-mediated loss of Flag-tagged TRAF2 proteins (Fig. 5B). Interestingly, probing the fate of endogenous TRAF2 suggested that it is not a substrate for Tir degradation, at least in cells expressing the Flag-tagged TRAF2 variant. Thus, similar levels of monomer and trimer-sized TRAF2 bands were evident in non-transfected and ptir-EGFP transfected cells (Fig. 5C) while the more prominent bands in pTRAF2-transfected cells correspond to the Flag-tagged variant (Fig. 5C versus 5B).
To investigate whether endogenous TRAF2 is a substrate for Tir-induced degradation, its fate was examined in cells that express Tir-eGFP, but not Flag-tagged TRAF2 proteins. Fig. 5D reveals that Tir expression can, in fact, induce the cellular loss of endogenous TRAF2. Interestingly, TNFα treatment of control cells reduced the level of multimeric TRAF2 with an increase in the monomer species (Fig. 5D) that, presumably, reflects intrinsic host mechanism(s) for down-regulating cytokine-induced signalling. By contrast, TNFα treatment of ptir-EGFP transfected cells produced a small pool of TRAF2 protein (monomer and trimer-sized forms; Fig. 5D) that may explain why TNFα triggered some p65 relocation to the nucleus of Tir-eGFP expressing cells (Fig. 2C). Ubiquitin-modified TRAF2 plays a key role in activating RIP1 which activates TAK1 [39] to, perhaps, explain why Tir inhibits NF-κB luciferase activity driven by plasmid-expressed RIP1 and, to a lesser extent, TAK1 (Fig. 4A). Collectively, the work implies that Tir inhibits TNFα-induced NF-κB activation by interacting (directly or indirectly) with TRAF2 - a key component of the TNFR signalling pathway - to induce its proteasomal-independent degradation.
Previous work [10] suggested that the Tir and NleH effectors are not required for EPEC to inhibit NF-κB activity, with recent studies reporting a non-essential role not only for NleH but also NleC [22κ27]. Indeed, HeLa cells infections with a tir mutant confirmed Tir's non-essential role in inhibiting TNFα-induced IL8 secretion [10] but also revealed a small, but statistically significant defect (Fig. 6A). To examine the relationship of this defect to the absence of Tir/Intimin-mediated intimate EPEC-host cell interaction, assays were carried out with an eae (Intimin-deficient) mutant. Unexpectedly, these studies indicated that Intimin (indirectly or directly) induces NF-κB activity or suppresses effector-mediated inhibitory mechanism(s), as the eae mutant inhibited TNFα-induced IL8 secretion to a greater extent than wild type EPEC (Fig. 6A). Intimin alters host cellular processes by Tir-dependent and -independent mechanisms [40]. Infection studies with an eaetir double mutant suggest that this Intimin function relates to Tir-dependent and -independent mechanisms (Fig. 6A).
Interestingly, time course infection studies support a Tir-TRAF2 relationship as EPEC induced a dramatic loss in the levels of activation-associated multimeric TRAF2 proteins by a process dependent on Tir and Intimin (Fig. 6B). The Intimin-dependent nature of this event, in contrast to that mediated by ectopically-expressed Tir (Fig. 2), implies that EPEC has evolved Intimin-dependent mechanisms for regulating this Tir activity. Interestingly, other effectors appear to contribute at early time points, with the eae mutant appearing to display an augmented ability to reduce TRAF2 levels (Fig. 6B). Intriguingly, confocal microscopy studies of disease-relevant polarised cells infected with EPEC only detect Tir at the apical (surface) membrane whereas a transient pool is evident within the cytoplasm of eae mutant-infected cells (Fig. S2). This suggests that Intimin promotes Tir's rapid association with the plasma membrane, with the transient cytoplasmic pool perhaps promoting Tir-TRAF2 interactions to explain the eae mutant's Tir-dependent augmented ability to reduce the level of TRAF2 multimers and inhibit TNFα-induced IL8 secretion.
Recent studies have described a prominent role for NleE1, promoted by NleB1 or NleC, in the EPEC NF-κB inhibitory process, as nleCnleE1 and nleB1nleE1 double mutants behaved like a T3SS-defective strain, compared with partial defects for single mutants [22], [25], [28], [29]. However, we and others have described T3SS-independent inhibitory mechanisms in the employed HeLa cell models (Fig. 1) [28], [29] that may obscure the contribution of effectors (Fig. 1) [28], [29]. Thus, an nleB1nleE1 double mutant was generated and evaluated in a small intestinal model where EPEC was confirmed to inhibit NF-κB function solely in a T3SS-dependent manner [10] (Fig. 6C). Indeed, the nleB1nleE1 double mutant behaved akin to the effector-delivery defective (T3SS) strain (Fig. 6C) despite displaying no obvious defect in delivering EspB or Tir effectors (not shown). By contrast, a nleB1nleE1tir triple mutant displayed a small (significant) capacity to inhibit NF-κB function - presumably due to remaining effectors. Interestingly, while EPEC and T3SS-mutant infected cells released ∼200 and ∼800 ng/ml of IL8, respectively, in response to TNFα treatment only ∼400ng/ml was secreted from tir mutant infected cells (Fig. 6C). Whilst these IL8 values support a non-essential contributory role for Tir in the inhibitory process, the difference between EPEC- and tir-infected cell was below the significance threshold (p = 0.075). Nevertheless, this work supports the idea that the NleB1/NleE1 effectors play a central role in enabling EPEC to inhibit NF-κB activity in intestinal cells, with the non-essential novel NF-κB inhibitory activities of NleC, NleH and Tir, presumably, playing evolutionary-advantageous roles in EPEC's lifecycle.
Given NleE1 and NleB1's key roles in the EPEC NF-κB inhibitory process, with only speculations on their targets [28], [29], we investigated where the blockage occurred by co-expressing them with signalling pathway components for NF-κB luciferase reporter assays. Expression of NleE1 inhibited NF-κB reporter activity driven by plasmid-expression of components from the TNFR (TRAF2, RIP1) and IL1R/TLR (TRAF6, MyD88) pathways (Fig. 7A) consistent with reports of it inhibiting NF-κB activation by multiple pathways [28], [29]. However, NleE1 failed to inhibit luciferase activity driven by plasmid-expression of TAK1 or IKK kinases (Fig. 7A) suggesting it inhibits TAK1 function to block signalling by TNFR, IL1R and TLR pathways. NleE1 may target TAK1 or factors needed for its activation, such as the TAB2/3 proteins which recruit TAK1 to ubiquitin-modified RIP1 and ubiquitin-modified TRAF6 proteins for activation in the TNFR and TLR/IL1R pathways, respectively [39]. By contrast, NleB1 inhibited luciferase activity driven by plasmid-expression of TRAF2 but not TRAF6 or IKKβ (Fig. 7B) supporting reports of it inhibiting signalling in TNFR but not TLR/IL1R pathways [28], [29]. Interestingly, NleB increased luciferase activity driven by TRAF6 suggesting that it has functions that (directly or indirectly) activate TRAF6-mediated signalling. Increases in TRAF6 signalling, despite NleB inhibition of TAK1 function (Fig. 7B), suggest that the effector may block RIP1-mediated activation of TAK1 to inhibit signalling in TNFR, but not TLR/IL1R pathways [39].
Interestingly, our screening program revealed that the NleE1 homologue, NleE2, induced NF-κB luciferase activity as effectively as TNFα(Fig. 7C) leading to a similar high level of p65 relocation into the nucleus (Fig. 7D). Indeed, both these NleE2-dependent alterations were inhibited by co-expressing Tir (Fig. 7C and 7D) suggesting that NleE2 activates NF-κB (either directly or indirectly) through a component at or upstream of Tir's target i.e. TRAF. While NleE2 is apparently not transferred into host cells [28], our finding supports the idea that EPEC effectors have features or functions that can activate NF-κB signalling and can be blocked by Tir.
In this study we describe a new property for the most-extensively studied EPEC effector by demonstrating that the Translocated Intimin receptor (Tir) protein has a potent and specific ability to prevent HeLa cells from activating NF-κB in response to the cytokine TNFα. Whilst this discovery involved ectopic expression of Tir and an indirect NF-κB reporter assay, its relationship to transcription factor function was demonstrated by several lines of evidence. Firstly, expression of Tir-eGFP, unlike eGFP, blocked the phosphorylation-associated activation of IKKα/β and the NF-κB component, p65 - events required for the nuclear import and transcriptional activity of NF-κB, respectively [13]. Importantly, absence of these modifications was not due to cell loss as Tir-eGFP expression had no observable impact on the total cellular level of IKK kinases or p65. Secondly, epifluorescent microscopy studies revealed that TNFα treatment induced some relocation of p65 into the nucleus of Tir-eGFP expressing cells but to a dramatically less degree than eGFP expressing cells. Indeed, as expected, these Tir-mediated inhibitory events translated into a dramatic deficiency in the NF-κB dependent event [35] of TNFα-augmented increases in IL8 secretion levels. Thirdly, use of a more physiologically relevant mechanism of introducing Tir into host cells (via the T3SS of another pathogen, Yersinia) inhibited TNFα-induced IL8 secretion to a similar level as control EPEC-infected cells, whereas the Tir-negative Yersinia strain had no inhibitory capacity. Finally, epifluorescent, biochemical and co-precipitation studies unearthed an inhibitory mechanism relating to Tir interaction with and subsequent cellular loss of a key component from the TNFR signalling pathway. These data illustrate that the EPEC Tir effector has a specific and potent ability to inhibit TNFα-induced NF-κB activation.
The absence of additional EPEC factors in these ectopic and Yersinia-delivery Tir experiments illustrate the Intimin-independent nature of the NF-κB inhibitory process. Whilst over a decade of studies has re-enforced the idea that Tir's subversive activities require it to interact with the EPEC surface protein, Intimin (and, thus, Tir's need to insert into the plasma membrane to act as a receptor for Intimin) a recent study described an Intimin-independent function [41]. Our discovery of a second such activity raises the possibility that Tir possesses additional Intimin-independent functions and the need to consider their contribution to Tir's critical role in the virulence of attaching and effacing pathogens that include strains targeting humans (EPEC and enterohaemorrhagic E. coli; EHEC), ruminants (EHEC) and various small mammals.
Our work also revealed a novel mechanism for a pathogen effector to inhibit NF-κB activity as it demonstrated that Tir interacts, directly or indirectly, with TRAF2 proteins (with a preference for activation-associated multimers) inducing the proteasomal-independent loss of this adaptor protein from host cells. TRAF adaptor proteins play critical roles in signalling to NF-κB by multiple pathways including the TLR, TNFR and IL1R pathways inhibited by EPEC [10,13,19κ21]. Indeed, Tir inhibited NF-κB reporter activity driven by plasmid-expression of the TRAF2 (Fig. 4) and TRAF6 (not shown) proteins of the TNFR and TLR/IL1R pathways, respectively. Moreover, TRAF2 and TRAF5 possess functionally redundant roles in TNFR-mediated signalling to NF-κB [37] suggesting that Tir also inhibits TRAF5 activity. It is speculated that Tir can inhibit signalling through other TRAF-dependent pathways. Six of the seven TRAF members carry amino-terminal zinc-binding motifs involved in their function as E3 ubiquitin ligases for activating downstream kinases, while TRAF1 (lacks the ‘Really Interesting New Gene’ RING domain) has regulatory functions [37], [42]. Studies with a dominant-negative variant of TRAF2 [43] suggests that Tir is unable to induce its degradation (Fig. S3) implicating a need for the absent RING domain in the degradation process. There are several examples of pathogens targeting TRAF proteins, including the poxvirus MC159 protein preventing TRAF2 sequestration into a signalosome [44] and the Yersinia YopJ effector deubiquitinating TRAF2 to inhibit signalling to NF-κB [45]. The Yersinia strain used in our studies has no detectable YopJ activity (linked to a polar insertion mutation of the ypkA gene immediately upstream of yopJ; Prof Hans Wolf-Watz personal communication). As far as we are aware, the proteasomal-independent degradation of TRAF2 by Tir represents a novel pathogen-mediated mechanism for inhibiting NF-κB activity. Determining if Tir induces the cellular loss of all or a subset of TRAF members may provide insights on the breadth of cytokine- and antigen-signalling pathways it can inhibit and/or highlight conserved features involved in the TRAF interaction and/or degradation processes. Studies are underway to define the features and mechanism by which Tir induces the proteasomal-independent degradation of TRAF adaptors.
Consistent with previous findings [10], Tir was not required for EPEC to inhibit TNFα-induced NF-κB activation though infection studies revealed a small, but statistically significant inhibitory defect for a Tir-deficient strain. This defect was unlinked to Tir's role with Intimin in mediating intimate EPEC-host cell interactions, as an Intimin-deficient (eae) mutant inhibited NF-κB activity to a greater extent than EPEC. This Intimin-related activity was associated with Tir-dependent and -independent mechanisms thereby revealing a new property for this EPEC surface protein. Intriguingly, strains lacking Intimin or Tir displayed a dramatic deficiency in EPEC's ability to decrease cellular levels of activation-associated TRAF2 multimers suggesting that, in the context of an EPEC infection, Tir requires Intimin to reduce TRAF2 cellular levels. Interestingly, microscopy studies identified a transient pool of Tir within the cytoplasm of epithelia infected with the Intimin-deficient, but not wildtype EPEC strain. While Tir has been proposed to insert into the plasma membrane during the translocation process [46], it is clear that it can insert from the host cytoplasm [47], though an infection-associated cytoplasmic pool has, until now, only been supported by Western blot analyses [34], [48]–[50]. It appears that Tir delivery into the host cytoplasm is normally followed by its rapid (Intimin signalling-promoted) association with the plasma membrane. The extended presence of Tir within the cytoplasm of eae-mutant infected cells may promote Tir-mediated loss of activation-associated TRAF2 multimers, as supported by the time course studies (Fig. 6B), to perhaps explain the Tir-dependent increased capacity of the eae mutant to inhibit TNFα-induced IL8 secretion.
While Tir's non-essential role in the EPEC NF-κB inhibitory process, like that of NleC and NleH [22]–[27], could be due to functional redundancy with other effectors, this is not the case as illustrated by studies with an nleB1nleE1 double mutant. Thus, despite displaying no defect in delivering Tir into host cells, the double mutant had no significant ability to inhibit TNFα-induced IL8 secretion in HeLa or small intestinal models. This finding supports the reported key role for the NleE1/NleB1 effectors in blocking NF-κB function [25], [28], [29] and implies that the described novel NF-κB inhibitory activities of Tir, NleH and NleC effectors [22]–[27] (this study) are minor or transient during EPEC infections. As Yersinia-delivered and ectopically-expressed Tir proteins are potent inhibitors of TNFα-induced NF-κB activity, unlike EPEC-delivered Tir, this implies that EPEC possesses factors that suppress Tir's inhibitory function. Interestingly, ectopically-expressed Tir decreases TRAF2 cellular levels in an Intimin-independent manner (Fig. 5) while decreases mediated by EPEC-delivered Tir depend on Intimin (Fig. 6B). Ectopically-expressed (and Yersinia-delivered) Tir differs from the EPEC-delivered Tir by i) being mainly cytoplasmic ii) only undergoing partial host kinase-mediated modification and iii) failing to interact with Intimin [34] (Fig. S3). Thus, EPEC suppression of Tir's ability to decrease TRAF2 cellular levels (and presumably its ability to inhibit NF-κB activity) is linked to undefined EPEC factors enabling Tir to insert into the plasma membrane to interact with Intimin. Further studies are required to define the putative EPEC factors and mechanisms involved in this regulatory process.
Our screening program revealed NF-κB inhibitory activities for NleC [23] and Tir (this study), a NF-κB activatory function of NleE2 (this study) and confirmed inhibitory activities [22]–[29] for NleH (not shown), NleE1 (this study) and NleB1 (this study) effectors. By contrast, no significant NF-κB modulatory activity was evident from screening other LEE or non-LEE effectors, though it is possible that the findings included false negatives due to effector expression problems or expressing effector-fusion proteins. Indeed, our work supports the premise [22]–[27] that NleC and NleH inhibit NF-κB function by targeting components downstream of the IKK complex [23] (not shown), with the inhibitory activities of NleE1, NleB1 and Tir linked, respectively, to blocking the function of the TAK1, RIP1 and TRAF components upstream of IKK. Interestingly, while NleB1 inhibited signalling by TNFR pathway components, it promoted that mediated by the TLR/ILIR pathway protein, TRAF6 (but not downstream RIP1) suggesting that it has properties that induce TRAF6-mediated NF-κB activation. Indeed, the idea that EPEC effector features or properties can activate NF-κB is supported by the finding that ectopically-expressed NleE2 was as effective as TNFα at inducing p65 nuclear relocation. Interestingly, this NleE2-mediated event was blocked by co-expression of Tir suggesting that its activatory property, as per NleB1, is transmitted through TRAF proteins - a defined target of Tir to inhibiting signalling to NF-κB.
Our findings on LEE and Nle effectors, in light of published work, lend themselves to speculations on the evolution of EPEC's capacity to inhibit NF-κB. Genome sequencing projects suggest that pathogenic E. coli evolved from commensal E. coli through the horizontal-acquisition of new functions encoded on mobile genetic elements. Thus, enterotoxigenic E. coli (ETEC) virulence is linked to strains acquiring functional enterotoxins and an enterocyte-binding pilus [51], whilst that of EPEC is linked to the acquisition of the effector-encoding mobile genetic elements [4], [52]. As Nle effector genes are generally missing from non-pathogenic E. coli strains and require the LEE T3SS for delivery into host cells, it is reasonable to assume that the progenitor EPEC strain possessed the LEE, but not Nle-encoding genetic elements. EPEC factors (including flagella) and LEE subversive functions (eg disrupting cell-cell interactions) can activate NF-κB to induce the expression of anti-microbial and inflammatory molecules that inhibit EPEC's virulence-critical ability to colonise epithelia [10], [21]. Thus, the LEE region presumably encoded factor(s) to inhibit this event, with our screening program defining Tir as the only LEE effector with significant NF-κB inhibitory activity. Indeed, our definition of Tir-TRAF interactions within the cytoplasm to inhibit NF-κB activity may explain why Tir transits through this host compartment prior to inserting into the host membrane. It is possible that Tir's inability to completely block signalling-induced relocation of NF-κB into the nucleus (Fig. 2C) provided a selective advantage to strains acquiring mobile genetic elements expressing effectors that promote the inhibitory process (eg NleC degradation of nuclear NF-κB complexes). Undoubtedly, the key point in the evolutionary process relates to the acquisition of the NleB1/NleE1-encoding mobile genetic element (Integrative element 6; IE6) given their critical roles in blocking NF-κB activation in HeLa cells [25], [28], [29] and disease-relevant small intestinal models. This premise is supported by the nleB1nleE1 genes being among the subset of nle genes found in all sequenced LEE-encoding pathogens [52] and their presence at the 3′ end of the LEE region in some enterohemorrhagic E. coli strains [53]. It is possible that this LEE/IE6 hybrid represents a minimal genetic unit required to provide strains with EPEC-like enteric pathogenic properties.
These studies used nalidixic acid resistant EPEC strains, specifically wild-type EPEC (E2348/69), eae (Intimin-deficient) and espA (T3SS-deficient) isogenic strains [54], [55]. The nleB1nleE1 double mutant was generated using described standard allelic exchange procedures [56], [57] to remove (confirmed by PCR analyses) the entire gene sequence (and inter-gene region) of the adjacent nleB1 and nleE1 genes. The nleB1nleE1tir triple mutant was generated using an available tir-deletion suicide vector as described [57]. Strains were grown in Luria-Bertani (LB) broth containing nalidixic acid (25 ug/ml final conc.) from single colonies, without shaking, at 37°C in a 5% CO2 incubator overnight. The Yersinia strains and their usage was as previously described [34]. Hela cells (ATCC CCL2) were grown at 37°C with 5% CO2 in Dulbecco's Minimal Eagles Medium (DMEM) supplemented with 10% heat-inactivated foetal calf serum and 2 mM L-glutamine. Caco2 parental or TC7 subclone cells were seeded at confluence onto Transwells (Corning) and polarised over 12–15 days as previously described [5], [6].
Prof Luke O'Neill (Trinity College, Dublin) kindly provided plasmids relating to the NF-κB luciferase-reporter construct and expression of IKKα, IKKβ, TRAF2, TRAF6, and MyD88 [30] with those for RIP1 [31] and TAK1 [32] kindly provided by Prof's Jürg Tschopp (University of Lausanne, Switzerland) and Martin Dorf (Harvard, USA), respectively. Tir-eGFP, eGFP-NleB1, NleE1 and NleE2 proteins were expressed from pEGFP-N1 (Clontech), pEGFP-C1 (Clontech), pIRES (Clontech) and pcDNA3 (Invitrogen) vectors, respectively.
Hela cells (∼2×105) seeded in 24-well plates were transfected the following day using JetPrime reagent (PEQLAB Ltd, UK) with a total amount of 250 ng DNA, comprising 100 ng of the NF-κB firefly luciferase reporter plasmid [30], 40 ng of the Renilla reniformis luciferase plasmid plus 110 ng of empty, or effector gene-containing plasmid. Levels of firefly luciferase expression were normalised against Renilla luciferase activity as a control for transfection efficiency (expressed as fold increase in luciferase activity over unstimulated control cells). When transfection efficiency was routinely found to be ∼65–80%, the Renilla luciferase plasmid was replaced with empty plasmid. High transfection efficiencies for pEGFP and/or pEGFP-tir experiments were routinely verified by visualising the eGFP signal. Twenty four hours post transfection, cells were incubated with or without TNFα (10 ng/ml) for 30 minutes, lysed in 100 µl of passive lysis buffer (Promega Ltd, Southampton, UK) for 15 minutes at room temperature with cell extracts taken for assessment of firefly luciferase activity following standard protocols and a FLUOstar Optima 413-3266 plate reader (BMG Labtech, Germany).
LB grown EPEC cultures were first diluted (1∶10) in DMEM and incubated for 3 hours at 37°C in a 5% CO2 incubator. The typical optical density (OD600) was between 0.2–0.3 with infections carried out at a multiplicity of infection, MOI, of ∼100∶1. The HeLa cell medium was replaced with DMEM at least 2 hours prior to infection (routinely 3 hours unless stated otherwise), with studies on transfected cells normally 24 hours post-transfection. Yersinia YIII MEKA strains were grown in modified brain-heart medium supplemented with 20 mM MgCl2 and 5 mM EGTA at 26°C without shaking, and used for infections as previously described [34]. When appropriate, cells were incubated with bactericidal levels of gentamycin (100 µg/ml final conc.) for 1 hour prior to adding TNFα (10 ng/ml) for between 30 minutes (the routine) and up to 2 hours. EPEC infections did not induce significant cell detachment under the employed experimental conditions.
HeLa cells were washed with cold Phosphate Buffered Saline pH 7.4 (PBS) and lysed with 1% Triton X-100 in the presence of protease inhibitors (1/1000 dilution, Sigma cocktail), sodium fluoride, sodium orthovanadate and PMSF (1.2, 1.2 and 1 mM final concentration, respectively). When appropriate centrifugation (13000 x g 5 minutes) was used to separate insoluble (contains host nuclei and cytoskeleton as well as adherent bacteria) and soluble (contains host cytoplasmic and membrane proteins as well as T3SS-delivered proteins). Samples were resolved on 10% SDS PAGE, transferred onto nitrocellulose, blocked in 5% Blotto milk powder/PBS/0.02% Tween and probed with antibodies against IKKα/β (Santa Cruz), phospho IKKα/β (Cell Signaling), NF-κB p65 (Santa Cruz), phospho p65 (Ser536), TRAF2 (Cell Signaling), actin (Sigma), FLAG tag (Sigma), Myc tag (generous gift; Prof D. Mann, Newcastle University) or GFP (Zymed). Absence of reducing agents allowed the detection of TRAF2 multimeric bands. Primary antibodies were incubated overnight in a 5% bovine serum albumin (BSA)/PBS solution, washed extensively. Bound antibodies were detected using horseradish peroxidase-conjugated secondary antibodies and Super Signal West Pico chemiluminescent substrate (Pierce) with Hyperfilm ECL (Amersham Biosciences) following the manufacturer's recommendations.
Supernatants (0.5 ml) were taken from above the HeLa cells and assayed for the level of IL8 using an ELISA kit (DB Biosciences) following the manufacturer's recommendations.
Immunoprecipitation of Tir-eGFP was performed using GFP-Trap A beads (Chromotek) according to the manufacturer's instructions. Briefly, 24 hours post-transfection, Hela cells were lysed in RIPA buffer, centrifuged and the supernatant incubated with GFP-Trap beads for 30 minutes at 4°C. Following centrifugation, the unbound material was harvested and the beads washed before being resuspended in sample SDS buffer for Western blot analyses.
Following experimentation, Hela cells seeded on glass coverslips were washed three times with PBS prior to fixing (2.5% Paraformaldhyde - Sigma - in PBS) for 30 minutes and permeabilisation of host membranes in Triton-X100/BSA (0.1% and 2.5% final conc, respectively)/ PBS solution for 30 minutes. Cells were then incubated overnight in the fridge with an appropriate primary antibody in 2.5% BSA/PBS solution, followed by multiple PBS washes and incubation with Alexa 488 or Alexa 555-conjugated secondary antibodies in a 2.5% BSA/PBS solution (1 hour; room temperature). Washed cells were mounted in DAPI-Vectashield (Vector Laboratories) and examined on a Zeiss Axioskop Epifluorescent or a Leica TCS SP2UV confocal microscopy. Nuclear p65 and TRAF clusters were counted in a semi-blind fashion i.e. slides were assessed without considering the slide order or orientation with obtained data mapped to labelling. For co-localisation studies, cells were visualised using the confocal microscope using an x63 objective lens with serial optical slices taken along the z-axis of cells within a field of view (∼20 cells), with signals analysed by Leica software and plotted to illustrate the degree of overlap.
|
10.1371/journal.pgen.1006601 | Excess of genomic defects in a woolly mammoth on Wrangel island | Woolly mammoths (Mammuthus primigenius) populated Siberia, Beringia, and North America during the Pleistocene and early Holocene. Recent breakthroughs in ancient DNA sequencing have allowed for complete genome sequencing for two specimens of woolly mammoths (Palkopoulou et al. 2015). One mammoth specimen is from a mainland population 45,000 years ago when mammoths were plentiful. The second, a 4300 yr old specimen, is derived from an isolated population on Wrangel island where mammoths subsisted with small effective population size more than 43-fold lower than previous populations. These extreme differences in effective population size offer a rare opportunity to test nearly neutral models of genome architecture evolution within a single species. Using these previously published mammoth sequences, we identify deletions, retrogenes, and non-functionalizing point mutations. In the Wrangel island mammoth, we identify a greater number of deletions, a larger proportion of deletions affecting gene sequences, a greater number of candidate retrogenes, and an increased number of premature stop codons. This accumulation of detrimental mutations is consistent with genomic meltdown in response to low effective population sizes in the dwindling mammoth population on Wrangel island. In addition, we observe high rates of loss of olfactory receptors and urinary proteins, either because these loci are non-essential or because they were favored by divergent selective pressures in island environments. Finally, at the locus of FOXQ1 we observe two independent loss-of-function mutations, which would confer a satin coat phenotype in this island woolly mammoth.
| We observe an excess of detrimental mutations, consistent with genomic meltdown in woolly mammoths on Wrangel Island just prior to extinction. We observe an excess of deletions, an increase in the proportion of deletions affecting gene sequences, and an excess of premature stop codons in response to evolution under low effective population sizes. Large numbers of olfactory receptors appear to have loss of function mutations in the island mammoth. These results offer genetic support within a single species for nearly-neutral theories of genome evolution. We also observe two independent loss of function mutations at the FOXQ1 locus, likely conferring a satin coat in this unusual woolly mammoth.
| Woolly mammoths (Mammuthus primigenius) were among the most populous large herbivores in North America, Siberia, and Beringia during the Pleistocene and early Holocene [1]. However warming climates and human predation led to extinction on the mainland roughly 10,000 years ago [2]. Lone isolated island populations persisted out of human reach until roughly 3,700 years ago when the species finally went extinct [3]. Recently, two complete high-quality high-coverage genomes were produced for two woolly mammoths [4]. One specimen is derived from the Siberian mainland at Oimyakon, dated to 45,000 years ago [4]. This sample comes from a time when mammoth populations were plentiful, with estimated effective population size of Ne = 13,000 individuals [4]. The second specimen is from Wrangel Island off the north Siberian coast [4]. This sample from 4,300 years ago represents one of the last known mammoth specimens. This individual comes from a small population estimated to contain roughly 300 individuals [4]. These two specimens offer the rare chance to explore the ways the genome responds to pre-extinction population dynamics.
Nearly neutral theories of genome evolution predict that small population sizes will lead to an accumulation of detrimental variation in the genome [5]. Such explanations have previously been invoked to explain genome content and genome size differences across multiple species [6]. Yet, within-species comparisons of how genomes are changed by small effective population sizes remain necessarily rare. These mammoth specimens offer the unique opportunity for within-species comparative genomics under a 43-fold reduction in population size. This comparison offers a major advantage as it will be free from confounding biological variables that are present in cross species comparisons. If nearly neutral dynamics lead to an excess of detrimental variation, we should observe an excess of harmful mutations in pre-extinction mammoths from Wrangel Island.
We use these two ancient DNA sequences to identify retrogenes, deletions, premature stop codons, and point mutations found in the Wrangel Island and Oimyakon mammoths. We identify an excess of putatively detrimental mutations, with an excess of stop codons, an excess of deletions, an increase in the proportion of deletions affecting gene sequences, an increase in non-synonymous substitutions relative to synonymous substitutions, and an excess of retrogenes, reflecting increased transposable element activity. These data bear the signature of genomic meltdown in small populations, consistent with nearly-neutral genome evolution. They furthermore suggest large numbers of detrimental variants collecting in pre-extinction genomes, a warning for continued efforts to protect current endangered species with small population sizes.
We identified all SNPs in each mammoth genome as well as one Indian elephant specimen, Maya, using GATK [7]. We identified all non-synonymous and synonymous changes relative to the L. africana reference genome (https://www.broadinstitute.org/scientific-community/science/projects/mammals-models/elephant/elephant-genome-project) using r3.7 annotations lifted over to L. africana 4.0 genome sequences. We observe a significant increase in the number of heterozygous non-synonymous changes relative to synonymous changes in the Wrangel island genome compared with Oimyakon (χ2 = 68.799, df = 1, P < 2.2 × 10−16; S1 Table). There is also a significant increase in the number of homozygous mutations at non-synonymous sites relative to synonymous sites (χ2 = 9.96, df = 1, P < 0.0016; S1 Table). We further observe an excess of premature stop codons in the genome of the Wrangel Island mammoth, with 1.8X as many genes affected. There are 503 premature stop codons in the Oimyakon genome (adjusting for a 30% false negative rate at heterozygous sites) compared with 819 in the Wrangel island genome (Fig 1, Table 1). There are 318 genes that have premature stop codons that are shared across the two mammoths, and 357 genes that are truncated in both mammoths, including mutations that form at independent sites. A total of 120 of these genes have stop codons in the two mammoths as well as in Maya the Indian elephant, suggesting read through in the L. africana reference. Among truncated genes, there is a significant excess of olfactory genes and oderant binding receptors that appear to be pseudogenized with an EASE enrichment score of 9.1 (S2 Table) [8, 9]. We observe 85 truncated olfactory receptors and 3 vomeronasal receptors as well as multiple signal transduction peptides compared with 44 olfactory receptors and 2 vomeronasal receptors pseudogenized in the mainland mammoth.
It is possible that DNA damage in the archaic specimens could contribute to a portion of the observed stop codons. When we exclude A/G and C/T mutations, there is still a gross excess of premature stop codons, with 645 genes truncated in the Wrangel Island mammoth compared with 377 in the Oimyakon mammoth. Hence, the patterns are not explained solely by differential DNA damage in the two mammoths. Maya, the Indian Elephant specimen shows 450 premature stop codons, but 401 when A/G and T/C mutations are excluded. When putative damage to ancient DNA is excluded, Maya appears to house an intermediate number of premature stop codons, with a 6% increase compared to the Oimyakon mammoth.
We identify 27228 deletions over 1 kb long in the Wrangel island genome, and 21346 (correcting for a 0.5% false negative rate at heterozygous sites) in the Oimyakon genome (Table 1). There are 6147 deletions (23%) identified in the Wrangel Island mammoth that are homozygous (≤ 10% coverage) compared with 5035 (24%) in the Oimyakon mammoth. (S3 Table). A total of 13,459 deletions are identified in both mammoth genomes (S4 Table). Some 4813 deletions in the Wrangel Island mammoth and 4598 in the Oimyakon mammoth appear hemizygous but have stretches of zero coverage for at least 50% of their length. These sites may represent multiple independent homozygous deletions that cannot be differentiated via change point statistics. Alternatively, they might indicate smaller secondary deletions that appear on hemizygous haplotypes. Such secondary deletions are common when large loop mismatch repair attacks unpaired, hemizygous stretches of DNA [10, 11]. The Wrangel Island Mammoth has sharply increased heterozygosity for deletions in comparison with the Oimyakon mammoth (S3 Table). Some portion of the inflated heterozygosity for deletions in the Wrangel Island mammoth could be due to this difficulty in inferring genotypes in a high throughput setting. Alternatively, the effective mutation rate may have increased as fewer deletions were removed from the population via purifying selection, inflating θdel. It is also possible that there was an increase in the rate of deletions in the Wrangel Island lineage due to defective DNA repair mechanisms. An increase in non-homologous end joining after DNA breaks rather than double stranded break repair could putatively induce such a change in the deletion rate.
Maya the Indian elephant shows a larger number of deletions than the Oimyakon mammoth, but with different character from the Wrangel Island mammoth. The bulk of these are derived from 22,954 hemizygous deletions (S3 Table). Maya houses only 5141 homozygous deletions, similar to the mainland mammoth (S3 Table). There is an increase in the number of hemizygous deletions that affect gene sequences, but only a modest increase in the number of homozygous deletions that affect gene sequences (S3 Table). Competing pressures of higher Ne, longer time frames to accumulate mutations toward equilibrium frequencies, differences in mutation rates between the mammoths and elephants, differences in selective pressures, differences in the distribution of selective coefficients for deletions, different effective mutation rates due to different selective constraints, or differences in dominance coefficients might all contribute to differences in the number of deletions observed in elephants and mammoths. Additional samples would be necessary to determine the extent to which genetic declines may be influencing the diversity of deletions in modern Indian elephants. We currently have no basis for conclusions given this single sample, with no prior comparison.
There is a significant difference in the size distribution of deletions identified in the two mammoth samples, with a mean of 1707 bp in Oimyakon and 1606 bp in the Wrangel mammoth (Wilcox W = 304430000, P < 2.2e − 16; Fig 2). This difference could reflect either differences in DNA replication or repair mechanisms in the two mammoths, or altered selective constraints for different types of duplications. No significant difference is observed between the Wrangel island mammoth down sampled sequence data (W = 2004400, P = 0.3917) suggesting that the observed decrease in size is not due to differences in coverage. Some 1628 genes have deleted exons in the Wrangel Island mammoth compared to 1110 in Oimyakon (Table 1), a significant excess of genes deleted compared to expectations based on the number of deletions (χ2 = 12.717, df = 1,P = 0.0003623). Among these deleted genes, 112 in the mainland mammoth are homozygous compared to 133 homozygous exon deletions in the Wrangel Island Mammoth. Gene functions for affected genes in the Oimyakon mammoth include synapse functions, PHD domains, zinc fingers, aldo-keto metabolism, calcium dependent membrane targeting, DNA repair, transcription regulation, and development (S5 Table). Gene functions overrepresented among deletions in the Wrangel Island mammoth include major urinary proteins, lipocalins, and pheromones, pleckstrins, transcription regulation, cell transport, DNA repair, chromatin regulation, hox domains, and development (S5 Table).
Among the genes deleted in the Wrangel Island mammoth, several have phenotypes of interest in other organisms. We observe a hemizygous deletion in riboflavin kinase RFK in the Wrangel Island mammoth, but normal coverage in the Oimyakon mainland mammoth (S1 Fig). Homozygous knockouts of riboflavin kinase, essential for B2 utilization/FAD synthesis, are embryonic lethal in mice [12]. Finally, we identify a hemizygous deletion in the Wrangel island mammoth that would remove the entire gene sequence at the FOXQ1 locus (S2 Fig). The alternative haplotype carries a frameshift mutation that disrupts the FOXQ1 functional domain. FOXQ1 knock-outs in mice are associated with the satin coat phenotype, which results in translucent fur but normal pigmentation due to abnormal development of the inner medulla of hairs [13], with two independent mutations producing this phenotype [13]. FOXQ1 also regulates mucin secretion in the GI tract, a case of pleiotropic functions from a single gene [14]. If the phenotype in elephantids matches the phenotype exhibited in mice, this mammoth would have translucent hairs and a shiny satin coat, caused by two independently formed knock-out alleles at the same locus. These genes each have functions that are conserved across mammals, though there is no guarantee that they would produce identical phenotypes in other species.
Retrogene formation can serve as a proxy for retrotransposon activity. We identify retrogenes that display exon-exon junction reads in genomic DNA. We observe 1.3X more retrogenes formed in the Wrangel island mammoth. The Wrangel Island mammoth has 2853 candidate retrogenes, in comparison with 2130 in the Oimyakon mammoth and 1575 in Maya (Table 1). There are 436 retrogenes that are shared between the two mammoths, though some of these could arise via independent mutations. This excess of retrogenes is consistent with increased retroelement activity in the Wrangel Island lineage. During retrogene formation, highly expressed genes, especially those expressed in the germline, are expected to contribute to new retrogenes. To determine the types of loci that had been copied by retrotransposons, we performed a gene ontology analysis using DAVID [8, 9]. Functional categories overrepresented among candidate retrogenes include genes involved in transcription, translation, cell division/cytoskeleton, post translational modification, ubiquitination, and chaperones for protein folding (S6 and S7 Tables). All of these are expected to be highly expressed during cell divisions or constitutively expressed, consistent with expectations that highly expressed genes will be overrepresented. Gene ontologies represented are similar for both mammoths (S6 and S7 Tables). Although these retrogenes are unlikely to be detrimental in and of themselves, they may point to a burst of transposable element activity in the lineage that led to the Wrangel island individual. Such a burst of TE activity would be expected to have detrimental consequences, additionally contributing to genomic decline.
Under nearly-neutral theory of genome evolution, detrimental mutations should accumulate in small populations as selection becomes less efficient [5]. This increase in non-neutral amino acid changes and premature stop codons is consistent with reduced efficacy of selection in small populations. We attempted to determine whether the data is consistent with this nearly-neutral theory at silent and amino acid replacement substitutions whose mutation rates and selection coefficients are well estimated in the literature. Under nearly neutral theory, population level variation for non-synonymous amino acid changes should accelerate toward parity with population level variation at synonymous sites.
Given the decreased population size on Wrangel Island, we expect to observe an accumulation of detrimental changes that would increase heterozygosity at non-synonymous sites (HN) relative to synonymous sites (HS) in the island mammoth. Heterozygosity depends directly on effective population sizes. We observe HS = 0.00130 ± 0.00002 in the Wrangel Island mammoth, which is 80% of HS = 0.00161 ± 0.00002 observed in the Oimyakon mammoth (Table 2). The magnitude of the difference between HS in these two mammoths is 28 standard deviations apart, suggesting that these two mammoths could not have come from populations with the same effective population sizes. The specimens are well beyond the limits of expected segregating variation for a single population. To determine whether such results are consistent with theory, we fitted a model using PSMC [42] inferred population sizes for the Wrangel island mammoth, based on decay of heterozygosity of (1 − 1/2N)t H0. The observed reduction in heterozygosity is directly consistent theoretical expectations that decreased effective population sizes would lower heterozygosity to HS = 0.00131.
At non-synonymous sites, however, there are no closed-form solutions for how HN would decay under reduced population sizes. We observe HN = 0.000490 in the Wrangel Island Mammoth, 95% of HN = 0.000506 in the Oimyakon mammoth (Table 2). To determine whether such results could be caused by accumulation of nearly-neutral variation, we simulated population trajectories estimated using PSMC [42]. This trajectory shows ancient populations with Ne ≈ 104, followed by a population decline prior to extinction. These numbers are slightly lower than previous estimates of ancestral Ne based on mitochondrial DNA [43]. We were able to qualitatively confirm results that population trajectories from PSMC with previously described mutation rates and selection coefficients can lead to an accumulation of detrimental alleles in populations. However, the magnitude of the effects is difficult to fit precisely. The simulations show a mean HS = 0.00148 and HN = 0.000339 in Oimyakon and HS = 0.00126 and HN = 0.000295 for the Wrangel Island Mammoth (S3 Fig). In simulations, we estimate HN/HS = 0.229 both for the Oimyakon mammoth and directly after the bottleneck, but HN/HS = 0.233 in the Wrangel Island Mammoth at the time of the Wrangel Island mammoth. These numbers are less than empirical observations of HN/HS = 0.370 (Table 2). Several possibilities might explain the observed disparity between precise estimates from simulations versus the data. The simulations may be particularly sensitive to perturbations from PSMC population levels or time intervals. Similarly, selection coefficients that differ from the gamma distribution previously estimated for humans might lead to greater or lesser changes in small populations. Additionally, an acceleration in generation time on Wrangel Island is conceivable, especially given the reduced size of Wrangel Island mammoths [15]. Finally, positive selection altering nucleotide variation on the island or the mainland could influence diversity levels.
Founder effects during island invasion sometimes alter genetic diversity in populations. However, it is unlikely that a bottleneck alone could cause an increase in HN/HS. There is no evidence in effective population sizes inferred using PSMC to suggest a strong bottleneck during Island colonization [4]. The power of such genetic analyses may be limited, but these results are in agreement with paleontological evidence showing no phenotypic differentiation from the mainland around 12,000 years ago followed by island dwarfism much later [15]. During glacial maxima, the island was fully connected to the mainland, becoming cut off as ice melted and sea levels rose. The timing of separation between the island and mainland lies between 10,000 years and 14,000 years before present [3, 15–17], but strontium isotope data for mammoth fossils suggests full isolation of island populations was not complete until 10,000-10,500 years ago [18]. Forward simulations suggest that hundreds of generations at small Ne are required for detrimental mutations to appear and accumulate in the population. These results are consistent with recent theory suggesting extended bottlenecks are required to diminish population fitness [19]. Thus, we suggest that a bottleneck alone could not produce the accumulation of HN/HS that we observe.
E. maximus indicus specimen, Maya shows an independent population decline in the past 100,000 years, with current estimates of Ne = 1000 individuals (S4 Fig). This specimen shows a parallel case of declining population sizes in a similar species of elephantid. Maya houses hemizygous deletions in similar numbers with the Wrangel Island Mammoth. However, the number of stop codons and homozygous deletions is intermediate in comparison with the Oimyakon and Wrangel mammoths (Table 1). It is possible that Indian elephants, with their recently reduced population sizes may be subject to similar accumulation of detrimental mutations, a prospect that would need to be more fully addressed in the future using population genomic samples for multiple individuals or timepoints and more thorough analyses.
Nearly-neutral theories of genome evolution have attempted to explain the accumulation of genome architecture changes across taxa [5]. Under such models, mutations with selection coefficients less than the nearly neutral threshold will accumulate in genomes over time. Here, we test this hypothesis using data from a woolly mammoth sample from just prior to extinction. We observe an excess of retrogenes, deletions, amino acid substitutions, and premature stop codons in woolly mammoths on Wrangel Island. Given the long period of isolation and extreme population sizes observed in pre-extinction mammoths on Wrangel Island, it is expected that genomes would deteriorate over time. These results offer genetic support for the nearly-neutral theory of genome evolution, that under small effective population sizes, detrimental mutations can accumulate in genomes. Independent analysis supporting a reduction in nucleotide diversity across multiple individuals at MHC loci suggests a loss of balancing selection, further support the hypothesis that detrimental variants accumulated in small populations [20].
We observe two independent loss-of-function mutations in the Wrangel Island mammoth at the locus of FOXQ1. One mutation removes the entire gene sequence via a deletion, while the other produces a frameshift in the CDS. Based on phenotypes observed in mouse models, these two independent mutations would result in a satin fur coat, as well as gastric irritation [14]. Many phenotypic screens search for homozygous mutations as causative genetic variants that could produce disease. More recently, it has been proposed that the causative genetic variation for disease phenotypes may be heterozygous non-complementing detrimental mutations [21]. These data offer one case study of independent non-functionalizing mutations in a single individual, genetic support for independent non-functionalizing mutations at a single locus. Woolly mammoth outer hairs house multiple medullae, creating a stiff outer coat that may have protected animals from cold climates [22] (though see [23] for alternative interpretations). Putative loss of these medullae through loss of FOXQ1 could compromise this adaptation, leading to lower fitness.
One of the two specimens comes from Wrangel Island, off the northern coast of Siberia. This mammoth population had been separated from the mainland population for at least 6000 years after all mainland mammoths had died off. Prior to extinction, some level of geographic differentiation combined with differing selective pressures led to phenotypic differentiation on Wrangel island [15]. Island mammoths had diminished size, but not until 12,000 years ago when mainland populations had reduced and ice sheets melted [15]. One possible explanation for the poor fit of simulations is that generation time may have decreased. Previous work suggested a very high mutation rate for woolly mammoths based on comparisons between island and mainland mammoths. It is possible that an acceleration in generation times could cause the accumulation of more mutations over time, and that the real mutation rate is similar to humans (1 − 2 × 10−8 [24] rather than 3.8 × 10−8 [4]). Such changes would be consistent with island dwarfism being correlated with shorter generation times, and would explain the unusually high mutation rate estimate for mammoths based on branch shortening observed in [4].
We observe large numbers of pseudogenized olfactory receptors in the Island mammoth. Olfactory receptors evolve rapidly in many mammals, with high rates of gain and loss [25]. The Wrangel island mammoth has massive excess even compared to the mainland mammoth. Wrangel island had different flora compared to the mainland, with peat and sedges rather than grasslands that characterized the mainland [17]. The island also lacked large predators present on the mainland. It is possible that island habitats created new selective pressures that resulted in selection against some olfactory receptors. Such evolutionary change would echo gain and loss of olfactory receptors in island Drosophila [26]. In parallel, we observe a large number of deletions in major urinary proteins in the island mammoth. In Indian elephants E. maximus indicus, urinary proteins and pheromones ellicit behavioral responses including mate choice and social status [27]. It is possible that coevolution between urinary proteins, olfactory receptors, and vomeronasal receptors led to a feedback loop, allowing for rapid loss in these related genes. It is equally possible that urinary peptides and olfactory receptors are not essential and as such they are more likely to fall within the nearly neutral range [25]. Either of these hypotheses could explain the current data.
Many factors contributed to the demise of woolly mammoths in prehistoric times. Climate change led to receding grasslands as forests grew in Beringia and North America and human predation placed a strain on already struggling populations [2]. Unlike many cases of island invasion, Wrangel Island mammoths would not have continuous migration to replenish variation after mainland populations went extinct. Under such circumstances, detrimental variation would quickly accumulate on the island. The putatively detrimental variation observed in these island mammoths, with the excess of deletions, especially recessive lethals may also have limited survival of these struggling pre-extinction populations. Climate change created major limitations for mammoths on other islands [28], and these mammoths may have struggled to overcome similar selective pressures.
Many modern day species, including elephants, are threatened or endangered. Asiatic cheetahs are estimated to have fewer than 100 individuals in the wild [29]. Pandas are estimated to have 1600 individuals living in highly fragmented territories [30]. Mountain Gorilla population census sizes have been estimated as roughly 300 individuals, similar to effective population sizes for pre-extinction mammoths [31]. If nearly neutral dynamics of genome evolution affect contemporary endangered species, detrimental variation would be expected in these genomes. With single nucleotide changes, recovered populations can purge detrimental variation in hundreds to thousands of generations, returning to normal genetic loads [19]. However, with deletions that become fixed in populations, it is difficult to see how genomes could recover quickly. The realm of back mutations to reproduce deleted gene sequences will be limited or impossible. Although compensatory mutations might conceivably correct for some detrimental mutations, with small effective population sizes, adaptation through both new mutation and standing variation may be severely limited [32]. Thus we might expect genomes affected by genomic meltdown to show lasting repercussions that will impede population recovery.
All sequences are taken from publicly available sequence data in the ENA or SRA. Indian elephant specimens for previously published sequence data were handled by the San Diego Zoo.
We used previously aligned bam files from ERR852028 (Oimyakon, 11X) and ERR855944 (Wrangel, 17X) (S8 Table) [4] aligned against the L. africana 4.0 reference genome (available on request from the Broad Institute—[email protected]; https://www.broadinstitute.org/scientific-community/science/projects/mammals-models/elephant/elephant-genome-project). We also aligned 33X coverage of sequencing reads for one modern E. maximus indicus genome Maya (previously described as “Uno”) using bwa 0.7.12-r1044 [33], with parameters set according to [4] bwa aln -l 16500 -o 2 -n 0.01. The E. maximus indicus sample, previously labeled in the SRA as “Uno”, is from Maya, a former resident of the San Diego Zoo wild-born in Assam, India, North American Studbook Number 223, Local ID #141002 (O. Ryder, personal communication). We were not able to use two other mammoth sequences are publicly available, M4 and M25 from Lynch et al. [34]. These sequences display abnormal PSMC results (S4 Fig), high heterozygosity (S5 Fig), and many SNPs with asymmetrical read support (S6 Fig). The unrealistically high heterozygosity as well as abnormal heterozygote calls raise concerns with respect to sequence quality. For further description, please see Supporting Information.
We used the GATK pipleine [7] v3.4-0-g7e26428 to identify SNPs in the aligned sequence files for the Oimyakon and Wrangel Island mammoths. We identified and realigned all indel spanning reads according to the standard GATK pipeline. We then identified all SNPs using the Unified Genotyper, with output mode set to emit all sites. We used all CDS annotations from cDNA annotations from L. africana r3.7 with liftover coordinates provided for L. africana 4.0 to identify SNPs within coding sequences. We identified all stop codons, synonymous substitutions, and non-synonymous substitutions for the Wrangel Island and Oimyakon mammoths at heterozygous and homozygous sites.
We aligned all reads from the mammoth genome sequencing projects ERR852028 (Oimyakon) and ERR855944 (Wrangel) (S8 Table) against elephant cDNA annotations from L. africana r3.7. Sequences were aligned using bwa 0.7.12-r1044 [33], with parameters set according to [4] bwa aln -l 16500 -o 2 -n 0.01 in order to account for alignments of damaged ancient DNA. We then collected all reads that map to exon-exon boundaries with at least 10 bp of overhang. Reads were then filtered against aligned genomic bam files produced by Palkopoulou et al [4], discarding all exon-exon junction reads that have an alignment with equal or better alignments in the genomic DNA file. We then retained all putative retrogenes that showed signs of loss for two or more introns, using only cases with 3 or more exon-exon junction reads.
We calculated coverage depth using samtools [35] with a quality cutoff of -q 20. We then implemented change point analysis [36] in 20 kb windows. Change point methods have been commonly used to analyze microarray data and single read data for CNVs [37–39] The method seeks compares the difference in the log of sum of the squares of the residuals with one regression line vs. two regression lines [36]. The test statistic follows a chi-squared distribution with a number of degrees of freedom determined by the number of change-points in the data, in this case df = 1. We required significance at a Bonferroni corrected p-value of 0.05 or less. We allowed for a maximum of one CNV tract per window, with minimum of 1 kb and maximum of 10 kb (half the window size) with a 100 bp step size. We did not attempt to identify deletions smaller than 1 kb due to general concerns of ancient DNA sequence quality, limitations to assess small deletions in the face of stochastic coverage variation, and concerns that genotype calls for smaller deletions might not be as robust to differences in coverage between the two mammoths. Sequences with ‘N’s in the reference genome did not contribute to change point detection. We excluded all deletions that were identified as homozygous mutations in both mammoths and in E. maximus indicus specimen Maya, as these suggest insertion in the L. africana reference rather than deletion in other elephantids. To determine the effects that coverage differences would have on deletions, we downsampled the sequence file for the Wrangel Island mammoth using samtools to 11X coverage, using chromosome 1 as a test set. We observe a reduction in the number of deletions for chromosome 1 from 1035 deletions to 999 deletions, resulting in an estimated false negative rate of 0.5% at reduced coverage for deletions greater than 1 kb. Highly diverged haplotypes with greater than 2% divergence might prevent read mapping and mimic effects of deletions, but this would require divergence times within a species that are greater than the divergence between mammoths and L. africana. Mutations were considered homozygous if mean coverage for the region was less than 10% of the background coverage level. Otherwise it was considered to be heterozygous. These methods are high-throughput, and it is possible that multiple small homozygous deletions interspersed with full coverage sequences might mimic heterozygote calls. Whether such mutations might meet the conditions for significant change-point detection would depend on the deletion length, placement, and background coverage level.
We identified SNPs that differentiate Mammoth genomes from the reference using samtools mpileup (options -C50 -q30 -Q30), and bcftools 1.2 consenus caller (bcftools call -c). The resulting vcf was converted to fastq file using bcftools vcf2fq.pl with a mimimum depth of 3 reads and a maximum depth of twice the mean coverage for each genome. Sequences were then converted to psmc fasta format using fq2psmcfa provided by psmc 0.6.5-r67. We then ran psmc with 25 iterations (-N25), an initial ratio of θ/ρ of 5 (-r5), and parameters 64 atomic time intervals and 28 free parameters (-p “4+25*2+4+6”) as was done in previous analysis of woolly mammoths [4]. Effective population sizes and coalescence times were rescaled using previously estimated mutation rates of 3.8 × 10−8. Using the population size estimates from PSMC, we calculated the expected reduction in heterozygosity at synonymous sites according to ( 1 - 1 2 N ) t for each time period in PSMC output. We compared the number of deletions, number of premature stop codons, proportion affecting gene sequences, and number of putative retrogenes between the two mammoth genomes using chi squared tests.
To determine expectations of sequence evolution at non-synonymous sites under population crash, we ran simulations using SLiM v. 2.0 population genetic software [40]. We modeled two classes of sites: neutral and detrimental. For detrimental mutations we used a gamma distributed DFE with a mean of -0.043 and a shape parameter of 0.23 as estimated for humans [41], assuming a dominance coefficient of 0.5 and free recombination across sites. Mutation rates were set as 3.8 × 10−8 based on previously published estimates [4]. The trajectory of population sizes was simulated according to estimates from PSMC, omitting the initial and final time points from PSMC, which are often subject to runaway behavior. We then simulated the accumulation of HN/HS in the Wrangel Island Mammoths. Simulations were run with a burn-in of 100,000 generations. We simulated 460 replicates of haplotypes with 100 sites for each mutation class.
To gather a portrait of functional categories captured by deletions, retrogenes, and stop codons, we identified all mouse orthologs based on ENSEMBL annotations for L. africana 3.7 for affected gene sequences. We then used DAVID gene ontology analysis with the clustering threshold set to ‘Low’ (http://david.ncifcrf.gov/; Accessed April 2016) [8, 9]. S2–S7 Tables include all functions overrepresented at an EASE enrichment cutoff of 2.0. Full gene ontology data is included in Supplementary Information.
|
10.1371/journal.ppat.1003038 | HCMV-Infected Cells Maintain Efficient Nucleotide Excision Repair of the Viral Genome while Abrogating Repair of the Host Genome | Many viruses subvert the host cell's ability to mount and complete various DNA damage responses (DDRs) after infection. HCMV infection of permissive fibroblasts activates host DDRs at the time of viral deposition and during replication, but the DDRs remain uncompleted without arrest or apoptosis. We believe this was in part due to partitioning of the damage response and double strand break repair components. After extraction of soluble proteins, the localization of these components fell into three groups: specifically associated with the viral replication centers (RCs), diffused throughout the nucleoplasm and excluded from the RCs. Others have shown that cells are incapable of processing exogenously introduced damage after infection. We hypothesized that the inability of the cells to process damage might be due to the differential association of repair components within the RCs and, in turn, potentially preferential repair of the viral genome and compromised repair of the host genome. To test this hypothesis we used multiple strategies to examine repair of UV-induced DNA damage in mock and virus-infected fibroblasts. Comet assays indicated that repair was initiated, but was not completed in infected cells. Quantitative analysis of immunofluorescent localization of cyclobutane pyrimidine dimers (CPDs) revealed that after 24 h of repair, CPDs were significantly reduced in viral DNA, but not significantly changed in the infected host DNA. To further quantitate CPD repair, we developed a novel dual-color Southern protocol allowing visualization of host and viral DNA simultaneously. Combining this Southern methodology with a CPD-specific T4 endonuclease V alkaline agarose assay to quantitate repair of adducts, we found efficient repair of CPDs from the viral DNA but not host cellular DNA. Our data confirm that NER functions in HCMV-infected cells and almost exclusively repairs the viral genome to the detriment of the host's genome.
| Human cytomegalovirus (HCMV) is a leading cause of birth defects. This may be due in part to this virus' ability to inflict specific damage to its host's DNA, combined with the disruption of an infected cell's ability to repair damage. Earlier studies found that components of the cell's repair machinery were differentially associated with the HCMV viral replication centers in the nucleus. Experiments here extend this observation to include components of the machinery involved in UV lesion repair. We hypothesized that association of components of the DNA repair machinery within the viral replication centers could favor the repair of viral DNA, but more importantly, be detrimental to the repair of cellular DNA. Infected cells were irradiated and examined for repair by three different methods. In the course of this study, we developed a new technique allowing simultaneous evaluation of both the viral and host genomes in an infected cell. These experiments found rapid, selective removal of UV lesions from the viral and not the cellular DNA within infected cells. Our results indicate the differential association of certain cellular repair proteins with this virus may have far-reaching implications in the disease pathogenesis of HCMV infection.
| Human Cytomegalovirus (HCMV) is among the leading causes of birth defects in the United States, affecting an estimated 8000 children per year [1]. Each year ∼1% of all newborns are congenitally infected with HCMV. Of these infants, 5–10% manifest signs of serious neurological defects at birth [2]–[5], with an additional 10–15% subsequently suffering consequences by age five. Recent literature also points to HCMV as a contributing agent for the development of certain types of cancers (for review see [6], [7]). Studies of HCMV infection in non-permissive cells indicate that HCMV can also act as a mutagen [8]–[10], inducing “hit and run” damage.
There is significant evidence that non-specific chromosomal aberrations and damage to the mitotic apparatus can occur in cells infected with a variety of human DNA and RNA viruses (see [11] for review). Yet, only two viruses, the oncogenic adenoviruses (Ad) and HCMV, have been found to cause site-specific chromosomal damage [11]–[13]. We have shown that HCMV is able to induce specific damage in chromosome 1 at two loci, 1q23 and 1q42 [12], [13], as early as 3 h post infection (hpi). In contrast to Ad type 12 [14], [15], induction of specific breaks by HCMV does not require de novo viral protein expression. Viral entry into the cell is sufficient to cause the specific breaks. It is also clear from the literature that many viruses interact with their hosts' DNA damage response (DDR) signaling molecules and repair machinery, often triggering responses upon initial entry and deposition of the genome in the nucleus or through successive rounds of replication. Some viruses are reported to utilize this initial DDR response to optimize infection, while others have been found to thwart it (as reviewed in [16], [17]). Work from our lab and others [18]–[20] has shown that host DDRs are activated both at the point of viral deposition and during late phase replication of HCMV in permissive fibroblasts, although the importance of this activation for establishing a fully permissive infection remains unclear. During HCMV infection, DDRs are not finished, resulting in incomplete repair without arrest or apoptosis. We have shown this is due, at least in part, to a differential association of the repair machinery components into the viral replication centers (RCs). After extraction of soluble proteins, we determined three categories of association: specifically associated within RCs, diffused throughout the nucleoplasm and excluded from the RCs [18].
These earlier studies demonstrated specific viral associations with key players in the DDR pathways. Other studies have examined the capability of infected cells (or cells expressing specific viral proteins) to repair different types of damage after infection and found both increases and decreases in the ability to repair induced damage [21]–[40]. However, these earlier studies looked at total cellular DNA and did not examine repair of the viral and host genomes separately within these cells.
We hypothesized that after the RCs were established, association of components of the DNA repair machinery within the RCs of the virus could favor viral repair, but more importantly, be detrimental to repair of the cellular DNA. To the best of our knowledge, our experiments in HCMV-infected cells are the first to examine repair in the host and viral DNA separately and the possibility of preferential repair in the viral DNA. To test our hypothesis, exogenous DNA damage was introduced into cells (UV dimers) via UVC (hereafter UV) irradiation. Analysis of damage repair used comet assays, immunofluorescent localization (IF) of UV-induced cyclopyrimidine dimers (CPDs) and T4 endonuclease V alkaline agarose (T4) assays. These studies found that HCMV-infected cells, although capable of mounting a damage response to UV irradiation, were unable to completely repair all of the exogenously introduced DNA damage. In situ localization of the CPDs clearly showed that the residual damage detected in these cells was found entirely within the cellular DNA. Moreover, dual-color T4 assays revealed proficient repair of CPDs from viral DNA but defective repair of host DNA within infected cells. Thus, there was selective repair of DNA damage in viral when compared to cellular DNA in permissively infected fibroblasts, indicating that association of the host's DNA repair machinery with HCMV RCs has detrimental consequences for the host.
Our previous work reported tight association of some, but not all, of the ATM-mediated double strand break (DSB) and ATR- mediated stalled replication fork response proteins with HCMV viral RCs within the nucleus of permissively infected cells by 48 hpi [18]. These studies were performed using “extraction first” procedures [41], which differ from the more common “fix first” technique which uses formaldehyde to initially fix proteins in place, followed by permeabilization in detergent to allow access of antibodies (Abs) into the cell. This “fix first” methodology allows for visualization of the entire complement of a given protein within the cell, regardless of how tightly or loosely it is associated with a given compartment. By contrast, an “extraction first” protocol initially treats cells with detergent and then fixes them in formaldehyde [41]. Initial extraction removes proteins that are not attached to the chromatin or scaffolding substructure of the nucleus, providing a clearer view of proteins associated with a given compartment or structure. It is often used in the visualization of DSB repair foci in damaged cells, as only a fraction of the entire protein complement will relocalize to sites of damage.
Performing these two fixation/extraction procedures provided valuable information regarding the nature of protein interactions in infected cells. First, we concluded that the majority of a protein was tightly associated with the RCs if it was distinctly localized within the RCs using only “fix first” protocols. Second, if we saw more distinct localization of a protein with the RCs after “extraction first” conditions we inferred that only a portion of the protein was tightly associated with these centers. It should be noted that the proportion not tightly associated with the RCs was also not tightly associated with the host DNA. Lastly, we concluded the protein was not specifically associated or excluded from these centers when no clear localization to the RCs occurred using either “fix first” or “extraction first” conditions.
A similar pattern of selective association of nucleotide excision repair (NER) proteins was found in permissively infected HFFs as had been observed for the DSB repair proteins [18]. Figure 1a shows an example of tight association of the Cockaine's Syndrome B (CSB) protein with the viral RCs within the nucleus (as evidenced by colocalization with the viral processivity factor, UL44). Figure 1b shows two other NER proteins, illustrating either tight association with the RCs (XPD) and an example of diffused nuclear staining (XPG). Figure 1 also illustrates the nuclear localization of these three repair proteins in mock-infected cells as a control. In addition, a summary of all the NER-associated proteins tested for localization after infection is given in Table 1. Table 1 indicates the localization of NER proteins using “fix first” or “extraction first” conditions, as indicated (note: it is often difficult to use rabbit primary Abs at 48 hpi using “fix first” conditions due to non-specific binding to the virus-encoded Fc receptor in the cytoplasm).
The differential localization we observed with the proteins crucial to NER within the RCs led us to hypothesize that the repair of viral DNA could be favored, potentially to the detriment of cellular DNA. This hypothesis was tested using UV irradiation of HCMV infected cells.
The first method used to test our hypothesis and visualize repair of UV-induced damage was the single cell gel electrophoresis or comet assay system [42]–[45]. The comet assay has been used historically in the literature to analyze repair of UVC-induced damage. Comet tails are visualized by staining with the fluorescent DNA intercalating agent SYBR Green, which binds to both ss- and dsDNA. Although comet tails could represent other alkali-labile forms of damage, the literature suggests that the very large proportion (>90%) of damage observed in cells irradiated with low doses of UVC irradiation (similar to our studies) are UV dimers ([46]–[48] and references within).
Tails at early timepoints are believed to be the result of initial incision events associated with NER processing of UV dimers. These incisions (strand breaks) allow uncoiling/relaxation of the chromatin to occur. Electrophoretically induced migration of the uncoiled/relaxed DNA is visualized as the formation of a comet tail (as reviewed in [49]). Therefore, only cells capable of initiating repair will have comet tails following UV irradiation [46]. Over a timecourse, successful repair is demonstrated by a decrease in both the number of cells with tails and the % DNA in the tails. Importantly, it has been shown convincingly that cells deficient in NER proteins, e.g. XP proteins, do not form tails above background levels in comet assays after UVC irradiation due to lack of incision events ([46], [47] and references within).
Two independent comet experiments were conducted on HFFs infected for 48 h and then irradiated with 50 J/m2 of UV and allowed to repair for 2, 6 or 24 h. It should be noted that this UVC dosage was non-lethal to the cells, with no cell loss in any of the experiments. Cell counts with and without UVC were essentially identical over the entire timecourse. The graph in Figure 2B shows the data from one of these experiments; data from the other was comparable. Comets were scored using VisComet software. The average % tail DNA for the population of cells scored in a given set is represented in each bar (error bars represent one SD). Four populations are represented in the graph and are shown in different shades of grey: mock infected cells plus or minus irradiation (M+UV and M alone, respectively) and virus-infected cells plus or minus irradiation (V+UV and V alone, respectively). A representative image of each group is shown in Figure 2A. As expected, M alone cells (white bars) produced very limited comet tails, with an average of less than 10% tail DNA. The M+UV cells (light grey bars) had significantly increased % tail DNA at early times, indicating their ability to begin NER incision events was intact. Twenty-four h of repair returned the M+UV cells' % tail DNA toward the M baseline. Surprisingly, the V alone samples (dark grey bars) had elevated % tail DNA throughout the timecourse. This seemingly perplexing result was investigated further (see below). The V+UV samples (black bars) had a substantially larger % tail DNA than the M+UV cells as early as 2 hp irradiation. In contrast to the M+UV samples, the % tail DNA in the V+UV samples did not decrease during the ensuing 24 h period. In fact, the % tail DNA remained high through 48 h of repair time (data not shown). Throughout the timecourse, the V+UV samples had statistically significant increases in % tail DNA over their M+UV counterparts (as measured by unpaired t-tests and indicated by asterisks in the graph). The persistence of comet tails in the V+UV samples was examined further in BrdU pulse/chase experiments below.
The distribution of % tail DNA within each sample type was plotted to distinguish whether changes were occurring over the timecourse (Figure 2C - four ranges of % tail DNA are represented in shades of grey). The distribution shown represents the experiment in Figure 2B. The distribution plot shows the increasing percentage of M+UV cells with less than 10% tail DNA over the timecourse of repair. In contrast virtually all cells in the V+UV samples have very high levels of tail DNA (greater than 50%) for the entire timecourse. As mentioned above, unexpectedly, the V alone samples increased in tail DNA percentage over the timecourse.
The source of comet tails in V alone samples was a conundrum. Electrophoretically induced migration of uncoiled/relaxed DNA is measured by the comet assay as the formation of a tail (as reviewed in [49]). The body of comet assay literature strongly suggests that the relaxation associated with open replication forks and Okazaki fragments connected with replicating genomes could appear as tail DNA in this assay ([50] and references within). Several studies have shown that HFFs infected in G0 with HCMV arrest at or near the G1/S transition, resulting in the replication of viral, but not cellular, DNA within these cells [51]–[54]. To determine whether the increase in % tail DNA in V alone samples was possibly attributable to the previously observed specific DSBs induced in a subset of cells by the incoming virus inoculum [12] or, more likely, the increase in viral DNA replication over time, the comet assays were repeated exactly as previously described in two parallel sets of samples (eight groups in total).
Ganciclovir, a viral replication inhibitor, was added to one of the parallel sets of cells beginning at 24 hpi and throughout the remainder of the experiment. Addition of the drug at 24 hpi interrupted the infection at the pre-replication foci stage and any further development of these foci (and associated viral replication) in the treated samples was halted. All samples were irradiated at 48 hpi and harvested 24 h later. We reasoned that, if comet tails in the V alone samples were due primarily to viral replication, ganciclovir treatment should reduce comet tail levels toward M alone background levels. As can be seen in Figure 3, inhibition of viral DNA replication in the V alone samples dramatically reduced % tail DNA back toward M alone baseline levels. This reduction was the only statistically significant difference observed with the addition of ganciclovir to the samples (as measured by unpaired t-tests and indicated with an *). This result led us to conclude that the majority of the DNA in the V alone comet tails was not due to specific DSBs, but rather primarily due to viral replication.
Interestingly, only nominal decreases in the % tail DNA were seen in the V+UV cells treated with ganciclovir. Our comet assay experiments with these cells were likely detecting the initiation of DNA repair in both host and viral DNA, not replicating viral genomes, however the ganciclovir experiments could not definitively distinguish if irradiation had inhibited viral replication during the repair cycle in the V+UV cells. Therefore, we assessed the extent of viral replication over time by BrdU-labeling the viral DNA before and at several points after UV-irradiation.
As described previously, cells were infected for 48 h on coverslips. Coverslips were then divided into two groups. One group was not irradiated and the second group received 75 J/m2 UV. In addition, one coverslip from each group was pulse-labeled with BrdU to provide a baseline of incorporation (and viral replication) (t = 0 h in Figure 4). The level of active viral replication in the two groups at each timepoint post irradiation was assessed by pulse labeling with BrdU (as described in Materials and Methods) just prior to harvesting coverslips (one each from the unirradiated and irradiated groups).
As seen in Figure 4, the unirradiated samples continued incorporating BrdU into the replicating viral DNA located in the RCs (upper panels, RCs are marked with arrows). However, after irradiation, viral replication essentially ceased in the irradiated samples (bottom panels, +6 and +11 h images show essentially no RC staining). A small amount of BrdU incorporation into the DNA in the RCs was seen in these V+UV cells after 24 h of recovery, but the amount of incorporation was nominal compared to their unirradiated counterparts. We also believe that the signal observable in the RCs at 24 hp irradiation in the V+UV samples is most probably due to BrdU incorporation into repair patches during unscheduled DNA synthesis (UDS) [55]. Small regions (of ∼20 nucleotides) must be resynthesized after removal of CPDs from the viral DNA. BrdU incorporation is commonly utilized in the DNA repair field to demonstrate repair has actually occurred in the nucleus of irradiated cells via UDS. These experiments demonstrated that viral replication was not responsible for the large % tail DNA in the V+UV comet assay samples.
There appeared to be residual damage in the V+UV cells in comparison to M+UV cells. We therefore investigated whether there was specific localization of these dimers within the nuclei. An Ab developed by Dr. Toshio Mori [56] and specific for the most prevalent form of UV-induced damage, CPDs, was used to immunofluorescently visualize induced dimers in situ at 0 and 24 hp irradiation. Removal of CPDs at doses ranging from 30–75 J/m2 UV was examined. Confocal microscopy found both M+UV and V+UV nuclei (Figure 5A) stained for CPDs across the entire nucleus at time 0 h post irradiation. CPD adducts were formed in both cellular and viral DNA, with minor variations in intensity seen across an individual nucleus and from cell to cell. Cells were pulse-labeled with BrdU prior to irradiation to allow localization of viral DNA within the RCs for further quantitative analysis. Cells were stained with CPD- (green) and BrdU- (red) specific Abs simultaneously. M+UV and V+UV cells both had residual CPDs 24 hp irradiation. However, in the V+UV cells, dimers were localized specifically to the periphery of the nucleus and dimers were largely absent from the viral RCs (as localized by BrdU staining).
It has previously been shown that cellular DNA is marginalized to the edges of the nucleus at late times pi using histone localization [57]. We stained both M and V cells at 0 and 24 hp irradiation with an Ab to detect the localization of histone H3 and found, much like Monier and colleagues, that this cellular histone associated almost exclusively with DNA at the edge of the nucleus and outside of the RCs in infected cells (as marked by UL44), but across the entire nucleus in mock-infected samples (Figure 5B). This confirmed that residual CPDs were located primarily within the cellular DNA. CPDs appeared to be specifically removed from the viral RCs at all three doses of irradiation tested (Figure 5a shows only images of cells treated with 75 J/m2; identical images were obtained at lower doses, which are not shown). These results led us to believe that in permissively infected HFFs irradiated at 48 hpi, there was preferential removal of CPDs from the viral DNA.
The removal of CPD signal from these infected cells was quantitated over the 24 h period of repair. Images of infected cells dually-labeled for viral DNA (BrdU) and CPDs were captured at 0 and 24 hp irradiation using confocal microscopy. All images were captured using exposure times below which any pixels were saturated, including the brightest areas at the 24 hp irradiation V+UV cells' peripheries. Data from three separate experiments were analyzed using Metamorph software as described in the Materials and Methods. Briefly, after finding the center plane of each image, the RC area and the total area of the nucleus were defined and the integrated intensities (INTINT) of both regions were recorded. An example of the regions created by MetaMorph are shown mapped onto the infected cells in Figure 5A to illustrate the process. The RC region is outlined in red and the entire nucleus in white. Subtracting the INTINT of the RC from that of the entire nucleus determined the INTINT of the cellular DNA for each cell. For example, the intensity data for the cells shown in Figure 5A was: for the 0 h cell, Nucleus Integrated Intensity (NII)- 21.4 Million counts (M), Replication Center Integrated Intensity (RCII)- 5.4 M, Host Integrated Intensity (HII)- 16 M; for the 24 h cell, NII- 5.7 M, RCII- 1.2 M, HII- 4.5 M. Initial comparisons found differences in the total CPD INTINT signal within the nucleus at the two timepoints post irradiation (0 and 24 h). A mixed-effects ANOVA model was used to test these data for statistical significance. Using the different experimental dates as blocking factors to control for technical variation among the dates on which the experiments were performed, the results showed that the CPD signal for the entire nucleus was significantly greater at 0 h than at 24 hp irradiation (F = 20.7; df = 1, 121; p-value<0.0001). The averages for the three separate experiments are plotted (and represented by different symbols) in Figure 5C. The grey bars represent an average of the three separate experiments for ease of interpretation.
The statistically significant decrease observed in the total CPD signal prompted further analysis of the component viral and host DNA signals. Two post hoc tests were performed to determine if the decrease in the CPD signal found in the entire nucleus was independently attributable to CPD signal changes in either the host or viral DNA. The change in CPD signal in the host DNA and in the viral DNA were analyzed separately, again using a mixed effect ANOVA model to control for technical variation among dates while testing the effect of time on removal. After correcting for multiple statistical tests on the same data, the difference between 0 and 24 hp irradiation in the signal intensity within the host DNA was not significant (F = 3.1; df = 1, 123; p-value>0.16), whereas the difference between 0 and 24 hp irradiation in the signal intensity in the virus DNA was highly significant (F = 64.5; df = 1, 109; p-value<0.0002). Again, the averages for each experiment were plotted in Figure 5D, with the grey bar representing the average of the three experimental points. Although the individual experiments showed different average raw intensities, the downward trend for each experiment demonstrated a statistically significant removal of CPD signal from the viral DNA. These results clearly indicated that the decrease in the CPD signal in the nucleus following 24 h of repair was due to a decrease in the CPD signal in the virus DNA with no parallel decrease in signal within the host DNA.
Others have reported that viral DNA could potentially be replicated, packaged and transported out of the nucleus within a 24 h period [58]. To determine if specific removal of CPDs from the viral RCs was due solely to normal egress of the virus during active infection, the actively replicating virus within the RCs of HCMV infected cells was pulse-labeled with BrdU at 48 hpi. One half of the coverslips were irradiated with 75 J/m2 UV. It should be noted that experiments at 50 J/m2 produced identical results and are therefore not shown. Cells from both irradiated and unirradiated groups were harvested at 0 and 24 hp irradiation. All planes of a confocal image were projected into a single plane to gain a view of the entire cytoplasm and nucleus of an infected cell at the different times post BrdU pulse. Using these projected images, we could observe some movement of pulse-labeled virus-containing virions out into the cytoplasm of the unirradiated cells by 24 h post irradiation (visualized as individual spots of BrdU in Figure 6, top right panel). It should be noted that a significant fraction of the labeled viral genomes still remained within the RCs at this point. Conversely, we detected negligible movement of pulse-labeled viral DNA out of the RCs in the irradiated samples over the 24 h period (Figure 6, bottom right panel). This indicated that the decrease in CPD signal from the RCs observed in these cells was not caused by virus egress, but rather was due to selective removal of CPDs from the viral DNA.
Global genomic repair of CPDs can be estimated using T4 endonuclease V cleavage analysis [59]. T4 makes a highly specific single strand nick 5′ of UV-induced CPD adducts [60]. Samples are then separated via electrophoresis on an alkaline agarose gel. In these T4 gels, DNA that is either undigested or unirradiated is visible as a distinct high molecular weight (HMW) band at the top of the lane. In contrast, UV-irradiated DNA subsequently digested with T4 and electrophoresed yields a smear of lower molecular weight fragments down the lane, indicating nicking at CPD lesions. Over the course of repair, the smear returns to a HMW band indicative of repaired, full length DNA. Until now, visualization of the cleavage products has typically been via 32P labeled probes [59], [61], [62] or ethidium bromide [63], [64]. In Figure 7A, we show an example of mock-infected samples irradiated at 50 J/m2 and then digested with T4, run on an alkaline agarose gel and stained with SYBR Gold, a ss and dsDNA binding dye, to illustrate these gels. It can be clearly seen that the undigested samples remain as HMW bands and the digested samples run as a smear in the gel. As repair occurs the length of the smear decreases and the HMW band returns, indicating removal of CPDs and religation of the DNA.
When quantitating the extent of DNA damage, the average fragment length of DNA within the lane is inversely proportional to the number of CPD lesions present within the sample, i.e.- the smaller the average fragment length the more T4 cleavage sites, and therefore CPD lesions, present within the sample. These techniques have been used extensively to study genomic [59], [63], [64] and gene-specific [61], [62] CPD repair in a variety of organisms. However, determining repair of virus and host genomic DNA independently within HCMV-infected cells proved more challenging. A traditional approach would probe, strip and reprobe a single Southern blot for host and viral DNA; however stripping introduces the potential loss of signal. We developed a new method to visualize both virus and host genomic DNA simultaneously on a single blot using a Li-Cor Odyssey infrared imager.
To develop the dual-color Southern technique, we ran DNA isolated from mock-infected HFFs, pelleted viral particles and infected HFFs on a native gel before blotting to nitrocellulose and probing with digoxigenin-labeled host probes and biotin-labeled viral probes (Fig. 7B). Host DNA was visualized in the red channel (685 nm) and viral DNA was visualized in the green channel (785 nm). In the merged image, the DNA from the infected cells appeared yellow, since infected cells contained both host and viral DNA, while the uninfected cellular DNA was red and the purified viral DNA was green.
After validating this dual-color Southern technique's ability to distinguish viral from host DNA, it was used to probe experimental T4 blots (Fig. 7C). The far left panels of Figure 7C illustrate unirradiated DNA digested with T4 on these gels. The visible bands are equivalent to the HMW bands in the irradiated, undigested samples shown in Figure 7A. The right-hand panels in Figure 7C show the irradiated samples run on alkaline agarose gels. An analysis of the single channel blots and the overlay in these right-hand panels display several readily discernible features. First, after 48 h of repair, there was substantially more HMW DNA in the mock +T4 lane when compared to the infected +T4 lane in the red channel indicating decreased repair in the host DNA of infected cells (cellular DNA- top panel). Second, analysis of the viral DNA in the infected cells also showed a substantial return of a HMW band after 48 h of repair, indicating efficient repair of UV-induced DNA lesions (+T4 lane, green channel, middle panel). Lastly, analysis of the 48 h infected cell +T4 lane in the overlay blot clearly showed a gradient of colors, with the HMW band being predominantly green (viral DNA) and substantially more red signal (cellular DNA) within the smaller molecular weight fragments of the smear (bottom panel). It is important to realize that although the decrease in the DNA smear was subtle in the mock-infected and viral DNA lanes, the reappearance of a “full length” product/band at the top of the lanes as the timecourse progresses was more significant and indicated dimer removal and completed repair. This band reappears convincingly in the mock-infected and viral DNA lanes, but is nominal in the host DNA within the infected cells.
In Figure 7D, the results from five biological replicate experiments are plotted using different symbols, with the average of these experiments represented by bars for ease of comparison. The data is represented as “percent repair” of the dimers in this graph using the quantitation protocol described in Bespalov et al [59]. There is considerable variability in the results for these five T4 experiments. The variability is on par with that found in both the comet assays and the CPD removal experiments. Use of a Stratalinker for UV irradiation may have contributed to this variability, as the data shows that the initial induction of CPDs was not entirely consistent across experiments. However, rather than confound our results, we found highly statistical differences between groups as detailed below.
As depicted in Figure 7D, mock-infected HFFs repaired an average of ∼50% of CPD adducts by 48 hp irradiation (Figure 7D, blue bars). In contrast, host CPD repair in HCMV-infected cells plateaued at an average of ∼10% by 24 h and remained constant through 48 h of repair (red bars), while within the same cells, an average of ∼60% of CPDs were repaired from within the viral DNA (green bars). Statistical analyses were performed for each time point comparing the mean repair of the host DNA versus the viral DNA (or versus mock DNA) using one-tailed paired t-tests. A paired t-test controls for variation between experiments as well as unequal variances between the two measures (host versus viral or mock DNA). At each time point, the amount of viral DNA repair was statistically greater than the amount of host DNA repair (p<0.001, p<0.01, p<0.01, for timepoints 6, 24 and 48, respectively using one-tailed paired t tests). Statistically significant differences between the host and mock DNA repair were only observed at 48 hp irradiation (p<0.26, p<0.30, p<0.02, for timepoints 6, 24 and 48, respectively). These significant differences are indicated by asterisks in 7D. Therefore, repair of viral DNA was initiated more quickly and progressed more rapidly than repair of the host DNA within the same cells. The differential in repair of host and viral genomic DNA in infected cells confirmed our IF observation that CPDs were selectively removed from the viral DNA, but remained in the host genomic DNA.
The work reported here was based on the observed differential association of cellular repair proteins with viral RCs within the nucleus. We hypothesized that this association could favor viral repair and more importantly, be detrimental to repair of cellular DNA. To test this premise we UV-irradiated infected cells and then analyzed the removal of UV dimers by three methods; comet assays, IF localization of CPDs and dual-color T4 assays. Comet assays revealed that although infected cells were capable of mounting a repair response, they were unable to complete repair of all of the exogenously introduced damage. In situ localization of the CPDs showed that residual damage was confined to the cellular DNA. Lastly, dual-color T4 assays revealed faster and more significant repair of CPDs in the viral DNA than the host DNA within infected cells.
Over the past decade a great deal of work has focused on interactions of viruses and their host's DNA damage signaling molecules and repair machinery. Many of these studies (including our own [18]) have examined the triggering of ATM- and ATR-mediated DDRs by both DNA viruses and retroviruses (as reviewed in [16]). Certain viruses (for example, Adenovirus) actively thwart these damage responses, while other viruses (like HIV) require a DDR to replicate to full capacity. These studies have been informative and have discovered specific viral interactions with key players in these repair pathways; however they have not assessed the ramifications of infection upon the cell's subsequent ability to repair further insult to its DNA.
A number of studies have analyzed a cell's repair capabilities following infection. These studies include the repair of exogenously introduced damage in the cellular DNA in the context of single viral protein expression [21]–[32] and the effects of a complete infection [33]–[40], [65]. These papers have examined the capacity of the cell's homologous recombination, base excision, nucleotide excision and non-homologous endjoining repair machinery to function, with the very large majority of the investigations finding decreased capacity of the cell to repair damage after viral protein expression (or full infection) commenced. Only four of these studies have reported evidence of an increase in repair capacity of the cell after infection or viral protein overexpression [21], [23], [36], [40]. Our results extend this analysis and separate the two genomes within an infected cell. We demonstrate that, at least in the context of HCMV-infected fibroblasts, there is increased repair of UV-induced CPDs in the viral DNA, without a corresponding increase in repair of the host DNA.
In the next few paragraphs we will focus on the above studies most pertinent to our own results, emphasizing studies examining interactions with the NER machinery and/or with HCMV infection's influence on cellular damage repair. Several studies have utilized expression of single viral proteins in the analysis of UV damage. Expression of the Hepatitis B X protein (HBX) in different cell types [22], [25], [27], [29], [32] or expression of the Epstein Barr virus proteins EBNA3C or LMP1 in transfected cells [26] decreased repair efficiency of UV-induced damage in transfected cells.
More pertinent to our study was that of Liang and colleagues [28], which used a herpesviral protein (γ herpesvirus 68 protein M2) and methodology similar to our own. Mouse 3T3 cells expressing M2 were assessed for the ability to repair exogenously induced UV damage. At low dosage (2.5 J/m2) M2-expressing cells' capacity to remove dimers was decreased, which was most pronounced at 24 hp irradiation. More dramatically, at 30 min post irradiation at very high dose (5000 J/m2) M2-expressing cells formed no comet tails, indicating they did not even initiate repair. Using a dimer-specific Ab they saw dramatically reduced dimer removal in the M2-expressing cells. Liang's results indicate that an M2-expressing cell had impaired ability to repair exogenous damage in host DNA via NER. We wonder if viral DNA would have been preferentially repaired if it had been present in these experiments?
An additional three studies have looked at NER repair in the context of full infection. Duong and colleagues [35] found reduced efficiency of Hepatitis C-infected cells to reactivate (and therefore repair) transfected UV-irradiated reporter plasmids (compared to uninfected control cells). Similarly, Philpott and Buehring found that multiple HTLV- and bovine leukemia virus-transformed lines (as well as cells transformed with just the HTLV Tax protein) had a decreased ability to repair a reporter construct damaged by UV [39]. Bowman and colleagues [65] looked at the removal of CPDs from host DNA during SV40 infection using dimer-specific Abs in slot blot analysis and found a decreased removal of these adducts. As in our studies, they utilized T4 assays to examine removal of damage from both the transcribed (transcription-coupled NER) and the non-transcribed (global genomic NER) strands of a cellular gene, DHFR. Interestingly, they found that repair of only the non-transcribed strand of DHFR was affected by SV40 infection, indicating that repression of p53 by SV40 might be involved (discussed further below). Once again, the question remains whether analysis of the SV40 DNA would have revealed increased and more rapid repair of the viral DNA in these cells.
The last set of papers that should be addressed deal specifically with repair in HCMV-infected cells. The literature has revealed varying effects of damage, depending upon the system being examined. Ranneberg-Nilsen and colleagues examined the capability of HCMV-infected human embryonic lung fibroblasts (infected under conditions similar to our study) to carry out BER [40], and found approximately twofold changes in repair, with different substrates being removed with greater or lesser efficiency. Studies from our own lab [36], using the same fibroblasts and HCMV isolate (Towne) as used in the current study, found that homology directed repair (HDR) was more efficient after infection, regardless of whether the reporter construct was integrated into the host cell genome or expressed transiently. Thus, neither BER nor HDR was affected as significantly as we have found NER to be.
Two additional works address the effects of HCMV infection on the introduction and frequency of DNA chromosome anomalies induced by subsequent exposure to genotoxic agents. The first [33] infected non-permissive peripheral blood lymphocytes (PBLs) with HCMV at low MOI in the presence of camptothecin and observed a synergistic increase in chromosome damage (including chromosome breaks), even in the absence of viral gene expression. These findings support our supposition that the repair of multiple forms of damage is inhibited in HCMV-infected cells. A separate study by Deng and coworkers [34] used freshly stimulated PBLs infected with HCMV at a higher MOI of 4. Their findings suggested that HCMV infection sensitized the chromosomes to drug-induced damage. Deng and coworkers' observation indicated that chromosome anomalies were present even without de novo viral gene expression in the non-permissive PBLs. This result is consistent with our earlier findings [12] that de novo viral protein expression was not required to induce site-specific chromosome damage. Our earlier results also indicated that certain virion-associated proteins cannot only induce damage, but may also interact specifically with the damage machinery to inhibit its operation.
These last studies have suggested experiments we intend to pursue in the future. First, does the same decrease in repair of cellular DNA occur if there is no replication of viral DNA within cells? This could be determined in non-permissively or semi-permissively infected cells by ascertaining whether a set of viral proteins and/or viral RC association of cellular proteins needs to occur for this effect to be observed. The results of the ganciclovir experiments shown in Figure 3 suggest that establishment of fully functioning replication centers may not be required for negative effects on cellular NER repair. Second, would the same decreases in repair capacity be seen in latently infected cells or cells with limited viral replication (such as long-term infected neurons [66])? Third, does the presence or absence of the p53 protein play a role in repair of different types of damage within infected cells? Certainly the reports of others [29], [32], [65] mentioned above indicate that, at least in the context of repair of UV-induced damage, interactions of the virus with p53 might influence global genomic repair within the cellular DNA. Our earlier studies have shown clear interactions with, and the importance of, p53 to HCMV replication [67]–[69], indicating p53 may play a role in the selective repair of viral over cellular DNA.
Our study is not the first to look at the capacity of an infected cell to repair exogenously introduced DNA damage. However, utilizing novel techniques, our experiments assessed initiation of repair, removal of CPDs and repair of the DNA substrate in both the cellular and viral DNA separately. Comet assays indicated that infected cells were fully capable of initiating repair, but still retained residual damage 24 hp irradiation. Confocal images of infected cells with separately labeled viral DNA (using BrdU pulse-labeling) showed definitive removal of CPD signal from viral DNA in the RCs but no statistically significant removal from the host genome. Importantly, this indicated the residual comet tail damage observed in the V+UV samples was due to persistence of CPDs in the host DNA and not in the viral genome. Additionally, development of a dual-color Southern methodology has allowed utilization of the well-established T4 assay to analyze two separate DNA genomes simultaneously. These dual-color T4 assays demonstrated faster and more significant repair of CPDs from the viral DNA than the host cellular DNA within the same cell. It is our belief that the compromised capability of infected cells to repair damage may ultimately be manifested in the induction of CNS defects in the HCMV-infected neonate. Future studies will extend this avenue of investigation.
Primary human foreskin fibroblasts (HFFs) (a gift from Steven Spector, UCSD) were isolated from tissue and propagated in Earle's minimal essential media (MEM) supplemented with 10% heat inactivated fetal bovine serum (FBS), L-glutamine (2 mM), penicillin (200 U/ml), streptomycin (200 µg/ml), and amphotericin B (1.5 µg/ml). Cells were grown in humidified incubators maintained at 37°C and 5% CO2.
G0 synchronized HFFs were trypsinized, counted, reseeded at a lower density and allowed to settle for approximately 2 h. Cells were infected at a multiplicity of infection (MOI) of 5 with the Towne strain of HCMV, obtained from ATCC (#VR 977). Two to four hpi, virus inoculum was removed and cells were refed with media and allowed to incubate as described below. The virus was propagated under standard procedures [70].
HFFs were mock- or virus-infected as described above. At 48 hpi, coverslips were harvested for colocalization of cellular NER proteins with the viral processivity factor, UL44. Coverslips were treated in one of two ways. In the first method, cells were extracted-first in a CSK buffer solution (10 mM Pipes, 100 mM NaCl, 300 mM sucrose, and 3 mM MgCl2) containing 0.5% Triton X-100 [41]. Cells were then rinsed in CSK twice and fixed with 3% formaldehyde in PBS (with 0.5 mM MgCl2, and 0.5 mM 3 mM CaCl2) for 10 min. In the alternate method, coverslips were extracted using standard formaldehyde fixation and Triton X-100 extraction as described previously [69]. See the Results section for further discussion of “fix first” versus “extract first” conditions and the information that can be gleaned from use of these different methods. Incubation of coverslips with Abs and mounting for examination were as described previously [69]. Nuclei were counterstained with Hoechst dye. The images of NER protein localization were obtained using a Nikon Eclipse E800 fluorescence microscope equipped with a Nikon DXM camera and Metavue software. Primary antibodies (Abs) used in Figure 1 and Table 1: mouse monoclonal Abs to XPB and XPD were kind gifts of Jean Marc Egly [71], [72]; mouse monoclonal Abs to XPA (2A4), XPG (8H7) and ERCC1 (3H11) and rabbit polyclonal Abs to XPC (RW028) and XPF (RA1) were kind gifts of Rick Wood [73]–[76]; mouse monoclonal Ab to UL44 (1202S - Rumbaugh Goodwin Institute); rabbit polyclonal Ab to CSB (Santa Cruz Biotechnology). Secondary Abs used in Figure 1A were donkey anti-rabbit TRITC-coupled Ab (Jackson Immunoresearch) and goat anti-mouse IgG1 alexafluor 488-coupled Ab (Molecular Probes). Secondary Ab used in Figure 1B was goat anti-mouse IgG FITC-coupled Ab (Jackson Immunoresearch).
HFFs were infected as described above. At 48 hpi, cells were washed in PBS and one set of mock and viral plates were irradiated in a Stratalinker 1800 at a dose of 50 J/m2. A second set of plates was left unirradiated. Irradiated cells were rinsed again, re-fed with media and allowed to recover for different periods of time (2, 6 and 24 hp irradiation). At the given timepoints, cells were washed once in cold PBS then scraped into cold PBS in microfuge tubes. Cell suspensions were adjusted to 1.5×105 cells/ml. 50 µl of suspension was added to 500 µl of low melting point agarose (1% in PBS) and 75 µl of this suspension was placed in a thin layer on a coated glass slide (Trevigen). The agarose was allowed to gel at 4°C for 15 min. Cells were then lysed for 30 min in situ in a high salt/detergent solution (2.5 M NaCl, 1% sodium lauryl sarcosinate, 1% Triton X-100) at room temperature. DNA was denatured by treatment in alkali solution (pH>13) for 40 min. Prepared slides were placed in an electrophoresis tank filled with the above alkali solution. Very low current (280–290 mA) was applied to the tank for 20 min. Slides were dehydrated in EtOH, stained with Sybr Green (which binds to both ss and ds DNA) and visualized/photographed using a Nikon E800 Eclipse microscope equipped with a Nikon DXM camera and Act One software. VisComet software was used to analyze 50–100 cells/sample set (except where noted) of mock (M alone), viral (V alone), mock+UV (M+UV) and viral+UV (V+UV) at the given timepoints post irradiation. Comets were analyzed for % DNA in the tail. Data in Figures 2 and 3 are represented as the average of % tail DNA for the given sample set. Error bars represent one SD from that average. Each experiment was performed twice, with the data from a representative experiment shown in the figures. Unpaired t-tests were performed to assess the statistical significance between sample sets using GraphPad statistical software as noted. To distinguish whether changes were occurring over the timecourse, the distribution of % tail DNA within each sample type (M alone, V, alone, M+UV, V+UV) at the three different timepoints was plotted. In this plot, the percentage of DNA in the tail for each comet analyzed in a sample set was assessed and assigned to one of four categories (<10% tail DNA, 11–25% tail DNA, 26–50% tail DNA or >50% tail DNA). The number of comets in each category was converted to a fraction of 100% and plotted.
Synchronized cells were reseeded into plates containing glass coverslips and infected as described above. At 48 hpi, cells were irradiated (or unirradiated) with 75 J/m2 UV and harvested at the indicated times post irradiation. Cells were also pulse-labeled with BrdU just prior to irradiation. BrdU labeling enabled viral RC visualization (as described previously [68]- 30 min pulse followed by 30 min chase in fresh media). After harvesting, cells were treated according to the methods in [77], which exposes both UV dimers and BrdU residues. Briefly, cells were fixed in ice cold MeOH: Acetic Acid (3∶1) for 20 min and subsequently washed in cold 100% EtOH. DNA was denatured for 3 min at room temperature (RT) using 70 mM NaOH dissolved in 70% EtOH. Finally, cells were washed extensively in PBS and stored at 4°C until staining.
Incubation of coverslips with Abs and mounting for examination were as described previously [69]. Cells were counterstained with Hoechst dye to visualize the nuclei. Mouse monoclonal Ab specific for CPDs has been described previously [56]. BrdU residues incorporated into viral DNA were stained with anti-BrdU rat monoclonal Ab (Harlan Sera-Lab). Cells stained for CPDs (detected with goat anti-mouse IgG2A Alexafluor 488 from Molecular Probes) and BrdU (detected with donkey anti-rat TRITC secondary Ab from Jackson Immunoresearch) were analyzed and photographed on an Olympus Fluoview 1000 confocal microscope using a 60× Plan Apo oil objective lens (1.42 NA). Care was taken to avoid the presence of saturated pixels within the images. Samples were excited using 405 nm (for BrdU), 488 nm (for CPD) and 561 nm (for Hoechst) laser lines. Images showing unirradiated samples stained for CPDs and BrdU were captured using the Nikon E800 Eclipse and Metavue software mentioned above.
In parallel, coverslips were harvested using “fix first” conditions as described above. These coverslips were stained with a polyclonal rabbit Ab to histone H3 (Millipore #06-755 detected using donkey anti-rabbit TRITC-coupled secondary Ab (Jackson Immunoresearch). They were also stained with the above-mentioned Ab to UL44 (detected using goat anti-mouse IgG1 AlexaFluor 488 (Molecular Probes)) to localize the RCs. These coverslips were blocked in 30% human IgG (instead of FBS) to inhibit non-specific binding of the rabbit Ab to the viral assembly complex within the cytoplasm of infected cells.
Image preparation and data generation were performed using MetaMorph (MM) Software (Universal Imaging). Stacked confocal images were captured as TIFF images on an Olympus Fluoview 1000 using 0.41 µm stepping. Twenty to thirty cells were analyzed per experiment, per time point as described below. Three separate experiments were analyzed. Using MM software, the center plane of each cell was identified from the stack of confocal images. The center plane was defined as the largest cross-sectional area of the virus RC. The image containing the center plane for each cell was color separated. The red (BrdU) and green (CPD) channels were saved as new images. This was performed for each individual cell, including all cells from images containing multiple cells. Thresholding of each color-separated image was used to define contiguous regions (in MM defined as Object(s)) for each nucleus and RC (many cells contained multiple RCs). Regions of Interest (ROIs) were created/saved surrounding these Objects (using the MM create ROI around Objects function). The CPD Integrated Intensity (INTINT) for each entire nucleus ROI was recorded. The ROI(s) of each RC(s) was mapped onto its corresponding CPD nucleus image. The associated CPD INTINT of each RC(s) region was recorded. A total RC CPD INTINT for cells containing multiples RCs was summed from that cell's multiple RC CPD INTINTs. The CPD total for the host cellular DNA was defined as all CPDs outside of the RC(s) (e.g. entire Nucleus CPD (-) Virus RC CPD = Host CPD). This data was analyzed using a mixed-effects ANOVA model (SAS, Cary, NC) comparing total CPD INTINTs between the 0 and 24 h post irradiation time points as described in the text.
Cells were treated as described above for comet analysis (irradiation of 50 J/m2 at 48 hpi). However, sample sets (M alone, M+UV, V alone, V+UV) were performed in duplicate. One of the sets continuously received 45 µM ganciclovir (after 24 hpi to inhibit viral DNA replication) and the second set a vehicle control. Cells were harvested at 24 h post irradiation for comet analysis as described above. For these experiments, 25–50 comets were scored per sample set. The experiment was repeated twice, and a representative sample set is shown in Figure 3.
Two plates of HFFs on glass coverslips were infected at an MOI of 5. After 48 h, one coverslip from each plate was removed and pulse-labeled with BrdU for 30 min and then chased for an additional 30 minutes in fresh media. These two coverslips served as time +0 h for the irradiated and unirradiated plates, respectively. After the chase period, one of the BrdU-labeled coverslips (and the remaining coverslips from its plate of origin) was irradiated at 75 J/m2. The other BrdU-labeled coverslip (and its partners) were left unirradiated. Time +0 h coverslips were then harvested. One h before each subsequent timepoint (at 5,10 and 23 hpi, respectively), an additional coverslip from each plate was removed and pulse-labeled with BrdU in preparation for harvesting at the appropriate timepoints (6, 11 and 24 hp irradiation). Cells were fixed and stained for BrdU incorporation into viral RCs as described previously [68]. Images were captured using the Nikon E800 Eclipse and Metavue software mentioned above.
HFFs were infected on coverslips as described above. After 47 hpi, one h prior to irradiation at 75 J/m2, infected cells were pulse-labeled with BrdU (and then chased for 30 min as described above) to label viral DNA within the RCs. Half of the coverslips were then irradiated with 75 J/m2; the second half was not irradiated. Timepoints were taken at 0 and 24 h post irradiation (or control treatment). Coverslips were fixed and processed for BrdU localization as described previously [68]. Cells were analyzed on the Olympus confocal microscope described above. Each Z-series was subsequently projected using the Olympus FSW software option of ‘Duplicate as displayed’ to create a single plane, 8-bit image for Figure 6.
Viral supernatants were centrifuged through a 25% sucrose (in PBS) cushion at 23, 000 rpm for 70 min at 10°C to pellet viral particles. Genomic DNA was extracted from HFF cells and viral particles as described previously [78]. HindIII-digested viral and HFF DNA were labeled with biotin-16-dUTP (Roche) and digoxigenin-11-dUTP (Roche), respectively, using the BioPrime Array CGH genomic labeling module (Invitrogen).
The Li-Cor Odyssey Southern protocol was modified as follows. DNA was separated on a 1% native agarose gel. The DNA was depurinated for 15 minutes in 0.25 N HCl then denatured in 0.5 M NaOH and 1.5 M NaCl prior to transferring by capillary action onto 0.45 µm Magnacharge nylon membrane (GE water and process technologies) in 20× SSC (pH 7.0). After UV crosslinking, the membrane was prehybridized in a solution containing 5× SSPE, 2% SDS, 10% dextran sulfate, 1× Denhardt's solution and 10 µg/ml sheared, denatured salmon sperm DNA for 2–4 hours at 65°C. Labeled probes were boiled for 5 min and then rapidly chilled on ice for 10 min before addition to the prehybridization buffer and hybridization for 16 h at 65°C. The membrane was washed twice for 5 min in 2× SSPE at RT, twice for 15 min in 2× SSPE with 1% SDS at 60°C, and twice for 15 min in 0.1× SSPE at 60°C. The blot was blocked in 0.6% cold water fish skin gelatin (Sigma) in TBS with 0.5% Tween-20 (TBST) and 1% SDS for 1 h at RT. Anti-digoxigenin Ab (Sigma) was diluted 1∶1000 in 0.6% cold water fish skin gelatin in TBST and the blot was probed for 1 h at RT. The blot was washed at RT for 5 min in TBST, 10 min in TBST with 1% SDS, and three times with TBST for 5 min. Anti-mouse IRdye700 and streptavidin IRdye800 (Rockland) were diluted 1∶4,000 and 1∶20,000, respectively in 0.6% cold water fish skin gelatin in TBST with 0.02% SDS and the blot was incubated 45 min in the dark at RT. The blot was washed at RT for 5 min in TBST, 15 min in TBST with 1% SDS, three times with TBST for 5 min, and twice in TBS for 5 min. The blots were scanned using a Li-Cor Odyssey infrared imager (Li-Cor Bioscience).
HFFs were infected and irradiated at 50 J/m2 as described above. Cells were harvested at 0, 6, 24, and 48 h post irradiation. DNA was extracted as described above. 150 ng of DNA was digested with T4- or mock-digested and the digestions were loaded on a 1% alkaline agarose gel and separated at 25 V for 18 h as described previously [59]. The gel was neutralized for 45 minutes in 0.5 M Tris HCl pH 7.5 and 1.5 M NaCl prior to depurination, denaturation and capillary transfer as described above. Analysis of T4 Southerns for CPD removal was performed as described previously [59]. SYBR Gold stained gels were performed in the same fashion with the following exceptions: one microgram of DNA was loaded in each lane and gels were stained with SYBR Gold after neutralization. For statistical analysis, one-tailed paired t-tests were performed for each time point comparing the mean repair of the host DNA versus the viral DNA (or versus mock DNA) as described in the text.
|
10.1371/journal.pgen.1002348 | Foxn1 Regulates Lineage Progression in Cortical and Medullary Thymic Epithelial Cells But Is Dispensable for Medullary Sublineage Divergence | The forkhead transcription factor Foxn1 is indispensable for thymus development, but the mechanisms by which it mediates thymic epithelial cell (TEC) development are poorly understood. To examine the cellular and molecular basis of Foxn1 function, we generated a novel and revertible hypomorphic allele of Foxn1. By varying levels of its expression, we identified a number of features of the Foxn1 system. Here we show that Foxn1 is a powerful regulator of TEC differentiation that is required at multiple intermediate stages of TE lineage development in the fetal and adult thymus. We find no evidence for a role for Foxn1 in TEC fate-choice. Rather, we show it is required for stable entry into both the cortical and medullary TEC differentiation programmes and subsequently is needed at increasing dosage for progression through successive differentiation states in both cortical and medullary TEC. We further demonstrate regulation by Foxn1 of a suite of genes with diverse roles in thymus development and/or function, suggesting it acts as a master regulator of the core thymic epithelial programme rather than regulating a particular aspect of TEC biology. Overall, our data establish a genetics-based model of cellular hierarchies in the TE lineage and provide mechanistic insight relating titration of a single transcription factor to control of lineage progression. Our novel revertible hypomorph system may be similarly applied to analyzing other regulators of development.
| The thymus is the specialized organ responsible for generating T cells, which are required to regulate and effect immune responses. The unique functions of the thymus are mediated by a diverse array of specialized epithelial cells found only within this organ. These specialized, functionally mature thymic epithelial cells are generated from immature epithelial progenitor cells present in the fetal and adult thymus through a highly regulated process, termed differentiation, that is tightly controlled by specific genes. Foxn1, a protein that is expressed in thymic epithelial cells, is a transcription factor—a protein that regulates how other genes are expressed. Here, we have investigated the role of Foxn1 in generating mature thymic epithelial cells from immature progenitors. We find that Foxn1 is required throughout this process, from the onset of differentiation in progenitor thymic epithelial cells in the developing fetus to the final differentiation steps through which thymic epithelial cells mature to acquire their full functionality. We further find that Foxn1 controls the expression of a variety of genes with different functions in thymic epithelial cells. Overall, our study defines the role of Foxn1 in thymus development at the cellular level and provides insight into how it mediates these functions.
| T cell development occurs in the thymus and depends on progressive interactions with the thymic stroma. Critical to this function is the thymic epithelium (TE), which comprises a diverse array of phenotypically and functionally distinct cell types broadly organized into cortical and medullary regions [1].
During development the thymus arises from the endoderm of the third pharyngeal pouches [2] which, from day 9 of embryonic development in the mouse (E9.0), contain cells specified to the TE lineage [2]. Until E11.5, this region uniformly expresses the cell surface determinant Plet-1 [3]-[5] and the earliest currently identified founder cells for the lineage are thus Plet-1+ third pharyngeal pouch cells. By E12.5 the thymic primordia, now colonized by haematopoietic progenitor cells and surrounded by a mesenchymal capsule, have separated from the pharyngeal endoderm and contain a predominant EpCam+Plet1+ epithelial population [3], [6], [7]. Recent evidence supports the existence within this population of a common thymic epithelial progenitor cell (TEPC) capable of generating both cortical and medullary TEC (cTEC and mTEC respectively) [7], [8]. Further studies indicate that a medullary sub-lineage specific progenitor cell is present at least as early as E13.5 [9], [10], and suggest the existence of a cortical sub-lineage specific progenitor [11]. However, the extent to which these progenitor activities persist in late organogenesis and the postnatal organ has not been determined and furthermore, the timing of emergence of the sub-lineage progenitors during organogenesis remains unclear.
The forkhead transcription factor Foxn1 [12], the gene mutated in the classical mouse mutant nude (nu) [13], is a pivotal regulator of TE lineage development. Nude mice are hairless and athymic [14], [15]. In Foxn1 null animals, the earliest stages of thymus development appear to occur normally, but development is arrested after initial formation of the organ primordium (around E12.0 in the mouse) and the primordium is not colonized by hematopoietic precursors [16]. Nonetheless, the Foxn1 null thymic primordium separates from the parathyroid component of the common primordium and migrates to the mid-line, where a small, alymphoid cystic thymic remnant persists postnatally [14]. Null alleles of Foxn1 in mouse, rat and human all exhibit essentially the same pleiotropic phenotype [17].
Forkhead family members play important roles in specification and progression of a number of cell lineages, and recent evidence reveals roles for some Forkhead proteins in chromatin modification/epigenetic regulation of cell fate [18]. Analysis of postnatal nude:wild-type chimeric mice demonstrated the cell autonomous requirement for Foxn1 in development of all major TEC sub-lineages and suggested that TE lineage cells unable to express functional Foxn1 might undergo maturational arrest at a common TEPC stage [19]. A revertible Foxn1 null model, Foxn1SA2/SA2 provided strong support for this hypothesis, as neonatal clonal reversion of this allele resulted in the generation of small units of functional thymus tissue containing both cortical and medullary compartments [8]. Further support for a role in differentiation comes from studies on keratinocytes, which implicate Foxn1 in regulating initiation of terminal differentiation [20], [21], and from analysis of a hypomorphic Foxn1 allele which generates a transcript lacking exon 3 and thus the N-terminal domain of Foxn1 [22]. In mice homozygous for this allele (Foxn1Δ), the postnatal Foxn1Δ/Δ thymus was highly cystic, contained no discernable cortical or medullary regions and could sustain only highly impaired thymocyte differentiation, suggesting that Foxn1 is actively required for TEC differentiation at stages beyond initiation of the TEC programme [22]. Evidence also supports roles for Foxn1 in TEC proliferation [23] and in regulating the balance between proliferation and differentiation in skin [24]. In addition, a requirement for Foxn1 for maintenance of the postnatal thymic microenvironment has recently been demonstrated [25]–[27], with evidence pointing to differential sensitivity of different TEC subsets to changes in Foxn1 dosage [26]. Collectively, these studies suggest that Foxn1 plays a complex role in regulating TEC lineage development. However, precisely how Foxn1 regulates the transit from the earliest fetal thymic epithelial progenitor cell to the fully functional postnatal thymic epithelium, and at which stages in this process it is required, remains undetermined. In addition, the molecular mechanisms regulated by this transcription factor in the thymus have not yet been addressed.
In this study, we have addressed the functions of Foxn1 throughout thymus ontogeny, via generation and analysis of a novel revertible hypomorphic allele of Foxn1, Foxn1R, which expresses only low levels of Foxn1 mRNA and protein. A particular advantage of our system is the revertible nature of the allele, which affords the capacity to test the relationship of cell states identified via analysis of mutant mice to states occurring in normal ontogeny. Our studies establish Foxn1 as a powerful regulator of differentiation in both the cTEC and mTEC sub-lineages. We find no evidence for a role for Foxn1 in regulating cell fate choice in the cortical or medullary TEC sub-lineages. Rather, we find that Foxn1 is required for progression of differentiation at multiple stages in cTEC and mTEC sub-lineage development in both the fetal and postnatal thymus, and show that different Foxn1 dosage is required to execute its function(s) at different differentiation stages. We further establish that Foxn1 regulates, either directly or indirectly, a suite of genes known to effect TEC function - including Dll4, CCL25, Cathepsin L, CD40, Pax1, and MHC Class II, and that these exhibit different response patterns to changes in Foxn1 dosage. Collectively, these findings significantly advance understanding of the role of Foxn1 in TEC, demonstrating for the first time its direct involvement in generating multiple TEC sub-types in both the cortical and medullary lineages in the fetal and postnatal thymus, and suggesting it acts effectively as a master regulator of the entire TEC programme.
We set out to generate a conditionally revertible null allele of Foxn1 as a tool for generating TEPC lines. Our rationale was that lack of Foxn1 expression would impose an early block on TEC lineage differentiation, effectively trapping TEC in an undifferentiated progenitor cell state, while reversion of the allele would remove this block and allow progression to terminal differentiation. Since the extent to which Foxn1 is required for TEC proliferation is unknown, SV40 T antigen was used to uncouple potential roles of Foxn1 in proliferation and differentiation. We thus generated the revertible Foxn1 allele, Foxn1R, by inserting a LoxP flanked cassette into intron 1b of the Foxn1 locus by homologous recombination in ES cells (Figure 1A, Figure S1) and used this ES line to generate the Foxn1R mouse strain. Initial characterization of postnatal Foxn1R/+ mice revealed thymus hyperplasia as expected, due to expression of SV40 Tag under the Foxn1 promoter. However, Foxn1R/R mice developed severely hypoplastic thymi rather than exhibiting the expected phenotype of complete thymic aplasia (Figure 1B). In keeping with this, immunoblotting for Foxn1 revealed low-level Foxn1 protein in Foxn1R/R thymi compared to wild-type (WT; Figure 1C) and RT-PCR analyses demonstrated that the transcripts produced from the Foxn1R allele contained either Exon1a-SV40Tag-IRES-eGFPneo elements or the full-length Foxn1 mRNA (Figure 1D). Thus some residual Foxn1 expression from Foxn1R occurred, and resulted from splicing around the targeted insertion subsequent to transcription proceeding beyond the transcriptional pause. No overt skin phenotypes were apparent in either Foxn1R/+ or Foxn1R/R mice. Collectively, these data establish that Foxn1R is a hypomorphic allele that expresses a low level of Foxn1 compared to WT and that expression of SV40Tag does not compensate for decreased levels of Foxn1.
The revertibility of the allele was demonstrated by crossing Foxn1R/+ mice with the ZP3Cre deletor strain, in which Cre recombinase is expressed in oocytes [28]. Foxn1R/+ZP3Cre/+ mice excised the LoxP flanked cassette in most if not all cells (Figure 1E), and showed normal thymus development and function.
The Foxn1 null phenotype of severe athymia results from a complete early block in thymus development and has limited understanding of the role of Foxn1 later in ontogeny. We reasoned that the Foxn1R allele might be informative in this light. However, it was first necessary to test whether expression of SV40Tag under the Foxn1 promoter affected TEC differentiation or function. Therefore, we characterized the Foxn1R/+ thymus phenotype, in particular seeking evidence for impaired or perturbed TEC development and/or function. The two thymic lobes in Foxn1R/+ mice exhibited a greater than ten-fold increase in both TEC and thymocyte numbers compared to WT littermates (Figure 1), but were otherwise histologically normal (Figure 1B, 1F). In both fetal and postnatal Foxn1R/+ thymi, TEC and thymocyte subset distributions were indistinguishable from Foxn1nu/+ and WT littermates and no evidence of malignant transformation could be found (Figure 1G-1I). We therefore concluded that the thymic hyperplasia observed resulted from a uniform expansion of the TEC compartment due to expression of SV40Tag under control of the Foxn1 promoter, but that this did not affect TEC differentiation or function in either the fetal or postnatal thymus. This is consistent with the action of SV40TAg in a variety of other tissues [29]–[31] and its known function in regulating cell proliferation. Of note is that the Foxn1 promoter drives relatively low levels of gene expression in TEC, based on QRT-PCR analyses relative to a range of housekeeping genes. This demonstrated that Foxn1 is expressed in the same range as alpha-tubulin, which is recognized to be expressed at low levels [32] (Figure S2). Furthermore, SV40Tag was not detectable by immunohistochemical analysis of thymus sections from Foxn1R mutant mice (Figure S2).
Next, we tested the effect of reduced Foxn1 dosage on postnatal TEC by analyzing the Foxn1R/R phenotype. Although small, Foxn1R/R thymi were grossly histologically normal, containing clear cortical and medullary regions (Figure 2A, 2B). Postnatal Foxn1R/R thymi contained all normal thymocyte subsets (Figure 2C, 2D). However, the CD4+8+ double positive (DP) subset was proportionally larger than in WT littermates while the CD4+ and CD8+ single positive (SP) subsets were proportionally reduced (DP proportions: Foxn1R/R, 91.2±0.85%; WT 82.9±0.21%, p = 0.0004) suggesting that the DP to SP transition and/or subsequent thymocyte maturation was perturbed in Foxn1R/R mice. The CD4-8- (Double negative, DN) cell subset distribution was also altered in postnatal Foxn1R/R thymi, with increased proportions of CD44+CD25- (DN1) and CD44-CD25+ (DN3) cells (Figure 2D). Most DN1 cells in postnatal Foxn1R/R thymi were B cells, as shown by B220 staining (Figure 2E), suggesting impaired commitment of hematopoietic progenitors to the T cell lineage [33], [34]. Furthermore, thymocyte development was delayed in Foxn1R/R mice early in ontogeny (Figure 2F). Thus, rather than simply resulting from reduced numbers of functionally normal TECs, the hypoplastic Foxn1R/R phenotype was characterized by impaired TEC functionality.
Interrogation of TEC subset distribution and marker expression in the Foxn1R/R thymus revealed apparently normal morphologies for medullary and cortical TEC (mTEC and cTEC respectively), and normal expression patterns for cytokeratin (K) 5, K8 and CD205 (Figure 2B, Figure 3A and 3Ai), indicating the presence of cortical and medullary compartments. Further analysis however revealed abnormal gene expression patterns in both compartments.
The cTEC marker CDR1 is acquired only postnatally in cTEC and thus represents a terminal differentiation marker for this TEC sub-lineage. Ly51 is another commonly used marker that is cTEC-specific within the thymus. Expression of both of these markers was severely down-regulated in Foxn1R/R thymi (Figure 3B-3Ci).
MHC class II, Cathepsin L and Dll4 are TEC markers with known functions in thymocyte development [11], [33]–[37]. MHC Class II staining on Foxn1R/R cTEC was dramatically reduced compared to Foxn1R/+ and WT littermates (Figure 3D and 3Di), and similarly, Cathepsin L expression was down-regulated in Foxn1R/R cTEC (Figure S3). Dll4 expression was dramatically reduced on both fetal and postnatal Foxn1R/R TEC compared to aged-matched WT controls (Figure 3E and 3Ei, see also Figure S3). These data provide a possible explanation for the altered thymocyte subset distributions present in Foxn1R/R mice.
CD40 is expressed on TEC and thymic dendritic cells and is required for medullary development [38], [39]. Expression of CD40 was also substantially down-regulated on postnatal Foxn1R/R cTEC but was largely unaffected on mTEC (Figure 3F). Expression of the recently identified cTEC-specific thymic proteasome subunit ß5t [40] was affected only slightly if at all (Figure 3G and 3Gi; proportion of TEC that are ß5t+: Foxn1R/R, twenty four percent; WT, twenty four percent).
Aire is a key regulator of thymus function, and within the thymus is specifically expressed in a subset of mTEC. Significantly fewer Aire+ mTEC were present in Foxn1R/R compared to WT thymi (Figure 3H and 3Hi).
Collectively, these data established that differentiation of both cortical and medullary TEC was impaired in the postnatal thymus of Foxn1R/R mutants. In particular, since cTEC morphology appeared normal but several cTEC-specific markers were severely down-regulated in Foxn1R/R thymi, they strongly suggested the hypothesis that terminal differentiation was blocked in the cTEC sub-lineage.
To test whether this hypothesis was correct, or alternatively whether the postnatal Foxn1R/R phenotype reflected the absence of a normal cTEC population and/or overgrowth of a normally minor TEC population, we bred a constitutively expressed tamoxifen inducible Cre recombinase allele, ROSA26CreERt2 [41], onto the Foxn1R/R background. We then tested the outcome of reverting the Foxn1R allele via Cre-mediated excision of the LoxP-flanked cassette, such that WT levels of Foxn1 mRNA expression were restored. Thus, adult Foxn1R/R; ROSA26CreERt2 mice were injected with 4-hydroxy tamoxifen (4OHT) to induce Cre recombinase activity. Two days after 4OHT injection, Foxn1 mRNA expression was restored to at least WT levels (Figure S3) and Ly51, CDR1 and MHC Class II staining was observed in a high proportion of cTEC (Figure 3I, see also Figure S3). Dll4 mRNA expression was also up-regulated by two days post-injection, as was expression of both CCL25 and Cathepsin L mRNA (Figure S3). Since the half-life of cTEC has been established as 10-14 days [42] these data cannot be explained by proliferation of a previously inhibited cell type followed by differentiation, but rather must reflect phenotypic conversion of existing cells. They are therefore consistent with release of a block in terminal differentiation. The proportions of Aire+ mTEC were also restored to WT levels in this experiment (Figure 3I, 3J).
Collectively, these data reveal two important and previously unreported roles of Foxn1. First, although the requirement for Foxn1 for postnatal TEC maintenance has recently been demonstrated [26], this is the first report of a role for Foxn1 in regulating the differentiation of postnatal TEC. Our data establish the requirement for this transcription factor for terminal differentiation of both cTEC and Aire+ mTEC in the postnatal thymus. Since the Foxn1R/R phenotype is most severe in the cortical compartment, they further suggest differential requirements for Foxn1 in the postnatal cortical and medullary TEC compartments. Consistent with this idea, QRT-PCR analysis indicated higher levels of Foxn1 mRNA in cortical than medullary TEC in the normal postnatal thymus (Figure 3K). Second, they also establish that Foxn1 regulates TEC-specific expression of MHC Class II, Dll4, CD40 and Cathepsin L, which are all critically required in TEC for execution of the T cell development programme. While this regulation may be either direct or indirect, it appears to be lymphocyte-independent, as all thymocyte subsets are present in Foxn1R/R thymi.
We next crossed Foxn1R/R mice with mice carrying the null allele, Foxn1LacZ [13]. Foxn1LacZ phenocopies the original nude mutation and is referred to as Foxn1- herein. Postnatal Foxn1R/- mice had small, alymphoid thymic rudiments that were devoid of any cortical or medullary organisation, and lacked peripheral T cells (Figure 4A); no hematopoietic colonization of the Foxn1R/- thymus was seen at any stage analyzed (Figure S4). Similar to Foxn1nu/nu TEC [19] postnatal Foxn1R/- TEC were predominantly Plet-1+ and co-expressed Keratin 5 and Keratin 8, with occasional cells expressing the cortical and medullary TEC markers CDR1 and UEA1 respectively (Figure 4B). However, unlike the hairless nu/nu or Foxn1-/- mice, Foxn1R/- mice produced an overtly normal coat of hair up to six weeks of age, after which they showed sporadic hair loss. Reactivation of Foxn1 by tamoxifen-mediated induction of Cre recombinase activity in postnatal Foxn1R/-; ROSA26CreERt2 mice up to at least four months old resulted in development of an organized, functional thymus containing all major cortical and medullary TEC sub-types (Figure 4C), indicating that thymic epithelial progenitor cells can persist for prolonged periods in vivo in the absence of wild-type levels of Foxn1, and demonstrating that expression of SV40Tag in these cells did not affect their differentiative capacity.
The differences observed between Foxn1R/- and Foxn1-/- mice suggested that further functions of Foxn1 might be revealed by analysis of an allelic series. We therefore established a series of six strains comprising all possible combinations of Foxn1-, Foxn1R and WT alleles. Precise determination of Foxn1 mRNA expression levels across this series relative to wild type was achieved by analysis of a defined TEC population, E12.5 Plet-1+ TEC (Figure 4D). In E12.5 Plet-1+ TEC, the Foxn1R allele expressed approximately twenty percent of WT Foxn1 mRNA levels (Figure 4D).
The postnatal Foxn1 null phenotype is characterized by the presence of linear aggregates of cells that express markers associated with the earliest thymus-restricted epithelial progenitor cells found in ontogeny [19], suggesting that, in the absence of Foxn1, TE lineage progression might be blocked at the earliest progenitor stage [19], [13]. This idea was supported by the recent demonstration of a persisting common TEPC in postnatal Foxn1-/- mice [8]. However, the possibility remains that more restricted precursors arise in the absence of Foxn1 but are not maintained later in ontogeny.
To determine the stages in early thymus ontogeny at which Foxn1 is required, we analyzed all strains of the allelic series (Figure 5) using marker combinations that identify early cortex (CD205 and K5lo/-) and early medulla (Claudin [Cldn] 4, UEA1, K5hi and MTS10) [10], [11], [43], [44]. Plet-1 [5] was also included as an indicator of founder/early progenitor cell status. We focussed on the E13.5 and E15.5 time points since at these stages the TE is undergoing active patterning into cortical and medullary compartments [45]. Notably, functional analyses have shown that the Cldn4hi cells present at E13.5 in the WT thymic primordium are medullary TE lineage-restricted precursors [10].
By E13.5 WT, Foxn1+/- and Foxn1R/+ thymi contained central K5hiCldn4hi clusters surrounded by K5lo/-K8+Cldn4lo/- regions (Figure 5A), as previously reported for WT thymi [3], [10], [45]; this central Cldn4hi region has previously been demonstrated to correspond to prospective medulla [10]. These data thus confirmed the normal progression of TEC lineage development in Foxn1R/+ mice.
K5hiCldn4hi and K5-Cldn4lo/- regions were also present in the thymic primordia of Foxn1-/- mice at both E13.5 and E15.5 (Figure 5A; Figure S5). These data establish that divergence of the mTEC lineage from the common TEPC occurs in the absence of Foxn1 and is thus Foxn1-independent, in contrast to the existing model of TE lineage development.
cTEC and mTEC differentiation was further probed using markers expressed at various intermediate stages of TEC development. Consistent with recent reports, neither CD205 [11] nor ß5t [40], [46] were detected in Foxn1-/- mice at any stage analyzed, confirming the dependence of the cortical sub-lineage on Foxn1 at least from the onset of expression of these markers. However, a CD205+ population was detectable by E15.5 in all other strains tested, including Foxn1R/- mice, demonstrating that only low levels of Foxn1 mRNA expression are required to promote the early stages of cTEC development (Figure 5B).
By E15.5, UEA1+ and MTS10+ mTEC were present in all strains in the allelic series that express ≥50% WT levels of Foxn1 mRNA (Foxn1+/-, Foxn1R/+ and WT thymi). However, while strains that expressed ≤20% WT Foxn1 mRNA all contained some UEA1+ TEC, MTS10 expression was not detected (shown for Foxn1-/-, Foxn1R/R and Foxn1+/-, Figure 5C). We note that UEA1 staining was present only on rare cells in the Foxn1-/-and Foxn1R/- thymic primordium, but that this staining was consistent and above background. Since MTS10 staining was detected as normal on postnatal mTEC in Foxn1R/R mutants, but was only observed on very rare cells in postnatal Foxn1R/- and Foxn1-/- thymi, we conclude that mTEC differentiation beyond the earliest mTEC sub-lineage progenitor state is blocked in the absence, and delayed at intermediate levels, of Foxn1 expression.
Collectively, the data described above indicate that Foxn1 is not required for divergence of the mTEC sub-lineage from the common TEPC but is required subsequently, in a dosage dependent manner, to promote TEC differentiation. Of note is that high level Plet-1 expression is maintained in Foxn1-/- TECs, supporting the conclusion that these cells remain blocked in the earliest differentiation state despite having undergone lineage divergence (Figure S5).
The above analysis suggested that specific intermediate states of TEC development are dependent on discrete levels of Foxn1 expression. To further test this notion, we used flow cytometry to provide quantitative evaluation of TEC differentiation status in the different allelic variants at defined time-points. For this we analyzed representative markers of the founder TE cell-state (Plet-1 [3], [5]) and functional maturation (MHC Class II [47]); TEC were positively identified using EpCam in all analyses.
In order to identify intermediate phenotypes that occur during normal TEC lineage progression we initially analyzed expression of these markers in WT mice at E12.5, E13.5 and E15.5 (Figure 5D). This demonstrated that normal TEC lineage progression is characterized by the down-regulation of Plet-1 coupled with acquisition of MHC Class II expression.
We then analyzed the allelic series. This revealed that TEC differentiation was delayed when Foxn1 expression was reduced, and that the severity of the delay was inversely proportional to Foxn1 mRNA level. In Foxn1R/R mice the TEC subset profiles at E13.5 and E15.5 were equivalent to those of WT mice at E12.5 and E13.5 respectively (Figure 5E, 5F). Thus, normal TEC differentiation occurred in the Foxn1R/R strain, but with delayed kinetics.
The two functionally athymic strains (Foxn1R/- and Foxn1-/-) each exhibited TEC subset profiles consistent with a severe early block in TEC differentiation. Unexpectedly, the severity of this block differed between these two strains. In Foxn1-/- thymi, the ratio of Plet-1+ to Plet-1lo/- TEC was maintained between E13.5 and E15.5 and was similar to that observed at E12.5 during normal development (determined based on comparison of the mean percentage of Plet-1+ and Plet-1lo/- cells as shown in Figure 5E, 5F legend). However, in Foxn1R/- thymi, the proportion of Plet-1lo/- TEC increased from twenty six percent at E13.5, comparable to E13.5 Foxn1-/- thymi, to forty one percent at E15.5 (see mean percentages provided in Figure 5E, 5F legend). This suggested that by E15.5, at least some Foxn1R/- TEC had been able to exit from the earliest progenitor state and enter the TEC differentiation programme, but that Foxn1-/- TEC could not make this transition. This conclusion was supported by analysis of the differentiation marker MHC Class II, which was expressed by a small population of Plet-1+ and Plet-1lo/- cells in Foxn1R/- mice at E15.5, but was virtually undetectable in Foxn1-/- mice (Figure 5F). Importantly, induction of MHC class II expression occurred in the absence of colonizing haematopoietic cells (Figure S4). Thus, while TECs in the Foxn1-/- thymus remained in an undifferentiated state apparently equivalent to the earliest progenitor stages of each TEC sub-lineage, Foxn1R/- TECs could initiate the very early events associated with TEC lineage differentiation but could not progress beyond this initial differentiation step.
Collectively, the above data establish that Foxn1 is required for progression through successive stages of both cTEC and mTEC differentiation. To test whether Foxn1 might also regulate TEC proliferation and/or survival, we also quantified both founder/early progenitor cells of the lineage (Plet-1+MHC Class II– TEC) and cells that were further into the TE differentiation programme (Plet-1- TEC) for each allelic variant.
At both E13.5 and E15.5, the absolute numbers of Plet-1+ cells correlated with Foxn1 dosage (Figure 6A), consistent with a role for Foxn1 in regulating proliferation of Plet-1+ TEC and thus potentially overall thymus size [48]. However, the most marked effect was on the numbers of Plet-1- TEC present at different levels of Foxn1. At intermediate and low Foxn1 levels (i.e. ≤20% of WT mRNA levels; Foxn1R/R, Foxn1R/- and Foxn1-/- mice) the numbers of Plet-1- TEC were decreased by four-fold compared to Foxn1+/- thymi at both E13.5 and E15.5. Notably, Foxn1 dosage was more important in determining Plet-1+ and Plet-1- cell numbers than the presence or absence of SV40Tag (Figure 6A). These data establish that a threshold level of Foxn1 mRNA between that found in Foxn1R/R and Foxn1+/- TEC is required for generation/expansion of the Plet-1- TEC population with normal kinetics, indicating a role for Foxn1 in regulating thymus size.
The above data could be consistent with roles for Foxn1 in the production, proliferation and/or maintenance of Plet-1- TEC. To begin to discriminate between these possibilities we performed cell cycle analysis on TEC in all of the allelic variants at E15.5. Both Foxn1R/- and Foxn1-/- TEC exhibited slower cell cycle times than WT TEC (Figure 6B; note the proportions of cells in the G1/G0 and S/G2/M fractions). However, no differences in cell cycle were observed between E15.5 WT, Foxn1R/+ and Foxn1R/R TEC (Figure 6B: note the proportions of cells in G0/G1, S and G2/M), establishing that the reduced Plet-1- cell numbers in fetal Foxn1R/R thymi could not be explained by differences in proliferation rate. We therefore conclude that in the fetal thymus, Foxn1 regulates allocation of cells into the differentiating Plet-1- TEC compartment and/or the maintenance of Plet-1- TEC. At present, we cannot distinguish between direct and indirect effects of Foxn1 in regulating fetal Plet-1- cell numbers, as we also observed a significant delay in maturation of the thymic mesenchyme and immigration of mesenchymal cells into the thymic epithelium in Foxn1R/R mice (Figure S6), raising the possibility that the deficit in Plet-1- TEC in these mutants at E15.5 might reflect impaired mesenchymal-epithelial cross talk.
In order to relate our findings to molecular control of TEC differentiation and function, we investigated the response of a suite of genes with diverse functions in thymus biology – i.e. that are known to regulate particular aspects of thymus development or are required in TEC to support T cell development - to changes in Foxn1 dosage. In this panel, we included genes known to be down-regulated or absent from the Foxn1 null thymic primordium or identified in our analysis as being regulated by Foxn1 (Dll4 [34] and CCL25 [49]), or with defined roles in TEC development (Pax 1 [50]-[52]). We also included Fgf2, to date the only gene verified by chromatin immunoprecipitation in keratinocytes as a direct Foxn1 target [53]. Expression of each gene was analyzed at two time points. Initially, we analyzed expression in EpCam+ cells purified from the thymic primordium at E13.5. QRT-PCR analysis revealed that all of the genes tested except Fgf2 showed changes in expression in response to Foxn1 levels (Figure 6C-6E and 6C’-6D’; see also Figure S3). In this population, Dll4, CCL25 and Pax1 were each expressed in proportion to Foxn1 mRNA levels, though titrating to different threshold levels (Figure 6C, 6D). Our data indicate considerable heterogeneity among E13.5 TEC, raising the possibility that the observed expression patterns reflected altered TEC subset balances between the different allelic variants. Therefore, we also analyzed expression of these genes in a defined cell population, E12.5 Plet1+ TEC. The expression levels of all three genes were also proportional to Foxn1 expression in this population (Figure 6 C’ and D’), indicating that, in TEC, Foxn1 is genetically upstream of all three genes. In contrast, EpCam was expressed at comparable levels in the E12.5 Plet1+ TEC population in all of the allelic variants. Surprisingly Fgf2 expression was detected in thymic mesenchyme but not in TEC, indicating that Foxn1 regulates different targets in TEC and keratinocytes (Figure 6F). To test whether Foxn1 regulation of Dll4 and CCL25 was direct or indirect, we over-expressed either the full-length Foxn1 cDNA or GFP cDNA in a cell population that does not normally express Foxn1 or either Dll4 or CCL25. Both genes were significantly up-regulated in cells transfected with Foxn1 but not in control samples (Figure 6G). These data thus strongly suggest that Foxn1 directly regulates both Dll4 and CCL25.
In conjunction with the data presented above showing Foxn1 regulates multiple intermediate TEC states, these data suggest that Foxn1 is a master regulator of the overall ‘thymus programme’ rather than regulating a particular aspect of TEC biology.
We have investigated the role of Foxn1 in the TE lineage via analysis of an allelic series that expresses defined levels of Foxn1 mRNA relative to WT. We have demonstrated that, in early ontogeny, Foxn1 is not required for divergence of the medullary TE sub-lineage from the common thymic epithelial progenitor cell; indeed we found no evidence for Foxn1 regulation of the choice between two alternative cell fates at any point in TE lineage progression. We have further demonstrated the requirement for Foxn1 for stable initiation of differentiation in the earliest TEC progenitors present in ontogeny, for subsequent progression through intermediate progenitor stages in both the cortical and medullary TEC sub-lineages, and for terminal differentiation of both cTEC and mTEC in the postnatal thymus. Taken together, these data establish that Foxn1 is required to execute the TE lineage programme at multiple stages in both the cortical and medullary TEC sub-lineages, but does not appear to regulate cell fate choice in TE lineage development. A model for Foxn1 regulation of cellular hierarchies during TE lineage development, based on the genetic analyses presented herein, is proposed in Figure 7.
In addition, we have established that Foxn1 is genetically upstream of a number of genes with known roles in TE lineage development or TEC function, specifically Pax1, Dll4, CD40, Cathepsin L and MHC Class II, demonstrating that Foxn1 regulates these genes either directly or indirectly. Finally, we have confirmed and extended observations that Foxn1 is required indirectly to regulate thymic mesenchymal development [54]. These findings significantly extend understanding of regulation of thymus development and function by this key transcription factor, and pave the way for molecular dissection of these processes.
With respect to TE lineage development during thymus organogenesis, our data (Figure 5) establish that Foxn1 is not required for divergence of the medullary TE sub-lineage from the common TEPC that is currently thought to exist at the basis of the TEC hierarchy [8]. This finding was unexpected, particularly given the recent demonstration of a persisting common TEPC in Foxn1 null mice [8]. Taken together with those of Bleul and colleagues, our data strongly suggest that only the common TEPC can persist in the postnatal Foxn1 null thymus and thus that Foxn1 is required directly or indirectly for TEC survival at stages subsequent to the common TEPC (including sub-lineage restricted progenitors). However, further work is required to explore this hypothesis, as an alternative possibility is that the activity of sub-lineage progenitors could not be detected in the clonal assay employed by Bleul.
Our data further establish that, subsequent to lineage divergence, Foxn1 is required in the fetal thymus for stable exit from the earliest progenitor cell state/entry into the differentiation programme leading to generation of functional TEC in both the cortical and medullary sub-lineages. Strikingly, low-level expression of Foxn1 (as found in Foxn1R/- TEC) is sufficient to initiate or stabilize entry into the TEC differentiation programme, as evidenced by acquisition of MHC Class II staining (see Figure 6), but is not sufficient to promote differentiation to a stage at which TEC in either sub-lineage are able to support T cell development. These data collectively establish that entry into the TE differentiation programme can be genetically separated from execution of the full programme (i.e. the programme leading to TEC that are functionally competent to support T cell development) based on different threshold expression levels of Foxn1. This indicates the requirement for Foxn1 for mediating both processes, and will enable dissection of the molecular role of Foxn1 at early and late differentiation stages.
We have also demonstrated, to our knowledge for the first time, that Foxn1 regulates differentiation of both cTEC and mTEC sub-lineages in the postnatal thymus. Specifically, we have shown that at intermediate levels of Foxn1 expression (as found in Foxn1R/R TEC) differentiation of cTEC is blocked at a CD205+CDR1- developmental stage and in addition, the number of Aire+ mTEC are significantly decreased. That the block in cTEC differentiation is relieved within two days of reversion to WT Foxn1 levels provides genetic evidence for the existence of the CD205+CDR1- intermediate stage in cortical TEC differentiation previously suggested by indirect analyses [11] and extends those findings from the fetal to the postnatal thymus. That the number of Aire+ mTEC is also rapidly restored to WT levels upon reversion of the Foxn1R allele suggests that differentiation of mTEC is also blocked at intermediate Foxn1 levels. From these data, we conclude that Foxn1 regulates similar mechanisms in each of the major TEC sub-lineages.
Our data further indicate MHC Class II as a direct or indirect target of Foxn1 in both fetal and postnatal TEC. Notably, TEC are unique outside the hematopoietic system in constitutively expressing MHC Class II, and the link between Foxn1 and MHC Class II expression is therefore significant.
Finally, we have demonstrated that Foxn1 directly or indirectly regulates a panel of genes that mediate diverse aspects of thymus development or function. These include Pax1, an essential mediator of TEC differentiation/survival [52]. Pax1 is expressed in the third pharyngeal pouch from E9.5 and remains expressed throughout thymus ontogeny, being restricted to cTEC in the postnatal thymus [50]. Although regulation of this Pax gene is only poorly understood, from E11.0 sustained expression of Pax1 requires Hoxa3 [55]. Our data show, to our knowledge for the first time, that expression of Pax1 in the thymic primordium is Foxn1-dependent. They thus place Foxn1 and Hoxa3 together in a network or cascade that regulates Pax1 expression, providing the first demonstration of a genetic interaction between Hoxa3 and Foxn1 (Figure 6H). In this regard, it is striking that Hoxa3+/-Pax1-/-compound mutant mice share some phenotype characteristics with Foxn1R/R mice; both mutants have hypomorphic postnatal thymi and reduced levels of MHC class II expression on TEC [51]. As Hoxa3 itself is expressed in TEC in all of the Foxn1 mutants at E13.5 (data not shown), our data could support either of two alternative explanations: Hoxa3 may regulate Foxn1, which in turn regulates Pax1 expression in the thymic primordium, such that Hoxa3 regulation of Pax1 is Foxn1-dependent (Figure 6Gi), or Hoxa3 and Foxn1 may be required independently to maintain Pax1 expression in the third pharyngeal pouch/early thymus primordium (Figure 6Gii). We note that since Hoxa3 is expressed both in TEC and neural crest-derived mesenchymal cells in the thymic primordium, its regulation of Pax1 could be either direct or indirect.
We further show that Foxn1 regulates CCL25 and Dll4, which each play critical roles in thymocyte development – CCL25 regulates colonization of the fetal thymus [49] while Dll4 is the obligate Notch ligand controlling commitment of haematopoietic progenitors to the T cell lineage [34] respectively. It has previously been suggested, based on antibody staining, that although Dll4 and CCL25 are absent from the Foxn1 null thymus [56] their expression is Foxn1-independent in TEC [57]. However, this view has been challenged in a recent report, which placed Foxn1 upstream of dll4a and ccl25a expression in medaka fish [58]. Our data show that in both early fetal TEC and in the postnatal thymus Dll4 expression is proportional to Foxn1 expression, and further demonstrate that over-expression of Foxn1 in embryonic stem cells results in expression of both Dll4 and CCL25. They therefore extend the findings in medaka to demonstrate Foxn1 regulation of Dll4 and CCL25 in the mammalian thymus, and provide strong evidence that Foxn1 regulation of both of these target genes is direct. Reduced expression of Dll4 in Foxn1R/R mutants also correlates with a proportional increase in intrathymic B cells, indicating that Foxn1 expression within the normal range is necessary to sustain full TEC functionality. Taken together, these data provide mechanistic insight into our findings that Foxn1 regulates multiple aspects of TEC development and function. As they demonstrate Foxn1 regulation of genes required for many different aspects of TEC biology, they suggest that this transcription factor functions a master regulator of the core TEC lineage programme. Furthermore, the finding that different Foxn1-regulated genes show distinct response patterns to changes in Foxn1 dosage may explain why different levels of Foxn1 are required for different functions in TEC.
Collectively, the data presented herein establish Foxn1 as powerful regulator of TEC differentiation in the fetal and adult thymus. Furthermore, they demonstrate that different threshold levels of Foxn1 mRNA are required for different functions. Together with the recent finding that reduction of Foxn1 expression postnatally causes premature thymic involution characterized by loss of TEC [25]-[27], this suggests that Foxn1 may be required in TEC from initiation of differentiation throughout the cell's lifetime. Further work is required to determine whether it is a maintenance factor for differentiated TEC as well as having roles in proliferation and differentiation. Improved understanding of the regulation of Foxn1 in thymic development and homeostasis, and of Foxn1 regulation of its targets in different TEC sub-types, is thus likely to become a priority for strategies aimed at protecting or regenerating the thymus for therapeutic ends.
All animal work was conducted according to UK Home Office guidelines, as established in the ANIMALS (SCIENTIFIC PROCEDURES) ACT 1986.
Rosa26CreERt2/+ [41] and ZP3-Cre [28] mice were maintained as homozygotes and crossed with Foxn1R/R mice as described. Foxn1-/- mice [59] were maintained as heterozygotes on a C57BL/6 background. Foxn1R mice were backcrossed onto the C57BL/6 background for at least 5 generations and subsequently maintained via intercrossing. For timed matings, noon of the day of the vaginal plug was taken as day 0.5.
A construct containing the targeting cassette shown in Figure S1 was generated by standard molecular biology techniques and verified by sequencing. Conventional subcloning was used for the majority of cloning steps. PCR cloning using a proof-reading Taq polymerase (Roche) and TOPO-TA cloning vector (Invitrogen) was employed to insert the SV40 T antigen cassette. RecET cloning [60] was used to isolate 4 kb of the mouse Foxn1 locus from PAC367b19.
Genomic DNA was processed for Southern Blotting as described [61].
Mouse sv129/ola ES cells (line E14tg2a) were electroporated with linearized targeting vector and grown under Blasticidin selection. Three correctly targeted clones, identified by Southern analysis (Figure S1), were expanded and transfected with a Cre recombinase expression plasmid to induce deletion of the BSD cassette. Clones in which this cassette had been deleted, but which retained the rest of the targeting construct including the two remaining LoxP sites, were identified by Southern analysis, verified by sequencing and injected into C57BL/6 blastocysts to generate chimeric mice. Germ-line transmission was confirmed by genotyping for two independently targeted ES cell clones.
E14Tg2a cells were cultured in GMEM containing FCS and LIF. Transfection was carried out in 12 well plates, at a cell density of 90-95% confluence, using GeneJuice transfection reagent (Novagen). The ratio of transfection reagent to DNA was 5:1 and 1 µg of DNA was used for each well. The medium was changed 24 hrs after transfection and 1.5 µg/ml Puromycin was added to the new medium to select for transfected cells. Cells were harvested 48 hours after transfection.
The sequences of primers used for conventional PCR and QRT-PCR are shown in Table 1 and Table 2 respectively.
Foxn1F and Foxn1R were used to detect the wild-type Foxn1 allele. TF and TR were used to detect the SV40 T antigen cDNA.
MTS20 (IgM) and MTS24 (IgG2a), rat mAbs that recognise Plet-1 [5], and MTS10, a rat mAb that recognises medullary TEC, were kind gifts from R.L. Boyd; anti-EpCAM (G8.8, rat IgG2a, DSHB); anti-Cytokeratin 8 (Troma 1, rat IgG2a, DSHB); anti-Cytokeratin 14 (LL002, mouse IgG3 was a kind gift from E.B. Lane); anti-Cytokeratin 5 (AF138, rabbit IgG, Covance); anti-Claudin 4 (rabbit IgG was a kind gift from S. Tsukita); anti-Dll4 was a kind gift from F. Radtke; anti-CD3-FITC (145-2C11, hamster IgG1); anti-CD4-FITC or PE (H129.19, rat IgG2a); anti-CD8-FITC (53-6.7, rat IgG2a); anti-CD11b-FITC (M1/70, rat IgG2b); anti-CD11c-FITC (HL3, hamster IgG1); anti-α-CD19-FITC (1D3, rat IgG2a); anti-CD25-PE (3C7, rat IgG2b); anti-CD44-APC (1M7, rat IgG2b); anti-Ly76-FITC (Ter119, rat IgG2b); anti-CD31-FITC or purified (390, rat IgG2a); anti-CD45-APC (30-F11, rat IgG2b); anti-PDGFRalpha (APA-5, rat IgG2a); anti-SV40 T antigen (PAb100, mouse IgG1)(all Pharmingen); ERTR7 (rat IgG2a was a kind gift from W. van Ewijk); anti-Cytokeratin (rabbit IgG polyclonal, DAKO); CDR1 (Rat IgG2a, was a kind gift from B Kyewski); anti-Foxn1 C-terminal (IMG-3744, polyclonal, Imgenex); Ly51 (Clone 6C3, Rat IgG2a, Biolegend); anti-ß5t (Rabbit polyclonal, purified, IgG, MBL International); biotinlyated UEA-1 (Vector Laboratories). For detection of unconjugated primaries the following secondary antibodies were used; goat anti-rabbit IgG-alexa488; goat anti-rat IgG-alexa647; goat anti-rat IgG-alexa488; goat anti-mouse IgG-alexa647; donkey anti-rat IgG-alexa488; Streptavidin-alexa647 (all Molecular Probes); mouse anti-rat IgM-PE (G53-238, Pharmingen).
Whole E13.5 – E15.5 embryos or adult thymi were processed for immunohistochemistry as described [2]. Isotype controls (not shown) were included in all experiments. Staining was analyzed using a Leica AOBS confocal microscope (Leica Microsystems GmbH). The images presented are either single optical sections or projected focus stacks of serial optical sections. The number of Aire+ TEC per mm medullary area was established by analyzing three medullary areas on each of two non-sequential sections from each of three individual mice (i.e. the number of Aire+ cells was counted for six different medullary areas for each of three individual mice per condition). The size of each medullary area was calculated using Adobe Photoshop and the number of Aire+ cells per square mm was then calculated. Statistical analysis was on the average number of Aire+ cells per condition.
Mice were treated with a single intraperitoneal injection of 1.5 mg 4-hydroxy tamoxifen (4OHT) prepared in ethanol and diluted appropriately in Cremophor (Sigma)/PBS.
RNA was prepared using Tri-reagent and RNAeasy (both Qiagen) according to manufacturer's instructions. All samples were DNase treated.
cDNA was prepared using the Superscript II first strand synthesis kit (Invitrogen) with Oligo-dT primers, according to manufacturer's instructions. For data shown in Figure 4, the IQ SYBR Green Supermix (Bio-Rad) was used for quantification and the relative expression level of the target genes was normalized to the geometric mean of three control genes (Hprt, Ywhaz, Hmbs). For data shown in Figure 3, Figure 6C, 6D, 6F and 6G, Figure S2 and Figure S3, relative expression levels were determined using the Roche Universal Probe Library (Foxn1, Probe 68; SV40Tag, Probe 32; alpha-tubulin, probe 58; Dll4, probe 106; CCL25, probe 9; Pax1, probe 105; Fgf2, probe 4; EpCam, probe 52; EVA, probe 100) with the Roche Lightcycler 480. Relative expression levels are shown after normalization to alpha-tubulin expression using Roche LC480 Relative Quantification software. Technical duplicates or triplicates were run for all samples and no RT and no template controls were included in all experiments. For data shown in Figure 6C’, 6D’ and 6E, relative expression levels were determined using the Roche Universal Probe Library (as above) using microfluidic QRT-PCR (Fluidigm). Pre-amplification was carried out using target specific primer pairs according to the manufacturer's protocol. Samples were loaded onto a BioMark 48.48 Dynamic Array (Fluidigm) and thermal cycling was performed using a BioMark instrument (Fluidigm) according to the manufacturer's protocol. Data analysis was carried out using BioMark Real-Time PCR Analysis Software v2.0 (Fluidigm) and the ΔCt method [63]. Expression levels are shown relative to WT after normalization to Epithelial V-like Antigen (EVA)[64].
Pregnant female mice were injected with 1mg BrdU at E15.5 and embryos were collected 1 hour after injection. Thymi were immediately microdissected, dissociated and stained with G8.8 (Pharmingen), anti-CD45-PE, anti-BrdU-APC and 7-AAD using the Pharmingen APC BrdU Flow Kit according to manufacturer's instructions.
Nuclear protein fractions were prepared using the Active Motif Nuclear Extract Kit according to manufacturer's instructions, separated by electrophoresis on a SDS-PAGE gel (Novex, Invitrogen), and processed for Western Blotting as described [5].
Statistical analysis was performed using the one-way ANOVA test (two tailed), as appropriate for normally distributed data (normal distribution was tested using Chi2 goodness of fit). The alpha level is taken as 0.05. Errors shown are standard deviations throughout. Sample sizes of at least n = 3 were used for statistical analyses.
|
10.1371/journal.pntd.0006373 | Diet and hygiene practices influence morbidity in schoolchildren living in Schistosomiasis endemic areas along Lake Victoria in Kenya and Tanzania—A cross-sectional study | Since 2011, cohorts of schoolchildren in regions bordering Lake Victoria in Kenya and Tanzania have been investigated for morbidity caused by Schistosoma mansoni infection. Despite being neighbouring countries with similar lifestyles and ecological environments, Tanzanian schoolchildren had lower S. mansoni prevalence and intensity and they were taller and heavier, fewer were wasted and anaemic, and more were physical fit compared to their Kenyan peers. The aim of the present study was to evaluate whether diet and school-related markers of socioeconomic status (SES) could explain differences in morbidity beyond the effect of infection levels.
Parasitological and morbidity data from surveys in 2013–2014 were compared with information on diet and school-related markers of SES collected in 2015 using questionnaires. A total of 490 schoolchildren (163 Kenyans and 327 Tanzanians) aged 9–11 years provided data. A higher proportion of Tanzanian pupils (69.4%, 95% CI: 64.3–74.5) knew where to wash hands after toilet visits compared to Kenyan pupils (48.5%, 95% CI: 40.9–56.1; P<0.0005). Similar proportions of children in the two countries ate breakfast, lunch and dinner, but the content of the meals differed. At all three meals, a higher proportion (95% CI) of Tanzanian pupils consumed animal proteins (mostly fish proteins) compared to their Kenyan peers (35.0% (28.3–41.7) vs. 0%; P<0.0005 at breakfast; 69.0% (63.9–74.1) vs. 43.6% (35.8–51.4); P<0.0005 at lunch; and 67.2% (62.1–72.3) vs. 53.4% (45.8–61.0); P = 0.003 at dinner). Multivariable analyses investigating risk factors for important morbidity markers among individuals revealed that after controlling for schistosome and malaria infections, eating animal proteins (fish) and knowing where to wash hands after toilet visits were significant predictors for both haemoglobin levels and physical fitness (measured as VO2 max).
These results suggest that the differences in morbidity may be affected by factors other than S. mansoni infection alone. Diet and hygiene practice differences were associated with health status of schoolchildren along Lake Victoria in Kenya and Tanzania.
Trials Registration numbers: ISRCT 16755535 (Kenya), ISRCT 95819193 (Tanzania).
| Millions of school-age children in Kenya and Tanzania are at risk for infection with Schistosoma mansoni, which has an impact on their physical health. A total of 490 schoolchildren (163 from Kenya and 327 from Tanzania) aged 9–11 years living along the shores of Lake Victoria in Kenya and Tanzania provided data on S. mansoni and malaria infections, weight, height, anaemia and physical fitness. The pupils in Tanzania had lower prevalence and intensity of S. mansoni infection compared to pupils in Kenya, but more had malaria parasites in their blood. In addition, Tanzanian pupils were taller and heavier, fewer were anaemic and they scored higher in a physical fitness test. Questionnaire data showed that more of the Tanzanian pupils knew where to wash hands after toilet visits, and more consumed animal protein (mostly fish protein) for breakfast, lunch and dinner. Multivariable analyses revealed that eating animal protein and knowing where to wash hands after toilet visits were significant predictors for haemoglobin levels and physical fitness. These results suggest that the differences in morbidity parameters may be affected by factors other than S. mansoni infection alone. Diet and hygiene practices seem to contribute to the health status of schoolchildren along Lake Victoria in Kenya and Tanzania.
| Schistosomiasis, also known as bilharzia, is an infectious disease caused by parasitic flatworms of the genus Schistosoma. Schistosomiasis is considered as one of the Neglected Tropical Diseases (NTDs) and is estimated to affect at least 230 million people annually [1] with a majority in sub-Saharan Africa [2]. In terms of public health impact, schistosomiasis is second only to malaria as the most important parasitic disease in developing countries [3]. Schistosome infections can result in anaemia, stunted growth, malnutrition, impaired physical fitness and numerous other complications [1,4]. Although praziquantel has been available for decades and is effective for the treatment of Schistosoma infections, schistosomiasis remains a major health concern. Schistosomiasis control efforts usually focus on reducing the prevalence and intensity of infection by Preventive Chemotherapy (PC).
Socioeconomic status (SES) has been linked to various health issues, such as nutritional status, disease burden and mortality, as well as accessibility and affordability of health services [5]. As schistosomiasis typically occurs in rural areas where the majority of the population is highly affected by poverty, the impact of schistosomiasis in a given area may be exacerbated by low SES. Collecting socioeconomic information may help identify potential risk factors that contribute to schistosomiasis as well as associated morbidity, and therefore help improve the impact of the national schistosomiasis control programmes. Addressing both schistosomiasis and these other factors would help direct resources to areas most in need.
The current investigation was conducted as a part of two cohort studies within a larger multi-country Schistosomiasis Consortium for Operational Research and Evaluation (SCORE) research project on gaining and sustaining control of schistosomiasis [6]. The SCORE projects in Kenya and Tanzania were implemented in areas near Lake Victoria, where prevalence of S. mansoni infection was 25% or greater [7,8]. The nested cohort studies were conducted to assess S. mansoni infection markers of morbidity over five years and compare the effect of different treatment strategies. Morbidity data collected from these cohorts during the year 3 assessment (2013–2014) showed that Tanzanian schoolchildren were taller and heavier, fewer were nutritionally wasted and anaemic, and they scored higher in a physical fitness test compared to Kenyan schoolchildren. The Tanzanian pupils also had lower S. mansoni prevalence and intensity despite assumed similar exposure risks and ecological setting. The aim of the present study was to evaluate whether diet and school-related markers of SES could explain differences in morbidity across study sites beyond the effect of infection levels.
Approval for the SCORE Kenya and Tanzania gaining and sustaining control of schistosomiasis and cohort studies was obtained from Institutional Review Boards at the Scientific and Ethical Review Committees of the Kenya Medical Research Institute (Nairobi, Kenya) and the Medical Research Coordination Committee (MRCC) of the National Institute for Medical Research (Tanzania). Trials Registration numbers: ISRCT 16755535 (Kenya), ISRCT 95819193 (Tanzania). The parasitological and morbidity data (year 3 cross-sectional data) were previously collected by the SCORE team in Kenya and Tanzania and are covered by the above mentioned approval, while the questionnaire data were collected from children who separately assented to participate and had written informed consent from parents or legally authorized representatives.
The study took place in Nyanza Province (Bondo District) of Kenya and Mwanza Region (Sengerema District) of Tanzania, along the Lake Victoria shoreline.
For the SCORE cohort studies, communities were randomly selected from the two arms with the most intense level of treatment (annual community-wide treatment over four years) and the less intense treatment strategy (biannual school-base treatment) as part of the larger cross-sectional study (Fig 1). In each of these two study arms, 4 (Tanzania) and 6 (Kenya) out of 25 communities were randomly selected in order to achieve a baseline cohort of 800 schoolchildren from each country. These children were 7–8 years of age at the initiation of the intervention study. The children were enrolled for 4 years of intervention succeeded by a 5th year follow-up testing.
Morbidity parameters were measured at baseline, and in year 3 and 5 prior to PC with praziquantel. The year 3 cross-sectional data on parasitology and morbidity were collected in 2013 in Kenya and 2014 in Tanzania. The one year difference was due to a one year delay in the larger cross-sectional study in Tanzania and do not reflect different intervention periods. Questionnaire data were collected in Kenya in January and February 2015 and in Tanzania from March to April 2015. Sample size calculations were not performed specifically for this section of the larger study, but all the cohort pupils available at the time of our visits were enrolled.
The methods for collection of cohort data are described in detail elsewhere [9,10], but are briefly presented below.
Participants were given stool containers and asked to bring fresh stool specimens to the school on three consecutive days. Specimens were processed using duplicate Kato-Katz thick smears with a 41.7 mg template [11] from each specimen in the school (Tanzania) or after being transported to the KEMRI/CDC laboratory (Kisumu, Kenya). All slides were examined at the laboratories in Mwanza or Kisumu for S. mansoni eggs. The number of S. mansoni eggs was multiplied by 24 and expressed as eggs per gram of stool (epg). Intensity was reported as the arithmetic mean of epg from the total number of slides per person.
Infections with STHs were not investigated in Tanzania as Ascaris lumbricoides and Trichuris trichiura have been recorded to be seldom present [12,13] and because eggs of hookworms were not visible because of the time span between preparation and reading.
A 5 mL of venous blood sample (Kenya) or finger-prick blood sample (Tanzania) was collected from each individual as part of the larger study design within each country and haemoglobin (Hb) measured using a portable HemoCue photometer (Ängelholm, Sweden). Hb level was reported in g/L and final values used in analysis were adjusted for altitude by subtracting 2 g/L from the raw values for both study sites [14]. Anaemia was defined as Hb values below 115 g/L according to the World Health Organization guidelines [14]. Infection with Plasmodium falciparum was determined by examination of blood smears by experienced microscopists in Kenya and by rapid diagnostic test (RDT) (SD Bioline, Republic of Korea) in Tanzania. All children were asymptomatic for malaria.
Height was measured on barefooted children using a wooden stadiometer. The child stood on the base of the stadiometer with their heels, buttocks, shoulder blades and back of the head touching the vertical backboard and looking straight ahead. When correctly positioned, the ruler was lowered and the height measured in centimetres with one decimal. Weight was measured on a digital scale on barefooted children having removed any excess clothing. Weight was measured in kilograms to one decimal. Height and weight were measured twice by the same examiner and the mean recorded. Z-scores were calculated using the WHO growth reference table [15]. Wasting was defined as a BMI-for-age Z-score of <-2 SD.
Physical fitness was assessed using the 20 metre shuttle run fitness test (20mSRT) as described in Bustinduy and colleagues [16]. In brief, during the test, children run continuously between two lines 20 meters apart at increasing speeds, turning when signalled to do so by recorded beeps. A “shuttle” is defined as a run from one line to the other. The running field was prepared in the school compound and runners were separated with at least one meter. Recorders were placed at each end of the field and every recorder was responsible for taking notes of three to five children. The recorder noted the level at which the test subject stopped and how many shuttles the child completed within that level. These numbers are correlated to a maximal oxygen uptake, the VO2max in mL/kg/min as described in Müller and colleagues [17].
The questionnaire was generated by the different research scientists, including social scientists, involved in the SCORE project [18–21]. The questionnaire was developed in English first and translated into Dholuo/Kiswahili language then verbally administered to participants. Prior to data collection, the questionnaires were pilot-tested in groups of children that were not part of the research project to ensure that the tool was feasible to administer and easy to understand for the respondent. The questionnaire contained questions on diet and a range of school-related markers of SES and is attached as supporting information (S1 Questionnaire).
Statistical analyses were performed using SPSS version 24 (IBM, Armonk, NY). Summary statistics were calculated and all statistical tests used were two sided, and P<0.05 was considered significant. To assess differences between the two countries, Pearson Chi-square was used to assess differences in proportions. The Student’s t-test and one-way ANOVA were used to assess differences between normally distributed means (two and more than two, respectively), while the Mann-Whitney U test was used to compare means that were not normally distributed.
To further analyse the potential determinants for morbidity differences seen in the pupils, we performed a more detailed statistical analysis using uni- and multivariable linear regression analyses. The regression analyses were applied with the two important morbidity indicators: ‘Haemoglobin (Hb)’and ‘Maximum oxygen uptake (VO2 max)’ as dependent variables. As too few individuals were nutritionally wasted, we did not perform an analysis with ‘Wasting’ as dependent variable. First, univariable analyses were performed on each of the two dependent variables on the following independent variables: ‘gender’, ‘age’, ‘height’, ‘Schistosoma infection’, ‘treatment arm’, ‘malaria infection’, ‘anaemia’ (but only for VO2 max), ‘transportation to school’, ‘distance to school’, ‘health information from teachers’, ‘knowing where to find a place to wash hands after toilet visits’, ‘wearing shoes at school’, ‘having animal proteins at meals’, and ‘having vegetables at meals’. The variable ‘having grain at meals’ was omitted as almost all pupils had grain for all meals. As height and weight are highly correlated (S1 Data); only height was included in the analyses. Age was divided in two groups; 9 years and a combined 10–11 years as only 11 children were 11 years old. Height was divided in two groups with the mean height as cut-off value. The presence of STH infections were recorded in Kenya and the effect of these infections on Hb and VO2max were performed on the Kenyan data only. All variables with P≤0.10 (plus ‘Schistosoma infection’ and ‘malaria infection’ to control for the two infections) were then included in the subsequent multivariable analyses with Hb and VO2 max as dependent variables using a stepwise strategy (criteria for enter ≤0.05 and for remove ≥0.10).
A total of 163 pupils from the Kenyan cohort and a total of 327 pupils from the Tanzanian cohort provided information for the current study (Table 1 and Table 2).
Table 1 shows the year 3 cross-sectional measurements from both countries. Apart from gender distribution, all parameters were significantly different between the two countries. The pupils in Tanzania had lower prevalence and intensity of S. mansoni infections compared to pupils in Kenya. More Tanzanian pupils were diagnosed with malaria but they were tested with the slightly more sensitive RDT, while the Kenyan students were diagnosed by microscopic detection of parasites in their blood. In addition, the Tanzanian pupils were taller and heavier, and fewer were nutritionally wasted compared to their peers in Kenya. Finally, fewer Tanzanian children were anaemic and in the physical fitness test they scored a higher VO2 max. In Kenya, information on STH infections was obtained from 150 individuals. Of those, eight (5.3%) were A. lumbricoides and T. trichiura positives and two (1.3%) had hookworm infections.
The results of the questionnaire are shown in Table 2. The majority of the pupils, 83.4% in the Kenyan cohort and 97.2% in the Tanzanian cohort put on shoes when they went to school, and the difference between the two cohorts was statistically significant (P<0.0005). While less than half (48.5%) of the Kenyan pupils knew where to wash their hands following a toilet visit, more than two-thirds (69.4%) of the Tanzanian pupils were aware (P<0.0005). More Kenyan pupils (92.0%) reported that they received health-related information from their teacher compared to Tanzanian pupils (54.7%, P<0.0005).
The pupils were asked specifically what they had for each meal the day before, categorised as grain, animal protein and vegetable. Grain (carbohydrate rich) was the main source of nutrients in both cohorts for all three meals of the day. However, the intake of animal protein was significantly different between the two cohorts for all three meals as more Tanzanian pupils consumed animal proteins. By contrast, a higher proportion of Kenyan pupils consumed vegetables. Of the animal protein consumed, fish proteins contributed to more than 80% of all meals for both cohorts.
Table 3 shows the univariable associations of demographic, anthropometric, parasitological (including treatment history), and diet and school-related markers of SES with mean Hb level (g/L) of the schoolchildren in both countries combined. The following variables were significantly associated with Hb: ‘age’, ‘height’, ‘know where to wash hands after toilet visit’ and ‘meals with animal protein’.
In the multivariable analysis where all variables with P≤0.10 were included and after controlling for schistosome and malaria infections; ‘age’, ‘height’, ‘know where to wash hands after toilet visit’ and ‘meals with animal protein’ were retained in the model (Table 4). The regression coefficient of ‘knowing where to wash hands after toilet visits’ means that those pupils who knew had on average 5.33 g/L higher Hb than their peers who did not know. The regression coefficient of ‘animal protein at meals’ corresponds to a 2.98 g/L increase in Hb for every step between having no animal proteins, having animal proteins once, twice or three times a day. The regression coefficient of ‘height’ means that pupils with a height of 135 cm or taller had a 4.50 g/L higher Hb compared to the pupils lower than 135 cm. Finally, the coefficient of ‘age’ means that the pupils of the age of 10–11 years had a 4.34 g/L higher Hb compared to the 9 year old pupils.
For separate analysis on the Kenyan data where STH infections were included, none of the three infections were retained in the multivariable analysis.
The analyses for VO2 max showed more significant predicators compared to the Hb (Table 5). The following were found to be significant: ‘gender’, ‘height’, ‘Schistosoma infection’, ‘where to wash hands’, ‘animal protein at meals’ and ‘animal protein at meals (grouped)’. In the multivariable linear regression analysis, which included all variables from the univariable analyses with P≤0.10 and after controlling for schistosome and malaria infections, the following variables were significant for VO2 max: ‘gender’, ‘height’, ‘animal protein at meals (grouped)’ and ‘know where to wash hands after toilet visit’ (Table 6).
The regression coefficient of ‘gender’ means that boys had a 2.80 mL/kg/min higher VO2 max compared to girls. The regression coefficient of ‘height’ means that pupils with a height of 135 cm or taller had a 1.69 mL/kg/min higher VO2 max compared to those pupils shorter than 135 cm. The regression coefficient of ‘meals with animal protein (grouped)’ means than those pupils having at least one meal with animal protein had a 1.42 mL/kg/min higher VO2 max compared to those having no meals with animal protein. Finally, pupils knowing where to wash hands after toilet visits had on average 1.04 mL/kg/min higher VO2 max than their peers who did not know.
For separate analysis on the Kenyan data where STH infections were included, none of the three infections were retained in the multivariable analysis.
Differences in diet and hygiene practices may explain differences in morbidities commonly associated with schistosomiasis. Schoolchildren in Tanzania more often consumed animal proteins compared to their Kenyan peers and this difference could possibly explain the difference in several morbidity markers between the two populations. Thus, when the two populations were combined and analysed at the individual level, consumption of animal proteins was a significant predictor of both Hb levels and physical fitness.
The consumption of animal proteins was associated with increased Hb levels with 3.0 g/L for every increase in number of meals with animal proteins per day. In a nation-wide survey in 2004/2005, the most common cause of anaemia among children in Tanzania was identified to be nutritional anaemia resulting from inadequate dietary intake of nutrients [22]. Furthermore, Mboera and colleagues [22] demonstrated in their study in central Tanzania that anaemia was most prevalent among communities with low prevalence of malaria, suggesting that the anaemia was most likely to be a result of dietary deficiency or caused by other infections than malaria. The consumption of vegetables was not associated with Hb levels despite the fact that these vegetables contain iron. This is probably because this iron is of the non-heme type and that non-heme iron is not as easily absorbed as the heme iron. Furthermore, iron-rich vegetables also contain oxalates and phytates, which impair iron absorption, and boiling decreases the content of iron in vegetables [23]. By contrast, fish contains high amounts of the more easily absorbed heme iron and cooking does not significantly reduce the content of iron in animal products [23].
A total of 69.4% of the Tanzanian pupils knew where they could wash their hands after toilet visit compared to 48.5% of the Kenyan pupils. The question ‘After toilet visit is there a place to wash your hands’ is meant to reveal whether there are possibilities for hand washing in the school, but it is also reflecting on the pupil’s knowledge about personal hygiene. These two parts cannot be separated based on the answers. Although hand washing after toilet visits is not one of the important tools in the toolbox of schistosomiasis control, it certainly has an effect on other infections such as the STHs A. lumbricoides and T. trichiura and other pathogens causing diarrhoeal diseases. These infections may also have an impact on Hb level. Unfortunately, information on STH infections is only available for the Kenyan cohort, but not for the Tanzanian cohort for reasons described in the method section. Although literature reports that A. lumbricoides and T. trichiura are more common in the Kenyan part compared to the Tanzanian part [24], the prevalence of the two infections in Kenya in this study was less than 6%. Thus, it is not likely that the difference in Hb levels between the two countries can be explained by a difference in prevalence of either of these infections.
Hookworm infections, and especially high intensity infections, have an impact on Hb levels. In the present study, hookworm prevalence was only 1.3% in Kenya (only two individuals), while information from Tanzania is lacking. However, a recent study reported hookworm infections in 16% of school and pre-school children in Magu District, which is a neighbouring district to Sengerema District, where the present study took place [13]. It is therefore not plausible that the lower Hb levels in Kenya compared to Tanzania can be explained by a higher prevalence or intensity of hookworm infections in Kenya. Lack of shoes was not a risk factor for Hb levels in this study probably because of the assumed low levels of hookworm infections.
Schistosoma mansoni infection was not a risk factor for Hb levels in this study. This is in contrast to the results from the baseline survey in Kenya where heavy S. mansoni infections were predictors of anaemia [9]. In a recent systematic review and meta-analyses on the effect of treatment on mean Hb levels there was no consistent or significant changes between pre- and post-treatment surveys in ten different studies in schoolchildren [25]. However, only one of these studies was on S. mansoni infection and although it documented a considerable impact on Hb and anaemia after two years of treatment through the Ugandan National Control Programme [26], it is not possible to attribute the improvements to praziquantel treatment alone. This is because the population was also treated with albendazole, which decreased the prevalence of hookworms from 50.9% to 10.7% and the mean hookworm intensity from 309 epg to 22 epg.
Our data showed that infection with Plasmodium had no significant association with Hb levels. Malaria and iron have a complex but important relationship. Plasmodium proliferation requires iron, both during the clinically silent liver stage of growth and in the disease-associated phase of erythrocyte infection [27]. Interestingly, human iron deficiency appears to protect against severe malaria, while iron supplementation may increase risks of infection and disease [27]. This could explain why the Tanzanian pupils had higher prevalence of malaria compared to the Kenyan pupils. However, it is important to note that the diagnostic techniques differed in the two countries; infection with Plasmodium falciparum was determined by examination of blood smears via microscope in Kenya and by rapid diagnostic tests in Tanzania. The rapid diagnostic test is known to be only slightly more sensitive compared to microscopy [28] when microscopy is performed by experienced technicians; still, the prevalence of malaria could be underreported in the Kenyan pupils.
Schistosome infection was negatively associated with physical fitness although only in the univariable analysis. In the multivariable analysis other measured parameters seem to be more important for physical fitness resulting in the exclusion of schistosome infection in the final model. Thus, being a boy, being taller than average, having animal proteins at least once a day and knowing where to wash hands after toilet visits were significant predictors of physical fitness. The lack of association between physical fitness and infection in the multivariable analysis is in accordance with other recent studies using the 20mSRT [16,17,29] and with the baseline results of the two cohorts investigated in this study [9,10]. However, in line with the present study, boys had better physical fitness compared to girls in the three studies reporting the associations [16,17,29], an association which was lacking at baseline in Tanzania [10] and not investigated in Kenya [9]. As age, height and weight were similar between genders in the present study (S1 Table), this difference is reflecting gender-specific differences and is less related to physical features. Besides the gender differences, Bustinduy et al. [16] found anaemia and growth stunting to be predictors of physical fitness, while only age was predictor in the study of Müller et al. [17]. The role of T. trichiura infections for the physical fitness of school-age children is unclear according to two Chinese studies [30,31]. The impact of awareness of personal hygiene and/or content of diet intake on physical fitness has not been assessed previously in the two study areas.
Reduced physical fitness is a manifestation of the body's inability to maintain adequate oxygen supply to the tissues and may have many different causes. In developing countries, low physical fitness is often the result of anaemia and under nutrition, which have multifactorial aetiologies such as poor diet and chronic infections. Most important of these infections are malaria, hookworm and schistosomiasis [16]. Having animal protein at meals was a strong predicator for VO2 max uptake in the multivariable linear regression analysis, suggesting that animal protein is a decisive factor for fitness. This is consistent with Bustinduy and colleague’s study, which demonstrated malnutrition parameters as strong predicators of decreased fitness [16]. As eating animal protein also was a strong predictor for Hb level, the physical fitness might be affected in two ways. The animal protein increased the Hb level which again increased the fitness, but at the same time the mere consumption of animal proteins might better satisfy the child’s nutritional needs, resulting in higher fitness.
There were limitations in our study, which need to be taken into account during interpretation of the results. Initially, the baseline study enrolled 800 pupils from each cohort, however, through the years, many pupils dropped out and the present study is thus implemented in a sub-group of the baseline cohort. The design of the questionnaire was useful; however, the question ‘After toilet visit is there a place to wash your hands’ could have been more specific so the presence of hand washing facilities could be clearly separated from the pupils’ knowledge on hygiene practices. It would have been valuable if information on prevalence and intensity of STH infections had been available from both study areas. In addition, it was difficult to compare prevalence of malaria infections between the two countries as the two diagnostic tests for malaria have slightly different sensitivities.
In conclusion, these results suggest that the differences in morbidity parameters between Kenyan and Tanzanian schoolchildren living near Lake Victoria may be due to factors other than S. mansoni infection alone. Knowing where to wash hands after toilet visits and having a diet rich in fish were associated with higher haemoglobin levels and a better physical fitness. The consequence of these results is that control programmes may improve their interventions by encouraging the communities to provide hand washing facilities in schools, strengthen education on good personal hygiene and promoting a healthy and nutritional diet rich in protein and iron.
|
10.1371/journal.ppat.1002759 | Polyfunctional Type-1, -2, and -17 CD8+ T Cell Responses to Apoptotic Self-Antigens Correlate with the Chronic Evolution of Hepatitis C Virus Infection | Caspase-dependent cleavage of antigens associated with apoptotic cells plays a prominent role in the generation of CD8+ T cell responses in various infectious diseases. We found that the emergence of a large population of autoreactive CD8+ T effector cells specific for apoptotic T cell-associated self-epitopes exceeds the antiviral responses in patients with acute hepatitis C virus infection. Importantly, they endow mixed polyfunctional type-1, type-2 and type-17 responses and correlate with the chronic progression of infection. This evolution is related to the selection of autoreactive CD8+ T cells with higher T cell receptor avidity, whereas those with lower avidity undergo prompt contraction in patients who clear infection. These findings demonstrate a previously undescribed strict link between the emergence of high frequencies of mixed autoreactive CD8+ T cells producing a broad array of cytokines (IFN-γ, IL-17, IL-4, IL-2…) and the progression toward chronic disease in a human model of acute infection.
| The emergence of a large population of mixed polyfunctional (type-1, -2, -17) CD8+ T cell effector responses specific for apoptotic T cell-associated self-epitopes rather than the dysfunction or altered quality of virus-specific CD8+ T cells is associated with the progression toward chronic disease in the human model of acute HCV infection. The chronic evolution is associated with the selection of autoreactive CD8+ T cells with higher T cell receptor avidity, whereas those with lower avidity undergo prompt contraction, as seen in patients undergoing infection resolution. We suggest that these autoreactive responses are secondary to the viral persistence and can participate to the HCV-related immunopathology. This data has implications for the prognosis and therapy of infections undergoing chronic evolution.
| The fate of the enormous number of apoptotic cells that derive from effector Tcells undergoing apoptosis after performing their functions during acute or chronic infections remain to be determined [1], [2]. Phagocytosis of apoptotic cells by dendritic cells (DCs) leads to the processing of apoptotic cell-associated antigens and the cross-presentation of the resulting peptides on major histocompatibility complex (MHC) class I molecules [3]–[6]. This phenomenon seems crucial for inducing either cross-priming or cross-tolerance of CD8+T cells, based on the presence or absence of various infectious or danger signals influencing the switch from tolerogenic immature (i)DCs to mature (m)DCs with high stimulatory and migratory capacities [3]–[7]. In previous studies, we found that the proteome of apoptotic T cells includes prominent caspase-cleaved cellular proteins and that a high proportion of distinct epitopes in these fragments (apoptotic epitopes) can be cross-presented by DCs to a wide repertoire of autoreactive CD8+ T cells [8]. Recent reports have confirmed the role of caspase cleavage in the processing and presentation of epitopes that are derived from apoptotic cells in different models [9]–[11]. In chronic HIV infection, these autoreactive CD8+ T cells correlate with the proportion of apoptotic CD4+ T cells in vivo and are involved in establishing polyclonal T cell activation that in the long run results in generalized T cell dysfunction/depletion [8]. In addition, apoptotic cells derived from activated T cells (in contrast to those derived from resting T cells or from non-lymphoid cells) retain the expression of CD40 ligand (L) and can then condition CD40+ DCs to acquire high capacities to prime or cross-prime autoreactive T cells [12], [13]. This mechanism is consistent with the evidence that the signals provided by CD40L+ apoptotic cells and not those provided by conventional apoptotic cells facilitate the emergence of autoreactive T cell responses to apoptotic self-antigens [12], [13].
Successful priming of naïve CD4+ or CD8+ T cells results in the generation of both effector memory T (TEM) cells expressing various differentiation programs (type-1, -2, -17), according to the environment in which they are exposed [14]–[21], and central memory T (TCM) cells that promptly proliferate and generate new waves of effector cells on demand [22]–[24]. The transcription factor T-box-containing protein expressed in T cells (T-bet) is the master regulator of the type-1 cell differentiation program that is associated with the production of IFN-γ, which is required for the development of protective immune responses against intracellular pathogens [15]. GATA-binding protein 3 (GATA-3) controls the development of the type-2 cell lineage that is characterized by the production of IL-4, -5, and -13, which is critical for immunity against helminths and other extracellular pathogens [15]. Retinoid acid-related orphan receptor (ROR)-γt in mice and the human ortholog RORC in humans represent the master regulators of type-17 cell differentiation that leads to the production of IL-17, which is specifically required for protection against several types of extracellular and intracellular bacterial infections [14], [16]–[18]. All these (type-1, -2, -17) functions can elicit either protective or harmful effects, depending on whether they are executed by pathogen-specific or autoreactive T cells or whether the pathogen-specific are involved during an acute resolving infection or a chronic infection, respectively.
Here we used the hepatitis C virus (HCV) infection as a human model of acute infection that generally undergoes chronic progression to verify whether CD8+ T cells that are specific for apoptotic self-epitopes have a distinct effector type-1, -2, or -17 phenotype, to distinguish which of them is associated with the fate of a viral infection (recovery versus chronicity), and to ascertain the mechanisms whereby these responses are induced and maintained.
We analyzed longitudinally the responses of 18HLA-A2+ patients with acute HCV infection. The follow-up ranged from the onset of acute disease (clinical onset) to 15–24 months (the sixth month being considered the time of conversion from an acute to a chronic infection). Of the 18 patients, 6 patients had a self-limited infection and 12 patients exhibited a chronic evolution of infection (Table 1). Initially, the effector responses were determined by the capacity of freshly isolated CD8+ T cells from either HLA-A2+ patients or healthy controls to form IFN-γ spots (in an enzyme-linked immunospot [ELISPOT] assay) within 4 to 6 hours (h) of contact with nine pools of synthetic apoptotic peptides (Table S1A–C), eight pools of HCV genotype 1c, or genotype 2c peptides selected for their capacity to bind the HLA-A2 molecule [8], [25], or nine pools of overlapping peptides spanning the entire sequence of the HCV genotype 3a (Table S2A–E). The different HCV genotype-related peptides were matched with the viral genotype infecting the single patients. Each peptide pool was tested in triplicate. The synthetic apoptotic peptides used were prepared according to the sequence of caspase-cleaved proteins that had been previously identified by the proteomic analyses of apoptotic T cells (i.e., fragments of actin cytoplasmic 1 [ACTB], heterogeneous nuclear ribonucleo protein [ROK], lamin B1 [LAM1], non muscle myosin heavy chain 9 [MYH9], vimentin [VIME], or proteasome component C2 [PSA1]) [8]. We found that the apoptotic (but not the viral) epitope repertoire recognized by IFN-γ+CD8+ TEM cells was significantly larger in patients undergoing chronic infection than in those undergoing recovery (Fig. 1A,B and Fig. S1). Interestingly, the mean number of IFN-γ spots promptly formed by CD8+ TEM cells in response to each pool of apoptotic epitopes (but not viral epitopes) was directly correlated with the viral (plasma HCV-RNA) load, thus supporting the relationship between these responses and chronic evolution (Fig. 1C). None of the 21 HLA-A2+ healthy donors exhibited significant effector responses against any of the apoptotic or viral peptides ex vivo (data not shown). The HLA-restriction of these responses was demonstrated both by blocking responses with an appropriate anti-class I mAb and by determining that no response was observed in HLA-A2− patients (data not shown).
We enumerated specific CD8+ T cells directly in the peripheral blood of HLA-A2+patients or healthy donors by using pentamers of HLA-A*0201 molecules complexed to either apoptotic (MYH9478–486, MYH9741–749, VIME78–87) or viral epitopes (NS31073–1081, NS31406–1415, Core132–140) that had been previously identified as the most immunogenic among all patients tested (Fig. 1D). Control HLA-A*0201 pentamers complexed to a non-natural irrelevant peptide were unable to stain CD8+ T cells in all peripheral blood mononuclear cells (PBMCs) tested (data not shown). The pentamer values were significantly higher in both in patients experiencing chronic infection and in patients with self-limited infection (at all the time points tested) than in 20 HLA-A2+ healthy donors (Fig. 1E,F). However, in contrast to the ELISPOT assay showing frequencies of IFN-γ+CD8+Tcells specific to apoptotic peptides significantly higher in patients experiencing chronic infection than in patients with self-limited infection (Fig. 1A), the total frequencies of either apoptotic or viral epitope-specific CD8+ T cells, as detected by pentamers, did not differ between patients undergoing chronic or recovery evolution at all the time points tested (Fig. 1D–F). This difference may be explained by the finding that each single pentamer+ cell population can simultaneously contain (rare) naïve T cells, many TCM cells and several types of TEM cells with the same epitope specificity, as well as T cells with a “stunned phenotype” (representing the reducing capacity of cells to perform effector functions) [26], whereas ELISPOT assay only identifies IFN-γ+ cells in our system. To detect different effector functions within the CD8+pentamer+ T cells, we analyzed the frequencies of freshly isolated CD8+pentamer+ T cells that produced a wide array of cytokines (IL-17, IFN-γ IL-4, IL-2 within a few h of contact with the relevant peptides and optimal concentrations of anti-CD28 mAb, which served as a surrogate costimulatory signal. Irrelevant cytokine production was observed when either apoptotic epitope- or viral epitope-specific CD8+pentamer+ T cells of 20 HLA-A2+ healthy individuals were stimulated with this procedure (data not shown). Importantly, apoptotic epitope-specific CD8+pentamer+ TEM cells promptly produced notable and sustained amounts of all the cytokines tested within a few h of contact with the relevant epitopes, much more in patients experiencing chronic infection than in those undergoing infection resolution (Fig. 2A,B), in all time points tested (Fig. S2). By contrast, the virus-specific CD8+pentamer+ TEM cells produced lower amounts of the same cytokine in both categories of patients without any differences between them(Fig. 2A,B and Fig.S3). Peptide dose-response curves of cytokine-producing CD8+pentamer+ cells emphasized this difference (Fig. 2C,D). Time course analyses, performed longitudinally throughout the follow-up in all patients, revealed that the frequencies of polyfunctional apoptotic epitope-specific CD8+ TEM cells were significantly higher in patients experiencing chronic infection (Fig. 3A,B). These responses were sustained over time in relation to the sustained viral load (HCV-RNA copies) and alanineaminotransferase (ALT) levels only in patients who evolved into chronic infection (Fig. 3A,B). Then, the majority of these cell frequencies, as well as the serum biomarkers of viral hepatitis, tended to decline considerably later in patients who evolved into chronic infection than in those resolving infection (Fig. 3A,B). By contrast, no substantial difference was revealed in the time course of the virus-specific effector response between the two categories of patients (Fig. 3C, Fig. S3). Notably, the polyfunctional responses in the majority of patients were maintained by the parallel presence of different antigen-specific CD8+ T cell subsets, each of which produced a single cytokine in all time-points tested (mixed polyfunctional populations) (Fig.S4A,B). Therefore, the minority of patients showed cells simultaneously producing significant amounts of IFN-γ and IL-17 (type 1/17 cells), or cells simultaneously producing significant amounts of IL-17 and IL-4 (type 2/17 cells) (Fig. 2A,B and Fig. S4A,B). Importantly, the frequencies of CD8+pentamer+ TEM cells promptly producing IFN-γ or IL-17 in response to the relevant apoptotic epitopes, but not to the viral epitopes (data not shown), were directly correlated with the plasma viral load or the serum ALT levels(Fig. 4A–D).
Fresh apoptotic epitope-specific CD8+pentamer+ TEM cells promptly produced IFN-γ or IL-17 ex vivo within a few h of contact with DCs that had been pulsed with apoptotic T cells (i.e., through the cross-presentation mechanism) (Fig. 5A,B). The cross-presentation resulted in a marked decrease in IFN-γ or IL-17 production when apoptotic cells had been previously treated with a selective caspase-3 inhibitor (C3I) (Fig. 5A,B). This phenomenon was confirmed in five independent patients (Fig. 5B). DCs alone, despite known to endogenously express high levels of the ubiquitous (long-lived) cellular proteins (vimentin, non-muscle myosin, actin, heterogeneous nuclear ribonucleoprotein, lamin B1…) (14),were unable to directly stimulate the related specific CD8+ T cells (Fig. 5A,B).The frequencies of apoptotic epitope-specific CD8+ T cells (but not those of viral epitope-specific CD8+ T cells [data not shown]) correlated with the number of circulating apoptotic T cells (Fig. 5C).The percentage of apoptotic T cells in PBMCs was significantly higher in patients than in the 20 healthy donors tested (11.0±7.7 versus 3.9±3.9; P<.001).
To verify if type-17 CD8+ TEM cells specific to apoptotic self-antigens in the long run acquire functional plasticity in vivo [15], [18], [27]–[29], we monitored (from the clinical onset of infection up to 24 months) selected patients showing a notable number of CD8+ TEM cells promptly producing IL-17 within few h of contact with the relevant apoptotic epitopes at the clinical onset. During the course of the follow-up, the frequency of type-17 CD8+ TEM cells exhibited a progressive increase, followed by the emergence of type-1/17 cells in response to apoptotic epitopes (Fig. 6A). These responses were associated with both the maintenance of the type-17 transcription factor RORC and the appearance of the type-1 transcription factor T-bet (Fig. 6B,C). This scenario was observed both in the 3 patients showing type-17 CD8+ TEM cells specific for the MYH9741–749 epitope (Fig. 6), and in additional 3 patients showing type-17 CD8+ TEM cells specific for different self-epitopes epitope (data not shown). By contrast, representative fully polarized type-1 CD8+ TEM cells strictly maintained this phenotype throughout the follow-up period in all patients studied (Fig. S5online). To determine whether antigen-specific type-17 CD8+ T cells can reprogram their phenotype and convert into type-1/17 CD8+ TEM cells in vitro(situation which may mimic the type-17 conversion into type-1/17 phenotype in vivo), we used anti-CCR6 and anti-CCR4 mAbs [30] to sort IL-17–producing cells from antigen-stimulated CD8+ T cells (purity >98% type-17 pentamer+CD8+ T cells) (Fig. 6D). These cells were then restimulated in vitro with irradiated autologous PBMCs (acting as antigen-presenting cells [APCs]) that had previously been pulsed with the relevant peptide in the presence of either a mixture of IFN-γ and IL-12 (polarizing toward the type-1 phenotype) or a mixture of TGF-β, IL-6, IL-23, and IL-1β (polarizing toward the type-17 phenotype) [14], [31]. After 10–12 days of culture in IL-2 conditioned medium, the cells were tested for their capacity to produce both IL-17 and IFN-γ in response to the peptide plus APCs. Interestingly, CD8+ T cells that had been cultured in the presence of the type-17 polarizing cytokines maintained or increased the original type-17 phenotype, whereas CD8+ T cells that had been cultured in the presence of the type-1 polarizing cytokines switched (in a notable proportion) toward the type-1 phenotype (Fig. 6D).
To understand why the self-epitope-specific cells of patients undergoing resolution display significantly lower polyfunctional functions than patients experiencing chronic infection (Fig. 2A–D and Fig. 3A,B), we performed a series of functional experiments. First, we ruled out the possibility that apoptotic epitope-specific CD8+ T cells from patients undergoing recovery expressed intrinsic defects of effector cell functions. Indeed, apoptotic epitope-specific CD8+ T cells from the two categories of patients yielded similar cytokine responses after stimulation by polyclonal mitogens (i.e., phorbol 12-myristate 13-acetate [PMA] and ionomycin [iono]) (Fig. S6). Second, the majority of CD8+pentamer+ T cells (both apoptotic epitope-specific and viral epitope-specific) in both categories of patients were either CD45RO+CD127+ (TCM cells) or CD45RO+CD127− (TEM cells) (Fig. S7A,B). This finding suggested that the CD127−CD8+ TEM cells, which promptly produced the vast array of cytokines tested within a few hours of contact with the relevant peptides (Fig. S7C), were likely derived from the CD8+ TCM cells rather than naïve cells in both categories of patients in vivo.
Then, we evaluated whether the apoptotic epitope-specific CD8+ TEM cells were less polyfunctional in patients undergoing infection resolution because they were conditioned by a more severe programmed death (PD)-1-dependent exhaustion [32] in comparison to those from patients experiencing chronic infection. We found a similar PD-1 expression in apoptotic epitope-specific CD8+ T cells between patients undergoing infection resolution and those experiencing chronic infection (Fig. S8A).This result might argue against the possibility of a more severe PD-1-dependent exhaustion of apoptotic epitope-specific CD8+ TEM cells from patients resolving infection. However, we cannot exclude that the two groups of patients might express different levels of PD-1 ligands (i.e., in inflamed liver) that might provide different threshold of PD-1 dependent exhaustion in vivo. PD-1 upregulation was also shown in HCV-specific CD8+ T cells without any significant difference between the two categories of patients (Fig. S8B). To determine the functional capacity of PD-1 expression, we first selected PBMCs containing either viral epitope-specific PD-1+CD8+ T cells or apoptotic epitope-specific PD-1+CD8+ T cells from patients undergoing infection resolution or those experiencing chronic infection. In the presence or absence of a blocking anti-PD-L1 mAb or isotype control mAb in vitro, these cells were stimulated with the relevant peptide and anti-CD28. After 10 days of culture in IL-2-conditioned medium, cells were double-stained with the appropriate pentamers and anti-CD8 mAb, stimulated or not with autologous APCs plus the peptide, processed for intracellular cytokine staining (ICS) with mAbs to IFN-γ, IL-17, and IL-4, and analyzed by flow cytometry. Apoptotic epitope-specific pentamer+CD8+ T cells produced notable amounts ofIFN-γ or IL-4 after 10 d of stimulation, and even more in the presence of a blocking anti-PD-L1 mAb (Fig. 7A). By contrast, IL-17 production under the same conditions failed to increase but rather decreased during the 10 d of antigen stimulation in vitro, irrespective of the presence of anti-PD-L1, emphasizing the possible functional instability of this cell population (Fig. 7A). Cumulative experiments with PBMCs from a total of eight patients confirmed that the PD-1/PD-L1 blockade resulted in an increase of IFN-γ or IL-4 production by both apoptotic epitope-specific pentamer+CD8+ T cells (Fig. 7B) and viral epitope-specific pentamer+CD8+ T cells (data not shown), whereas the production of IL-17 was not affected. Importantly, the degree of increase in the responses exhibited by both apoptotic epitope-specific pentamer+CD8+ T cells and viral epitope-specific pentamer+CD8+ T cells was virtually the same between patients undergoing infection resolution and those evolving into chronic infection (data not shown). Taken together, these findings suggest the following. First, the apoptotic epitope-specific PD-1+CD8+ T cell responses are gently modulated by PD-1 because they are highly polyfunctional in patients experiencing chronic infection ex vivo. Second, the PD-1 blockade does not seem a principal cause of the decreased responsiveness exhibited by apoptotic epitope-specific CD8+ T cells from patients undergoing infection resolution in comparison to responses from patients who have developed a chronic infection, given that the degree of increase in the effector responses upon PD-1/PD-L1 blockade was very similar between the two categories of patients. However, we cannot exclude that other PD-1 ligands or the differential PD-L1 expression by inflamed liver can intervene in favoring T cell exhaustion or dysfunction in vivo, thus explaining the decreased responsiveness exhibited by apoptotic epitope-specific CD8+ T cells from patients undergoing infection resolution [33].
Finally, we hypothesized that differences in T cell receptor (TCR) avidity might account for the different apoptotic epitope-specific CD8+ T cell responsiveness between patients experiencing chronic infection and those undergoing infection resolution. To assess TCR avidity, we evaluated the dissociation kinetics of peptide/HLA-A*0201 pentamer binding to antigen-specific CD8+ T cells that were isolated from the two groups of patients [34]. Specifically, we stained fresh CD8+ T cells with saturating amounts of HLA-A*0201 pentamers that were complexed to either apoptotic or viral epitopes and an anti-CD8 mAb for 45 min at room temperature. Cells were then washed and incubated at 4°C with saturating amounts of an anti-HLA-A2 mAb to prevent rebinding of pentamers during the pentamer dissociation assay. The rate of decay was measured by flow cytometry at appropriate time points. We obtained linear decay plots of the natural logarithm of the normalized fluorescence versus time in all experiments performed, indicating that the pentamer decay was occurring stochastically and that the resulting pentamer staining half-lives(t1/2) should be proportional to the t1/2 of respective pentamer/TCR complexes (Fig. 8A). The t1/2 for apoptotic epitope-complexed pentamer staining to fresh CD8+ T cells from patients experiencing chronic infection was significantly longer than the decay of pentamer staining to CD8+ T cells from patients undergoing infection resolution (Fig. 8A,B). By contrast, the t1/2 for viral epitope-complexed pentamers to CD8+ T cells did not differ between patients experiencing chronic infection and those undergoing infection resolution (Fig. 8A,B). Control experiments in which HLA-A*0201 pentamers were complexed to a non-natural irrelevant peptide showed that the t1/2 for staining to CD8+ T cells was undetectable (data not shown).
Here we demonstrate for the first time that the multispecificity, magnitude, and polyfunctional (type-1, -2, -17)strength of CD8+ TEM cell responses directed to apoptotic self-epitopes were wide and robust during the acute phase of HCV infection, particularly in patients experiencing chronic progression compared with those undergoing infection resolution. The responses were directly correlated with the plasma viral load, the serum ALT levels or the number of circulating apoptotic T cells, and were then sustained over time in relation to the viral persistence. Our parallel study still in progress indicates that similar autoreactive CD8+ T cell responses in chronically infected patients are recruited in the inflamed livers (Fig. S9), are related with the signs of hepatic damage, and decrease in relation with the decline or the disappearance of the viral load upon antiviral therapy (interferon plus ribavirin@) (data not shown). Altogether these results suggest that strong CD8+ T cell responses against apoptotic self-epitopes arise and are maintained in HCV infection and may potentially contribute to the liver immunopathology through the production of high levels of inflammatory cytokines.
Recently, several models of chronic viral infection demonstrated that virus-specific CD4+ or CD8+ T cells producing elevated levels of IL-17 correlated with either viral persistence or a wasting syndrome with a multiple organ neutrophil infiltration [20], [35], [36]. Currently, our data suggest that the emergence of high frequencies of mixed autoreactive CD8+ T cells producing a broad array of cytokines (including IL-17) is prominent in patients undergoing chronic progression in the human model of acute HCV infection. By contrast, the frequencies of virus-specific effector cells (producing the different cytokines analyzed) were extremely low as compared with the apoptotic epitope-specific. Our data are coherent with the majority of studies revealing that the magnitude of HCV-specific CD8+ T cell effector responses does not correlate with the clinical or viral outcome in acute HCV infection [37]. In particular, HCV-specific CD8 T cells have been reported to express increased levels of PD-1 and an exhausted phenotype (weak proliferation, IFN-γ production, and cytotoxicity) [37]–[41].Although depletion studies in the chimpanzee model are consistent with a role of CD8+T cells as primary effector of protective immunity [42], studies in natural HCV infection were unable to find clear correlations between HCV-specific CD8+ T cell responsiveness and outcome of infection [37]–[41], [43], [44]. It is possible that the mechanisms that control HCV in the long term lie not exclusively on these conventional functions, but they are also displayed by some other subset of immune mediators, including HCV-specific antibodies [45]or CD4+ T cells [26], [40], [46].Consistent with this finding, in vivo depletion of CD4+ T cells from HCV-recovered chimpanzees abrogates protective CD8+ T cell–mediated immunity upon rechallenge [47], which suggests that CD4+ T cell help is required for the generation and maintenance of protective CD8+ T cells. Therefore, the viral immunological correlates of infection should be detected by multiparametric analyses (antibody, CD4+, CD8+ responses…) rather than individual analysis that may underestimate the multiple immunological variables related to infection outcome. In this respect, our study suggests that the apoptotic epitope- more than the virus-specific CD8+ cell responses discriminate patients with different infection outcome.
The observation that cross-presentation of apoptotic T cells by DCs requires caspase-dependent cleavage of apoptotic self-antigens to promptly activates specific CD8+ TEM cells ex vivo indicates that this mechanism might be operative in the induction of the related polyfunctional autoreactive responses in vivo. This possibility is emphasized by the finding that the frequencies of apoptotic epitope-specific CD8+ T cells correlated with the number of circulating apoptotic T cells. Consistently with our previous observations (14), cross-presentation of apoptotic cells plays a key role in activating autoreactive CD8+ T cells, as caspase-dependent cleavage of cell-associated (long-lived) proteins (such as vimentin, non-muscle myosin, actin, heterogeneous nuclear ribonucleoprotein, lamin B1…) is required to efficiently target the related fragments to the processing machinery of DCs. By contrast, live DCs alone, despite known to express the whole form of the same ubiquitous (long-lived) cellular proteins(14), are unable to stimulate the related specific CD8+ T cells by direct presentation mechanism, likely because they do not possess the caspase-cleavage program required for the presentation of these proteins (14).Collectively, these data suggest that these autoreactive CD8+ T cells may perform their functions through the by-stander effect of the pro-inflammatory cytokines upon cross-presentation of apoptotic cells rather than by the direct killing of cells endogenously expressing the related self-antigens. The strong production of IFN-γ and IL-17 may favor the triggering of recruitment of inflammatory cells, which contribute to the immunopathology [48]–[50].Our study provides a possible explanation for why the enormous expansion of activated T cells, during persisting viral infections, is only minimally attributable to virus-specific T cells [8]. Inflamed tissues (including the HCV-infected liver) are generally infiltrated by several billions of activated lymphocytes and the rate of apoptotic cells derived from them by far exceeds that originated by the turn-over of epithelial cells (i.e., hepatocytes) [51]. The demonstration that apoptotic cells derived from activated T cells (in contrast to those derived from epithelial cells) are CD40L+ and then condition CD40+ DCs to prime T cells [12], [13], suggest that they are the most important source of apoptotic self-antigens capable to cross-prime CD8+ T cell responses in an inflamed microenvironment. However, we cannot exclude that also apoptotic hepatocytes may amplify this phenomenon in an inflammatory context, because they might potentially generate the same caspase-cleaved antigenic fragments described in apoptotic T cells [8], and be cross-presented by DCs.
Recent data have clearly stressed the importance of infections in inducing and maintaining autoimmunity [52]. In particular, the initial emergence of apoptotic antigen-specific T cells in acute HCV infection may be dependent on virus-specific T cells that can provide the first waves of apoptotic substrates, upon performing their effector function. This mechanism may be maintained in patients evolving towards the viral persistence, and be further amplified by the apoptotic antigen-specific T cells themselves providing further waves of apoptotic antigens. Additional studies, even in appropriate experimental models, are required to ascertain the role of these autoreactive responses in the chronic evolution of infection. Moreover, it could be of interest to investigate if an expansion of mucosal associated invariant type-17 CD8+ T cells may participate in the high IL-17 production [53], as well as if other additional cytokines including IL-10 [54] may increase the polyfunctionality of apoptotic epitope-specific CD8+ T cells.
IL-17 production can account for the transient intra-hepatic infiltration of neutrophils found only in the early phase of acute HCV infection [55]. Then, these responses timely decline likely through the simultaneous presence of autoreactive type-1, -2, and -17 CD8+ TEM cells, which may regulate each other, and even the capacity of type-17 CD8+ TEM cells to express a certain degree of plasticity [15], [18], [27]–[29], and to convert to type 1/17 CD8+ TEM cells. Therefore, the environmental setting during an acute inflammatory disease seems to be addressed to guarantee the coexisting polarization of type-1, -2, -17, -1/17 CD8+ TEM cells, and even type-2/17 CD8+ TEM cells, likely to limit excessive damage by fine-polarized type-1 or type-17 CD8+ TEM cells. In this context, it is intriguing the observation that the polyfuctional autoreactive CD8+ T cells express high PD-1 levels in vivo and are limited only partially by the inhibitory PD-1 capacity in vitro. A possible explanation of this is that these autoreactive CD8+ T cells are primed later and thus submitted to a less prolonged antigenic stimulation than the virus-specific, that in contrast show a profound exhaustion in patients with HCV infection [32]. Recently, the persistent antigenic stimulation has been demonstrated to cause down regulation of T-bet, which results in more severe exhaustion of virus-specific CD8+ T cells [56]. However, the tuning of autoreactive CD8+ TEM cell functions by PD-1, as well as the high frequencies of apoptotic epitope-specific CD8+ TEM cells producing IL-4, may contribute to limit excessive functional responses over time [19], [54], [57], [58].In support of this hypothesis, our study in patients with long-term chronic HCV infection demonstrates that liver-infiltrating CD8+ TEM cells specific to apoptotic self-epitopes produce levels of cytokines significantly lower than patients with acute hepatitis (data not shown).
An important facet of our findings is that they demonstrate a link between the TCR avidity of autoreactive CD8+ T cells and the difference in the responsiveness of apoptotic epitope-specific CD8+ T cells exhibited by patients experiencing chronic infection and those undergoing infection resolution. The dissociation kinetics of peptide/HLA-A*0201 pentamer binding to antigen-specific CD8+ T cells clearly demonstrated that the t1/2 for apoptotic epitope-complexed pentamer staining to CD8+ T cells from patients experiencing chronic infection was significantly longer than the decay of pentamer staining from patients undergoing infection resolution. The t1/2 for pentamer staining was detected on freshly isolated CD8+ T cells, suggesting that TCR avidity measured by this system likely reflects what occurs in vivo. In the original study, this methodological approach made it possible to postulate that T cells with TCRs that bind peptide/MHC complexes for a longer duration are selectively preserved in comparison to T cells that express TCRs with lower avidity [34]. Recent studies suggested that in response to different microbial infections, initially naïve T cells with a wide range of avidity are efficiently recruited and expanded [59], [60]; subsequently, those with lower avidity undergo premature contraction, whereas those with higher avidity are selected because of a more prolonged expansion and correlate with protection [59], [60]. Our results provide an additional challenge to this model, and demonstrate that the TCR avidity of autoreactive CD8+ T cells specific for apoptotic self-epitopes was significantly higher in patients undergoing chronic infection than in those resolving infection. The selection of the autoreactive CD8+ T cells with higher avidity likely occurs because of a sustained stimulation by apoptotic antigens [8]. By contrast, lower avidity CD8+ T cells in the presence of weaker stimuli would undergo rapid contraction, as seen in the peripheral blood of patients with self-limited HCV infection. The viral persistence may provide the conditions that influence the availability of sustained apoptotic antigenic stimuli. However, our model does not exclude the possibility that cross-reactive CD8+ T cells, even expressing dual TCR [61], may intervene in this process.
Finally, our results may provide an important platform for the design of innovative therapeutic strategies to re-engineer protective immune responses in persisting infections. In addition, further studies will ascertain whether polyfunctional CD8+ T cells that are specific to apoptotic epitopes could predict chronic infection in other acute (i.e., HBV or HIV) infections that develop viral persistence. The detection of these autoreactive CD8+ T cells may also be relevant in determining whether the contraction or the quality variation of the polyfunctional responses can be used as biomarkers to verify the protective effects of conventional or innovative antiviral therapies [62], [63].
The study cohort included 18HLA-A2+patients with acute HCV infection (5 women, 13 men, median age 34 years, range 22–57 years), according to the ethical guidelines of the 1975 Declaration of Helsinki and priori approval by the Ethics Committee of the Italian National Institute of Health: written informed consent was obtained from all patients. Diagnosis of acute HCV infection was based on (1) high levels of serum ALT; (2) seroconversion assessed by third generation enzyme linked immunosorbent assay, or anti-HCV positivity at the time of the diagnosis with an anti-HCV negative test in the previous 12 months; (3) presence of HCV-RNA in at least the first serum sample, and (4) sudden onset of liver disease symptoms. Alternative causes of acute hepatitis, such as other viral infection, autoimmunity, alcohol, drugs, and toxins were excluded. Patients with concomitant immunological disorders or with HIV coinfection were also excluded from the study.
PBMCs were isolated and T cell clones were generated as previously described [64]. CD8+ T cells were purified from PBMCs by positive selection coupled to magnetic beads (MiltenyiBiotec) as previously described [54]. Flow cytometry analysis demonstrated >99% CD8+ cells in the positively purified population and <5% in the CD8-depleted population. To purify antigen-specific type-17 CD8+ cells, PBMCs were stimulated with the relevant peptide plus anti-CD28 (4 µg/ml) (BD Pharmingen). Then, cells were stained with allophycocyanin (APC)-labeled anti-CCR6 (R&D System) and phycoerythin-cyanine 7 (PE-Cy7)-labeled anti-CCR4 (BD Pharmingen) and processed with FACSAria (Becton Dickinson) to sort CCR6+CCR4+ cells: >98% of these cells both produced IL-17 in response to the relevant peptide and were susceptible of staining with the relevant pentamers, as detected below. Spontaneous apoptosis of PBMCs from patients was determined by staining with Annexin-V (ApoAlert Apoptosis Kit, Clontech Laboratories Inc), propidium iodide (PI) (Sigma-Aldrich) and PE-Cy7-labeled anti-CD3 (BD Pharmingen) before and after 18 h incubation at 37°C. Immature (i)DCs were derived from peripheral monocytes that had been purified by positive selection with anti-CD14 mAb coupled to magnetic beads (MiltenyiBiotec). CD14+ cells were incubated for 5 days in RPMI 1640 medium containing 5% FCS, 2 mM glutamine, 1% nonessential amino acids, 1% sodium pyruvate, 50 µg/ml kanamycin (Gibco BRL), 50 ng/ml rGM-CFS (Novartis Pharma), and 1000 U/ml rIL-4 (gently provided by A. Lanzavecchia, Bellinzona, CH). Mature DCs were obtained by a 40-h stimulation of iDCs with soluble rCD40L molecules (Alexis Biochemicals, Alexis Corporation). The definition of monocyte-derived DCs was based on their surface phenotype profile by staining with anti-CD14, anti-CD86 (Caltag Laboratories), anti-CD1a, anti-CD1c, anti-CD11c, anti-CD32, anti-CD80 (BD PharMingen) mAbs, Annexin-V (ApoAlert Apoptosis Kit, Clontech Laboratories Inc), PI (Sigma-Aldrich), and the appropriate secondary labeled antibodies (BD PharMingen).
Highly purified CD8+ T cells (1×105) from PBMCs were stimulated for 4–6 h with nine independent pools of apoptotic peptides (Table S1A–C), eight independent pools of viral-peptides (genotype 1b), eight independent pools of viral-peptides (genotype 2c) [8], [25], or nine pools of overlapping peptides spanning the entire HCV genotype 3a (Chiron Mimotopes) (Table S2A–E), and irradiated autologous CD8-depleted PBMCs, used as APCs, and tested by an ELISPOT assay, as described [8]. Each peptide pool contained 5 µg/ml of each single peptide.
PBMCs were incubated with APC-labeled–HLA-A*0201 pentamers (complexed to vimentin78–87 [LLQDSVDFSL], non-muscle myosin478–486 [QLFNHTMFI], or non-muscle myosin741–749 [VLMIKALEL] peptide) for apoptotic epitopes and APC-labeled–HLA-A*0201 pentamers (complexed to HCV-NS31073–1081 [CINGVCWTV], HCV-NS31406–1415 [KLVALGINAV] or HCV-Core132–140 [DLMGYIPAV] peptide) for viral epitopes (ProImmune Limited, Oxford, United Kingdom), in FACS buffer (PBS 1×, 2% human AB serum) for 10 min at 37°C, followed by washing and further incubation with APC-Cy7-labeled mAb to CD8 (BD Pharmingen, San Diego, CA), fluoresceinisothiocyanate (FITC)-labeled anti-PD-1 (BD Pharmingen), PE-labeled anti-CD127 (BD Pharmingen), FITC-labeled anti-CD45RO (Caltag Laboratories, Burlingame, CA) for 20 min at 4°C. Negative controls were obtained by staining cells with an irrelevant isotype-matched mAb. Cells were washed, acquired with a FACSCanto flow cytometer and FACSDiva analysis software (Becton Dickinson) or FlowJo software version 7.5.5 (Tree star, Inc. San Carlos, CA).
PBMCs were stained with pentamers and mAb to CD8, and then stimulated with or without different concentrations of the corresponding uncomplexed peptide (ProImmune Limited) plus anti-CD28 mAb (4 µg/ml) (BD Pharmingen), or with PMA (50 ng/ml) plus ionomycin (1 µg/ml) (Sigma Aldrich, Milan, Italy), for 6 h at 37°C. At the 2nd h, 10 µg/ml Brefeldin A (Sigma-Aldrich) was added. Cells were washed, fixed and permeabilized using Cytofix/Cytoperm solution (BD Pharmingen) at 4°C for 20 min, re-washed with Perm Wash Buffer (BD Pharmingen), and intracellularly stained with different combinations of Alexa Fluor 488-labeled anti-IL17A (eBioscience San Diego, CA), PE-labeled anti-IFN-γ, PE-labeled anti-IL-2, FITC-labeled anti-IL-4 (BD Pharmingen), PE-labeled anti-RORC (eBioscience) or purified anti-T-bet (Santa Cruz Biotechnology Santa Cruz, California) for 30 min at 4°C. When stained with unlabeled specific antibody to detect T-bet, cells were washed and stained with the appropriate secondary FITC-labeled antibody. Cells were washed, acquired with a FACSCanto flow cytometer and FACSDiva analysis software (Becton Dickinson) or FlowJo software (Tree star).
Cloned CD8+CD95+ T cells (10–100×106) were incubated in the presence or absence of 14 µg/ml C3I (Z-DEVD-FMK), or a negative caspase control (K, Z-FA-FMK) (BD Pharmingen) for 1 h at 37°C in a 24-well plate. Then, cells were induced to apoptosis by incubation with 500 ng/ml anti-Fas (anti-CD95 mAb [clone CH11], Upstate Biotechnology) for at least 6 h. Apoptotic cells were determined by staining with Annexin-V (ApoAlert Apoptosis Kit, Clontech Laboratories Inc), PI (Sigma-Aldrich) and flow-cytometry analysis. PBMCs were double-stained with pentamers and mAb to CD8 and cultured with iDCs that had been pulsed or not with apoptotic cloned T cells. After 6–8 h, cells were tested for their capacity to produce IL-17 and IFN-γ by ICS as described above. Cells were washed, acquired with a FACSCanto flow cytometer and analyzed with FACSDiva analysis software (Becton Dickinson) or FlowJo software (Treestar).
PBMCs were stained with saturating amounts of APC-labeled-HLA-A*0201 pentamers and APC-Cy7-labeled-CD8 (BD Pharmingen) for 45 min at room temperature [34]. Then, cells were washed three times with buffer (2% FCS, 0.01 sodium azide in PBS) and resuspended in 500 µl of buffer with saturanting amounts of mAb to HLA-A2 (BB7.2, ATCC). At various time points (0, 30 min, 1 h, 2 h and 3 h), an aliquot cells was washed and the fluorescence intensity was determined by flow cytometry analysis. Double staining using an anti-human TCRα/β (BD Pharmingen) and pentamers was performed in parallel to normalize pentamer fluorescence against the expressed TCR. The values were then normalized to percent of the total fluorescence at the initial time point and plotted on a logarithmic scale. t1/2 are determined by calculating the (ln2)/mean slope value of plots of the natural logarithm (ln) of the pentamer fluorescence normalized for the TCR fluorescence. The slope is equivalent to ln(Fa/Fb)/t, where Fa is the normalized fluorescence at the start of the interval, Fb is the normalized fluorescence at the end of the interval, and t is the length of the interval (minutes).
PBMCs were incubated for 10 days at 37°C with specific peptides (apoptotic or viral peptides), human rIL-6 (50 ng/ml), rIL-1β (10 ng/ml), rIL-23 (50 ng/ml) and rTGF-β (10 ng/ml) (R&D Systems) for the Th17 cell polarization. For the Th1 cell polarizing condition, PBMCs were antigen-stimulated in the presence of recombinant human rIL-12 (10 ng/ml) and rIFN-γ (100 U/ml) (R&D Systems). Recombinant IL-2 was added on day 4 of culture (50 U/ml). On day 10, cells were stained with surface antibodies, pentamers, anti-IL-17A, anti-IFN-γ, anti-IL-2 and anti-IL-4 mAbs. Cells were washed, acquired with a FACSCanto flow cytometer and analyzed with FACSDiva analysis software (Becton Dickinson) or FlowJo software (Tree star).
All statistical analyses were performed with Prism 4 (GraphPad) software using nonparametric Spearman's correlation test, nonparametric Mann-Whitney U-test for unpaired data and Wilcoxon test for paired data. The differences were considered significant at P<0.05.
actin cytoplasmic 1 [ACTB] P60709
heterogeneous nuclear ribonucleoprotein [ROK] P61978
lamin B1 [LAM1] P20700
non muscle myosin heavy chain 9 [MYH9] P35579
vimentin [VIME] P08670
proteasome component C2 [PSA1] P25786
|
10.1371/journal.pgen.1001063 | Transcriptional Regulation by CHIP/LDB Complexes | It is increasingly clear that transcription factors play versatile roles in turning genes “on” or “off” depending on cellular context via the various transcription complexes they form. This poses a major challenge in unraveling combinatorial transcription complex codes. Here we use the powerful genetics of Drosophila combined with microarray and bioinformatics analyses to tackle this challenge. The nuclear adaptor CHIP/LDB is a major developmental regulator capable of forming tissue-specific transcription complexes with various types of transcription factors and cofactors, making it a valuable model to study the intricacies of gene regulation. To date only few CHIP/LDB complexes target genes have been identified, and possible tissue-dependent crosstalk between these complexes has not been rigorously explored. SSDP proteins protect CHIP/LDB complexes from proteasome dependent degradation and are rate-limiting cofactors for these complexes. By using mutations in SSDP, we identified 189 down-stream targets of CHIP/LDB and show that these genes are enriched for the binding sites of APTEROUS (AP) and PANNIER (PNR), two well studied transcription factors associated with CHIP/LDB complexes. We performed extensive genetic screens and identified target genes that genetically interact with components of CHIP/LDB complexes in directing the development of the wings (28 genes) and thoracic bristles (23 genes). Moreover, by in vivo RNAi silencing we uncovered novel roles for two of the target genes, xbp1 and Gs-alpha, in early development of these structures. Taken together, our results suggest that loss of SSDP disrupts the normal balance between the CHIP-AP and the CHIP-PNR transcription complexes, resulting in down-regulation of CHIP-AP target genes and the concomitant up-regulation of CHIP-PNR target genes. Understanding the combinatorial nature of transcription complexes as presented here is crucial to the study of transcription regulation of gene batteries required for development.
| Different cell types in multi-cellular organisms are determined by the repertoire of genes active in each cell. This repertoire, or transcriptome, is established by the coordinated activity of transcription factors and cofactors that form modular transcription complexes. The modular nature of transcription complexes complicates our understanding of how transcription factors shape the transcriptome. CHIP/LDB transcription complexes direct formation of various cell types including blood and nerve cells. CHIP/LDB malfunction leads to developmental defects and cancer. The function of these complexes depends critically on the docking of specific transcription factors and co-factors at a specific time and in a specific cell type, making them outstanding models for intricate transcriptional regulation. Here we demonstrate that loss of SSDP, a key regulatory component of CHIP/LDB transcription complexes, alters transcription of a large group of genes. We used bioinformatics tools and genetic tests to examine the function of additional components of CHIP/LDB transcription complexes and their target genes during the development of specific organs. We demonstrate how differences in the availability of transcription factors in different cells can affect the function and composition of CHIP/LDB transcription complexes.
| The intricate regulation of gene expression in multi-cellular organisms involves an elaborate collaboration between repertoires of cis-regulatory sequences and modular, multi-protein transcription complexes that bind them (reviewed in [1]). Transcription complexes are now viewed as being composed of relatively ubiquitous core elements and a variety of context-dependent cofactors that interact with the core elements to regulate context-specific transcription (reviewed in [2]). An increasing number of such cofactors are being identified and the diverse roles of each transcription complex is thought to depend on the unique combination of associated cofactors (reviewed in [1]–[3]).
A prime example for this combination of general and specific factors are complexes formed by transcription factors that interact with cofactors of the CHIP/LDB family. CHIP is a Drosophila gene product that is closely related to the LDB (alias CLIM or NLI) proteins that have been well preserved in evolution all the way from Caenorhabditis elegans to man. These multi-adaptor proteins mediate interactions between different classes of transcription factors and additional co-regulators of transcription (reviewed in [4]). One of the best studied CHIP/LDB complexes is the Drosophila CHIP-APTEROUS complex (Figure 1A). APTEROUS (AP) is a LIM-homeodomain (LIM-HD) transcription factor [5] homologue of mammalian LHX2 and LHX9 [6]. The CHIP-AP complex is composed of a dimer of CHIP molecules [7], each of which binds one molecule of AP [8], [9] through a LIM interacting domain (LID) [7], [8] and one molecule of single-stranded DNA-binding protein (SSDP) through a CHIP/LDB conserved domain (LCCD) [10]. In the fly, this complex triggers a signaling cascade that specifies the dorsal compartment of the wing imaginal disc and serves to define the dorsal/ventral boundary at the adult wing margin (reviewed in [11]).
CHIP-AP complex function is negatively regulated by the Drosophila LIM-only (dLMO) protein (Figure 1B). dLMO binds CHIP in vitro and competes with AP for binding to CHIP [9]. This cofactor exchange is crucial for the proper function of the CHIP-AP complex during wing imaginal disc development as evident from the analysis of mutant and transgenic flies [7]–[9], [12]–[15].
An additional level of regulation is introduced by concomitant protein-protein interaction and cofactor exchange with non-LIM transcription factors (Figure 1C). Specifically, CHIP and dLMO form an alternative complex together with a GATA family transcription factor, PANNIER (PNR), and the beta-HLH transcription factors ACHAETE (AC), SCUTE (SC), and DAUGHTERLESS (DA) [16]. We refer to this complex as CHIP-PNR. One function of the CHIP-PNR complex is directed toward thoracic macrochaete (sensory bristles) differentiation (Figure 1D). The pattern of sensory bristles reflects the distribution of precursor sensory mother cells in the wing imaginal disc. These precursors are specified during the third larval instar and early pupal stages from a restricted group of cells that express ac and sc [17]. The expression of ac and sc, in turn, is regulated in part by the CHIP-PNR complex [16].
In the context of the CHIP-PNR complex, dLMO is a positive regulator [18], [19] and DNA binding is mediated through the GATA and beta-HLH transcription factors. There is a complex antagonistic relationship between CHIP-PNR and CHIP-AP, as the interaction between CHIP and PNR prevents CHIP from forming the homodimer that is crucial for the function of the CHIP-AP complex. Indeed, the function of the CHIP-PNR complex is antagonized by AP [16].
Like the CHIP/LDB encoding genes themselves, the components, assembly, and function of CHIP/LDB-based complexes appears to be highly conserved [6], [13], [20], [21]. For example, complexes containing SSDP1, LDB1 and LHX2 or LHX3 (termed LDB-LHX) are found in the mouse pituitary cell line alfaT3-1 [22], and a complex containing LDB1, GATA-1, LMO2, TAL1 and E47 (termed LDB-GATA) regulates erythropoiesis in mice [23]-[25].
SSDP proteins play a crucial role in the formation, stability and function of CHIP/LDB-based complexes in flies and mice [10], [13]. SSDP proteins promote assembly of LDB-LHX and LDB-GATA complexes and contribute to their transcription activity. Moreover, proteasome-mediated turnover of LDB1, LHX and LMO proteins is inhibited by formation of a complex with SSDP proteins [22], [26], [27]. Thus, the functional interaction between LDB and SSDP proteins appears to be independent of the specific composition of LIM or non-LIM proteins within the complex.
While the function of CHIP/LDB complexes depends on SSDP, the function of SSDP proteins in turn depends on interaction with CHIP/LDB complexes: both in flies and in mammals SSDP proteins do not contain a nuclear localization signal and have to bind CHIP/LDB in order to enter the nucleus [10]. Thus, SSDP proteins are key components of CHIP/LDB complexes in both functionality and specificity. CHIP/LDB and SSDP are therefore a valuable model for studying the intricacies of transcriptional regulation at the genomic level. Here we address genome-wide effects of Drosophila SSDP on the transcriptional activity of CHIP/LDB-based complexes. Using a combination of microarray analysis and genetic interaction tests we identified novel genes downstream of SSDP that affect the development of wing and thoracic bristle development. Using transcription factor binding site analysis, we were able to show that SSDP makes distinct contributions to the transcriptional activity of the CHIP-AP and the CHIP-PNR complex.
We have conducted a genomic search for putative SSDP target genes using Drosophila microarrays [28] to report expression of 14,142 predicted transcripts. Poly-A+ RNA was extracted from third instar larvae (males only to avoid potentially confounding sex-biased gene expression). We used two different heteroallelic combinations of ssdp hypomorphic alleles, ssdpneo48/ssdpBG1663 and ssdp31/ssdpBG1663, which allow survival up to the pupal stage [13]. We opted to use heteroallelic combinations of ssdp on different genetic backgrounds rather than homozygotes, in order to minimize inadvertent homozygosity for extraneous mutations. The heteroallelic mutant pairs were compared to each of the corresponding single heterozygotes (Table S1). We identified 189 candidate target genes that were differentially expressed between experimental and control samples (FDR corrected p<0.05; Table S2). Since SSDP is believed to be a positive transcriptional regulator of the CHIP/LDB complex [10], [13], we expected most of the target genes to exhibit lower expression in the ssdp mutants compared to the heterozygous controls. Interestingly, only a third of the 189 genes met this expectation (Table S2). These results might suggest that SSDP has a hitherto unidentified negative transcriptional regulatory effect on certain genes. Alternatively, secondary targets may be negatively regulated by direct targets of SSDP.
One way of testing for direct targets of SSDP is to look for enrichment for SSDP binding sites in the upstream regions of the 189 putative target genes. SSDP was first identified due to its ability to bind a single stranded poly-pyrimidine sequence present in the promoter of the chicken alfa-2(I) collagen gene [29]. Our gel shift experiments showed that this binding site is specifically recognized by fly SSDP (Figure 2).
We searched for enrichment for putative SSDP binding sites in the 500 bp upstream region of the 189 candidate genes identified in the microarray work, using two algorithms, PRIMA [30] and DEMON, We found SSDP binding site enrichment upstream the 189 candidate genes (p = 0.037) using DEMON. Interestingly, the SSDP binding site was even more significantly enriched (p = 0.02) among the genes down-regulated in the mutants, while there was weak significance among the genes up-regulated in the mutants (p = 0.17). This is consistent with the accepted role for SSDP as a positive transcriptional regulator. These data suggest that a significant number of the genes down-regulated in mutants are indeed direct targets of SSDP.
In order to determine whether SSDP target genes are also likely CHIP/LDB target genes, we searched the same upstream regions for binding sites of AP and PNR, transcription factors known to function in the CHIP/LDB complex. Binding site matrices for all available insect transcription factors (including the AP binding site) were obtained from TransFac [31], and a matrix of PNR binding sites, not included in TransFac, was constructed [32]. Strikingly, the PRIMA algorithm detected impressive enrichment for both AP (p = 0.04) and PNR (p = 4.64E-07) binding sites. Enrichment for the latter was also detected by DEMON (p = 2.64E-05). Interestingly, the enrichment for the AP binding site was lost when the down- (PRIMA: p = 0.085 and DEMON: p = 0.67) and up- regulated (PRIMA: p = 0.33 and DEMON: p = 0.21) gene groups were analyzed separately. This suggests that both groups harbor genes that are targeted by AP. The enrichment for SSDP and AP binding sites in the genes down-regulated in ssdp mutants is in agreement with SSDP functioning as a positive cofactor of the CHIP-AP complex. In contrast, the PNR binding sites were significantly enriched in the genes up-regulated in the mutants (PRIMA: p = 2.59E-05 and DEMON: p = 6.1E-06) but not in the genes down-regulated in the mutants (PRIMA: p = 0.085 and DEMON: p = 0.27). This suggests that a significant number of the genes up-regulated in ssdp mutants are direct targets of PNR. Given that AP and PNR bind to CHIP competitively during Drosophila thorax formation [16], we suggest that loss of SSDP disrupts the normal balance to favor CHIP-PNR complex formation. This would result in the down-regulation of CHIP-AP target genes and the simultaneous up-regulation of the CHIP-PNR target genes. Furthermore, up-regulated AP target genes may be regulated by both complexes. For example, AP and PNR are both known to positively regulate the expression of stripe, a key gene regulating development of the wing imaginal disc [33], [34].
In addition to the expected enrichment for the SSDP, AP and PNR binding sites upstream of the candidate target genes, we found enrichment for several other binding sites in the upstream regions of these genes (see Table S3 for p-values and binding sites information). Whether the function of all of these transcription factors is dependent on, or independent of, SSDP and/or of the CHIP/LDB transcription complexes remains to be determined. However, several of them have already been implicated in CHIP/LDB complex function (see Discussion).
The fact that the 189 putative SSDP target genes identified in our microarray experiments are enriched for binding sites of SSDP itself and its known partners in transcription is an independent orthogonal validation of the microarray results. These data encouraged us to ask if these putative targets have a genetic function in developmental events mediated by CHIP/LDB.
The analysis of SSDP target genes suggested that they are targeted by both CHIP-AP and CHIP-PNR complexes. To simplify the interpretation of genetic tests, we chose to begin looking for functional interactions between SSDP target genes and the CHIP-AP complex in the wing, where pnr is not expressed [35].
In the wing imaginal disc the CHIP-AP complex is involved in determination of the dorsal compartment. The edge of the CHIP-AP domain is the dorsal/ventral (D/V) boundary which will later give rise to the adult wing margin. Subtle disruption of the transcription activity of the CHIP-AP complex causes irregularities in the D/V boundary, which are evident as notches in the adult wing margin [12], [36], [37]. Indeed, such disruptions occur in the over-expression allele, DlmoBx which encodes a negative regulator of the CHIP-AP complex.
DlmoBx mutants have been previously shown to genetically interact with various ssdp loss-of-function alleles [13]. Thus, the DlmoBx2 allele provides a sensitized background to determine whether SSDP target genes function in D/V boundary formation. An example of the assay is depicted in Figure 3. Since Dlmo resides on the X chromosome, heterozygous females have a considerably less severe notching than hemizygous males (Figure 3). The wing notching phenotype displays a characteristic distribution of severities [12] allowing us to delicately determine the extent of genetic interactions by scoring enhancement or suppression of the wing notching phenotype by the Wilcoxon signed-rank test. The DlmoBx2 wing phenotype was subdivided into six severity classes, where Class 1 represents flies with the least severe (wild type wings) and Class 6 represents the most severe wing notching. The control distributions were of DlmoBx2/+ females and DlmoBx2/Y males (Figure 3A and 3B, respectively). As expected, when the DlmoBx2 mutation was combined with a heterozygous null mutation of ap, such as apUGO35 (DlmoBx2/+; apUGO35/+ or DlmoBx2/Y; apUGO35/+), the wing notching phenotype was enhanced, as evidenced by a shift of the distribution towards the more severe phenotypic classes in the double-heterozygous flies. Flies heterozygous for apUGO35 alone (apUGO35/+) had normal wings. As expected due to the lack of pnr expression in this tissue, the pnr loss of function allele, pnrV1, did not interact genetically with DlmoBx2 in our assay (Figure S1A and S1B).
If CHIP-AP transcriptional activity was synergistically reduced by mutations in Dlmo and ap, leading to the down-regulation of target genes of the CHIP-AP complex, then loss of function mutations in the target genes themselves (i.e. DlmoBx2/+; “target gene−”/+ and DlmoBx2/Y; “target gene−”/+) might have a similar effect on the DlmoBx2 wing notching phenotype. This is indeed the case with fringe (fng), a known CHIP-AP target gene in the wing disc [38], which shows reduced expression in DlmoBx2 mutant larvae dorsal wing pouch cells [9]. Double heterozygotes for DlmoBx2 and fng80 [38] exhibit a more severe wing notching phenotype than DlmoBx2 alone, just as observed for the interaction of DlmoBx2 and apUGO35 (Figure S1C and S1D). Control fng80/+ flies have normal wings.
We tested 39 genes from our original set of 189 SSDP candidate target genes in this genetic interaction assay with the DlmoBx2 mutation (Table 1). These genes had publicly available mutant strains and their differential expression were evenly distributed (ranging between 0.00019 and 0.049 FDR-corrected p-values, Figure S2) in our array experiments. The mutations used were usually single transposable elements insertions, and where possible two independent mutant strains per gene were tested (allele-specific interactions are shown in Table S4). Strikingly, twenty eight of these genes (72%) interacted genetically with DlmoBx2 (Table 1). This is a very high rate of agreement between the microarray results and the genetic interaction assay. In comparison we observed only 30% genetic interaction between DlmoBx2 and a random set of 20 chromosomal deletions. These chromosomal deletions encompass 322 genes that are not included in the 189 SSDP target genes, such that the “background” interaction rate per gene is considerably less than 30%. These results indicate that a large number of the genes identified by the microarray are bona fide SSDP targets and have genetic functions in the CHIP-AP transcription complex pathway during wing development. As expected, most of the interacting target genes (25, i.e. 89%) enhanced the wing notching phenotype of DlmoBx2 and only three (11%) suppressed it. In comparison, the interactions observed with the random set of deletions always suppressed DlmoBx2. Thus, loss-of-function mutations in SSDP target genes have a similar effect on DlmoBx2 as loss of function mutations in ap and in its previously known target gene fng. This is consistent with a negative regulatory role for dLMO with respect to the CHIP-AP complex [7]-[9], [12], [14], [15]. SSDP target genes that failed to interact with DlmoBx2 may be targets that are not dose sensitive, interact in different temporal or spatial contexts, or false positives.
Genetic interactions between ssdp and Chip or DlmoBx in a double heterozygous state are readily detected in the wing [13], but analogous genetic interactions between ssdp and loss of function alleles of ap are not. Therefore, to study the interactions between SSDP and CHIP-AP we needed another assay. We therefore explored using the only available dominant allele of ap, apXa, as a sensitized background. This mutant exhibits severe wing notching in a heterozygous state. We examined apXa/+ versus apXa/+; ssdpL7/+ flies, and observed augmentation of wing notching phenotype in the double heterozygous flies (Figure 4). In a population of apXa/+ flies, two classes of wing notching phenotypes can be distinguished (Figure 4A, 4B, and 4G) whereas the apXa/+; ssdpL7/+ flies exhibited three more severe wing notching classes (Figure 4D–4G). The apXa mutant is a gain of function allele [39], but its exact effect on the activity of the CHIP-AP complex is unknown. Our observation that apXa/+; ssdpL7/+ flies exhibit more severe wing notching than apXa/+ flies suggests that apXa causes reduced activity of the CHIP-AP complex, similar to DlmoBx2. These results clearly establish a genetic interaction between ssdp and ap, and indicate that apXa is useful for examining genetic interactions between candidate SSDP target genes and ap.
We tested seven of the SSDP target genes in the apXa/+ background (apXa/+; target gene−/+). Mutations in the katanin-60, CG12163 and Myofilin genes ameliorated the wing notching phenotype of apXa/+ flies, whereas CG11893 and Xbp1 mutations exacerbated wing notching (CG1518 and Cyp6d4 did not show an overt genetic interaction with apXa). These data indicate that both SSDP and SSDP target genes interact with AP and are therefore likely to act in a common pathway. Interestingly, the SSDP target genes enhanced the apXa wing notching less severely than ssdp itself, suggesting that the effect of SSDP is distributed among a large number of SSDP targets.
The genetic interactions with DlmoBx2 and apXa demonstrated that the SSDP target genes we identified are likely regulated by the CHIP-AP complex. Next we used genetic interactions to directly test our hypothesis that loss of SSDP disrupts the balance between the CHIP-AP and CHIP-PNR complexes in favor of the latter. To look at this balance between complexes, we examined thoracic bristles where both complexes function [16].
The CHIP-PNR complex positively regulates formation of thoracic sensory bristles via direct binding to the ac/sc enhancer. This CHIP-PNR function should be antagonized by AP since PNR and AP compete for binding of CHIP [16]. Consistent with our hypothesis, that loss of SSDP disrupts the balance between these two complexes, we found that both ssdpL7 and Chipe5.5 mutants display duplication of scutellar bristles as heterozygotes (<30% and <20% penetrance, for ssdpL7/+ and Chipe5.5/+, respectively, data not shown), a phenotype similar to gain of function alleles of pnr [35]. Flies heterozygous for pnrVX6 alone have normal number of scutellar bristles. We therefore expected that double heterozygous flies (ssdpL7/+; pnrVX6/+) would exhibit reduced occurrence of scutellar bristle duplications due to the lower levels of pnr. Indeed, duplicated scutellar bristles phenotype was abolished in ssdpL7/+; pnrVX6/+ flies. Thus, reduced levels of pnr rescued the duplicated bristle phenotype of a loss of function ssdp mutant, supporting the antagonistic model for CHIP-PNR and CHIP-AP complex formation.
This model predicts that mutations in the SSDP target genes will have a similar phenotypic effect as altering the balance between CHIP-PNR and CHIP-AP complexes. To test this prediction we used the ssdpL7 and Chipe5.5 mutations as a sensitized background to screen the SSDP target genes for modifiers of scutellar bristle formation (ssdpL7/+; target gene−/+ and Chipe5.5/+; target gene−/+). Given the opposing roles of the CHIP-AP and CHIP-PNR complexes in this tissue we expected SSDP target genes to either enhance or suppress the duplicated scutellar bristles phenotype of ssdpL7/+ and Chipe5.5/+ depending on which of the two complexes regulates that particular SSDP target.
Mutations in twenty eight SSDP target genes were tested as double heterozygotes with either ssdpL7 or Chipe5.5 (allele-specific interactions are shown in Table S5). A total of 23 of them were found to interact with either ssdpL7 or Chipe5.5 (Table 1). Fourteen genes (52%) interacted genetically with ssdpL7 and the same number of genes interacted genetically with Chipe5.5. Five genes (17.8%) interacted with both. This impressive rate of interaction suggests that SSDP targets are regulated by either or both CHIP complexes. The rate of interaction with CHIP and SSDP mutations in bristles is somewhat lower than what we observed for interaction with DlmoBx2 in the wing. However, this is not surprising as SSDP target genes may be regulated by either AP or PNR or both, which might make bristles more robust to perturbation and thus make it harder to detect genetic interaction in the thoracic bristles compared with the wing, where only AP is present.
Among the 23 interacting SSDP target genes, mutations in 12 were found to partially suppress the duplicated scutellar bristle phenotype suggesting that they are positive regulators of scutellar bristle formation (Table 1). Conversely, mutations in 11 interacting SSDP target genes enhanced the duplicated scutellar bristle phenotype, suggesting that they are negative regulators of bristle formation (Table 1). Interestingly, ten of the 12 suppressors affected the Chipe5.5 bristle phenotype and only five affected the ssdpL7 bristle phenotype (three genes suppressed both Chipe5.5 and ssdpL7 phenotypes). In contrast, nine of the enhancers affected the ssdpL7 bristle phenotype while only four enhanced the Chipe5.5 phenotype (two genes enhanced both Chip and ssdp bristle phenotypes). Thus, it appears that loss of ssdp has a predominant effect on genes that negatively regulate scutellar bristle formation. This finding is consistent with our microarray and transcription factor binding site enrichment analyses which showed that loss of ssdp function resulted in down regulation of the CHIP-AP target genes, and with the antagonistic effect of AP on bristle formation. In contrast, although CHIP functions as a cofactor for both AP and PNR, the Chipe5.5 mutation was more useful than the ssdpL7 mutation for identifying genes that are positive regulators of scutellar bristle formation. The reason for this difference is unknown, but given the complexity evident when comparing the interactions and function of CHIP/LDB complex in just two tissues, it is likely that further complexity remains to be discovered in other contexts. The salient point is that our genetic interaction results demonstrate a clear modularity of the regulation of SSDP target genes by CHIP/LDB complexes in different tissues. Understanding this type of context-dependent component shuffling in transcription complexes will be required for a full understanding of transcriptional networks.
Our genetic screens described above tested the ability of heterozygous mutations in SSDP target genes to cause subtle changes in the dominant phenotypes of DlmoBx2, apXa, ssdpL7 and Chipe5.5 in the wing and scutellar bristles, respectively. Next we wished to determine whether the SSDP target genes identified are essential for proper development of these structures. The simplest way is to examine mutations in SSDP target genes in a homozygous state. Unfortunately, those mutations which were homozygous viable did not exhibit any wing or thorax morphological defects. For example, the CG2604EY05974 mutation enhanced the DlmoBx2 wing notching in a double heterozygous state (Table S4). Yet, in an otherwise wild type background, CG2604EY05974 homozygous flies are viable and do not have any wing or thoracic morphological abnormalities (not shown). It is possible that these genes participate in, but are not essential for, wing and thorax formation, or that the mutations used to test for function were weak hypomorphs. For example, CG2604EY05974/Df(3R)ED5147 exhibit ectopic wing veins (Figure S3) indicating that at least some of the failure to find homozygous mutant phenotypes is due the use of classic hypomorphic mutations.
Several of SSDP target gene mutations we used were homozygous lethal prior to adulthood, precluding examination of wing or thorax phenotypes. To avoid difficulties due to pleotyropic affects on viability, we utilized the transgenic GAL4/UAS system for targeted silencing of the SSDP target genes [40]. This approach offered two advantages: First, the UAS-RNAi constructs that were used are gene-specific. Second, expression of the UAS-RNAi can be targeted to a subset of cells depending on the GAL4 driver used while the rest of the cells maintain normal expression of the target gene, thus avoiding lethality. The ap-Gal4 [41] and pnr-Gal4 [35] drivers drive reproducibly high levels of UAS-lacZ transgene expression in cells known to express ap and pnr respectively, within the wing disc. Thus, by combining the transgenic constructs (ap-Gal4/+; UAS-RNAi-target gene/+ or pnr-Gal4/+; UAS-RNAi-target gene/+) we silenced SSDP target genes in either ap- or pnr-expressing cells. We knocked down nine SSDP target genes that interacted with DlmoBx2, apXa, ssdpL7 and Chipe5.5. Silencing of two of them had profound effects.
Silencing of Xbp1 (a.k.a. CG9415) in ap-expressing cells resulted in semi-lethality. Survivors reaching adulthood developed severely disrupted wings which appeared as small amorphic inflated structures, accompanied by marked excess of bristles on the wing and scutum, while the scutellum was not affected (Figure 5B, 5C, and 5E). As expected by the pattern of pnr expression in the adult fly [35], silencing of Xbp1 in pnr-expressing cells caused a similar excess of bristles that were limited to the mid-line of the scutum while the wings were not affected. Interestingly, no extra bristles were observed on the scutellum, and some of the flies even exhibited a reduced number of scutellar bristles (Figure 5D). These observations indicate that Xbp1 has opposing roles in regulating bristle development in the scutum and scutellum.
Silencing of G-salpha60A (a.k.a. CG2835) in ap-expressing cells caused a curled wing phenotype (Figure 5F). In addition, silencing of this gene in pnr-expressing cells reversed the orientation of the posterior pair of scutellar bristles (Figure 5G). It is therefore obvious that these two SSDP target genes are essential for normal wing and thorax development. The remaining seven SSDP target genes tested in this manner exhibited variable effects on the number of scutellar bristles and at very low penetrance. Given the large number of SSDP target genes and the likely robustness that this facilitates, some weak effects are expected. Combinatorial knock down experiments, much like the large set of double heterozyote tests we report here, will be required to piece these genes together into a more developed model. Importantly, like the CHIP-AP and CHIP-PNR complexes themselves, SSDP target genes show context-dependent effects on development.
Drosophila SSDP was identified on the basis of its ability to bind the nuclear adaptor protein CHIP/LDB [10], [13]. Both nuclear localization of SSDP [10] and its ability to modulate the transcription activity of the CHIP-AP complex during wing development [10], [13] depend on its interaction with CHIP/LDB. In the present study we have implemented a combination of molecular, bioinformatic and genetic approaches that allowed us to gain insight into the effect of SSDP on the transcriptional activity of CHIP/LDB complexes and their role in development. We have conducted a genome wide screen for SSDP target genes in Drosophila using expression microarrays with mRNA isolated from larvae bearing hypomorphic alleles of ssdp. Our analysis of transcription factor binding site enrichment served as an orthogonal assay that validates and extends the microarray results and thus contributes to our understanding of the relation between the CHIP-AP and CHIP-PNR transcription complexes in specific tissues (e.g. wing and thorax).
SSDP proteins directly bind DNA [29], and mouse SSDP1 activates the expression of a reporter gene in both yeast and mammalian cells indicating that it is capable of regulating transcription activity [21], [42]. We found enrichment for SSDP binding sites [29] upstream of the genes identified in the microarray experiments on flies lacking SSDP. Moreover, in agreement with the positive transcriptional role of SSDP, enrichment for SSDP binding sites was restricted to the genes showing decreased expression in mutants. This strongly suggests that a significant number of these genes are bona fide SSDP target genes.
Consistent with the involvement of SSDP with the CHIP-AP complex, we found that upstream regulatory regions of the SSDP putative target genes are also enriched for the AP binding site [31] and the SSDP binding site. These sites are likely to be functionally significant, since loss of ssdp enhances the wing notching phenotype of a dominant allele of ap. Additionally, over-expression of Dlmo, whose product negatively regulates the CHIP-AP complex, also interacts with mutants of SSDP target genes, demonstrating that SSDP target genes are involved in the CHIP-AP pathway. The efficiency of finding genetic interactions among the genes differentially expressed in the microarray experiments, demonstrated the power of this approach. Specifically, 72% of the loci we tested with DlmoBx2 is more than an order of magnitude higher than an EP insertion screen (1.3% interacting) in a DlmoBx1 sensitized background [43]. Our combined microarray and genetic loss of function screen allowed the identification of a similar number of Dlmo-interacting genes by screening a much smaller group of putative target genes [43]. Of the 35 genes identified by Bejarano and colleagues only CG1943 was found in the 189 genes identified in our microarray screen. Our study specifically identified down-stream targets of SSDP, while those researchers searched for any modifiers of the Dlmo wing notching phenotype and thus uncovered genes that function in other regulatory pathways or genes that are upstream of the CHIP-AP complexes. This may explain the limited overlap between their results and ours.
In contrast to the enrichment of SSDP binding sites in the genes down-regulated in ssdp mutants we found the PNR binding site to be enriched specifically in the genes up-regulated in the ssdp mutants. We therefore present a model in which loss of SSDP disrupts the balance between the CHIP-AP and CHIP-PNR complexes. Mammalian SSDP proteins protect LDB, LHX and LMO proteins from ubiquitination and subsequent proteasome-mediated degradation by interfering with the interaction between LDB and the E3 ubiquitin ligase, RLIM. It is therefore possible that in the absence of SSDP proteins, CHIP/LDB and LMO can escape degradation by interacting with GATA and beta-HLH proteins that are not subjected to proteasome-mediated regulation [27]. The N-terminus of CHIP/LDB proteins is responsible for interaction with both PNR [16] and RLIM [22]. Thus, PNR/GATA proteins may partially interfere with the interaction between CHIP/LDB and RLIM making the CHIP/LDB-PNR/GATA complex more resistant to proteasome regulation and less dependant on the levels of SSDP proteins then the CHIP/LDB-LHX/AP complex.
According to our model, in cells where both the CHIP-AP and CHIP-PNR complexes are active, loss of SSDP should result in the same phenotype as over-expression of PNR. Indeed, we found that ssdpL7/+ flies display duplications of scutellar sensory bristles, similar to gain of function mutations in pnr. In addition, lowered levels of pnr in ssdpL7/+; pnrVX6/+ flies suppresses scutellar bristle duplication. This indicates that the duplicated scutellar bristle phenotype of ssdpL7/+ flies depends on the presence of PNR. As predicted by our model, since both AP and PNR regulate bristle formation, the functional interactions between SSDP target genes and ssdpL7 and/or Chipe5.5 resulted in either suppression or enhancement of the duplicated scutellar bristle phenotype.
Our results in flies indicate that SSDP contributes differentially to CHIP/LDB complexes containing AP versus PNR. By contrast, mouse SSDP proteins positively contribute to the transcription activity and assembly of both LDB-GATA and LDB-LHX complexes [13], [21], [22], [26], [27], [44], but the relative contribution of mammalian SSDP proteins to LDB complexes containing LHX proteins versus GATA proteins has not been specifically examined. It is possible that SSDP alters the balance of LIM-based CHIP/LDB complexes and GATA-containing CHIP/LDB complexes in the development of mice, as occurs in flies.
Our search for enrichment of transcription factor binding sites upstream of the putative SSDP target genes identified additional transcription factors that may warrant future study. Some of these factors are associated with SSDP and CHIP/LDB complexes. For example, the binding sites for PNR and ZESTE (Z) were both enriched in the up-regulated putative SSDP target genes. This is in agreement with previous studies showing that Z can recruit the BRAHMA (BRM, the Drosophila homolog of the yeast SWI2/SNF2 gene) complex [45] via its member OSA [46], which together negatively regulate the CHIP-PNR complex during sensory bristle formation through direct and simultaneous binding of OSA to both CHIP and PNR [47].
Some of the additional regulatory inputs at SSDP target genes may be evolutionarily conserved. For example, we found enrichment of STAT92E and SSDP binding sites in the down-regulated SSDP target genes. This may be significant, as a known role of ssdp is regulation of the JAK/STAT pathway during Drosophila eye development [48]. Interestingly, mammalian STAT1 confers an anti-proliferative response to IFN-γ signaling by inhibition of c-myc expression [49]. Similarly, expression of mammalian SSDP2 in human acute myelogenous leukemia cells [50] and prostate cancer cells [51] leads to cell cycle arrest and inhibits proliferation accompanied by down-regulation of C-MYC. These findings indicate that both in Drosophila and in mammals SSDP and STAT proteins have similar functions and may share common target genes.
While our transcription factor binding site analysis utilized all of the 189 putative SSDP target genes, our genetic screens were conducted on a subset of them due to the availability of mutants. This suggests that more genetic interactions will be found among the untested genes. Even among this more limited subset, there are interesting new stories that suggest future experimental directions. For example, an insertion mutation in the Xbp1 gene suppressed the duplicated scutellar bristle phenotype characteristic of ssdpL7/+ and Chipe5.5/+ flies, indicating that XBP1 contributes positively to bristle formation. In contrast, when Xbp1 was silenced in ap-expressing cells both the wings and the scutum displayed a marked excess of sensory bristles while the scutellum was not affected. These results suggest that in the wing and scutum XBP1 acts as a negative regulator of bristle formation. Silencing of Xbp1 in pnr-expressing cells caused a similar excess of bristle on the scutum, accompanied by a reduced number of scutellar bristles, further emphasizing the opposing effects of XBP1 in these two distinct parts of the thorax. Such contrasting phenotypes have been previously documented for several pnr mutants as well [35]. In flies and mammals XBP1 regulates the ER stress response, also termed the unfolded protein response (UPR, reviewed in [52], [53]). Since one of the functions of the ER is the production of secreted proteins, UPR-related pathways are widely utilized during the normal differentiation of many specialized secretory cells (reviewed in [52]). In this respect it would be interesting to examine whether SSDP and CHIP/LDB complexes affect the production of secreted morphogens, such as WINGLESS (WG), the secreted ligands of the EGFR receptor, SPITZ (SPI) and ARGOS (AOS), or the secreted NOTCH binding protein SCABROUS (SCA) (reviewed in [54]) via XBP1 during wing and sensory bristle formation. Alternatively, the transcription factor XBP1 may directly regulate the expression of genes required for differentiation of the wing and sensory bristles. Indeed, carbohydrate ingestion induces XBP1 in the liver of mice, which in turn directly regulates the expression of genes involved in fatty acid synthesis. This role of XBP1 is independent of UPR activation and is not due to altered protein secretory function [55]. Curiously, the two GO function categories ‘cellular carbohydrate metabolism’ and ‘cellular lipid metabolism’ which are enriched among Xbp1 target genes in mouse skeletal muscle and secretory cells [56] were also enriched in our list of putative SSDP target genes (Table S6). Whether this reflects a secondary effect due to the down-regulation of Xbp1 in ssdp mutants or a direct regulation of these processes by SSDP is yet to be determined.
Additional novel functions for CHIP/LDB complexes are implied by our results regarding the Gs-alpha60A (a.k.a. CG2835) gene. G protein coupled receptors are important regulators of development by for example, signaling via the protein kinase A (PKA) pathway [57]. Activation or inhibition of PKA signaling during pupal wing maturation perturb proper adhesion of dorso-ventral wing surfaces resulting in wing blistering. This phenotype may be due to miss-regulation of wing epithelial cell death [58] in ap-expressing cells [59]. Interestingly, similar wing blisters occur in the wing of DlmoBx2 flies. Moreover, we found that mutant alleles of Gs-alpha60A enhanced the wing blistering phenotype of DlmoBx2 (data not shown). Silencing of G-salpha60A in ap-expressing cells caused a curled wing phenotype. Such a phenotype can result from differences in the size of the dorsal and ventral wing blade surfaces. In addition, silencing of this gene in pnr-expressing cells caused the posterior pair of scutellar bristles to form in reversed orientation. Bristle orientation have been proposed to be regulated by planar cell polarity genes [60]. Taken together these results point to novel aspects of regulation of wing and sensory bristle development by SSDP and CHIP/LDB complexes mediated by G-alpha proteins.
Our genome-wide expression profiling and bioinformatics analysis of ssdp mutant larvae, combined with genetic screens enabled us to gain insight into the intricate context-dependent transcriptional regulation by CHIP/LDB complexes. We were able to identify 28 putative SSDP target genes that are involved in wing development and 23 putative SSDP target genes that play a role in scutellar bristle formation. Examination of two of these, xbp1 and Gs-alpha60A, suggests novel aspects of developmental regulation such as the involvement of SSDP and CHIP/LDB complexes in ER function and PKA signaling. Furthermore, we showed for the first time that SSDP proteins contribute differentially to transcription activity, and probably to the balance in formation of CHIP-AP and CHIP-PNR complexes. Furthermore we identified potential novel partners of SSDP in regulating transcription of downstream genes during fly development. It stands to reason that an extension of our genetic analysis to mammals and other vertebrates will reveal a host of additional functions of SSDP and CHIP/LDB during the multifaceted process of transcriptional regulation that underlies the development of multicellular organisms.
Unless otherwise stated, flies were grown on a standard medium containing cornmeal, yeast, molasses, and propionic acid at 25°C. The ssdp mutant strains (i.e ssdpBG1663, ssdpneo48 and ssdp31) used for the microarray experiment were previously described [13], all three were balanced on TM3-GFP (FBba0000338). The rev(ssdpneo48) line is a precise excision of the P element inserted in ssdpneo48. UAS-RNAi lines 18873, 38686, 38186, 24959, 24959, 6367, 40871, 9026, 12823 and 15347 were obtained from VDRC [61]. Chromosomal deletions Df(2L)ED49, Df(2L)ED548, Df(3L)ED231, Df(3L)ED4284, Df(2L)ED1109, Df(2L)ED299, Df(1)ED7067, Df(2R)ED2222, Df(3R)ED5156, Df(3L)ED4528, Df(2L)ED270, Df(2L)ED774, Df(2L)ED746, Df(3R)ED5187, Df(2L)ED673, Df(2L)ED120, Df(1)ED6957, Df(2L)ED19, Df(3R)ED5657 and Df(3R)ED10257, were obtained from the DrosDel collection [62]. All other fly stocks were obtained from the Bloomington Drosophila Stock Center (http://flystocks.bio.indiana.edu). Oregon-R flies were used as wild type.
RNA handling was performed exactly as described [65]. Briefly, larvae were flash frozen. Total RNA was extracted using Trizol (Life Technologies, Carlsbad, USA), followed by mRNA isolation using an Oligotex poly(A) extraction kit (Qiagen, Valencia, USA). RNA concentration was determined using RiboGreen dye (Molecular Probes, Oak Ridge, USA). RNA quality was determined by capillary electrophoresis using the 6000 Nano Assay kit (Agilent). All procedures were carried according to the manufacturer's instructions.
All genomic sequences were obtained from the UCSC genome browser (http://genome.ucsc.edu/, assembly Apr. 2006 for the D. melanogaster genome) [70]. The 500 bp upstream of the 189 candidate genes scanned using two algorithms termed PRIMA [30] and DEMON, for identifying enrichment of transcription factors binding sites in a set of co-regulated genes. Both methods require a background set for comparison (in this case all the annotated genes in Drosophila).
Analysis for enrichment of GO functions was conducted using the database for annotation, visualization and integrated discovery (DAVID) [71], [72]. Default setting were used and the enrichment cut off was set to p = 0.05 after FDR correction.
Fly ssdp was PCR amplified, cloned into pZEX plasmid and expressed with a GST tag in E.coli BL-21. Crude cell extract or purified GST-SSDP fusion protein was used for binding assays. GST-SSDP was purified on a glutathione agarose column (Sigma G4510). The ssdp single stranded CT oligonucleotide [29] was used as prob. Binding assays were carried out using the DIG Gel shift kit 2nd generation (Roche, Mannheim, Germany) according to the manufacturer instruction in a final volume of 20 µl containing labeled DNA (150 fmoles), 1 µl of poly-L-lysine and 3 µl poly-[d(I-C)], 140 ng cell extract. For competition experiments 90 or 360 ng of unlabeled probe were added. Following a 20 min incubation at room temperature, the binding reaction products were separated on a native 6% polyacrylamide gel in 0.5% TBE (pH = 8). The gel was contact blotted onto a Hybond-N+ membrane (Amersham Biosciences). The chemiluminescent detection was performed following the manufacturer's instructions (Roche, Mannheim, Germany). The membrane was exposed to X-ray film (FUJI) for 15 min at 37°C.
|
10.1371/journal.pgen.1007410 | Evolution via recombination: Cell-to-cell contact facilitates larger recombination events in Streptococcus pneumoniae | Homologous recombination in the genetic transformation model organism Streptococcus pneumoniae is thought to be important in the adaptation and evolution of this pathogen. While competent pneumococci are able to scavenge DNA added to laboratory cultures, large-scale transfers of multiple kb are rare under these conditions. We used whole genome sequencing (WGS) to map transfers in recombinants arising from contact of competent cells with non-competent ‘target’ cells, using strains with known genomes, distinguished by a total of ~16,000 SNPs. Experiments designed to explore the effect of environment on large scale recombination events used saturating purified donor DNA, short-term cell assemblages on Millipore filters, and mature biofilm mixed cultures. WGS of 22 recombinants for each environment mapped all SNPs that were identical between the recombinant and the donor but not the recipient. The mean recombination event size was found to be significantly larger in cell-to-cell contact cultures (4051 bp in filter assemblage and 3938 bp in biofilm co-culture versus 1815 bp with saturating DNA). Up to 5.8% of the genome was transferred, through 20 recombination events, to a single recipient, with the largest single event incorporating 29,971 bp. We also found that some recombination events are clustered, that these clusters are more likely to occur in cell-to-cell contact environments, and that they cause significantly increased linkage of genes as far apart as 60,000 bp. We conclude that pneumococcal evolution through homologous recombination is more likely to occur on a larger scale in environments that permit cell-to-cell contact.
| Bacteria shuffle their genes far less often than humans do and genes or traits are more directly linked with the singular bacterial parent cell rather than the two parents that are involved in sexual reproduction. However, bacteria do occasionally have sex in the form of homologous recombination by taking up external DNA and incorporating it into their genomes. This happens far less regularly than sexual reproduction happens in human generations but is a known way that bacteria undergo ‘Horizontal gene transfer’. This means that genes can be acquired without being inherited. In this study we show that this form of horizontal gene transfer is more likely to happen in certain environments over others in Streptococcus pneumoniae. In particular, we show that this is more likely to happen in environments that closely mirror the nasopharynx which is the natural habitat of S. pneumoniae.
| Streptococcus pneumoniae (the pneumococcus) is a paradigm for genetic transformation, in which cells become ‘competent’ for uptake of DNA through binding of a cognate competence stimulating peptide. The single stranded DNA that is taken up is then incorporated into the genome of the competent cell and can be detected as scattered segments of heterologous DNA. The consequences are of great relevance to human health, as they enable rapid adaptation to interventions such as antibiotic therapy and vaccination [1–4].
The pace of adaptation available through genetic transformation is limited both by the number of interactions and by the amount of DNA such an interaction can transfer, but neither of these parameters is well understood for any streptococcal species. Early estimates of the average size of recombined segments formed during transformation of pneumococcus in vitro varied over a range of 2–6 kb [5–7], although rare larger transfer events have since been detected in vitro with selective pressure [8]. WGS has enabled more comprehensive estimates of the sizes of DNA segments that can be transferred. In the first such global analysis in pneumococcus, determining the extent and frequency of replacements during transformation in vitro [9] with saturating concentrations of DNA, recombination events were limited to an average size of 2.3 kbp.
There is evidence however that recombination events observed in nature are larger. This could be a result of selection; for example, large-scale events were linked to vaccine escape variants with altered capsular loci, which emerged in the US from 2000–2007 [10]; one escape lineage (“P1”) displayed at least 9 replacements (sizes: 0.3, 0.5, 0.6,1.2, 2.3, 3.8, 26, 28, and 44 kb), all apparently from the same donor strain. A study of co-carriage of two strains in a single patient reported two large-scale multiple replacement events, each occurring within a 3-month time span [11]; one transferred three segments (sizes: 4.3, 13, 30 kb), while another transferred 6 segments (sizes: 0.8, 1.1, 4.3, 5.3, 6.5, and 20 kb). A survey of recombination events within a single geographically expansive clone of pneumococcus over 24 years mapped tracts of exchanged DNA over a wide range of sizes, from 2 to ~72,000 bp [1]. Similarly, two pneumococcal strains (PMEN3 and CGSP14) were identified that had acquired, respectively, 9 and 15 regions from PMEN1 or relatives, encompassing segments of 1.6–32 kbp or 4.6–22 kbp [4]. A survey of 173 isolates of a single lineage (CC320), mapped numerous recombination events, including an exchange of 78.8 kbp [12]. 228 ST127 strains from 3 lineages were reported to display 239 recombination events, with mean sizes varying widely among the lineages (4650, 21150, and 75156 bp) [13]. These examinations of products of gene transfer in nature are all retrospective analyses that are difficult to interpret clearly, as the strains were recovered after extended but poorly defined periods in which multiple exchange events might have accumulated. However, direct experimental studies of transformation in the natural environment are both rare and challenging to design.
To compare reports of recombination in nature with what occurs in vitro, we collected length distributions documented in six studies of events in natural populations and in one study of events in vitro (Fig 1). Although there is variation in event-calling algorithms and biological histories of the strains examined, two common features of the natural population studies stand out. First, small recombination events are most common. Second, the bulk of the DNA transferred is carried in larger transferred segments. The proportion of DNA transferred in segments above 10 kb ranges between 70 and 90% in 5 of the 6 collected studies. In contrast, less than 12% of the DNA transferred in vitro is found in larger fragments.
Some of the larger recombination segments observed in nature transfer cps cassettes, which are as large as 30 kbp, make one-step shifts in serotype, and are likely to be under strong selection. However, it is not known if strong selection is also the basis for other large transfers persisting in wild lineages. While transfer of a cps locus in vitro is certainly possible, as demonstrated previously [8] for segments as large as 47.8 kbp, the low yield of such replacements (1 per 1000 closely linked single-gene transfers) leads us to expect that they should be more rare relative to small events than we observe in nature.
An increase in WGS of the pneumococcus has made it possible to map many recombination events occurring during the evolution of a lineage; the larger of these inferred events has often attracted comments on lengths that seem long for products of genetic transformation. The largest events have led to speculation about unknown transfer mechanisms or special circumstances of transformation not reproduced in the classical laboratory genetic transformation experiments. Two classes of explanation have been advanced to account for the contrast between the prevalence of large transfer segments in nature and their paucity in vitro. One attributes it to the selective filtering recombination, in which events must survive in order to be fixed in a lineage. The other proposes that there are unidentified special conditions of transformation in nature that generate an enrichment in large transfer segments, prior to selective filtering. Distinguishing between these explanations would require characterization of gene transfer products immediately after transfer in nature, or in suitably natural conditions, before intervention of selective filtering.
Bacterial physiology variation in nature, such as reduction in mismatch repair efficiency or altered levels of recombination proteins, could account for the observed difference. However, the recent recognition that gene transfer is especially efficient within biofilms [14] led us to speculate that the key distinction could be that in natural populations gene transfer often occurs in biofilms. A notable difference between the classical transformation of pneumococcus with purified DNA and the natural environment of the nasopharynx is the potential in the latter for direct cell-cell contact. The ability of competent pneumococcal cells to kill non-competent pneumococci on contact, through action of a competence-specific surface-bound lysin, CbpD, and other lysins [15] might create an enriched micro-environment affording a source of closely related, concentrated, and largely intact DNA. Our hypothesis is that transfers during S. pneumoniae cell-to-cell encounters are often qualitatively different from the cell-DNA encounters previously studied in the laboratory. This might yield larger recombination events, and that they more closely resemble those occurring during the habitual growth of pneumococci on host tissue surfaces in multicellular aggregates (biofilms) where interacting cells are anchored in close proximity. If correct, this hypothesis would offer explanations for both perplexing features of natural pneumococcal gene exchange–the increased size of individual recombination events and the simultaneous modifications at multiple genomic sites.
To examine the mechanism of pneumococcal gene transfer in settings that can allow experimental manipulation but better model the natural biofilm growth habitat, we designed pairs of densely labeled recipient/donor strains with genetic markers that enable precise monitoring of cell-cell interactions, detection of gene transfer events, and recovery of recombinant progeny for comprehensive mapping of the products of those events. We report here that global mapping of exchanges in recombinants recovered from two models of cell-cell interaction indicates significantly increased rates of co-transfer over large genetic distances during such encounters.
To trace recombination events with high resolution, a pair of well-marked pneumococcal strains was designed. Features comprised (1) >10,000 SNPs distinguishing the strains, (2) a donor that is incapable of developing competence, (3) a recipient whose competence development depends on exogenous competence pheromone peptide (CSP), and (4) markers to allow selection of recombinants. The recipient CP2204 was a comA derivative of the reference strain R6 that is incapable of secreting CSP and therefore depends on exogenous pheromone for competence development. It also carries the robust selective marker, RifR. The donor was a competence defective comE deletion mutant that cannot sense CSP, with two selective markers, SpcR and NovR, introduced by transformation from R6 derivatives (S1 Table). The SpcR and NovR markers are unlinked during transfer by transformation, as they are separated by ~800 kbp. The construction of this strain pair is summarized in Materials and Methods. After this pair proved incompatible in biofilm culture, a second competent (but comA+) recipient strain, R36AKan (a low-passage ancestor of R6) was substituted for use in mixed biofilm cultures, where competence would depend on endogenous elaboration of CSP by the recipient. Resequencing of the new derivatives showed that both recipients were distinguished from the donor by ~16,000 SNPs (15,954 in the case of CP2204 and 16,065 in the case of R36AKan). In the fratricide model of transformation, the cell taking up the DNA is termed the ‘attacker’ as it has lysed the other, which is termed the ‘victim’. Here we use the more general terms ‘recipient’ and ‘donor,’ which are independent of the mode of interaction in which transfer occurs. However, the reader will appreciate the different cellular roles in the interaction contemplated and that the donation of DNA is not ‘voluntary’.
Transformation events in nature likely occur among mixed populations such as those growing on an epithelial surface, often forming a biofilm. It is not clear what to expect for recombination in a biofilm. Biofilms frequently contain large quantities of DNA, which might compete with any DNA released during fratricidal attacks [16], but if they also include nucleases this DNA may be highly fragmented. High rates of gene transfer have been reported within experimental pneumococcal biofilms [14], but the nature of the responsible recombination events was not investigated in detail.
Crosses were carried out under three different conditions: transformation with saturating purified DNA (similar to previous studies, [9]), transfer within mixed biofilm cultures (similar to Marks [14]), and transfer using a novel artificial biofilm-like assemblage (Fig 2). For direct comparison with DNA mediated transformation, a control experiment was carried out using the SNP-marked strain pair–CP2215 as source of 30-40-kb DNA fragments, and CP2204 as recipient. In transformation with saturating CP2215 DNA, copious NovR single-gene transformants were recovered, including detectable numbers of NovRSpcR double transformants (S2 Table). 22 independent double transformants were retained from the cross for mapping of transfer events. Although the transformation frequency for a single marker was only ~10−4, 0.5% of the SpcR recombinants were also recombinant for the unlinked marker, NovR. This high rate of co-transfer of two distant markers, termed congression [17], suggests that a minority of cells were highly competent, as this frequency would be expected to be no more than 0.01% for a recipient population of uniformly competent cells.
To ask whether large-scale recombination might be more common during cell-cell contact, we developed a synthetic biofilm-like assemblage in which competence could be controlled and recombinants directly and rapidly recovered. The assemblage consisted of culture mixtures collected from log-phase cultures onto a Millipore filter to a depth of ~20-cell thicknesses and incubated on the surface of an agar-solidified chemically defined medium (CDM). In this “filter assemblage”, recombinants were recovered after only an hour’s exposure to CSP (S3 Table). Among 2 x 1010 viable recipient cells recovered from the assemblage were 240,000 SpcR transformants (a frequency of 1.2 x 10−5), among which 1% were also NovR. The high congression rate again suggests that a minority of cells were highly competent. 22 independent NovRSpcR double transformants were retained from the cross for mapping of transfer events.
To ask whether large-scale recombination events also arise in biofilms, we recovered recombinants arising in the environment described previously [14], which uses co-culture in CDM to establish biofilms on a substrate of fixed confluent cells of a human lung epithelial cell line. We used a pair of strains with the known SNP differences described above (CP2215 and R36AKan), as initial trials showed that formation of biofilms with this pair of strains regularly yielded recombinants. Mature biofilms were developed after 48 h following inoculation of wells with 10,000 cells of each strain; but the recovered viable cells were a mixture of R36AKan and rare recombinants, with few remaining donors, consistent with widespread fratricide of CP2215 by R36AKan (S4 Table). Recovery of recombinants from the biofilm wells was variable, ranging from none to 0.1% of recipients (S4 Table). Despite the low rate of transformation, a high rate of congression, near 1%, again suggests the presence of rare competent cells. 22 independent recombinants were recovered for mapping of recombination events at the genomic scale. The retained clones represent selection for the recipient marker KanR in combination with NovR (14), SpcR (3), or CmR (3) markers individually, as well as two cases of NovRSpcR double transformation.
Recombinants’ DNA was sequenced at 300X coverage and SNPs were identified by alignment to a finished genome of the donor strain (CP2215). Circos [18] plots of SNPs transferred from donor to recipient (Figs 3, 4 and 5) show that most of the SNPs transferred from donor to recipient occurred as clusters. The locations of the recombination events in each recombinant strain were identified by mapping contiguous sets of donor SNPs as described in the methods (Fig 2) and are listed in S5 Table. To compare the outcomes of recombination in the three types of cross, we assembled statistics on three readily quantified aspect of the results: the number of individual transfer events per recipient, the mean and standard deviation of sizes of individual transfer events, the number of events >10,000 bp, the total amount of DNA transferred to a recipient, and the percentage of the total differentiating donor/recipient SNPs that were transferred from the donor into the recombinant (S6 Table). These are minimum estimates, because in those genetic regions where the donor and recipient are identical, no recombination can be detected. There was substantial variation in the total amount of recombinant sequence acquired within all three types of experiments, with over 5% of the genome transferred in some recombinants, but only 0.04% in others. The number of events per recombinant also varied considerably, ranging from a low of 2 to a maximum of 20.
The recombination segment lengths were exponentially distributed in all three environments (S1 Fig); the corresponding per base probabilities of recombination are listed in Table 1. The majority of events were small, with rarer large events, consistent with previous work on transformation in nature and in vitro [9]. However, the mean sizes of events in the filter assemblage (4051 bp) and biofilm co-culture (3938 bp) crosses were larger than those in crosses using saturating DNA (1815 bp). Standard pairwise T-tests of the normally distributed mean recombination event size in the three equally sampled datasets showed that the filter assemblage and biofilm samples were not significantly different from each other (p = 0.7926), but the mean recombination event size of the saturating DNA experiments was significantly different to both the filter assemblage (p = 5.602 x 10−9) and biofilm (p = 3.885 x 10−8) cases.
While the average number of recombination events per cell increased only modestly under conditions for cell-cell interaction, the mean length of events doubled, and the total bp transferred per recombinant approximately tripled (Table 1). A clue to the source of this difference in total DNA transfer is the 8-10-fold increase in segments longer than 10 kbp.
To visualize the size dependence of these recombination events, the dependence of event number and transferred DNA content on segment size is shown in Fig 6. The frequency of smaller segments is the same for transfers in all three modes, but the increase in the amount of transferred DNA during cell-cell encounters occurred entirely through greatly (8-10-fold) increased numbers of larger (8–30 kbp) segments. We interpret this pattern as indicating that the cell-contact mode of genetic transformation increases the incorporation of long segments (and tracts of nearby segments) because the competent cells encounter longer DNA molecules, reducing premature interruption of processive transport by encounters with DNA ends.
Inspection of the recombination maps revealed exchange events near genome position 700,000 bp in 8 out of 22 recombinants recovered from biofilm co-culture, but in none of the recombinants from the other crosses. Notably, such events were not restricted to specific antibiotic selections, as they were observed in recombinants selected separately on novobiocin, chloramphenicol, and spectinomycin plates. These events were in the region of CtpA (a copper transporter P-type ATPase) and SpxB (pyruvate oxidase). It is possible that growth in biofilms may have imposed strong selective forces, such as oxidative stress and stationary phase cycles, which may have selected for recombinants in this region. Effects of sequence polymorphisms in SpxB have been investigated in relation to hydrogen peroxide production in S. pneumoniae [19]. However, effects of polymorphisms in CopA, SpxB, or both that could explain such a selection in biofilm cultures remain unknown.
In crosses accomplished by genetic transformation, co-transfer of two distant markers into a single competent cell can occur independent of their proximity on the chromosome, by encounters with two separate DNA fragments, in the process termed congression. Linkage due to genetic proximity of two markers is revealed in practice by a frequency of co-transfer greater than the background rate of co-transfer by congression. In the present dataset, ~16,000 SNPs all served as genetic markers to evaluate the genomic extent of linkage to any selected marker. The incorporation of selective markers was typically accompanied by multiple additional transfers at widely separated sites throughout the genome, but tracts of tightly spaced recombination events were also observed (Figs 4 and 5). To visualize the effect of transfer mode on the span of linkage created by such tracts, Fig 7 displays the frequency with which each base in the vicinity of the nov-1 marker was inherited by any of the NovR transformants in the DNA cross, the filter assemblage cross or the biofilm crosses. It is immediately apparent that the span of linkage to NovR is much greater in the cell-cell contact crosses than with saturating donor DNA.
It has been suggested that differences between experimental recombination in vitro and what we see in nature may reflect the results of selection. For example, the major antigen in pneumococcus is determined by the capsule biosynthetic locus, which usually exceeds 10 kb in length. Large events at this locus are therefore more likely to change the serotype. In the PMEN1 dataset, containing over 240 isolates of a major intercontinentally disseminated resistant clone [1] the mean recombination event size was ~6kb. The largest event size detected in the PMEN1 dataset was 72,830 bp, while 20.3% of events were 10,000 bp or greater. We performed a similar analysis using the Gubbins algorithm [20] on 240 isolates collected by the CDC Active Bacterial Core (ABC) Surveillance system [21] and found a similar mean inferred recombination event size, 5589 bp (median event size 2963 bp, with a standard deviation of 7571). The largest event detected in the ABC dataset was 64,581 bp and 14.7% of events exceeded 10,000 bp.
To compare Gubbins-detected rates of recombination with our experiments, we must first investigate whether Gubbins can detect events as accurately as our mapping method. It should be noted that there are at least three systematic sources of bias toward detecting larger events in Gubbins [20]. First, as the algorithm detects elevated frequencies of SNPs in an isolate compared with an inferred ancestor, smaller regions would be less likely to be detected if donor/recipient SNPs are less dense in these samples than in our designed pairs of strains. This will raise the mean event size calculated both due to missing smaller events and due to omission of small intervening non-recombinant segments. Second, as discussed previously [20], the Gubbins algorithm is not able to differentiate clusters of recombination events that are separated by a few intervening recipient SNPs in the same way that our method does. Finally mismatch repair may have played a more important role in our study by preventing larger events. This could explain the clustering of events with intermediate recipient SNPs. Natural selection could take advantage of any larger recombination events arising from the failure of mismatch repair to provide a selective advantage that then survive in the population.
Comparison of Fig 1 with Figs 3, 4 and 5 shows that while cell-cell interaction increased mean segment size through greatly increased numbers of events in the 8–30 kbp range, it did not result in the even longer events reported elsewhere using Gubbins mapping tools. However, S5 Table details the results of applying the Gubbins algorithm to our dataset, with frequent clustering of individual events into very large events. The Gubbins algorithm repeatedly fails to recognize tracts of clustered individual events and instead groups them together, as seen, for example, in recombinant F4 where Gubbins groups 10 individual recombination events into one event over 90 kbp long. This provides one explanation as to why events are so long in retrospective evaluation of recombination events in published datasets: they may simply be groupings of smaller events that can be more accurately detected when the precise donor and recipient strains are known.
When Mostowy et al. [22] examined recombination events in a single pneumococcal clone (PMEN1), they pointed out that three of the four known mechanisms of horizontal gene transfer were poor candidate explanations for the contrast between in vitro transformation and the apparent frequency of large recombination events in population samples. Transduction and transfer by gene transfer agent (GTA) particles [23] are strictly limited by the carrying capacity of the transfer particle, and, although integrated conjugative elements (ICEs) can transfer multiple linked genes, such transfer is usually restricted to a specific region of the chromosome. The remaining more likely responsible mechanism is natural genetic transformation, which is widespread in pneumococcus, transfers genes equally well from all parts of the genome, and uses a mechanism without inherent size limits or sequence specificity. It was suggested that large events could arise from a sporadic suppression of mismatch repair mechanisms. We now see that the cell-cell interaction previously missing in in vitro transformation experiments can itself explain most or all of the contrast.
The distribution of recombination events described here raises the question of what mechanism(s) might lead both to the widely-spaced events and to the closely clustered ones. The widely-scattered events must represent multiple parallel entry initiations, as cells remain competent for a limited time, and no single transport pore could import over half of a complete genome within this time. The process generating longer tracts of recombination events is less clear. They may be the result of the uptake of a single long strand of DNA, portions of which are then separated and independently inserted by homologous recombination. Alternatively, multiple pores may initiate parallel uptake events from multiple initiating scissions within a single 50-100kb domain of DNA. However, coordination to achieve the consistent strand choice would be difficult to explain.
A simple model that could account qualitatively for the major features of these patterns and is consistent with previously established properties of the transformation mechanism posits just 4 mechanistic steps:
Such a model would explain both the excess of small replacement events, due to nuclease activity within an unprotected cell lysate, and the existence of an upper limit to the sizes of the larger tracts of recombination events. Under this model, it is expected that different factors affect the yield of recombinant events in different ways, according to fragment size. At the low end, for example, it is known that the efficiency of recombination declines progressively for fragments below ~2 kb, but is constant above ~8 kb [24, 25]. Gaps in tracts of recombination segments would be a natural consequence of competition by daughter chromosome arms for segments of a single internalized donor strand. At the high end, in contrast, an upper limit on the size of recombination event tracts would be set by a finite rate of transport combined with a limited temporal competence ‘window’, a window that appears to be determined quite stringently both by retro-inhibition of the CSP response by the competence-specific protein DprA [26] and by lability of some competence effector proteins. Such a strand could approach 50–100 kb, if transport continued for 20 min at 80 bp/sec [27, 28].
These results show that large-scale gene transfer events do occur in vitro when the primary source of donor genes is living target cells, and supports the hypothesis that cell-to-cell contact facilitates larger recombination events.
Complex recombination events may be key to understanding the exceptionally large events reported for the PMEN1 and CC180 lineages [22]. Clusters of recombination events mapped near the comE and nov-1 loci illustrate the potential for recombination episodes that simultaneously extensively modify the recipient genome over neighborhoods as large as 40–80 kb. Such recombination clusters appeared within recombinants from both filter assemblages and biofilm, but rarely or never during transformation with saturating DNA. Some recombination events identified within the PMEN1 and CC180 lineages were in the 20–80 kb range of sizes. Indeed, 50% of the total DNA transfer reported for each of these lineages was in events classified as 20 kb or larger. Many or all of these may represent clusters similar to those observed here but in which short internal recipient sequence blocks were obscured by the Gubbins detection algorithm in regions that are not densely populated enough with SNPs [20]. The same algorithm was used to re-analyze our experimental dataset to test event tract detection. The Gubbins software grouped many of these tracts into single large recombination events by ignoring the intervening recipient SNPs (S5 Table). This confirms that our mapping method provides a finer scale and more accurate measure of recombination events when the recipient and donor are known.
We note that the work here suggests that the frequency with which large recombination events occur, and hence the supply rate of selectable large-scale variation, may be higher than previously assumed within these antibiotic selected experiments for a limited number of samples (22 in each environment). Indeed, if these events are so relatively frequent that we readily observe them experimentally, why do they appear so comparatively infrequently in natural populations? One possible reason is that large insertions are more likely to have a negative fitness impact as a result of epistatic interactions with other parts of the genome [29]. In this scenario, large transfers happen at a high rate, but most such events are lost, leaving only the subset that provided some selective benefit or were neutral. There is previous evidence for this, in that in some cases making larger changes to the capsule locus appears to reduce growth rate [30], but not enough to indicate that it is a general mechanism. We note that taking up DNA from the environment may also serve purposes beyond adaptation, such as providing scarce nutrient [31] or helping with DNA repair [32]. If the majority of DNA is taken up for metabolism and rarely inserted into the genome, then competence would be maintained despite potential negative epistatic events of rare large recombination events. It is also important to remember that when cells come into contact with strains that are not as distantly related to themselves as the donor/recipient pair used here, recombination would introduce fewer SNPs and negative epistatic interactions would be less likely.
While this work suggests that larger fragments can be acquired under experimental conditions that more closely approximate natural co-colonization in the nasopharynx, it should be noted that the great majority of events remained relatively short, even in the biofilm. Importantly, this means that recombination is generally more likely to lead to the loss of accessory genes than their gain–consistent with a proposal that the function of homologous recombination is to remove parasitic elements [33]. However, our finding that larger recombination events occur alongside smaller ones suggests that while parasitic mobile elements can be effectively removed, this will happen at the same time as the acquisition of larger recombination events as part of the same process and from the same donor. If the majority of large recombination events have negative fitness consequences, the removal of the parasitic mobile element would be accompanied by the fitness cost of the large recombination event.
Does the pattern of large-scale recombination in population samples reflect the accumulation of separate transfer episodes, or coordinated simultaneous transfer of multiple DNA segments? Previous work has left this question open, as events had occurred within time windows of years [22] or months [10, 11]. This study shows that multiple and large transfers can occur within very much shorter temporal windows—a single 60-min competence cycle in the filter assemblages, and within 2 days in the experimental biofilm model. Thus, an inference that historical multiple large recombination events occurred within single or very few competence cycles is not unreasonable and explains instances where a single strain provided the source for multiple recombination events.
Within the wider context of other transformable bacteria, it is interesting to compare our results with those from similar genomic approaches to the study of recombination in Gram-negative pathogens. In particular, apparent clustering of insertion events into tracts has been reported in Haemophilus influenzae [34, 35] and Helicobacter pylori [36] (as well as previously in the pneumococcus [9]). However, the roles of biofilms have yet to be probed in experimental studies of these pathogens.
The present evidence that multiple recombination events can happen simultaneously means that we should not necessarily expect that recombination events are small, and occur at a rate of one per generation [37–40]. It is also clear that, although there was an obvious bias toward recombination events near the selected markers, simultaneous recombination events were widespread. This has resulted in a large proportion of the genome being transferred; indeed, in one isolate, as much as 5.8% (123,190 total bp) of the genome was transferred from donor to recipient, apparently in a single step, involving the insertion of 20 recombination segments. This amount is more than double the previous estimate of maximum recombination effect of recombination on the genome in S. pneumoniae (2.5% [9]); however a clinical long-term carriage study found that 7.8% of the genome could recombine [11]. The large amount of DNA that can transfer has obvious evolutionary implications and implies that evolutionary models for S. pneumoniae that attempt to encompass recombination should incorporate this scale of potential transformation.
The rate of recombination of a species is important for estimates of mutation and genomic change, and therefore the capability of a species to adapt to important selective pressures such as vaccines or antibiotics [41]. Current models of pneumococcal evolution incorporate regular short recombination events but not regular large recombination events [33, 42, 43]. While recent work on the impacts of recombination on epistasis indicated the proportion of the genome transferred in recombination as an important parameter [29], this work suggests it may be larger than would be assumed from previous experiments. There are several caveats of this observation that should be duly noted, measurements in nature will be effected by the loss of recombinants by negative selection, the potential enrichment of some recombinants by positive selection and a variable amount of growth of recipients that could also have biased these results. Nevertheless, this work suggests that the rate with which large recombination events occur is higher than has been suspected, and given that they appear to happen readily in experiment, but are comparatively rarely observed in nature [22], this may suggest that they are more likely to be selected against for reasons that will require further examination.
There are several limitations to this work. Our experimental model remains chemically and physically different from the natural environment of the nasopharynx, and we have worked with strains tractable to our purpose that may not be representative of the great diversity of pneumococcal lineages [44]. Indeed, our original pair did not effectively form a joint biofilm, and another suitable recipient strain had to be generated. This may suggest a reason for the observed and striking variation in recombination rate among pneumococcal lineages, where some seem to undergo no recombination at all despite possessing the necessary molecular machinery [45, 46]. This has been found to be associated with duration of carriage [13], but mixed biofilm formation is another variable that should be studied. Pneumococci that form mixed biofilms poorly with other strains may have fewer opportunities to acquire ‘foreign’ DNA in general and larger fragments in particular. Similarly if some lineages are more likely to co-colonize this may lead to a higher than expected rate of transfer between them [47].
Other limitations relate to aspects of the differing environments that vary in an unquantifiable way. For example, in the filter assemblage crosses, the ratio of donor cells to recipient cells was 1:5 (see S3 Table). Considering that the experiment was conducted for only 1h, the DNA ratio in the entire assemblage would remain mostly unchanged. In the biofilm context, this is harder to estimate, as the experiment lasted longer (48h). We can only take the input cell ratio as a guide to possibilities. The input ratio was 1:1 (see S4 Table). It has also occurred to us that the percentage of competent cells could influence the results; however, by our best estimates, competent cells were actually a minority in all three cross formats examined; so, if this aspect affects results, it may be expected to have the same effect in all three formats compared here.
The use of short read sequencing of the recombinants is a further limitation to the work. We have used long read sequencing and a finished genome of the donor strain to try to limit this but areas with too much genetic divergence (e.g., completely absent from the recipient genome) may encounter some mapping problems. Incomplete mapping can lead to misinterpretation of recombination event location or recombination start/stop position errors but cannot produce false positive recombination events.
This study provides significant evidence that cell-to-cell contact, such as occurs naturally in a biofilm, significantly increases the likelihood of pneumococcal strains acquiring larger recombination events. We have also shown that these events are more likely to be clustered and characterized by short inter-recombination event regions in cell-to-cell contact environments. This has important implications for the study of pneumococcal evolution and may explain the apparent increased rate of recombination of some lineages of S. pneumoniae by an increased tendency to form biofilms. This work provides an explanation of why we observe larger recombination events when measured in the nasopharynx than what is measured in classical in vitro studies. In general, the broader implication of our study is that larger recombination events may depend on the opportunity for cell-cell contact, with minimal degradation of the acquired DNA.
CP2204 is a CSP-dependent derivative of the R36A descendent CP2000 [48] (genotype: malM bgl hex Δcps rpsL1 hlpA::GFP::CAT comA::ermB rif; phenotype: Hex- Mal- SmR CmR Cps- GFP EmR RifR ComA-). R36AKan is a derivative of the R36A ancestor of CP2000 that was marked by a Kan insertion in rgg (genotype: R36A but rgg::kan; phenotype: Hex+ Com+ KanR). MD5037JANUS is a Δcps derivative of the clinical isolate MD5037 [30]. The donor was CP2215, a non-transformable derivative of MD5037, which was chosen as being separated from R6 by >10,000 SNPs (genotype: MD5307 but Δcps, str-1 hlpA::RFP::CAT comE::spc nov-1; phenotype: capsule-negative SmR CmR SpcR NovR Com-). The sources and identities of selective markers used are listed in S1 Table. THY was Todd-Hewitt Broth (Becton Dickinson and Company, Le Pont de Claix, France) supplemented with 5 g/L yeast extract. CDM medium was as described [49], but supplemented before use with 1% choline and 1/100 volume of THY. Synthetic CSP1 [50] was obtained from Eurogentec at 95% purity and stored frozen in water at 100 μg/ml. Genomic donor DNA was purified from strain CP2215 as described previously [24]. For selection of recombinants, dilutions were incorporated in multi-layer THY agar plates, which allow expression of new genes before application of selective pressure [48]. Top agar contained selective compounds at selective levels of 100 μg/ml spectinomycin, 5 μg/ml rifampicin, 10 μg/ml chloramphenicol or novobiocin, or 800 μg/ml kanamycin.
Genomic DNA extracts were purified from log-phase THY cultures by bead beating according to the manfucturer’s protocol (Quick-DNA Fungal-Bacterial Microprep kit, Zymo). DNA concentration was measured using a Qubit fluorometer (Invitrogen, Grand Island, NY). Genomic DNA was normalized to 0.2ng per microliter prior to library preparation utilizing the Illumina Nextera XT kit. Briefly, 1 ng of the normalized DNA was subjected to enzymatic shearing and adapter ligation with the Illumina tagmentation enzyme. Adapted DNA was then amplified during which unique index sequences are added as well as the sequences required for cluster formation in sequencing. The resulting amplified library was cleaned by AMPure XP beads to remove short library fragments. Purified libraries were then checked with Agilent Tapestation and normalized for size and mass before pooling into a sequencing run. Libraries were sequenced on Illumina NextSeq with 2x150 bp sequencing reads generating approximately 7 million sequencing clusters and each recombinant sequenced to an average coverage of 300X.
Nanopore sequencing was performed using the Ligation Sequencing Kit 1D. Briefly 1 μg of genomic DNA was sheared utilizing Covaris g-Tube to fragments of approximately 10 kb in length. The sheared DNA was end repaired and dA-Tailed according to protocol with NEBNext reagents. After AmpPure bead clean-up, adapters were ligated utilizing NEB Blunt T/A Ligase master mix. Following adapter ligation, the library is again bead cleaned and prepared for flow cell loading. Sequencing was performed for 48 hours on an R9.4 Spot ON flow cell. Total output for the Nanopore sequencing was > 40,000 passing reads. Illumina and Nanopore sequence data was combined for Hybrid Assemblies.
De novo assembly was performed using the Spades assembler [51] version 3.9 on both raw Illumina and Nanopore reads, with multiple k-mers specified as “-k 31,51,71,91”. Assembly of the strain resulted in one finished contig. Coverage levels were assessed by mapping raw Illumina reads back to the contig with BWA MEM [52] and computing the coverage as the number of reads aligning times the length of each read divided by the length of the contig. Automated annotations for the assembled contig were generated using Prokka [53].
Fastq files of recombinants and recipients produced by WGS were quality filtered using trimmomatic [54]. The filtered fastq files (SRA BioProject:PRJNA448170) were mapped to the finished assembly of donor strain CP2215 (genbank Accession:CP028436) using SMALT. The consensus alignment was used to find SNPs between CP2215 and the two recipient strains (CP2204 and R36KAN). A customized python script was written to identify high quality SNPs (QUAL Filter > 50, Depth (DP) filter > 5, DP4 ratio filter > 0.75 (i.e. >75% of the forward/reverse reads need to be supporting the base call), Mapping Quality (MQ) filter > 30, Allele frequency (AF) filter < 0.05, Alternate allele frequency filter > 0.95, Strand bias > 0.001, Base Quality Bias > 0, Mapping Quality Bias >0, Tail distance bias > 0.001) in the alignment that were shared between donor and recombinant but not original recipient (full SNP lists available at https://github.com/laurencowley/Environmental-dependence-testing-on-recombination-in-Streptococcus-pneumoniae). This script then attempted to extend the match between donor and recombinant to adjacent SNP positions until the two sequences were no longer identical. Once the beginning and end (the last matching SNP position between the donor and recipient) of the recombination event had been established, the script extracted the sequence (Fig 2).
To examine Gubbins detection of rates of recombination within the ABC dataset, quality filtered fastq files for the 240 strains were mapped to the R6 reference genome to produce an alignment. This alignment was used in the recombination detection program Gubbins [20] which looks for elevated numbers of SNPs in an alignment to detect likely regions of recombination. Gubbins implements an approach to detect recombination that assumes a null evolutionary model on an alignment that has been base called to high quality SNP thresholds, with recombination events being identified as anomalous clusters of SNPs found close together. Hence in our analysis any regions with lower SNP quality could produce false positives that result in falsely recognized recombination regions. SNPs were called in this alignment using the same high quality thresholds as listed above. A custom python script extracted the recombination sizes from the Gubbins output to produce size distribution statistics.
To examine Gubbins detection of rates of recombination for our dataset, the above described consensus alignment was used in the recombination detection program Gubbins [20], which looks for elevated numbers of SNPs in an alignment to detect likely regions of recombination. Gubbins results were parsed for concordance with our mapping method in identified recombination events and in their clustering.
|
10.1371/journal.ppat.1004583 | A Wild C. Elegans Strain Has Enhanced Epithelial Immunity to a Natural Microsporidian Parasite | Microbial pathogens impose selective pressures on their hosts, and combatting these pathogens is fundamental to the propagation of a species. Innate immunity is an ancient system that provides the foundation for pathogen resistance, with epithelial cells in humans increasingly appreciated to play key roles in innate defense. Here, we show that the nematode C. elegans displays genetic variation in epithelial immunity against intestinal infection by its natural pathogen, Nematocida parisii. This pathogen belongs to the microsporidia phylum, which comprises a large phylum of over 1400 species of fungal-related parasites that can infect all animals, including humans, but are poorly understood. Strikingly, we find that a wild C. elegans strain from Hawaii is able to clear intracellular infection by N. parisii, with this ability restricted to young larval animals. Notably, infection of older larvae does not impair progeny production, while infection of younger larvae does. The early-life immunity of Hawaiian larvae enables them to produce more progeny later in life, providing a selective advantage in a laboratory setting—in the presence of parasite it is able to out-compete a susceptible strain in just a few generations. We show that enhanced immunity is dominant to susceptibility, and we use quantitative trait locus mapping to identify four genomic loci associated with resistance. Furthermore, we generate near-isogenic strains to directly demonstrate that two of these loci influence resistance. Thus, our findings show that early-life immunity of C. elegans against microsporidia is a complex trait that enables the host to produce more progeny later in life, likely improving its evolutionary success.
| Infectious diseases caused by microbes create some of the strongest forces in evolution, by killing their hosts, and impairing their ability to produce progeny. Microsporidia are very common microbes that cause disease in all animals, including roundworms, insects, fish and people. We investigated microsporidia infection in the roundworm C. elegans, and found that strains from diverse parts of the world have differing levels of resistance against infection. Interestingly, a C. elegans strain from Hawaii can clear infection but only during the earliest stage of life. This resistance appears to be evolutionarily important, because it is during this early stage of life when infection can greatly reduce the number of progeny produced by the host. Consistent with this idea, if the Hawaiian strain is infected when young, it will ultimately produce more progeny than a susceptible strain of C. elegans. We find that this early life resistance of Hawaiian animals is due to a combination of genetic regions, which together provide enhanced immunity against a natural pathogen, thus enabling this strain to have more offspring.
| Infectious disease is one of the strongest drivers of evolution, generating diversification in hosts and pathogens through a dynamic co-evolutionary process of adaptation and counter-adaptation. The dynamism of these relationships is apparent in emerging infectious diseases, whereby an interaction between organisms changes from being benign to being harmful for the host [1]. Emerging diseases can have devastating effects on their hosts, and understanding how infectious diseases emerge is therefore a major goal for medicine, agriculture, and evolutionary biology.
Microsporidia are emerging pathogens that comprise a diverse phylum of more than 1400 species of fungal-related obligate intracellular parasites that are able to infect virtually all animals [2–5]. Encephalitozoon intestinalis is one of the many species known to infect humans, and stands out as having the smallest eukaryotic genome identified to date [6]. One consequence of the genomic reduction observed in microsporidia is their reliance on host metabolic machinery for propagation. Microsporidia commonly infect intestinal epithelial cells and can be transmitted via a fecal-oral route, although tissue tropism varies broadly. Incidences of microsporidia infection in humans were previously thought to be restricted to immunodeficient patients, but several recent studies have found an unexpectedly high prevalence among healthy people in developed countries, although the overall impact of microsporidia on the health of immunocompetent people is poorly defined [7–9]. In addition to their previously underappreciated disease-causing potential in humans, microsporidia are considered emergent pathogens of agriculturally important animals including fish and honeybees [10–12]. Despite such ubiquity, little is known about the genetic and molecular basis for pathogen defense in any host-microsporidia interaction.
Immune defense against pathogens such as microsporidia will provide evolutionary benefit if it enables hosts to produce more progeny. As such, evolutionary theory predicts that there will be less selection for immunity in post-reproductive animals [13]. The decline of immune function is termed immunosenescence, and has been observed in post-reproductive animals ranging from humans to invertebrates [14–16]. In the human population immunosenescence has been shown to be a complex trait regulated by several genetic loci [17]. Several outstanding questions remain regarding the process of immunosenescence, including its precise timing over the lifetime of an organism and how it has been shaped by pathogens over evolutionary time.
We use the nematode Caenorhabditis elegans as a convenient host to characterize resistance to a natural microsporidian pathogen. This pathogen is called Nematocida parisii, or nematode-killer from Paris, because it was isolated from wild-caught C. elegans from a compost pit near Paris [18]. The life cycle of N. parisii is similar to those of other microsporidia species, which use a specialized infection apparatus called a polar tube to invade directly into host cells, where they undergo their life cycle (S1 Fig.). In the case of N. parisii, spores are consumed by C. elegans, enter the intestinal lumen, and then invade intestinal cells. This N. parisii ‘sporoplasm’ becomes a meront, which replicates in direct contact with host cytosol, and then differentiates back into spores. These spores enter the host trafficking system, exit host cells via apical exocytosis back into the intestinal lumen, and return to the outside environment via defecation [19] (S1 Fig.). Wild-caught nematodes infected with Nematocida species have been isolated from many distinct environmental locations [18, 20], suggesting that microsporidia have imposed widespread evolutionary pressure on the defense system of C. elegans. C. elegans has no known professional immune cells and relies predominantly on epithelial cells as ‘non-professional’ immune cells for defense against infection [21, 22]. The C. elegans intestine is a relatively simple structure composed of just 20 non-renewable epithelial cells that share structural and functional similarity with human intestinal epithelial cells [23]. Thus, the natural C. elegans-N. parisii host-pathogen pair provides an excellent system in which to investigate epithelial defenses shaped over evolutionary time.
Here, we show that there is natural variation in C. elegans defense against microsporidia. We find that a C. elegans strain from Hawaii has enhanced resistance to N. parisii compared to other C. elegans strains. Interestingly, immunity in the Hawaiian strain occurs via clearance of intracellular infection from intestinal epithelial cells. This clearance of N. parisii represents an impressive example of non-professional immune cells being able to not just resist but eliminate microbial infection. Intriguingly, only very young (first larval stage L1) animals are able to clear infection; Hawaiian animals rapidly lose this ability even before they reach reproductive age. We find that N. parisii infection impairs progeny production only when animals are inoculated at the L1 stage, and not when they are inoculated at the later L4 stage, providing a likely evolutionary explanation for why there is enhanced resistance only in L1 animals. Enhanced resistance confers a selective advantage, allowing Hawaiian animals to outcompete a susceptible host strain in only a few generations. Finally, we determine that Hawaiian resistance to N. parisii is a complex multigenic trait that maps to at least four quantitative trait loci (QTL), and we show with near-isogenic lines (NILs) how two of these loci contribute to resistance. These results demonstrate that intestinal epithelial cells in a wild C. elegans strain can eliminate intracellular microsporidia infection. Interestingly, this complex trait acts only at a developmental stage in which it promotes progeny production, and thus likely provides an evolutionary benefit to the host.
To determine whether there is natural variation in the ability of C. elegans to defend against its natural intracellular pathogen N. parisii, we investigated infection in a collection of geographically diverse C. elegans strains. N. parisii has been shown to shorten the lifespan of C. elegans due to a lethal intestinal infection [18], and so we first examined survival upon infection using six strains that represent diverse haplotypes from a global collection of C. elegans [24]. We infected populations of synchronized first larval stage (L1) animals with N. parisii spores and quantified the percentage of animals alive over time. In these experiments, we observed variation in the survival during infection with time to 50% of animals dead (TD50) ranging from 90 to 120 hours among the various C. elegans strains (Fig. 1A). The standard C. elegans N2 laboratory strain from Bristol, England had a relatively short TD50 of about 90 hours. This strain has been passaged under laboratory conditions for decades, and interestingly, did not have significantly different longevity than the C. elegans wild-caught strain ERT002 from Paris, France, which has been passaged very little under laboratory conditions. Also of note, ERT002 harbored the original isolate of N. parisii [18], indicating that it had been exposed to pressure from microsporidia in the wild in the recent past. Strains JU778 from Portugal and JU258 from Madeira had intermediate levels of survival upon infection. By contrast, strain ED3046 from South Africa and strain CB4856 from Hawaii, USA (hereafter designated HW) survived significantly longer than the other strains. Furthermore, we observed that all strains lived longer in the absence of infection (S2 Fig.). N2, HW, and JU258 had similar lifespans in the absence of infection, which were on average slightly longer than those of ERT002, JU778, and ED3046. Thus, there is natural variation in survival of C. elegans upon infection by its natural intracellular pathogen, N. parisii.
Variation in survival upon infection could be due to variation in resistance (the ability to control pathogen load) or tolerance (the ability to cope with infection). To discriminate between these possibilities, we measured pathogen load 30 hours post-inoculation (hpi), which corresponds to the meront stage of N. parisii development, before spores have formed (see S1 Fig. for N. parisii life cycle). To measure pathogen load, we developed a quantitative PCR assay whereby levels of N. parisii small subunit ribosomal RNA are measured and normalized to levels of C. elegans small subunit ribosomal RNA as a control. Using this assay, we observed variation in pathogen load among strains (Fig. 1B) and found that most strains that survived longer had lower pathogen load (S3 Fig.). These results demonstrate that there is natural variation in C. elegans resistance against infection, i.e. the ability of C. elegans to control levels of N. parisii pathogen load.
Given the phenotypic extremes exhibited by N2 and HW, we further investigated the variation in pathogen resistance between these two strains. In the experiments described above, we found that HW was highly resistant to a strain of N. parisii that was isolated in the state of Hawaii (strain ERTm5—See Materials and Methods). We next infected N2 and HW with a strain of N. parisii that was isolated in Paris, France (strain ERTm1) to determine whether HW C. elegans were also more resistant to a strain of N. parisii isolated from a distant geographical location. Indeed, we found that HW also lived longer and was more resistant than N2 when infected with the N. parisii strain from France (S4 Fig.). All subsequent experimentation was performed with the N. parisii strain from Hawaii.
To confirm via a different assay that HW animals are more resistant to infection than N2 animals, we examined pathogen load in N2 and HW animals using a fluorescence in situ hybridization (FISH) assay with a fluorescent probe that targets the N. parisii small subunit rRNA. Consistent with the qPCR results (Fig. 1B), we found that pathogen load 30 hpi was much lower in HW animals compared to N2 animals (Fig. 2A–B).
One potential reason for decreased pathogen load of HW animals is that they may simply feed less than N2 animals and thereby ingest a lower initial inoculum of N. parisii spores. To examine this possibility, we compared the feeding rate between N2 and HW at the L1 stage by inoculating animals with GFP-labeled E. coli together with N. parisii spores and measuring fluorescent accumulation in the intestinal lumens of individuals over time. These experiments revealed that HW L1 animals did not feed less than N2 L1 animals and in fact fed slightly more (S5 Fig.). Thus, lower pathogen load in HW animals is not simply caused by a lower rate of feeding and a lower initial inoculum of pathogen.
The experiments described above were performed with animals infected at the first larval stage of development, although previously we had described that N2 C. elegans are susceptible to infection by N. parisii at all four larval stages (L1 through L4), as well as the adult stage [18]. Interestingly, we found that the difference in pathogen load between N2 and HW was vastly reduced when animals were inoculated with N. parisii at the L2 stage, compared to animals that were inoculated at the L1 stage (Fig. 2C–D). We quantified pathogen load by FISH staining and COPAS Biosort analysis of a population of animals 30 hpi at each of the four larval stages and found that the differences between N2 and HW were restricted to infections initiated at the L1 stage (Fig. 2E). These results indicate that young HW animals are much more resistant to infection than young N2 animals, but this enhanced pathogen resistance of HW animals is rapidly lost with age.
The pathogen resistance of HW animals could be caused by an inability of N. parisii to invade and establish an infection in these animals, or by the ability to limit or clear an infection once it has been established. The results from our feeding experiments with fluorescent E. coli indicated that HW animals receive a similar initial inoculum of pathogen in their intestinal lumens (S5 Fig.), but it remained possible that the pathogen may be less able to invade and establish an infection inside intestinal cells of HW animals. To investigate this possibility, we analyzed intracellular infection at a very early stage. Previously, we had identified the earliest signs of N. parisii invasion and intracellular growth at 8 hpi [25] (and see life cycle in S1 Fig.), and here we show that intracellular N. parisii parasite cells can be identified even earlier at 3 hpi, by visualization of small, mono-nucleate N. parisii ‘sporoplasms’ inside C. elegans intestinal cells (Fig. 3A). These sporoplasms then develop into larger, multi-nucleate meronts by 20 hpi (Fig. 3B). We quantified this infection and found that approximately 90% of animals in a population of either N2 or HW animals had at least one intracellular pathogen cell in their intestines (Fig. 3C). To further quantify this initial invasion and infection, we counted the number of parasite cells per animal at 3 hpi and found that this number was slightly lower in HW animals (Fig. 3D). The fact that a similar percentage of N2 or HW animals is infected at 3 hpi lends further support to the hypothesis that the variation in resistance is not caused by differences in the rate of pathogen exposure or invasion but rather by an enhanced resistance in HW animals that acts post-invasion to mediate clearance of infection.
Next, we directly assessed whether HW C. elegans can clear an infection by comparing infection at different time points. Because it is necessary to fix and stain infected animals to conclusively identify pathogen cells, we cannot track a single parasite cell over time in the same animal. Instead, we analyzed animals sampled from the same infected population over time. In the previously described experiments analyzing infection at 30 hpi (Figs. 1 and 2), C. elegans animals were inoculated with infectious N. parisii spores and were continuously exposed to these spores throughout the course of the experiment. Under these conditions, all animals in a population will eventually become infected. To create conditions in which it may be possible to observe an animal clear an infection that has already been established, we developed a ‘pulsed-inoculation’ assay. Specifically, we took half of the animals from a population at 3 hpi that had been analyzed as described above and re-plated them in the absence of spores. We then harvested these animals at 20 hpi, fixed, FISH-stained to label pathogen cells and then determined the percentage of animals exhibiting infection, where 0% means no animals in the population had infection and 100% means that all animals in the population had at least one pathogen cell present. Strikingly, the percentage of HW animals that showed any evidence of infection was dramatically decreased from 90% at 3 hpi to only about 20% at 20 hpi, indicating that most animals that were infected at 3 hpi were able to clear the infection by 20 hpi (Fig. 3C). By contrast, N2 animals did not show a lower percentage of animals infected at 20 hpi, indicating they were not able to clear infection. Furthermore, when animals were inoculated at the L4 stage, neither N2 nor HW animals were able to clear the infection (Fig. 3C). Thus, it appears that young HW animals can clear an intracellular N. parisii infection from their intestinal epithelial cells, but they lose this ability before reaching a reproductive age.
One potential driver of age-specific resistance could be variation in the selective pressure that is applied by infection at different ages. Thus, we investigated the relative fitness of N2 and HW animals exposed to pathogen at different ages, focusing first on survival as a measure of fitness. In our results described above, HW animals lived about 33% longer than N2 animals during infection (Fig. 1A). In these experiments, animals were inoculated as L1 animals and then exposed to pathogen throughout their lifetimes. In order to more closely compare differential immunity to exposure at different ages, we performed the ‘pulsed-inoculation’ for three hours, removed animals from pathogen and then measured lifespan. With this ‘pulsed-inoculation’ introduced at the L1 stage, HW animals lived two and half times longer than N2 animals (Fig. 4A). Strikingly, HW animals inoculated as L1 animals had relatively little decrease in survival compared to uninfected HW animals (Fig. 4A). By contrast, N2 animals inoculated as L1 animals had dramatically decreased survival compared to uninfected N2 animals. Thus, the early life immunity of HW L1 animals was sufficient to nearly eliminate the negative impact of pathogen exposure on survival during this time. Interestingly, no significant difference in survival between N2 and HW animals was observed when pathogen inoculation occurred at the L4 stage. In this experiment, both N2 and HW animals died much more quickly than uninfected controls.
To further investigate how age-specific resistance of HW animals may affect fitness, we investigated overall progeny production, which is a key driver of evolutionary success. We found that progeny production for both the N2 and HW strains was not significantly different in animals inoculated with pathogen at the L4 stage compared to animals that were not exposed to pathogen. However, inoculation at the L1 stage led to a significant reduction in lifetime fecundity. In particular, N2 had drastically fewer progeny, while HW had only slightly fewer progeny (Fig. 4B). Thus, HW immunity at the L1 stage improves lifetime fecundity and is likely to improve evolutionary success. By contrast, resistance at the L4 stage does not appear to be important for evolutionary success, given that progeny number is not significantly reduced by pathogen inoculation at this stage.
Both N2 and HW have reduced lifetime fecundity when infected at the L1 stage, but infected HW animals have significantly more progeny than infected N2 animals. We tested to see if this difference confers a competitive advantage to HW in an environment shared with N2. L1 stage animals were inoculated with spores for three hours and then grown to the L4 stage in the absence of spores, followed by plating of equal numbers of N2 and HW animals on a shared plate. The population was then expanded to saturation and analyzed for the relative abundance of each C. elegans strain within the population. In uninfected populations the ratio of N2 to HW animals was 58% to 42%, respectively (Fig. 4C). In the presence of pathogen, the ratio of N2 to HW animals shifted to 19% and 81%, respectively (Fig. 4C). Taken together, these experiments demonstrate that the enhanced resistance of young HW animals confers a selective advantage over N2 animals in a laboratory setting.
Having established phenotypic variation in resistance to microsporidia infection, we sought to characterize the underlying genetic variation. Several phenotypic differences between N2 and HW have previously been investigated, and the causative genes responsible for those differences have been identified [26–34]. In particular, a variant in the npr-1 gene, which encodes a neuropeptide Y-like G-protein-coupled receptor, is known to mediate several fitness-related differences between N2 and HW in a laboratory setting, including lifetime fecundity and avoidance of the human pathogen Pseudomonas aeruginosa [26, 34]. To determine whether npr-1 is responsible for the differences in resistance to microsporidia, we measured pathogen load in N2 and HW strains harboring an introgressed npr-1 locus from the other strain and found no significant differences between the introgressed strains and the parental strains (S6 Fig.). Furthermore, a deletion mutation for npr-1 in the N2 background had similar pathogen load as the N2 strain (S6 Fig.). Thus, the npr-1 gene does not appear to be responsible for the enhanced resistance to microsporidia infection of HW animals compared to N2 animals.
The increased resistance of HW animals to infection by N. parisii could be caused by the absence of a host factor important for N. parisii growth or by the presence of an increased host immune response. To distinguish between these two models, we examined whether the HW resistance phenotype was dominant or recessive to the N2 phenotype. We tested the F1 heterozygous progeny from a cross between N2 and HW for pathogen load by FISH and found that heterozygotes were as resistant as HW homozygotes (S7 Fig.), indicating that resistance is dominant. Together with the data on clearance of infection, these results support the model that HW has an increased immune response to N. parisii infection compared to N2.
Next, we sought to identify the number and location of the genetic regions contributing to the variation in immunity between N2 and HW animals. We used quantitative genetic analyses to map the causal quantitative trait loci (QTL) by infecting 179 recombinant inbred advanced intercross lines (RIAILs) between the N2 and HW strains [35] and measuring pathogen load 30 hpi by qRT-PCR (S1 Table). Pathogen load values for RIAILs varied continuously and were generally well bounded by the parental values (S8 Fig.). Replicate data from the parents and all RIAILs indicated that the broad-sense heritability of resistance was 69%, signifying that much of the variation in resistance is caused by genetic factors. Single-marker regression revealed four loci on chromosomes II, III, and V that are associated with variation in resistance (Fig. 5A and S2 Table). RIAILs bearing the HW allele at these loci had significantly lower pathogen loads than those carrying the N2 allele. We named these loci Resistant Against Microsporidia Infection (rami): rami-1, rami-2, rami-3, and rami-4. Together, these four genetic loci account for 51% of the phenotypic variance. Thus, the rami QTL appear to explain the majority (51/69 = 74%) of the N2-HW genetic variance.
To confirm that the genetic loci identified by QTL analysis could influence pathogen resistance, we made and tested NILs, which bear either an interval from the N2 strain introgressed into the HW strain or an interval from the HW strain introgressed into the N2 strain. We investigated the rami-1 and rami-4 loci, which should account for about 15% and 12% of the phenotypic variance respectively (S2 Table). We generated NILs for rami-1 and rami-4 where the N2 interval was introgressed in the HW strain and vice versa. We then infected these strains with N. parisii and quantified pathogen load by qRT-PCR. Compared to the N2 strain, NILs in the N2 background with the rami-1 or rami-4 locus from HW had on average a 44% or 36% reduction in pathogen load, respectively (Fig. 5B). When both rami-1 and rami-4 from HW were present in the N2 background there was a 62% reduction in pathogen load compared to N2. The opposite effect was seen for NILs that were made in the HW background with rami-1 or rami-4 from N2; i.e., these animals were more susceptible than HW. Compared to the HW strain, NILs with rami-1 or rami-4 from N2 were 39% and 31% more susceptible, respectively, and 67% more susceptible when rami-1 and rami-4 were combined (Fig. 5B). Additionally, we tested the rami-4 NIL in the N2 background for pathogen load with the FISH assay and found that with this assay as well, the rami-4 locus made N2 significantly more resistant to infection (S9 Fig.). Altogether, our results indicate that the rami-1 and rami-4 loci both additively promote C. elegans resistance to N. parisii.
Our findings demonstrate that there is natural variation in C. elegans host defense against microsporidia infection. We used variation between the N2 and HW strains to characterize the phenotypic and genetic basis of resistance to N. parisii. Surprisingly, we found that intestinal epithelial cells can clear intracellular infection in the HW strain but only when infection occurs at a young age. We observed that infection has a large negative impact on progeny production if it occurs at a very young age but not at a later pre-reproductive age, delineating one potential evolutionary reason for the age-specific resistance we identified. We used RIAILs generated from crosses between the susceptible N2 strain and the resistant HW strain to identify four QTL that contribute to a complex genetic basis of resistance to a natural intracellular pathogen.
Age-related decline in immune response has been widely observed among animals [15], although our findings of loss of immune function at such an early, pre-reproductive stage are unusual. Most studies of immunosenescence focus on reproductive or post-reproductive animals. For example, a master regulator of immune defense in C. elegans is the p38 MAPK PMK-1, which has been shown to functionally decline around day six of adulthood [16], after reproduction has ended. In addition, the C. elegans JNK-like MAPK KGB-1 has a reversal in protective function beginning in adulthood [36]. Here, we made the surprising observation that the enhanced resistance of the HW strain to N. parisii infection is limited to very young animals, and that immunity to this pathogen declines well before animals have begun adulthood and production of progeny. Our analysis suggests that the absence of enhanced immunity in older, albeit pre-reproductive, HW animals may have been shaped by weakened selective pressure. Employing a strong immune response may have negative consequences, including metabolic costs and the potential for self-damage. In the absence of selective pressure imposed by infection on progeny production that we observed in older larvae, maintaining a robust immune response may be superfluous and costly to the evolutionary success of the individual. It is surprising that infection of older pre-reproductive animals led to sharp decreases in lifespan but not to significant decreases in production of progeny. Because older animals were not able to clear infection, perhaps resources at older age are reallocated from immunity to reproduction. Our data indicate a drastic decline in immune responses to pathogens that occur earlier than those results described in other studies.
As wild C. elegans strains infected by microsporidia have been isolated from around the world [18, 20], it is likely that co-evolution has contributed to genetic diversity and natural variation in resistance. Researchers have isolated strains of C. elegans from six continents, and the genetic diversity among these strains was recently documented [24]. The N2 and HW strains are highly divergent from each other, and we found that they vary in resistance to N. parisii. The enhanced resistance of the HW strain may incur costs that make it less fit in the absence of infection. Our data on relative fitness support this idea, in that HW has a shorter lifespan than N2 in the absence of infection (Fig. 4A). However, this difference may be explained by variation in the npr-1 gene [26], while the difference we see in resistance to N. parisii cannot (S6 Fig.). Regardless, variation between these two strains may not necessarily capture variation that is relevant to evolution in a natural setting due to adaptations that may have occurred in a laboratory setting. A case in point is the variation between N2 and HW in NPR-1-mediated behaviors, which were originally believed to be naturally derived but have since been convincingly shown to be due to a laboratory adaptation in N2 [32, 35]. As discussed above, variation in NPR-1 between N2 and HW gives N2 a fitness advantage in standard laboratory conditions. This variation confounds our ability to assess the potential costs of immunity that may be part of the resistance of the HW strain. We tested four additional wild isolates that span the geographic and genetic range of strains characterized so far and found equal proportions of relative resistance and susceptibility. The data from this limited set of strains suggest that natural variation in resistance to N. parisii is an ecologically relevant trait. Genetic association studies with additional strains may identify the resistance alleles that are segregating in the global population.
We found that increased survival upon N. parisii infection among different C. elegans strains generally correlated with increased pathogen resistance (ability to control N. parisii pathogen load). However, the C. elegans strain JU778 survived infection as long as the JU258 strain despite having higher pathogen load 30 hpi, suggesting that both tolerance and resistance vary among wild strains (S2 Fig.). Also supporting the variation in tolerance is the observation that the JU778 strain slightly outlived the N2 strain when infected but died faster in the absence of infection (S1 Fig.). These observations indicate that the longevity advantage of the JU778 strain may be specific to the context of N. parisii infection. Although there may be variation among strains in their ability to tolerate N. parisii infection, we focused on variation in resistance. We found that the enhanced resistance of HW is mediated by an active clearance of infection from intestinal epithelial cells. To our knowledge, clearance of intracellular pathogens by intestinal epithelial cells has not previously been demonstrated in any animal host. The cell-intrinsic immune capabilities of epithelial cells are increasingly appreciated in mammals [37]. For example, autophagy in epithelial cells can limit invasion and dissemination of bacterial pathogens [38]. Microsporidia commonly infect intestinal epithelial cells in humans. Interestingly, studies of microsporidia infection in humans suggest that intestinal infections by microsporidia might be cleared by immunocompetent people [2, 39]. It is known that the adaptive immune system is important for clearing microsporidia infections in humans, but it is attractive to speculate that human intestinal epithelial cells may also play a role in clearing infections, similar to C. elegans intestinal epithelial cells. Identifying the mechanisms of clearance in C. elegans may help elucidate the immune capacity of epithelial cells in general, which are the first line of defense against many microbial infections.
Although previous studies in C. elegans found that variation in resistance to the human pathogen P. aeruginosa was a simple trait determined predominantly by a single gene [34], we found that C. elegans resistance to N. parisii infection is a complex genetic trait. We mapped four loci that explain a large fraction of the genotypic variance and used NILs to directly confirm the effects of rami-1 and rami-4. Immunity-related genes have undergone exceptional amounts of positive selection in humans and other organisms [40]. For example, genes encoding major histocompatibility locus (MHC) proteins, immune signaling proteins and antimicrobial peptides have increased in diversity over recent evolutionary time. Hundreds of genes fall within the rami loci, and one approach to identifying candidates for further study may be to screen for genes that display signatures of positive selection. For example, the ubiquitin-dependent proteasome adaptors encoded by F-box and MATH-BTB genes are among the most rapidly diversifying genes in the C. elegans genome and have unparalleled rates of birth-death evolution [41]. These genes are under strong positive selection in their substrate-binding domains but not in their Cullin-binding domains, suggesting that they have evolved to detect and degrade foreign proteins as an immune defense mechanism [41]. Ubiquitin-mediated proteolysis has been implicated in host-pathogen interactions in both plants and animals [42] and is an attractive hypothesis for how C. elegans might combat an intracellular invasion such as N. parisii infection. In support of this hypothesis, we recently found that components of the ubiquitin-proteasome system are upregulated during infection and that disrupting the ubiquitin-proteasome system or autophagy in C. elegans during N. parisii infection increases pathogen load in the N2 strain [43]. Further refinement of our infection assays, together with genetic and molecular analyses, should uncover the specific genetic polymorphisms that have evolved to produce enhanced epithelial resistance in the HW strain. Epithelial cells are critical sites of host interactions with pathogens, and we find that they can directly eliminate intracellular infection based on several genetic loci that are tailored to the success of propagating the species.
C. elegans strains were maintained on nematode growth media (NGM) seeded with E. coli OP50–1 (which is a streptomycin-resistant OP50 strain) as previously described [44]. For simplicity, this strain is referred to as OP50 throughout. To obtain starved and synchronized L1 larvae, gravid adults were bleached to isolate eggs, which then were allowed to hatch overnight at 20°C [45]. The C. elegans strains N2, CB4856 (HW), JU778, JU258, and ED3046 were obtained from the Caenorhabditis Genetics Center. ERT002 is derived from strain CPA24, which was previously isolated from a compost pile in Franconville, France and was the original strain of C. elegans isolated with N. parisii ERTm1 infection [18, 25]. Strain CPA24 was subsequently bleached to remove the infection and renamed ERT002 to conform to C. elegans nomenclature conventions. For mapping, we used a set of advanced intercross recombinant inbred lines generated previously with the N2 and CB4856 strains [35]. The N. parisii strain we used in all infection experiments except S4 Fig. was ERTm5, a N. parisii strain isolated from JU2055, a Caenorhabditis briggsae strain found in a rotting breadfruit sampled in early April 2011 by Christopher Nelson in Limahuli Gardens, Haena, Kauai (Hawaii 22.219 North, -159.5763 West). Spores were prepared and quantified as previously described [46].
For survival measurements in the six C. elegans strains during infection (Fig. 1), synchronized L1 larvae were plated on 6 cm NGM plates seeded with OP50 and inoculated with 2 × 106 N. parisii spores at 25°C. At 48 hpi, 30 animals were transferred to 3.5 cm plates seeded with OP50 with three plates per experiment. Live animals were quantified every 24 hours and transferred to fresh plates. For survival in N2 and HW during infection and in the absence of infection (Fig. 4), 1200 synchronized L1 larvae were inoculated with 100 μl of a 10x concentrate of an overnight OP50 culture and 2 × 106 N. parisii spores on 6 cm NGM plates at 25°C for three hours. Animals were then washed several times to remove spores and re-plated with OP50 until 48 hpi at 20°C. Uninfected animals followed the same conditions in the absence of spores. L4-infected animals followed the same conditions but were infected for three hours at the L4 stage. 20 individuals from each condition were then plated on 3.5 cm NGM plates seeded with OP50, incubated at 20°C, and transferred to fresh plates every 24 hours until death or progeny production stopped. Mortality was recorded every 24 hours. Data were analyzed in Prism 6 with the Log-rank (Mantel-Cox) test.
Animals were infected in liquid culture or on solid media. For liquid culture infections, 2000 synchronized L1 larvae in 0.5 mL of M9 buffer were plated per well in a 24-well plate. 0.5 mL of M9 buffer containing OP50 and 2 × 106 N. parisii spores was then added to each well. Plates were incubated on a rocker at 25°C. For solid media infections, 1200 synchronized L1 larvae were inoculated with 100 μl of a 10x concentrate of an overnight OP50 culture and 2 × 106 N. parisii spores on 6 cm NGM plates. When plating, media was evenly distributed across the entire plate. Plates were then incubated at 25°C. For infections initiated at stages other than the L1 stage, animals were plated for 24 hours at 20°C before adding spores for the L2 stage, plated for 24 hours at 25°C before adding spores for the L3 stage or plated for 24 hours at 20°C followed by 24 hours at 15°C before adding spores for the L4 stage. Samples were fixed different times post-inoculation with Tri-Reagent (Molecular Research Center) to extract RNA or with acetone to stain by FISH.
RNA was isolated by extraction with Tri-Reagent and bromochloropropane (BCP) (Molecular Research Center). 250 ng of RNA from each sample was used to synthesize cDNA with the RETROscript kit (Ambion). cDNA was quantified with iQ SYBR Green Supermix (Bio-Rad) on a CFx Connect Real-time PCR Detection System (Bio-Rad). We measured pathogen load by measuring the relative abundance of an N. parisii rDNA transcript normalized to a C. elegans rDNA transcript with the following primer sets: Np_rDNAF1: aaaaggcaccaggttgattc, Np_rDNAR1: agctctctgacgcttccttc, Ce18S_F1: ttgcgtacggctcattagag, Ce18S_R1: agctccagtatttccgcagt. Primer efficiencies were measured, and fold difference was calculated using the Livak comparative Ct method (2-ΔΔCt).
We used the MicroB probe conjugated to a red Cal Fluor 610 dye (Biosearch Technologies) to stain infected animals for an N. parsii ribosomal RNA small subunit sequence as previously described [18]. Pathogen load was measured with the COPAS Biosort (Union Biometric) or by microscopy. For analysis with the COPAS Biosort, greater than 600 animals per condition were measured for time of flight (TOF, a measure of size) and red fluorescence. Pathogen load per individual was determined by normalizing the red signal to TOF. For microscopy, samples were mounted on agarose pads with VECTASHIELD mounting medium containing DAPI (Vector Labs) and imaged using fluorescent microscopy on a Zeiss AxioImager M1 upright microscope with a 10x or 100x oil immersion objective equipped with an AxioCam digital camera and AxioVision software. Sporoplasms at 3 hpi were imaged by confocal microscopy acquired on a Zeiss LSM700 at 630x magnification using ZEN2010 software.
For L1 experiments, 1200 synchronized L1 larvae were inoculated with 100 μl of a 10x concentrate of an overnight OP50 culture and 2 × 106 N. parisii spores on 6 cm NGM plates at 25°C for three hours. Animals were then washed several times to remove spores and half were fixed in acetone while the other half was re-plated with OP50 and incubated at 25°C until fixing 20 hpi. L4 experiments followed the same procedure, but infections were initiated at the L4 stage. Sample were stained by FISH and analyzed by microscopy. For the 3 hpi samples, 100 animals per condition were imaged at 1000x to count the number of infected individuals and the number of parasite cells per animal. For the 20 hpi samples, 100 animals were imaged at 100x to count the number of infected individuals.
1200 synchronized L1 larvae were inoculated with 100 μl of a 10x concentrate of an overnight OP50 culture and 2 × 106 N. parisii spores on 6 cm NGM plates at 25°C for three hours. Animals were then washed several times to remove spores and re-plated with OP50 until 48 hpi at 20°C. Uninfected animals followed the same conditions in the absence of spores. L4-infected animals followed the same conditions but were infected for three hours at the L4 stage. Twenty individuals from each condition were then plated on 3.5 cm NGM plates seeded with OP50, incubated at 20°C, and transferred to fresh plates every 24 hours until death or progeny production stopped. Progeny per animal were counted every 24 hours following the L4 stage. After transferring to fresh plates, the source plates were incubated at 20°C for 24 hours to allow all eggs to hatch, then incubated at 15°C for 24 hours before counting. Data were analyzed in Prism 6 by one-way ANOVA and Tukey’s multiple comparison test.
Infection was initiated as in the lifetime fecundity experiments. Once animals had reached the L4 stage, 15 N2 and 15 HW animals were added to 15 cm NGM plates seeded with 3 mL of a 10x concentrate of an overnight OP50 culture and incubated at 20°C. Animals were harvested once food was nearly depleted, which was approximately five days post-plating for uninfected populations and approximately seven days post-plating for infected populations. Each condition was repeated in triplicate per experiment over three total experiments. Genomic DNA was obtained by phenol-chloroform extraction. To determine the ratio of N2 to HW genomic DNA in the samples, we used qPCR to measure the relative abundance of a transcript in the zeel-1 locus (deleted in HW) normalized to a snb-1 transcript (present in both) with the following primer sets: zeel1_N2F1: gggcaattttcaaaagcaga, zeel1_N2R1: gttggtgtgctgaattttct, snb-F1: ccggataagaccatcttgacg, snb-R1: gacgacttcatcaacctgagc. Standard curves of measuring N2 and HW genomic DNA independently and at different known combined concentrations over several biological and technical replicates revealed that on average the observed ratio was 5% off from the expected ratio.
For each condition, 2000 synchronized L1 larvae in 0.5 mL of M9 buffer were plated per well in a 24-well plate. 0.5 mL of M9 buffer containing unlabeled OP50 or GFP-labeled OP50 and 2 × 106 N. parisii spores was then added to each well. Plates were incubated on a rocker at 25°C. Animals were collected each hour for three hours post-plating and mounted on agarose pads for imaging using fluorescent microscopy at a constant exposure time on a Zeiss AxioImager M1 upright microscope with a 40x oil immersion objective equipped with an AxioCam digital camera and AxioVision software. The relative amount of GFP-labeled bacteria in the intestinal lumens of animals was quantified by outlining individual animals and calculating the mean fluorescent intensity with AxioVision software.
For analyzing infection in F1 progeny of N2 and HW crosses, 50 L4 stage males were set up with 30 L4 stage hermaphrodites overnight. Gravid hermaphrodites were bleached on 3.5 cm plates to yield eggs that hatched in the absence of food to obtain synchronized F1 progeny. OP50 was added, and synchronized larvae were inoculated with 7 × 105 N. parisii spores and incubated at 25°C for 16 hours. Animals were fixed and stained by FISH, mounted on agarose pads, and imaged on a Zeiss AxioImager M1 upright microscope with a 10x objective equipped with an AxioCam digital camera and AxioVision software. 30 animals per condition were outlined and measured for mean fluorescent intensity in the red channel.
179 recombinant inbred advanced intercross lines (RIAILs) from a cross between the Bristol (N2) and Hawaii (CB4856) strain were phenotyped by isolating RNA from infected animals 30 hpi and measuring pathogen load by qRT-PCR (see above). N2 and HW were phenotyped in parallel for each experiment, and pathogen load in the RIAILs was normalized to N2. 21 RIAILs were phenotyped on solid media and 158 RIAILs were phenotyped in liquid media (see above for setup, data shown in S1 Table). The normalized fraction of N. parisii DNA of each RIAIL and the respective genotype data [35] were entered into the R statistical programming environment and processed using the qtl package [47]. The phenotypic distribution of the RIAILs had a long right tail, so QTL were mapped using non-parametric marker regression. The 5% genome-wide significance threshold was calculated based on 10,000 permutations of the phenotype data [48]. The most significant marker was used as a covariate to identify additional QTL until no more significant QTL were detected. The total phenotypic variance explained was calculated by squaring the rank-sum correlation of genotype and phenotype for each QTL. Broad-sense heritability was calculated as the fraction of phenotypic variance explained by strain from fit of a linear mixed-model of repeat phenotypic measures of the parents and some recombinant strains [49]. The total variance explained by each QTL was divided by the broad-sense heritability to determine how much of the heritability is explained by each QTL. Confidence intervals were defined as the regions contained within a 1.5 LOD drop from the maximum LOD score.
RIAILs were selected that contained N2 or CB4856 genomic regions spanning the QTL intervals for chromosome II or chromosome V. We backcrossed these regions to the appropriate parental strain at least 12 times for each line, genotyping at SNPs bounding the interval at each cross. To generate strain ERT246 jyIR1[CB4856 > N2] II, Qx228 males were crossed to N2 hermaphrodites and the F2’s that segregated CB4856 markers at SNPs corresponding to the physical locations 1,373,016 and 2,090,144 were selected and homozygosed. Male progeny homozygous for CB4856 markers were crossed to N2 hermaphrodites, which was repeated until the F12 generation. NILs were then genotyped at markers across the arms and centers of all autosomes to confirm that they were N2 outside of the interval. The same basic strategy was followed for generating the other three single NILs, with the chromosome V interval genotyped at physical locations 16,734,456 and 17,917,291: the Qx88 strain was used to generate strain ERT247 jyIR2[N2 > CB4856] II, the Qx217 strain was used to generate strain ERT248 jyIR3[CB4856 > N2] V, and the Qx239 strain was used to generate strain ERT249 jyIR4[N2 > CB4856] V. Double NILs bearing both QTL intervals from one parent in the reciprocal background were generated by crossing single NILs and genotyping at the bounding markers listed above for homozygotes in the F2 progeny. Double NIL strains are ERT250 and ERT251.
|
10.1371/journal.pgen.1005534 | Insect Resistance to Bacillus thuringiensis Toxin Cry2Ab Is Conferred by Mutations in an ABC Transporter Subfamily A Protein | The use of conventional chemical insecticides and bacterial toxins to control lepidopteran pests of global agriculture has imposed significant selection pressure leading to the rapid evolution of insecticide resistance. Transgenic crops (e.g., cotton) expressing the Bt Cry toxins are now used world wide to control these pests, including the highly polyphagous and invasive cotton bollworm Helicoverpa armigera. Since 2004, the Cry2Ab toxin has become widely used for controlling H. armigera, often used in combination with Cry1Ac to delay resistance evolution. Isolation of H. armigera and H. punctigera individuals heterozygous for Cry2Ab resistance in 2002 and 2004, respectively, allowed aspects of Cry2Ab resistance (level, fitness costs, genetic dominance, complementation tests) to be characterised in both species. However, the gene identity and genetic changes conferring this resistance were unknown, as was the detailed Cry2Ab mode of action. No cross-resistance to Cry1Ac was observed in mutant lines. Biphasic linkage analysis of a Cry2Ab-resistant H. armigera family followed by exon-primed intron-crossing (EPIC) marker mapping and candidate gene sequencing identified three independent resistance-associated INDEL mutations in an ATP-Binding Cassette (ABC) transporter gene we named HaABCA2. A deletion mutation was also identified in the H. punctigera homolog from the resistant line. All mutations truncate the ABCA2 protein. Isolation of further Cry2Ab resistance alleles in the same gene from field H. armigera populations indicates unequal resistance allele frequencies and the potential for Bt resistance evolution. Identification of the gene involved in resistance as an ABC transporter of the A subfamily adds to the body of evidence on the crucial role this gene family plays in the mode of action of the Bt Cry toxins. The structural differences between the ABCA2, and that of the C subfamily required for Cry1Ac toxicity, indicate differences in the detailed mode-of-action of the two Bt Cry toxins.
| Transgenic crops expressing the insecticidal protein Cry2Ab from Bacillus thuringiensis (Bt) are used worldwide to suppress damage by lepidopteran pests, often used in combination with Cry1Ac toxin to delay resistance evolution. Until now, the Cry2Ab mode of action and the mechanism of resistance were unknown, with field-isolated Cry2Ab resistant Helicoverpa armigera showing no cross-resistance to Cry1Ac. In this study, biphasic linkage analysis of a Cry2Ab-resistant H. armigera family followed by EPIC marker mapping and candidate gene sequencing identified three independent INDEL mutations in an ATP-Binding Cassette transporter subfamily A gene (ABCA2). A deletion mutation was identified in the same gene of resistant H. punctigera. All four mutations are predicted to truncate the ABCA2 protein. This is the first molecular genetic characterization of insect resistance to the Cry2Ab toxin, and detection of diverse Cry2Ab resistance alleles will contribute to understanding the micro-evolutionary processes that underpinned lepidopteran Bt-resistance.
| In recent decades, agriculture has increasingly come to rely on toxins encoded by the Gram-positive bacterium Bacillus thuringiensis (Bt) for the production of insect–resistant transgenic crops. Narrow spectrum insecticides such as the protoxin crystals produced by Bt during sporulation are highly specific for certain insect groups including the Lepidoptera, Diptera and Coleoptera (e.g., [1]). Bt sprays have been used for many years, gaining widespread acceptance in pest management due to their relative target-specificity and their safety for humans, most other organisms, and the environment. However, the increasing cultivation of Bt transgenic crops poses a significant risk with various field populations of major lepidopteran pests reported to have developed resistance [2–4], threatening the sustainability of this strategy for crop protection. Indeed a major reason for the uptake of Bt was the evolution of resistance to chemical insecticides such as organochlorides, synthetic pyrethroids, and organophosphates in pests such as the cotton bollworm Helicoverpa armigera. This species is one of the most damaging and economically important lepidopteran pests known worldwide in a variety of crops, and one of four major pests in the genus Helicoverpa, the others being H. punctigera, H. zea and H. assulta [5–8].
For control of lepidopteran pests, genes encoding members of the Cry1A family were the first to be used in transgenic crops. However resistance alleles to the Cry1Ac toxin have been reported in field-collected H. punctigera from Australia [9,10], and H. armigera from China [11–13]. Resistance to Cry1Ac has also been reported in Indian populations of Pectinophora gossypiella [14] and H. zea from the New World [15–17]. Genetic studies on the field-derived strains have provided critical insight into the mode of action of this toxin, by identifying key receptors present on the surface of midgut epithelial cells (e.g., [18–23]; see also [24–26]). This binding of the activated toxins to specific receptors is crucial for formation of pores in the affected cells, leading eventually to the death of the larvae.
Resistance to the Cry1Ac toxin in the Lepidoptera was first shown to be associated with mutation of a gene encoding a 12-cadherin domain protein. Deletions of different lengths were observed in various regions of the gene in, e.g., Heliothis virescens [27], P. gossypiella [28] and H. armigera [11–13,24], as well as insertions by transposable elements [12,29,30]. Down regulation of cis-mediated transcription of the trypsin gene HaTryR allele due to mutations at the promoter region, mis-splicing of the ABCC2 gene, and a deletion mutation of the Aminopeptidase N (APN) gene have also been demonstrated to lead to resistance to Cry1Ac in H. armigera [31–33]. Recently, mutations in the ABCC2 gene belonging to Family C of the ATP-Binding Cassette (ABC) transporter family have also been shown to confer resistance to Cry1Ab and Cry1Ac in H. virescens [34], Plutella xylostella [35], Bombyx mori [36], Spodoptera exigua [37], and H. armigera [32]. Heterologous expression of ABCC2 from resistant and susceptible B. mori has shown that it aids in pore formation [38], and modification or deletion of ABCC2 is hypothesized to block the final step in the toxin's mode of action [39].
To prolong the efficacy of individual Bt toxins as transgenic control agents, multiple Bt genes, encoding different toxins with different modes of action, have been incorporated into plants. The second Bt gene adopted in many countries in transgenic plants has been a member of the Cry2A family, Cry2Ab. Cry1A resistance due to mutations in the cadherin or ABCC2 genes is not known to confer cross-resistance to Cry2Ab [22,34]. However, Cry2Ab resistance has now been reported in various lepidopteran pests (e.g., P. gossypiella [40]; H. zea [41]). In Australia, field-derived Cry2Ab resistance alleles in H. armigera and H. punctigera were first isolated in the summer of 2002/2003 and 2004/2005 respectively, and were used to establish the homozygous resistant lines SP15 [42] and Hp4-13 respectively [43]. All Cry2Ab resistant H. armigera and H. punctigera alleles isolated from Australia to-date have been shown to be recessive [9,21,42,43]. Isolates were captured using the "F2 screen" method [44] with a discriminating dose of Cry2Ab toxin, and confirmed as allelic by complementation tests [83]. These isolates are being used in F1 tests which involved crossing them to a field-collected insect (of unknown genotype), and screening the F1 offspring for resistance [45–47]. These F1 screens have estimated the frequency of Cry2Ab-resistance-conferring alleles in H punctigera and H. armigera field populations to range between 0.010 and 0.047, and 0.015 and 0.044, respectively, with no significant linear trend over time (from 2007 to 2014) [48–50].
The F1 and F2 screen techniques do not directly reveal the molecular identity of such resistance alleles, and the molecular basis of the Cry2Ab resistance is likely due to specific target-site alterations located within the midgut [21]. Whether the H. armigera and H. punctigera Cry2Ab resistance genes are homologous is not known, although parallel evolution in orthologous ABCC2 genes leading to Cry1Ac resistance in different species has been reported (e.g., [34,35]). A third generation of transgenic cotton (Bollgard III (BGIII)) expressing three Bt toxin genes (Cry1Ac, Cry2Ab and Vip3A) will soon become available in Australia. In light of these developments and whilst the Australian industry adopts a pre-emptive strategy to manage resistance to Bt, several key assumptions of this strategy are theoretically sound but empirically untested. Marker-assisted detection of the resistance alleles in insect populations will therefore not only enable a more efficient monitoring effort but will also enable assumptions about the ecology of resistance to be rigorously examined.
In this paper we report on the identification of the Cry2Ab resistance gene in H. armigera using linkage mapping and a chromosome walk with the assistance of exon-primed intron-crossing (EPIC)-PCR markers. This gene, which is expressed in the midgut, encodes an ABC transporter in the A subfamily—ABCA2—and is the likely site of mutations conferring resistance to Cry2Ab. By screening additional lab-isolated resistant lines derived from field-collected materials, we show that resistance to the Cry2Ab toxin in H. armigera occurred through independent evolutionary events involving different mutations, all of which were located in different exons of the same ABCA2 gene in both species. This work therefore provides the first insight into the detailed mode of action of a Cry2A toxin, which is conserved across different lepidopteran species, and is of considerable significance for the management of Bt resistance globally.
The absence of crossing-over in female Lepidoptera makes it possible to map a recessive trait such as the Cry2Ab resistance in SP15 to a linkage group using biphasic linkage analysis with AFLPs as genetic markers [18]. Progeny from a female-informative backcross family were bioassayed with a discriminating dosage of Cry2Ab; 161 AFLPs segregating in this family were grouped into 31 independently assorting linkage groups. Linkage to Cry2Ab resistance was tested by comparing bioassayed survivors with untreated control progeny. Only AFLP linkage group (LG) 8 showed a significant association with resistance (χ2 = 19.44, P < 0.001; Fig 1); all 29 treated survivors were homozygous for the SP15 homolog of this linkage group.
Southern blot analysis of RFLPs in an unrelated Cry2Ab-susceptible H. armigera family showed that one AFLP from LG8 was linked to ribosomal protein gene RpL22. RpL22 in B. mori is located on chromosome 17 (BmChr17, KAIKObase [51]). Using specific probes for additional ribosomal protein genes, the Cry2Ab-resistance-associated linkage group in H. armigera was also shown to carry genes for RpL38 and RpS24, confirming homology with BmChr17 [52]. This assignment excludes a number of previously-identified genes as candidates for Cry2Ab resistance. Chromosomes harbouring homologs of previously-identified Cry1Ac resistance mutations in H. virescens include BmChr06 with the 12-cadherin-domain protein [27], BmChr15 with the ABCC2 protein [34], and BmChr21 with the BtR-5 gene [53]. Moreover, genes for previously identified Cry1Ac binding proteins map to chromosomes other than BmChr17: several aminopeptidase genes are located on BmChr09 [54], a membrane-bound alkaline phosphatase gene maps to BmChr03 [23], and the P252 glycoprotein gene is on BmChr25 [55,56]. Although different levels of cross-resistance between Cry1A and Cry2A toxins have been reported in H. armigera from China [57,58], in H. virescens [59], H. zea and P. gossypiella ([40], see also [60]); independent segregation of BmChr17 relative to all of these other chromosomes is nevertheless consistent with the absence of cross-resistance between Cry1Ac and Cry2Ab in both the SP15 H. armigera [42] and Hp4-13 H. punctigera [9,10] lines, thereby supporting the notion that the two toxins have different modes of action. However, we found that the ortholog of the bre-5 glycosyltransferase gene in a mutant of the nematode C. elegans resistant to the Cry4B toxin [19] is located on BmChr17 (Fig 2). This gene was therefore further investigated as a candidate gene for Cry2Ab resistance.
Additional linkage mapping in a male-informative backcross and two F2 families was performed to further localize the resistance locus. A preliminary map based on 72 progeny from these families gave the gene order and recombination values as follows:
Bre−5–(0.16)–Cry2Ab resistance locus–(0.10)–RpL22–(0.06)–RpS24
The order and spacing of the three marker loci was similar to that in B. mori on BmChr17. However, the large fraction of recombinants between the resistance locus and bre-5 ruled out the latter as a candidate (Fig 2).
The Cry2Ab resistance gene was further localised within BmChr17 using recombinational mapping in backcrosses with F1 males. For this work, markers were developed from H. armigera orthologs for genes mapped along BmChr17. Sequences allowing design of EPIC-PCR primers for the H. armigera orthologs for these genes were obtained from transcriptome sequencing of midgut RNA extracted from larvae of the GR susceptible colony.
Recombinational analysis of selected markers in H. armigera showed the linkage order of these markers to be the same as in B. mori (Fig 2), greatly assisting the subsequent analysis which employed the B. mori genome as a reference framework. Analysis of recombination rates between 3 of these markers (RpL38, Zip2, VGCal-A) and the Cry2Ab resistance allele placed it between Bre-5 (BGIBMGA005534) and RpL22 (BGIBMGA006986, at nt 14152986). The markers for the voltage-gated channel protein gene (orthologous to BGIBMGA007009, starting at 12341859 on BmChr17 –see S1 Table) further restricted the area containing the resistance gene. The target area could however be more narrowly defined, since the gene is ~10cM from BGIBMGA005534 and ~16cM from BGIBMGA006986; these genes are located at ~3Mbp and ~10Mbp respectively on the BmChr17 sequence (see S1 Table). In fine scale analysis of this region, markers for the genes BEACH, ANK_SAM, NaPT, DUF410, and AN_Peroxidase all showed recombination with the resistance trait. Of these, the marker for DUF410 (BGIBMGA007299, located at 7124451 on BmChr17) most closely restricted the target region on the proximal side, corresponding to less than 3Mbp of BmChr17, and containing fewer than 30 genes (S1 Table).
Two ABC transporter A subfamily genes are located adjacently between nts 8466000–8564000 on BmChr17. The first of these, termed BmABCA1, is well-predicted as BGIBMGA007221, while the other, BmABCA2, includes the partial predictions BGIBMGA007218 and BGIBMGA007217 (see [61,62] for analysis of the original uncorrected gene models). The sequence of HaABCA1, the H. armigera ortholog to the BGIBMGA007221 BmABCA1 gene, was obtained from RNAseq libraries, from total larvae of the susceptible GR colony. The EPIC-PCR marker for HaABCA1 (ABCA1; S2 Table) gave a genotype profile consistent with tight linkage to the Bt Cry2Ab resistance allele; the F2 bioassayed offspring (homozygous allele size of 272bp / 272bp) was identical to the SP15 grandmother (272bp / 272bp) in 100% of all samples tested (final n = 72). The GR grandfather was heterozygous with alleles 264bp / 282bp, leading to the F1 male being heterozygous with allele sizes 272bp / 282bp. The F2 control (n = 20) gave the expected 50:50 ratio with n = 11 being 272bp / 272bp homozygous and n = 9 heterozygous (272bp / 282bp). This tight linkage between the HaABCA1 gene and resistance made it a candidate for being the target of the resistance mutation.
To assess whether HaABCA1 was the real location of resistance mutations, we checked whether this gene is expressed in the midgut. No evidence for significant midgut expression of ABCA1 was found in either B. mori (Bm-MDB, B. mori Microarray Database [63]; see S1 Table) or H. armigera, making it unlikely that this gene is actually involved in resistance. Initially detected in total larval and pupal transcripts, its expression was more specifically evident in larval foregut, hindgut, trachea and haemocytes. However the adjacent ABCA2 gene was significantly expressed in the midgut of both Bombyx and H. armigera; of the genes in this region of BmChr17, it is among the most highly expressed in the midgut [63] (see S1 Table). The full-length transcript of the H. armigera ortholog HaABCA2 (Fig 3) encodes a protein of 1,742 amino acids with 67.28% identity to the BmABCA2 gene in B. mori (S1 Fig).
We therefore further explored whether any changes were evident in the transcripts of the ABCA2 gene in midgut RNA from larvae of the resistant line. The ABCA2 candidate gene cDNA was fully sequenced from a SP15 resistant individual that identified a 73bp deletion at exon 16 (‘ G CTA GGA GTT CTG CGT TAC GTC ATG TCT TTA TCA CCA ACC ATT AGA ACT AGG TGG TTG TCG TTG GAA GAA GGG’ from nucleotides 2,889–2,961) / 8bp (‘C GGT TAA G’) insertion mutation (= allele 1, Ha2Ab-R01) (Fig 4a and 4b) of the coding sequence which resulted in the replacement of leucine (L) by glycine (G) at position 964, followed by a stop codon downstream from the mutation site at position 965. The 8bp ‘C GGT TAA G’ nucleotides matched completely to part of intron 16 (intron 16 nucleotide positions 56 to 62) of the SP15 H. armigera line. From the sequence of this mutated cDNA, we designed primers to screen additional independently isolated field resistant lines from 2005 (line 5–405); 2006 (lines 6–364 and 6–798), 2009 (line 9–4802), 2010 (line 10–485), and 2012 (line 12–2169), four of which (i.e., 5–405, 9–4802, 10–485, 12–2169) showed the same 73bp deletion/8bp insertion mutation as identified in the SP15 individual. A second resistance allele (= Ha2Ab-R02) with a 5bp (‘ACA AG’) deletion mutation at nucleotides 3,127–3,131 of the coding sequence was identified in a homozygous individual from the Cry2Ab resistant lines 6–364. A heterozygous resistant individual from line 6–798 was identified to possess one Ha2Ab-R02 allele, and a third resistance allele (= Ha2Ab-R03) at nucleotides 4,104–4,108 that represented a 5bp (GAATA) nucleotide duplication (Fig 4b), similar to the target site duplication (TSD) signature that is widespread in the H. armigera genome due to transposable element transposition activities (e.g., see [64]). cDNA sequencing of both the 6–364 and 6–798 lines identified the presence of the Ha2Ab-R02 allele as homozygous in the 6–364 line, and also both Ha2Ab-R02 and Ha2AB-R03 alleles at exons 18 and 24 respectively in line 6–798, thereby confirming that this resistant line was heterozygous for the ABCA2 gene (see S2 Fig). All three mutations identified to date in the Cry2Ab resistant lines are located at the 3’ region of the 5.1Kb coding sequence, and result in truncation of the protein. A summary of the three Cry2Ab resistance alleles identified in H. armigera is presented in Fig 4a and 4b. Sequence analyses of the SP15, 6–364 and 6–798 resistant individuals confirmed that no other INDELs or nonsense mutations were present in coding regions of the ABCA2 gene. Similarly, additional sequence analyses from multiple susceptible H. armigera individuals showed the predicted fully functional (non-truncated) ABCA2 gene, while nucleotide variation (<1%) between Cry2Ab susceptible H. armigera ‘GR’ lines resulted in nine amino acid changes, six of which involved nonsynonymous substitutions between amino acids with hydrophobic side-chains (e.g., valine (V), isoleucine (I), tyrosine (Y), phenylalanine (F), methionine (M), and leucine (L)), and one amino acid substitutions of each between lysine (K) and threonine (T), alanine (A) and proline (P), and glycine (G) and glutamic acid (E) (S2 Fig).
To confirm the significance of the HaABCA2 mutations in resistant lines, we asked whether susceptible individuals collected from the field carried the same INDEL mutations or other inactivating mutations in the H. armigera ABCA2 gene. Starting with single pairs of field-collected insects, F1 pools from each mating pair generated F2 progenies (n = 90) that were screened against a discriminating dose of Cry2Ab and identified those pairs whose F2 progeny all died as carrying only susceptible alleles. This test is enough to exclude any resistance-conferring alleles (occurring as heterozygotes) amongst the grandparents. Ten individual 3rd instar larvae representing 10 different field-collected susceptible populations were RNA extracted and RT-PCR used to generate cDNA. To screen for any evidence of mutations in the ABCA2 gene, PCR of the cDNA and sequencing using appropriate primer pairs (see S1 Table) were performed. No evidence for any inactivating mutations was found in any of these 10 individuals, i.e. all contained untruncated transcripts at exons 16, 18 and 24 where Ha2Ab-R01, R02 and R03 alleles were detected, respectively.
A total of two transmembrane domains (TMDs; i.e., TMD 1, TMD 2) with each consisting of six transmembrane helices (TM I-VI in TMD 1; TM VII-XII in TMD 2, Fig 5) were predicted from the HaABC2 sequence that corresponded to those characteristic of the ABC transporter subfamily A. As for other members of this subfamily, N-glycosylation sites were also predicted for both of the extracellular domain (ECD) loops between TM I and TM II, and between TM VII and TM VIII (Figs 3 and 5). The intracellular loop between TMD 1 and TMD 2 (i.e., between TM VI and TM VII) and after TM XII contained the highly conserved regions for ATP Nucleotide Binding Fold 1 (NBF1) (including the Transporter signature Motif 1; TpM1), and NBF2 (and TpM2), respectively.
Each of the mutant alleles Ha2Ab-R01, Ha2Ab-R02 and Ha2Ab-R03 introduced stop codons into the reading frame, causing significant truncations of the HaABCA2 protein (Fig 4). The first two mutations introduced stop codons in the extracellular loop between TM VII and TM VIII of TMD 2, whereas the 5bp insertion that resulted in the Ha2Ab-R03 allele occurred just one amino acid after TM XII of TMD 2. All three mutations therefore truncated the protein before the second nucleotide-binding domain NBF2 (Fig 5), which would render the ABC transporter completely inactive, even if the protein were expressed and integrated into the cell membrane.
In the Cry2Ab-resistant H. punctigera Hp4-13 strain, the allele (Hp2Ab-R04) encoding the homolog of HaABCA2 was found to contain a 14bp deletion (Fig 4). This deletion disrupts the coding region of the transcript by introducing frame shifts that lead to a missense mutation and to the loss of the TpM2 transporter motif at the NBF2 (Figs 4 and 5). Although linkage mapping was not performed to conclusively associate this mutation with the resistant phenotype in H. punctigera, the fact that the same gene is mutated in a resistant strain in this species strongly supports the role of this gene as the target of mutations conferring resistance to Cry2Ab.
We next asked whether sequences for these ABCA proteins existed in genomes of other Lepidoptera, including some (H. virescens [65], Plutella. xylostella [66]; B. mori [67]) known to be susceptible to Cry2Ab. We examined the published genome sequences of Danaus plexippus [68], Heliconius melpomene [69], and P. xylostella [70,71]. Existing predictions of the genes were often inaccurate, e.g., as with the two partial predictions for BmABCA2, and ABCA1 was predicted as two separate partial proteins for the D. plexippus genome [68]. We used Scipio [72] and FGENESH [73] as well as additional transcriptomic data to generate complete predictions (GenBank Accession numbers KP219762-KP219770). In each species, the ABCA1 and ABCA2 genes were present, and situated adjacently in a tail-to-tail orientation just as in B. mori. Orthology was further confirmed by conserved flanking genes; in each species the ortholog of B. mori mitochondrial ribosomal protein S7 (GenBank Accession XP_004927481) occurred upstream of ABCA2, and the ortholog of B. mori ubiquitin carboxyl-terminal hydrolase (GenBank Accession XP_004927480) occurred upstream of ABCA1. An alignment of the predicted ABCA1 and ABCA2 proteins along with the respective Drosophila protein showing greatest similarity (S1 Fig) was used to construct a phylogenetic (ML) tree. This tree indicated that the gene duplication creating ABCA1 and ABCA2 likely occurred in the common ancestor of the Lepidoptera shown (Fig 6).
The importance of Bt toxins for insect pest and disease control has stimulated enormous interest in the study of their mode of action. For Cry1A toxins, there is specific and saturable binding to membrane targets, and a sequential mode of action has been proposed [25,39]. The toxin first binds to the 12-cadherin domain protein, resulting in processing and accelerated oligomerization before binding to membrane-bound glycosylated proteins such as aminopeptidases, alkaline phosphatase and other glycoproteins [25,26]; the integral membrane ABCC2 protein then facilitates pore insertion. Cry1A and Cry2A proteins have comparable three-domain structures [74,75], making them likely to act in similar ways as pore-forming toxins. Specific and saturable binding to membranes was also recently shown for Cry2Ab [22,75], and resistance is associated with a loss of binding [21]. Despite these similarities, toxicity of Cry2Ab in general is unaffected by mutations conferring Cry1Ac resistance. Specifically, mutations in APN or cadherin or ABCC2 do not render insects resistant to Cry2Ab, so that this toxin must be binding to one or more different targets.
The identification of ABCA2 suggests a mode of action of Cry2Ab differing slightly from that proposed for Cry1A toxins. ABCA2 carries two extracellular domains that are present as long loops between helices TM I and TM II, and between helices TM VII and TM VIII (Figs 3 and 5). Both of these loops are glycosylated in mammals [76], and six glycosylation sites are predicted for HaABCA2 (Fig 3). In contrast, for the lepidopteran ABCC2, the corresponding loops are very short and contain no glycosylation signals [34]. We hypothesize that Cry2Ab also has a sequential mode of action in which the ABCA2 protein itself is able to provide both binding and pore insertion functions. Specifically, Cry2A toxins would, upon activation [75] bind to the glycosylated ECD loops in TMD 1 and/or TMD 2. This binding could form the basis of oligomerization and bring the pre-pore structure close to the TMDs for pore insertion, as proposed for ABCC2 [34]. It is possible that other proteins may also be involved in Cry2Ab binding and pore formation, particularly since mammalian ABCAs have been suggested to occur in multi-protein complexes in the membrane [77]. Interestingly, the ABCA2 mutations confer resistance to very high concentrations of Cry2A [42], as would be expected if both receptor and pore insertion functions are simultaneously blocked. Similarly, in H. virescens the ABCC2 mutation results in higher levels of resistance to Cry1Ac than does the mutation in cadherin, but when both are homozygous in the same strain and both receptor and pore insertion functions are blocked, extremely high resistance levels result [34].
To what extent are these findings likely to apply to other Bt toxins of the Cry2A family? Phylogenetic analyses based on the shared common three-domain structure [78] showed that Cry2Aa and Cry2Ai are sister toxin groups and occupy a basal position to both Cry2Ab/Cry2Ag and Cry2Ae/Cry2Ah clades. Cross-resistance between Cry2Ab and Cry2Aa has been demonstrated in the SP15 strain of H. armigera [42], and between Cry2Ab and Cry2Ae in both H. armigera and H. punctigera [21]. Resistance to Cry2Aa has also been identified in H. virescens [53,79], in H. zea [15], in P. gossypiella [40], and in Ostrinia nubilalis [80]; ABCA2 remains to be investigated in these species. Cry2A toxins are also toxic to some Diptera [81,82] (but see [83]), with Cry2Ab recently shown to be effective against the malaria mosquito vector Anopheles gambiae [82]. Cry2Ab was ineffective against Aedes aegypti [82] but the similar toxin Cry2Ag was highly effective [84].
Seven ABC transporters of the A subfamily are present in the B. mori genome (three on BmChr17, and one each on chromosomes 5, 14, 16, and 19; see [61,62] for analysis of the initial gene models), and similar numbers have been found in other Lepidoptera with sequenced genomes. This subfamily has been well characterised in vertebrates; there are 12 members known in humans [85] and a similar number in the mouse; the nomenclature for these differs from that of insects. The HaABCA2 gene and the other insect genes shown in Fig 6 belong to an insect-specific clade within this subfamily, with no direct orthologs among the vertebrate genes [86]. The human ABCA genes have been extensively analysed and are expressed in a variety of tissues, with most being involved in lipid transport and trafficking. Mutations in human ABCA2 (not orthologous to lepidopteran ABCA2) are associated with early-onset of Alzheimer’s disease [87,88]. The mouse ABCA2 (an ortholog of the human ABCA2) has a possible role in regulating cholesterol homeostasis and low-density lipoprotein receptor metabolism in N2 neuroblastoma cells [89], with knock-out causing a ‘shaky’ (tremor) phenotype [90]. Mutations in the mouse ABCA3 gene, which is expressed in lung tissues, are associated with a foetal surfactant deficiency that is fatal. However, in H. armigera (but not H. punctigera), the Cry2Ab resistant line homozygous for ABCA2-inactivating mutation has no demonstrated substantial fitness costs compared to Cry2Ab susceptible insects [91]. Whether this is due to functional redundancy with another of the midgut-expressed ABCAs remains to be determined.
The frequency distribution of resistant ABCA2 alleles identified to-date is non-uniform. Seven resistant H. armigera lines, isolated independently from the field between 2002 and 2012, produced three resistant alleles. The Ha2Ab-R01 allele was present in five lines: SP15, 5–405, 9–4802, 10–485 and 12–2169; the Ha2Ab-R02 allele was homozygous in 6–364 and present in heterozygotes in 6–798; the Ha2Ab-R03 allele was the alternative allele in the heterozygous 6–798 line. A single mutant allele, Hp2Ab-R04, was found in the one H. punctigera resistant line HP4-13. Thus some alleles in H. armigera were common enough to be recovered several times from the field. Whether this is due to some selection by an unknown agent in the Australian environment as proposed [92] remains to be tested.
The still-rare resistance-conferring alleles identified in field populations occur at a limited number of locations in the gene. If confirmed by studies of further alleles, this raises the possibility that DNA-based screens will allow monitoring of the spread of Bt resistance in H. armigera and H. punctigera. Although PCR-based screens [29,93] for mutations in the 12-cadherin domain protein of H. armigera that confers resistance to Cry1Ab and Cry1Ac identified the same allele (r1) in field material from northern China as that originally identified as conferring resistance [12], this has not always been the case. For H. virescens [94], for example, different mutations in the same gene were identified by the F1 screen (i.e., mating field-caught individuals with an existing homozygous resistant strain and testing the F1 offspring). For P. gossypiella, screening in India [95] found different cadherin gene mutations to those originally identified in Arizona [96].
Every Cry2Ab-resistant line from an F2 screen that has been molecularly characterized has shown mutations in the ABCA2 gene. This confirms the value of using the less expensive F1 screen with ABCA2-mutant lines to extend the estimation of Cry2Ab resistance allele frequencies in Australia. Incorporating PCR-based screening will further improve detection efficacies of ABCA2-based resistance in the field, enabling more accurate and faster estimates of resistance allele frequencies, and is especially relevant for the analysis of historical field material generated through F1/F2 screening methods [43,92], and for tracking spatial and temporal movement patterns of resistance alleles across the landscape. Further characterisation of resistance-conferring ABCA2 alleles will also help to resolve the current discrepancy between the F2 screen and the F1 screen in estimating allele frequencies [91]. It will be important to determine whether Cry2Ab-resistance-conferring ABCA2 mutations occur in H. armigera elsewhere in its geographic range, including its recent incursions into the Americas [97,98]. Finally, examination of ABCA2 may provide insight in several species where the Cry2A resistance mechanism is still unknown, including H. virescens [53], P. gossypiella [40], H. zea [41], and Trichoplusia ni [99].
As a result of an F2 screen in 2002, the first H. armigera Cry2Ab resistant strain (Sp15) was established from a single pair of moths collected as eggs on corn near Griffith, New South Wales (NSW), Australia [42]. Detailed descriptions of the techniques employed have been provided [42,92]. F1 progeny from that pair were intercrossed and the resultant F2 larvae exposed to a screening concentration of Cry2Ab in ground leaf material of the cotton variety Sicala V-2 transformed with the B. thuringiensis variety kurstaki cry2Ab gene construct. Survivors among the F2 formed the basis of the resistant colony Sp15. Since 2003, F2 screens with H. armigera and H. punctigera performed as part of a resistance monitoring program have isolated additional lines [10]. The isolated H. punctigera Cry2Ab resistant line (Hp4-13) was established from eggs collected at St George, Queensland, Australia in 2004 [43]. Complementation tests for allelism established that one or more alleles at the same locus conferred resistance to Cry2Ab in five lines of H. armigera derived from the field in Australia from 2002 to 2006, including SP15 and 5–405 (previously named NA405) [100].
Assignment of the Cry2Ab resistance locus to an AFLP linkage group was carried out using the resistant line SP15 and the susceptible GR line. The initial cross was a SP15 Cry2Ab r/r ♂ x GR Cry2Ab s/s-♀, yielding family G. An F1 female from family G was crossed to an SP15 male to produce the female informative backcross family F2031. Backcross progeny were bioassayed using the discriminatory Bt Cry2Ab concentration [92] to select for homozygous resistant (Cry2Ab r/r) individuals. Additional backcross progeny were not exposed to Cry2Ab, to serve as controls. AFLPs [101] from genomic DNA of grandparents, parents and 59 progeny of family F2301 were analysed for linkage using the method of Heckel et al. [18]. Twenty-nine progeny were survivors of exposure to Cry2Ab and 30 were untreated controls. AFLPs were grouped using the program DBM3Lnk.p as in Heckel et al. [18]. As expected from achiasmatic oogenesis in female Lepidoptera, no recombinants were found within AFLP linkage groups. Linkage to resistance was tested for each linkage group using signed interaction chi-squared tests with one degree of freedom [102], with a Bonferroni correction for 31 linkage groups.
One AFLP band from the only linkage group with a significant association with resistance was cut out of the gel, reamplified, cloned and sequenced (GenBank Accession No. KJ419919). The insert was hybridized to a Southern blot made from an unrelated Bt-susceptible H. armigera family in which several ribosomal protein genes had previously been mapped, enabling comparison to the homologous ribosomal protein genes of Bombyx mori. This showed AFLP group 8 in family F2301 to correspond to B. mori chromosome 17 (BmChr17).
Additional linkage mapping in a male-informative backcross (G2016) and two F2 families (G2020, G2029) was performed to further localize the resistance locus. Offspring from these families that had survived the discriminating concentration and were presumably homozygous for the SP15-derived resistance allele were examined for recombinants at marker loci. H. armigera homologs of ribosomal protein genes RpL22 and RpS24 on B. mori BmChr17 were sequenced to identify polymorphisms to be used in mapping. The gene bre-5 on BmChr17 was considered a candidate for the resistance gene because of its role in Cry4B resistance in the nematode Caenorhabditis elegans [19], and was also mapped using sequence variation in the coding region and a PCR-RFLP using a polymorphic PstI restriction site.
To establish an appropriate mapping family, a GR Cry2Ab susceptible homozygous male (Cry2Ab s/s ♂) was mated with a SP15 Cry2Ab resistant homozygous female (Cry2Ab r/r ♀). The F1 susceptible heterozygous male (Cry2Ab r/s ♂) was back-crossed to a SP15 female to obtain F2 offspring of either homozygous resistant (Cry2Ab r/r) or heterozygous susceptible (Cry2Ab r/s) genotypes in equal proportions. Approximately 300 F2 offspring were bioassayed using the discriminatory Bt Cry2Ab concentration [92] to select for homozygous resistant (Cry2Ab r/r) individuals. Control (n = 100) F2 offspring were not bioassayed and were included in subsequent genotyping experiments using EPIC-PCR markers as described below.
EPIC PCR markers used in this study were designed using the primer designing criteria previously reported by [103] for H. armigera. Briefly EPIC-PCR primers were designed using the primer analysis software Oligo Version: 7.17 (Molecular Biology Insights, Inc., Cascade, CO 80809, USA) and avoiding false primer annealing sites for both forward and reverse primer, with no or minimal hairpin structures and primer dimmer formation. We also designed the EPIC-PCR primers with intron amplicon of typically less than 500bp such that polymorphisms in F2 cross can be easily scored. Intron sizes were estimated based on B. mori gene annotation. EPIC-PCR primers were optimised prior to having a fluorescent tag (FAM, HEX or TET) attached to the 5’ end of the forward primer. Amplicons of the mapping family from individual EPIC-PCR primer pairs were visualised on 1–1.5% agarose gels prior to being purified by acetic acid/ethanol precipitation and sent to Genetic Analysis Facility (GAF) at James Cook University (JCU) for genotyping. PCR conditions, and genotyping procedures were previously described [103,104]. A list of all EPIC-PCR primers used in this study can be found in S2 Table.
Genomic DNA was extracted using the Qiagen Blood and Tissue extraction kit (Qiagen Cat. #69506). For the founding grandparents (i.e., F0) and parents (i.e., F1) one leg each was used in gDNA extraction, with gDNA eluted in 200°L of the AE buffer. Bioassayed and control F2 samples were collected as 3rd instar larvae and gDNA was extracted as for the parents and grandparents. All genotyping with EPIC-PCR markers involved screening of grandparents, F1 parents, 72 bioassayed (Cry2Ab r/r) offspring and 20 control F2 offspring (i.e., either Cry2Ab r/r or Cry2Ab r/s). Under the linkage mapping pattern, genome/chromosome walking towards the resistance gene should generate reduced recombination rates in the resistant F2 as one approaches the genomic region of interest.
Messenger RNA sequencing was done in order to generate full-length transcripts in H. armigera for candidate genes, identify resistant alleles where cDNA amplification failed and to identify the homologous candidate genes in H. punctigera. Total RNA was extracted from the midgut of third-instar larvae or whole larvae using the TRIzo Plus RNA purification kit (Life Technologies, Cat # 12183555) and dried down for shipping with RNAstable Tube Kit (Biometrica Cat. # 93221–001). RNAseq library preparation, sequencing and bioinformatic analysis was done according to standard Illumina protocols by the Beijing Genomics Institute (BGI) in Shenzen, China.
Except for the resistant line 7–183 which used gDNA as a template for sequencing, candidate genes from the remaining resistant lines were completely sequenced using a cDNA template from 3rd instar larvae prepared using an RNA extraction kit (Qiagen RNeasy mini kit, Cat. # 74106), and trace genomic DNA contaminants removed using the Qiagen RNase-Free DNase set (Cat. # 79254). First strand cDNA was synthesised using the Invitrogen SuperScript III RT First Strand Synthesis System for RT PCR (Cat. # 18080–051), in the presence of RNase H. All sequencing was performed at the John Curtin School of Medical Research, Australian National University (ANU), and used the ABI BigDye v3 chemistry. Contig assembly used the Staden pregap4 and Gap4 software [105] and was visualised using Artemis (Release 12.0) [106]. Sequences generated and used in this study have been deposited in GenBank (Accession numbers KP259910, KP259911, KP259912).
The amino acid sequence predicted from a complete mRNA sequence from a Cry2Ab susceptible individual belonging to the GR-line was used to predict the domain structure of the H. armigera ABCA2 protein. The protein prediction software Split V3.5 <http://split4.pmfst.hr/split/4/> [107] was used to search for transmembrane protein secondary structure (i.e., transmembrane helices). In the mouse RmP ABC transporter, several N-glycosylation sites were predicted on the protein’s extracellular domains [76,108]. We used the NetNGlyc 1.0 Server <http://www.cbs.dtu.dk/services/NetNGlyc/> developed to predict N-Glycosylation sites in human proteins for the purpose of predicting N-Glycosylation sites in the protein sequences. The software uses artificial neural networks to examine for Asn-Xaa-Ser/Thr sequence context.
Sequences in the transcriptome databases corresponding to candidate genes were identified by standalone BLAST. Homologs in GenBank were identified using BLAST and homologous gene clusters identified in NCBI and in Ensembl. Orthologous genes from other lepidopteran genomes were retrieved using their online databases from public domains as cited in the appropriate sections below. Protein sequences were aligned using Multiple Alignment using Fast Fourier Transform (MAFFT) [109] <http://www.ebi.ac.uk/Tools/msa/mafft/> and phylogenetic tree (maximum likelihood (ML) with rapid bootstrapping) inference using RAxML-HPC2 on XSEDE (8.0.24) (available at the CIPRES Science Gateway V3.3) <http://www.phylo.org/sub_sections/portal/> [110–112], and redrawn using Dendroscope version 2.4 [113].
|
10.1371/journal.ppat.1004600 | Reprogramming of Yersinia from Virulent to Persistent Mode Revealed by Complex In Vivo RNA-seq Analysis | We recently found that Yersinia pseudotuberculosis can be used as a model of persistent bacterial infections. We performed in vivo RNA-seq of bacteria in small cecal tissue biopsies at early and persistent stages of infection to determine strategies associated with persistence. Comprehensive analysis of mixed RNA populations from infected tissues revealed that Y. pseudotuberculosis undergoes transcriptional reprogramming with drastic down-regulation of T3SS virulence genes during persistence when the pathogen resides within the cecum. At the persistent stage, the expression pattern in many respects resembles the pattern seen in vitro at 26oC, with for example, up-regulation of flagellar genes and invA. These findings are expected to have impact on future rationales to identify suitable bacterial targets for new antibiotics. Other genes that are up-regulated during persistence are genes involved in anaerobiosis, chemotaxis, and protection against oxidative and acidic stress, which indicates the influence of different environmental cues. We found that the Crp/CsrA/RovA regulatory cascades influence the pattern of bacterial gene expression during persistence. Furthermore, arcA, fnr, frdA, and wrbA play critical roles in persistence. Our findings suggest a model for the life cycle of this enteropathogen with reprogramming from a virulent to an adapted phenotype capable of persisting and spreading by fecal shedding.
| To establish infection and colonize within a host, infecting pathogens have to cope with a variety of destructive surroundings. The food-borne pathogen Y. pseudotuberculosis can cause persistent infection in mice. Upon infection, Y. pseudotuberculosis passes the anti-microbial gastrointestinal milieu and finally remains associated with lymphoid follicles in cecal tissue surrounded by polymorphonuclear leukocytes, indicating that the bacteria are exposed to multiple environmental cues. We performed complex RNA-seq of small cecal biopsies of infected mice to reveal Y. pseudotuberculosis gene expression in vivo. We found that Y. pseudotuberculosis underwent reprogramming from a virulent phenotype, expressing virulence genes during early infection, to an adapted phenotype capable of persisting in the harsh cecal environment. Persistence was characterized by a novel expression pattern with down-regulation of virulence genes and up-regulation of genes involved in anaerobiosis, chemotaxis, and protection against oxidative and acidic stress. Mutagenesis of selected genes revealed that the regulator rovA was critical for the establishment of infection, and that arcA, fnr, frdA, and wrbA play critical roles in maintaining infection for long periods of time. Our study shows the power of RNA deep sequencing, which can be used to reveal the in vivo expression patterns of small amounts of bacteria in complex intestinal environments.
| Yersinia pseudotuberculosis is a food borne pathogen that can penetrate the intestinal epithelium and cause gastroenteritis. This enteropathogen invades lymphoid follicles of Peyer’s patches and cecum, where it survives extracellularly before breaking the barrier and becoming systemic [1,2]. All pathogenic Yersinia species are capable of inhibiting important host immune mechanisms in local lymph nodes, and this essential virulence property is dependent on the plasmid-encoded Yersinia outer proteins (Yops) YopE, YopH, YopJ, YopM, YopT, YpkA, and YopK. Upon intimate contact with a target host cell, the Yops are delivered into the host cell via the Yersinia type three secretion system (T3SS) [3]. Inside the target cell, the Yop effectors interfere with several key mechanisms of the host immune defense; for example, YopH and YopE inhibit phagocytosis and YopJ interferes with the production of pro-inflammatory signaling molecules [4]. Polymorphonuclear neutrophils (PMNs), which are rapidly recruited to infection sites, are the main target cells for Y. pseudotuberculosis T3SS-mediated Yop translocation during infection [2,5]. Current knowledge of Y. pseudotuberculosis virulence mechanisms is based, to a great extent, on studies using the acute mouse infection model in which infection results in systemic infection.
We recently found that the enteric pathogen Y. pseudotuberculosis can cause persistent infection in mice, where it persists associated with the lymphoid follicles of the cecum [6]. In this model, low dose oral infection (106–107 colony-forming units (CFUs)) leads to asymptomatic infection in 20–30% of infected mice with an observed infection duration as long as 115 days. Even if no signs of disease are present, Y. pseudotuberculosis persistence is associated with an immune response in which the Y. pseudotuberculosis foci are surrounded by PMNs and bacteria are shed in feces [6]. Yersinia, Salmonella, and Campylobacter have all been reported to infect and affect the ileocecal area in humans [7], suggesting that the cecum is a beneficial niche for bacterial persistence.
Many pathogenic bacteria are capable of maintaining infection in mammalian hosts, giving rise to persistent infections [8]. Diagnosis of a persistent infection can be difficult, as symptoms are not always obvious. Prolonged persistent infections can cause chronic inflammation, which can lead to complications, or even precipitation of certain diseases in susceptible hosts [9]. In addition, persistent bacterial infections are a major cause of the overuse of antibiotics in both humans and animal husbandry. Increasing evidence indicates that persistence contributes to the development and spread of antibiotic resistance [10]. Therefore, the identification of bacterial mechanisms involved in the development of persistent infections is of great interest. One well-established model of persistence is Salmonella typhimurium; upon infection of Nramp1-expressing mice, this intracellular pathogen can persist inside phagocytic cells in classical granuloma lesions in the spleen, liver, and mesenteric lymph nodes [11,12]. Another model of S. typhimurium is the infection of antibiotic-treated DBA/2 and 129Sv/Ev mice, which results in colitis and chronic cholangitis [13]. Similar to the Y. pseudotuberculosis persistence model, the colitis phase is associated with PMN infiltration into the cecum and bacterial shedding in the feces. Studies of S. typhimurium persistence using these models have shown a variety of factors that contribute to persistent infection [14], including effectors encoded by the pathogenicity islands SPI1 and SPI2, which are required for initial invasion and intracellular growth. Two SPI2 factors, Sse1 and SseK2, have been implicated as being important for later stages of infection, with Sse1 affecting host cell adhesion and migration. Different adhesive proteins and factors that protect against host-derived antimicrobial peptides and factors that aid in coping with oxidative and nitrosative stress have been found important for sustained colonization of S. typhimurium in the gastrointestinal tract, as well as systemic persistence [14].
The mechanisms enabling Y. pseudotuberculosis to persist in cecal tissue in the presence of immune cells for a prolonged period of time are largely unknown. Our previous study showed that the T3SS effectors YopH and YopE contributed to Y. pseudotuberculosis persistence in the cecum, likely by enabling initial colonization in the presence of phagocytic cells [6]. One way to understand mechanisms and metabolic traits that are important for Y. pseudotuberculosis persistence in the cecum is to identify the genes involved. Several methods have been used to identify genes induced in vivo during infection, such as in vivo expression technology (IVET) [15], signature-tagged mutagenesis (STM) [16], cDNA microarray analysis [17,18], and the recently developed RNA sequencing technology with massively parallel cDNA sequencing (RNA-seq) [19]. IVET and STM, which are both based on infections with bacterial libraries, are not suitable for in vivo studies of enteropathogenic Yersinia due to restricted clonal invasion of the intestinal tissue by this pathogen [20,21]. Furthermore, because the intestinal tract is colonized by the intestinal microflora and harbors many commensal bacteria, the reliability of DNA microarray is greatly diminished due to cross-reactions between species-specific probes on the microarray chips. In contrast, RNA-seq provides a promising approach for monitoring gene expression in a specific organism in the presence or absence of others. This method is sensitive and allows accurate discrimination between similar RNAs originating from different species, offering an excellent opportunity to reveal the gene expression patterns of pathogens within host tissues, even in heavily colonized environments, such as the intestine, stomach, and cecum.
In this study, we performed RNA-seq on Y. pseudotuberculosis YPIII in small cecal tissue biopsies from mice at early and persistent stages of infection to reveal mechanisms of importance for persistent infection. We found that the bacteria underwent substantial transcriptional reprogramming. Initially, genes encoded on the virulence plasmid, including T3SS and associated effectors, were highly up-regulated. At the persistent stage these genes were found to be down-regulated, and other genes were up-regulated, including those encoding for anaerobic growth, motility, protection against acidic and oxidative stress, and genes indicating envelope perturbation, suggesting adaptation to the harsh environment in cecal tissue.
To identify mechanisms promoting Y. pseudotuberculosis persistence, RNA-seq was employed to determine the differential gene expression profiles of bacteria in the cecum during the early phase of infection and during persistence. To obtain infected tissue for the isolation of Y. pseudotuberculosis RNA, FVB/N mice were infected orally with bioluminescent wild-type (wt) bacteria at an infection dose of ∼2×107 CFUs. The infection was monitored in real time by an in vivo imaging system (IVIS) at certain intervals for 42 days. In agreement with that reported earlier [6], we found bacterial foci associated with the cecal lymphoid tissue (Fig. 1A–B), massive infiltration of PMNs surrounding the bacterial foci (Fig. 1C–D) as well as superficial destruction of the epithelial lining and mixed inflammatory infiltrates.
For sample preparation, isolated cecal tissues from 2 and 42 days post-infection (dpi) were analyzed by IVIS to verify that they contained Yersinia. The bioluminescent signal from Y. pseudotuberculosis allowed identification of the precise location of the bacteria in the tissue. Small biopsies (3 mm Ø) of cecal tissue containing bioluminescent bacteria were isolated using a hole punch. Total RNAs were extracted from biopsies from two mice infected for 2 days and two asymptomatic mice infected for 42 days. As a control, we extracted total RNAs from the cecal tissue of two un-infected mice. We also included RNA samples from bacteria grown in Luria broth (LB) in vitro at 26°C and in Ca2+-depleted LB at 37°C, a condition known to induce T3SS [22], hereafter referred to as T3SS-inducing conditions. The quality and quantity of all RNA samples were determined using an Agilent Bioanalyzer 2100, and all total RNA preparations had RIN values >7. This analysis revealed pure bacterial RNA in the in vitro samples (16S and 23S rRNAs) and mouse RNA (18S and 26S rRNAs) in samples from un-infected cecal tissue. As expected, both eukaryotic and prokaryotic RNAs were detected in the samples from infected cecal tissue and appeared as four distinct bands: 16S, 18S, 23S, and 26S rRNAs (Fig. 2A). However, as we previously recovered only 1×105 to 2×106 CFUs Y. pseudotuberculosis from cecal tissues [6], the amount of prokaryotic RNA was unexpectedly high in the infected tissues. Therefore, we performed qPCR to determine the Yersinia RNA abundance in infected tissues by comparing ymoA expression as an indication of Yersinia RNAs and GAPDH expression as an indication of host RNAs, finding ∼0.2% Yersinia RNA in the total RNA preparations. Therefore, the remarkably higher amount of prokaryotic RNA compared to that predicted for Yersinia RNA was assumed to reflect the presence of other microbial inhabitants in the samples.
To identify other bacteria in the cecum during the two different phases of infection, the in vivo-derived total RNA samples were analyzed by RNA-seq. The sequencing reads from these samples were mapped initially to the NCBI 16SMicrobial database with a full alignment parameter in order to map only unique reads to each 16S rRNA sequence. Next, matched 16S rRNA sequences were filtered with at least 80% coverage due to conserved regions in the 16S rRNA sequences. The number of reads mapped to each sample was normalized to the depth of sequencing in order to estimate the relative bacterial load in the tissue samples. Compared to un-infected samples, the bacterial RNA content of samples from early infection was 5.2-fold higher, and from persistent infection 3.7-fold higher, according to the normalized ratio of mapped reads for each tissue sample (Fig. 2B and S1 Table). The presence of other bacteria in uninfected samples likely reflects luminal bacteria, whereas the greater amount of bacteria in infected cecum samples may indicate that the infection leads to dysregulation of the luminal microbiota and/or that luminal bacteria gained access to the tissue. We identified 11 species in uninfected samples, 30 in early infection samples, and 11 in persistent infection samples (S1 Table). The identified species were grouped by phylum using RDP Classifier [23]. The abundance and composition of bacterial phyla in uninfected samples were similar to previous reports [24,25] but differed in samples from infected cecums. Samples from early infection (Firmicutes 60%, Bacteroidetes 27%, Proteobacteria 3.3%) differed from samples from persistent infection (Firmicutes 27%, Bacteroidetes 55%, Proteobacteria 9%, Verrucomicrobia 9%; Fig. 2C. The two independent replicates of the persistent sample contained similar species, and the strictly anaerobic Gram-negative bacterium Akkermansia muciniphila [26] was present in high abundance (75% of its transcriptome was revealed by RNA-seq; S1 Fig.).
Given the relatively small amount of Y. pseudotuberculosis in the cecal tissue, samples were enriched for bacterial mRNA by depleting fractions of poly(A)-tagged RNAs, rRNAs, and tRNAs prior to RNA-seq. The enrichment procedure was performed with both in vivo and in vitro total RNA samples.
Rigorous validations on mapped reads were required due to the presence of RNAs from other microbial inhabitants and homologous mRNAs that did not represent true Y. pseudotuberculosis transcripts. The genomes available for bacterial species found in the in vivo samples (42 genomes) were used as reference to optimize the alignment parameters for sequencing reads. Eventually, the optimization trials ended with a strict criterion of 95% alignment specificity to retrieve reads uniquely mapped to Y. pseudotuberculosis in the presence of other bacterial genomes. Finally, the uniqueness of these reads was double-checked with probabilistic variant detection, which searches for single nucleotide polymorphisms (SNPs) using CLC Genomic Workbench. With sample enrichment, optimized alignment parameters, and high sequencing depth, we revealed ∼92% of the metatranscriptome, which was composed of mouse and 42 other bacterial species in addition to Y. pseudotuberculosis. The reads mapped to Y. pseudotuberculosis were used in the subsequent RNA-seq analysis in order to calculate the expression of each open reading frame (ORF). The range of reads per ORF for in vitro and in vivo-derived samples was up to more than 200,000 and 490, respectively. Therefore, fewer ORFs were mapped (1551 ORFs, 36% transcriptome coverage) in vivo than in vitro, which had complete coverage (Table 1). Because we reached full coverage of the mouse transcriptome, and in some cases up to 75% coverage of other bacterial transcriptomes in the in vivo-derived samples (S1 Fig.), the relatively low coverage of Y. pseudotuberculosis was due to the very low abundance of its transcripts. Nevertheless, the RNA-seq results for Y. pseudotuberculosis showed very high correlation (Pearson and Spearman correlation R-values ≥0.98) of normalized RPKMO values (i.e., reads per kilobase pairs of a gene per million reads aligning to annotated ORFs) between biological replicates of all samples, which verifies the robustness of the analysis (Table 1). In addition, transcriptionally active regions were found to be the same for both in vitro and in vivo-derived samples (Fig. 3A, highlighted with gray borders). Even though the number of reads varied (53 to 111,817 reads) between different in vivo and in vitro samples, the distributions of the reads were similar for active regions (Fig. 3B). As qPCR requires more starting material than RNA-seq, we tested the differential expression of 9 genes that were highly expressed during persistent infection. The qPCR analysis confirmed the differential expression of all tested genes (Fig. 3C). Thus, the RNA-seq results for the in vivo samples provided valid information about differentially expressed genes in Y. pseudotuberculosis during early versus persistent infection in mice.
RNA-seq analysis of the in vitro samples revealed that 665 genes were differentially expressed (log2 fold change ≥0.7, p <0.05) under the different in vitro growth conditions. A total of 146 genes were up-regulated at 37°C (T3SS-inducing conditions), 55 of which were located on the 70-kb virulence plasmid that encodes the T3SS components; 519 genes, including 29 flagellar genes, were up-regulated at 26°C. This confirmation of T3SS induction by the temperature shift to 37°C, combined with Ca2+-depletion [22] and the motile phenotype of Y. pseudotuberculosis at 26°C [27], verifies the reliability of the RNA-seq analysis. The global expression patterns of cultured bacteria at 26°C and 37°C are shown on a histogram and heat map in Fig. 3A (see also S2 Table).
A total of 1288 genes were found to be differentially expressed (log2 fold change ≥0.7) in vivo. RPKMO values detected by RNA-seq are shown in a histogram and heat map in Fig. 3A to highlight the differences in individual ORFs during early and persistent infection (see also S3 Table). Surprisingly, the T3SS components encoded on the virulence plasmid that were highly expressed during the early stage of infection were distinctly down-regulated during persistent infection. Another conspicuous finding was the up-regulation of flagella and chemotaxis genes. T3SS is known to be induced at 37°C, but flagella are down-regulated at this temperature; in vitro, flagella are expressed only at 26°C (confirmed by qPCR in the same samples used in RNA-seq; Fig. 4A–B). In analogy, the T3SS master regulator lcrF was up-regulated during early infection and down-regulated during persistence (S3 Table), and the flagellar regulator flhCD was down-regulated during early infection and up-regulated during persistence (Fig. 4B). In addition, up-regulation of the gene encoding the adhesion protein invA, which is co-regulated with flagella [28], and its positive regulator rovA [29] suggested that other genes that are only expressed at 26°C in vitro could be up-regulated during persistence. Accordingly, a comparison of the expression patterns of in vivo and in vitro-derived samples showed that, during the early phase of infection, bacteria have an expression pattern similar to that seen in vitro at 37°C, whereas the expression pattern of persistent bacteria was much more similar to that of bacteria grown in vitro at 26°C (Fig. 5A and S4 Table). These results clearly indicate that, though increased temperature triggers T3SS and associated virulence genes during initial infection, other environmental cues are responsible for the observed transcriptional reprogramming of Y. pseudotuberculosis during prolonged infection.
Flagellar genes are known to be down-regulated at 37°C; therefore, the up-regulation of flagellar genes at later time points of infection at this temperature suggests that motility may be important for certain stages of persistence. However, Y. pseudotuberculosis is expected to remain flagellated for some time after the temperature shift and therefore flagella have also been assumed to participate in initial infection. To investigate this possibility, Y. pseudotuberculosis was grown at 26°C and then shifted to 37°C, followed by sampling at different time points for the detection of flagellated bacteria using atomic force microscopy. This analysis showed that the flagella remained for at least 2 hours after shifting temperature (S2 Fig.).
To obtain an overview of the diversity of metabolic pathways and other functional systems utilized by persistent bacteria, functional clustering was performed using KEGG pathway mapping [30] for the 1288 genes differentially expressed in vivo (Fig. 5B). Down-regulation of T3SS and up-regulation of motility genes during persistence were verified. Induction of genes involved in ribosome biogenesis, amino-acyl tRNA biosynthesis, and RNA degradation suggests an active metabolic state during persistent infection. Induction of DNA replication and repair, as well as purine and pyrimidine biosynthesis, indicates proliferation of persistent bacteria and correlates well with constant bacterial shedding from the tissue to luminal sites during infection. The induction of TAT secretion components, which are involved in the transport of proteins synthesized mostly under anaerobic conditions [31], and induction of genes indicative of oxidative phosphorylation reflect an anaerobic/microaerophilic environment. Moreover, up-regulation of genes associated with two-component signal transduction systems, type VI secretion system, and chemotaxis indicates the presence of various external stimuli within the cecal environment. In addition to the functional annotations, the up-regulation of other genes indicates that the bacteria are influenced by different environmental conditions, such as acidic, oxidative, and other forms of stress (Table 2). The up-regulation of genes encoding proteins involved in envelope biogenesis, including a variety of inner and outer membrane proteins, and lipopolysaccharide biosynthesis also suggests an environmental influence. Therefore, the expression pattern of persistent bacteria suggests adaptation to an environment with limited oxygen and oxidative and acidic stress, a need for motility/chemotaxis, and modulation of the bacterial surface (Table 2).
The up-regulation of several genes encoding proteins involved in oxidative phosphorylation in persistent bacteria (Fig. 5B) raised questions about energy metabolism and nutrient utilization. The cecal environment is expected to be anaerobic, and this was indicated by the sequencing data. The induction of genes involved in switching from aerobic to anaerobic respiration, such as the two-component system genes arcA-arcB, the fumarate nitrate reductase gene fnr encoding a global regulator of anaerobic growth, and other functional genes encoding proteins involved in microaerophilic/anaerobic respiration (Table 2), prompted us to investigate the influence of the anaerobic environment. We compared the expression profiles of Y. pseudotuberculosis grown in vitro under aerobic and anaerobic conditions at 26°C during the exponential and stationary phases using microarrays. Comparison of the differentially expressed genes identified in persistent bacteria by RNA-seq with the genes identified by microarray to be differentially expressed during anaerobiosis showed high similarity between persistent infection and in vitro anaerobic growth (Fig. 5C). Up to 42.5% of the genes up-regulated during persistence were also up-regulated during anaerobic growth (Fig. 5C and S4–S5 Tables). We found no obvious bias towards genes differentially expressed in the logarithmic or stationary phase (S3 Fig.). Thus, a substantial part of the expression profile of persistent bacteria is due to limited oxygen availability. The induction of some genes associated with anaerobic growth was also evident in samples from early infection (Fig. 5C), suggesting adaptation to the new environment at this stage.
The observed up-regulation of invA and its positive regulator rovA suggests that the RovA regulatory cascade contributes to the expression pattern of persistent Y. pseudotuberculosis. RovA is a regulator of the MarR/SlyA family, which controls different physiological processes [32]. Expression of rovA is controlled by the global regulators CsrA and Crp [33], which were both up-regulated during persistence. PhoP, which positively regulates rovA in Y. pestis and some Y. pseudotuberculosis strains, is not functional in the YPIII strain, where instead the Csr system via differential regulation of Csr RNAs influences production of RovA [34]. The RovA regulon of the Y. pseudotuberculosis YPIII strain (grown in vitro at 26°C) was recently revealed by microarray analysis [35]. A comparison of the gene expression patterns of persistent Y. pseudotuberculosis and the reported regulon [35] revealed 27.7% of the RovA regulon with 62 activated (lpp, fliC, and ftn verified by qPCR; Fig. 3C) and 26 suppressed genes in persistent bacteria (S4 Fig. and S4 Table). A comparison of the Y. pseudotuberculosis in vivo transcriptome and the Crp and CsrA regulons [35] revealed that more than 20% of the respective regulons represented genes identified as being differentially expressed during persistence. Some of the identified genes were shared and others were unique for Crp or CsrA (S4 Fig. and S4 Table). Notably, the fractions of Crp, CsrA, and RovA regulons observed in persistent bacteria appear to be relatively high with regard to the low coverage of the Y. pseudotuberculosis in vivo transcriptome.
To determine the importance of potential persistence genes identified in the RNA-seq analysis, we constructed a set of single gene deletion mutants to test in the mouse infection model of persistent infection. We selected genes implicated in different environmental responses (i.e., rovA, arcA, fnr, hdeB, uspA, napA, frdA, motB, cheW, and wrbA). Infection was achieved and monitored with IVIS for up to 42 dpi. Among the mice infected with the wt strain 31.6% had persistent infection, 31.6% cleared the infection, and 36.8% succumbed to severe disease (Fig. 6A), and this distribution is in accordance with that reported earlier [6]. The ΔrovA strain was completely attenuated and did not establish infection, indicating that it is indispensable for initiation of infection. Three of the mutants lacking genes involved in anaerobic respiration (arcA and fnr) or oxidative stress (wrbA) had a reduced capacity to establish persistence (Δfnr, 15.4%; ΔarcA, 7.7%; and ΔwrbA, 6.7%). These strains also gave less and later onset of severe disease than the wt strain (Δfnr 15.4%; ΔarcA, 7.7%; and ΔwrbA, 14.2%; Fig. 6A and S5 Fig.). A similar but less dramatic phenotype was observed for the ΔfrdA mutant, which had a reduced capacity to establish persistence (7.7%) but still caused severe disease (38.5%), with even earlier onset than the wt strain (Fig. 6A and S6 Fig.). ΔarcA, Δfnr, ΔfrdA, and ΔwrbA, established infections as wt, with almost similar levels of clearance during the first 7–14 days. However, after this, there were significant differences between these mutants and the wild-type strain with higher degrees of clearance, suggesting that these gene products are important during the persistent state (Fig. 6B). Interestingly, ΔhdeB was cleared to almost the same extent as the wt strain but was unable to establish persistent infection; this strain also appeared to be more aggressive, with infection resulting in more frequent and faster onset severe disease (68.8% vs. 31.6% for the wt strain; Fig. 6A and S5 Fig.). The ΔuspA mutant had a higher frequency of persistence than the wt strain (55.6% vs. 31.6%). The other mutants, ΔnapA, ΔmotB, and ΔcheW, had infection profiles similar to the wt strain (Fig. 6A).
The decreased virulence and persistence of the mutants related to anaerobic growth (ΔarcA, Δfnr, and ΔfrdA), combined with the down-regulation of T3SS and up-regulation of motility genes during persistent infection, prompted us to analyze T3SS secretion and motility in vitro in the presence and absence of oxygen. Secretion of T3SS effectors was reduced under anaerobic conditions for wt and all tested mutants (S6A Fig.). The reduction in T3SS effectors was also confirmed by qPCR (S6B Fig.). Moreover, to evaluate the contributions of the selected genes to bacterial motility, strains with mutations affecting the establishment of persistent infection were subjected to a motility test at 26°C and under T3SS-inducing conditions at 37°C in the presence and absence of oxygen. Neither the wt nor any of the mutants were motile at 37°C, independent of oxygen, but all strains except a ΔfliC mutant were motile at 26°C in the presence of oxygen. Interestingly, in the absence of oxygen at 26°C, the motility of Δfnr and ΔarcA mutants was significantly reduced (Fig. 6C).
Next, we investigated the influence of acid on T3SS and motility. Similar to what was observed for low oxygen, acidic conditions (pH 5.2) inhibited the induction of T3SS under inducing conditions at 37°C (S6A–S6B Fig.). Inhibition of T3SS under acidic conditions in Y. pseudotuberculosis was reported previously and suggested to involve pH-dependent inhibition of YscU proteolysis [36]. Furthermore, acidic conditions completely repressed the motility of the wt strain and all mutants tested in this study. Thus, both low oxygen and acidic conditions have a negative effect on T3SS induction and represent environmental cues that can contribute to the observed reprogramming of Y. pseudotuberculosis from a virulent to adaptive state.
In this study, we applied in vitro and in vivo-based RNA-seq to determine key players that enable Y. pseudotuberculosis to establish a persistent infection. We found that the bacterium undergoes reprogramming from a virulent phenotype, massively expressing T3SS components in the early invasive phase of infection, to an adapted phenotype capable of persisting in a microaerophilic and hostile environment. This finding has clear impact on future rationales to identify bacterial targets for new antibiotics.
As shown here, RNA-seq is a powerful method for retrieving robust information about bacterial gene expression profiles during in vivo infection. We show that pathogen gene induction can be detected even if the amount of infecting bacteria in the isolated tissue is very low and mixed with other bacterial species. Thorough work with varied optimization steps allowed us to discriminate Yersinia-specific reads at the resolution of a single base with partial coverage from in vivo-derived samples and full coverage from in vitro-derived samples. We found that very strict read mapping parameters should be used to discriminate Y. pseudotuberculosis-specific reads for in vivo data. This strategy, which was established and optimized in this study, provides a controlled solution for discriminating between species-specific transcripts in complex RNA populations. Using this methodology we obtained robust data, revealing 36% of the Y. pseudotuberculosis in vivo transcriptome and providing novel information about bacterial gene expression during infection.
A discrepancy in the level of coverage between in vitro and in vivo-derived bacteria was expected based on previous studies of Vibrio cholerae [19] and Campylobacter jejuni [37]. The starting materials in those studies were cecal or intestinal contents, not infected tissue, and contained 100–300 times more bacteria. In addition, the overall coverage in those studies was higher than what we obtained with our tissue biopsy approach. However, the overall coverage for the metatranscriptome was >92%, and the ratio of Y. pseudotuberculosis total RNA was ∼0.2%. Full coverage of a bacterial transcriptome in such a complex population with low abundance is estimated to require a sequencing depth of at least 1.5–2 billions reads [38]. This estimate was calculated based on only the host and Y. pseudotuberculosis RNA being present in the sample, whereas additional RNAs from many bacterial species were present in our samples. Therefore, the depth of sequencing for such biopsy samples may need to be several times higher than 1.5–2 billion reads.
The massive expression of T3SS virulence genes during the early phase of infection is likely necessary to break the epithelial barrier and defend against innate immune cells. This assumption is supported by previous data showing that yopH or yopE mutants are defective in establishing the initial infection and less able to cause persistent infection [6]. Later, the bacteria become persistent with a novel expression profile, suggesting substantial transcriptional reprogramming. At this stage, the T3SS components are down-regulated. Thus, the bacteria prefer to use other genetic resources to adapt to the environment instead of producing massive amounts of invasive T3SS components.
Taking the host temperature into consideration, the repertoire of up-regulated genes in persistent bacteria was remarkably different from the repertoire of bacteria grown in vitro at 37°C. One striking observation was the up-regulation of flagella at 37°C, as achieving the induction of flagella is impossible at this temperature in vitro. Therefore, the situation in the animal greatly differs from laboratory settings. The surprising finding that the expression pattern seen during persistence is similar to the pattern seen at 26°C in vitro, indicates that the pathogens during infection encounters multiple environmental cues, other than the temperature that substantially influences its gene expression. Consequently, the reprogramming likely enables bacteria to persist in the harsh environment in cecum lymphoid follicles, where the tissue-associated bacteria are surrounded by PMNs. The functional annotation analysis revealed genes indicative of a microaerophilic environment with acidic and oxidative stress factors. We found similarities between the repertoire of up-regulated genes in bacteria grown in vitro under anaerobic conditions and the repertoire of persistent bacteria in the cecums of infected mice. Given the presence of PMNs in the cecum during persistence, it is not surprising that adaption requires protection against acidic and oxidative stress, or that it involves modulation of the bacterial surface for protection. Reprogramming probably occurs after initial colonization that requires T3SS, and later on certain environmental cues force the pathogen to reprogram its transcriptome where the induced gene products aid in maintaining long-term survival in this particular niche. At this late stage of infection the pathogen resists elimination by PMNs with very low expression of its T3SS-associated virulence arsenal, which is puzzling. However, the pathogen may be less recognized by innate immune cells due to its adapted phenotype with altered surface, and maybe also secretion of protective factors. In addition, we cannot exclude that presence of other microbial inhabitant(s) contributes. The observed up-regulation of type VI secretion genes previously reported to participate in bacteria-bacteria communication [39], as well as up-regulation of genes involved in biofilm formation and quorum sensing, may reflect interactions with other bacteria.
Up-regulation of genes involved in DNA replication and repair, RNA degradation, tRNA biosynthesis, and ribosome biogenesis suggests a metabolically active state for persistent bacteria, but it is not a direct indication of whether persistent bacteria are in a dormant or replicative form. We hypothesize that bacteria have a restricted replicative form in order to maintain bacterial load with consistent bacterial shedding into the feces. Motility may also be required for efficient shedding and spread of the bacteria within a restricted host environment, as shown by reduce bacterial shedding into the cecal lumen in a flagellar mutant of avian pathogenic Escherichia coli in a chick persistence model [40].
We show that regulators of anaerobic growth, as well as genes involved in oxidative/acidic stress, are important for the establishment of persistent infection with Y. pseudotuberculosis. Both arcA and fnr mutants demonstrated a markedly reduced ability to establish severe or persistent infection in mice, demonstrating the importance of reprogramming to anaerobic respiration. The function and importance of arcA and fnr have not been studied extensively in Yersinia. Here, we show that these gene products control motility, which has also been shown for arcA in Salmonella enterica sv. Typhimurium and E. coli [41,42]. In an earlier study, arcA was reported to be dispensable for acute Y. pseudotuberculosis infection upon intragastric inoculation in BALB/c mice [43], whereas another more recent study showed that a Y. pseudotuberculosis arcA mutant had attenuated virulence [35]. Whether the former study is contradictory to the latter and our results or just reflects different requirements of arcA depending on infection dose or intestinal delivery of bacteria remains to be elucidated. However, in agreement with our data, arcA has been implicated in virulence for a variety of bacteria [44,45]–[46–48]. Importantly, our data also show that a reduced ability to cause persistent infection is not directly coupled to decreased virulence in general, as increased virulence and an absence or reduced level of persistence was observed for the hdeB and frdA mutants. However, the mechanisms responsible for the virulence phenotypes of the hdeB and uspA mutants, with the former resulting in increased severe disease and the latter increased persistence, is not obvious and requires further investigation.
The regulatory pathways responsible for the switch, where T3SS is down-regulated is a central question in this context. We show that T3SS cannot be induced under low oxygen or acidic conditions. Therefore, regulatory circuits mediating T3SS repression are active under these conditions. Notably, FNR indirectly represses the expression of T3SS effectors under anaerobic conditions in Shigella flexneri [49]. However, we showed here that FNR per se has no effect on T3SS secretion in Y. pseudotuberculosis, as T3SS was repressed at the same level in the fnr mutant and wt strains under anaerobic conditions. The same was observed with the arcA mutant.
We found that many of the genes (20%) that were differentially expressed during persistence overlapped with the Crp/CsrA/RovA regulons, indicating that this regulatory circuit contributes to persistence in the host. The global regulatory systems of Crp and CsrA involves energy metabolism, but also control of certain virulence functions [33]. The interplay between these regulators and RovA is delicate and incorporates a series of complex regulatory loops that can be influenced by other regulators including UvrY and Hfq (all up-regulated during persistence). CsrA can control RovA via RovM, and independently of RovA, positively regulate flagella/motility genes as well as arcA [33,35]. Induction of flagella/motility genes is also influenced by Crp, which can control CsrA and promote induction of rovA. We hypothesize that genes induced by the Crp/CsrA/RovA regulatory cascades, which are mainly down-regulated during the early phase when T3SS is on, participate in reprogramming of Yersinia physiology by promoting expression of genes necessary for persisting in the cecal environment. RovA has been shown to be critical for virulence in Y. enterocolitica and Y. pestis [29,50]. By analogy, we found that RovA was required for virulence upon low dose infection. Interestingly, in contrast to the arcA, fnr, frdA, and wrbA mutants that initially infected cecum, but thereafter were efficiently cleared, the rovA mutant did not establish infection at all upon oral infection of mice. Hence, not only is RovA required for the positive regulation of many genes expressed at 26°C in vitro and during persistence, but is important also for initial infection where the gene expression pattern actually resembles that seen at 37°C in vitro with expression of T3SS. This suggests that RovA contributes to adaption to the host environment also at early stages of infection. Regulators of the Mar/SlyA family have been implicated in the regulation of genes involved in coping with diverse environmental stresses [32]. As such, RovA could contribute to the initial infection by regulating genes important for resistance to low pH in the stomach and to reactive oxygen metabolites produced by innate immune cells in cecum.
The persistence route may reflect the life cycle of this enteropathogen. In such a cycle we hypothesize that, during the initiation of infection, Y. pseudotuberculosis still has flagella and expresses T3SS virulence genes for breaking the epithelial barrier (Fig. 7). Flagella expression is supported by in vitro data, which showed flagellated bacteria up to 2 hours after shifting the temperature to 37°C. T3SS components are expected to be instrumental for resisting the attack from arriving PMNs during the early phase. In later stages, the bacterium adapts to the environment by reducing the expression of T3SS components and increasing the expression of genes important for survival in the cecum lymphoid compartment, from where it can spread to other hosts by fecal shedding, possibly through motility. In this context, Y. pseudotuberculosis has been found in the colon of wild mice with hyperemic cecal membranes [51], suggesting that this compartment is a potential reservoir for this pathogen.
YPIII/pIBX, a kanamycin-resistant bioluminescent Y. pseudotuberculosis strain (S6 Table), was used in this study. The YPIII strain represents a well established model for Y. pseudotuberculosis that has been used for decades to elucidate various aspects of Y. pseudotuberculosis pathogenesis, including identification of Yops [2,4–6,33,36,52]. For in vitro total RNA preparation, the strain was cultured at 26°C or 37°C in brain heart infusion (BHI) broth or LB at acidic (pH 5.2) or normal pH supplemented with 50 μg/ml kanamycin, 5 mM EGTA, and 20 mM MgCl2 for T3SS induction at 37°C. For microarray analysis, wt Y. pseudotuberculosis YPIII was grown at 25°C in LB medium supplemented with 10 g/l glucose and 0.2 M HEPES buffer under aeration or under anaerobic growth conditions (in a nitrogen atmosphere). Additional glucose was added to maximize energy production and the growth rate under anaerobic growth conditions. The absence of oxygen in the culture medium was tested by a gas chromatograph (GC-WLD, Carlo Erba Vega Series 6000) coupled with a detector and integrator (Spectra-Physics, SP4270) using a Poropak QS (100–120 Mesh) column and helium (Westalen 4.6) at 300 kPa. See Supplemental Experimental Procedure for the mutant strains used in this study.
In order to generate an in-frame deletion mutant of the gene of interest, approximately 200 nucleotides from the 5’ and 3’ flanking regions of the gene were amplified by PCR and ligated together into SalI and BglII (New England Biolabs, Inc) linearized pDM4 [53] using the In-Fusion HD Cloning Kit (Clontech Laboratories, Inc) according to the manufacturer’s instructions. The plasmid was transformed into E. coli DH5αλpir and selected on a Cml (25 μg/ml)-containing agar plate. Positive colonies were confirmed by colony PCR. Plasmids purified from positive clones were sequenced to confirm insertion. Confirmed plasmid constructs were transformed into E. coli conjugation strain S17-1λpir for conjugal mating with Y. pseudotuberculosis YPIII-Xen04. Positive allelic exchange was selected as described previously [53]. Finally, in-frame deletion mutants of arcA, fnr, hdeB, uspA, cheW, frdA, and motB were confirmed by sequencing (S7 Table).
Mice were housed in accordance with the Swedish National Board for Laboratory Animals guidelines. All animal procedures were approved by the Animal Ethics Committee of Umeå University (Dnr A108-10). Mice were allowed to acclimate to the new environment for one week before the experiments.
Eight-week-old female FVB/N (Taconic Farms, Inc) mice were deprived of food and water for 16 hours prior to oral infection with ∼107 CFUs of wt or mutant Y. pseudotuberculosis YPIII-Xen04 strains, which were supplied in their drinking water for 6 hours. Bacteria were subcultured on LB agar plates supplemented with kanamycin (50 μg/ml). For infection, the bacteria were grown overnight in LB at 26°C and concentrations estimated by absorbance at OD600nm. Cultures were re-suspended to 107 CFUs/ml in sterilized tap water supplemented with 150 mM NaCl. The infection dose was determined by viable count and drinking volume. Mice were inspected frequently for signs of infection and to ensure that infected mice showing prominent clinical signs were euthanized promptly to prevent suffering. The infections were monitored using IVIS Spectrum (Caliper LifeSciences, Inc.) every third day after infection to 15 dpi, and then every week up to 42 dpi. Prior to imaging, the mice were anesthetized using the XGI-8 gas anesthesia system (Caliper LifeSciences, Inc), which allowed control over the duration of anesthesia. Oxygen mixed with 2.5% IsoFloVet (Orion Pharma, Abbott Laboratories Ltd, Great Britain) was used for the initial anesthesia, and 0.5% isoflurane in oxygen was used during imaging. To analyze bacterial localization within organs, mice were euthanized, the intestine, mesenteric lymph nodes, liver, and spleen removed, and the organs imaged by bioluminescent imaging (BLI). Acquisition and analysis were performed using Living Image software, version 3.1 (Caliper LifeSciences, Inc.).
Cecal tissue was fresh frozen in isopentane pre-chilled with liquid nitrogen and kept at −80°C. For detection of Yersinia in the tissue, 10-μm cryosections were fixed and stained with α-Yersinia serum and for immunohistochemistry sections were stained with rat-α-mouse Gr-1 (clone RB6-8C5, BD Biosciences Pharmingen) as described previously [6]. Cecal sections were also stained with hematoxylin-eosin using standard methods. Analysis was performed using a NIKON Eclipse 90i microscope and images captured with a Hamamatsu Orcha C4742-95 camera or NIKON DSFi1 camera and NIS-Elements AR 3.2 software (Nikon Instruments).
Total RNA was isolated as described previously [54,55] with small modifications. Dissected cecums were emptied by flushing the luminal contents several times with 1× PBS using a sterile syringe. Parts of the cecal tissue associated with Y. pseudotuberculosis-Xen4 (bioluminescent) was cut out using a sterile 3 mm hole punch and immediately transferred to RNAlater (Ambion) for overnight incubation at 4°C after IVIS confirmation of bacterial in the isolated tissue. The RNAlater solution was removed the next day and the tissue samples stored at −80°C. The tissue was homogenized using Dispomix Drive (Medic Tools AG, Switzerland) and all steps performed at 4°C. The samples were transferred to previously cooled Dispomix homogenization tubes (Medic Tools AG, Switzerland) containing 1 ml of Solution D and homogenized twice using homogenization program 9. Tissue lysates were spun down with a quick spin. Each sample was aliquoted (0.5 ml) into separate 2 ml bead beater tubes containing small (0.1 mm) and big (1 mm) glass beads and treated with Mini-Beadbeater (Biospec Products Inc, USA) at a fixed speed for 1 min. Samples were cooled on ice for 1 min and the following added sequentially: 50 μl 2M sodium acetate (pH 4.0), 500 μl water-saturated phenol (Invitrogen, CA, USA), and 100 μl chloroform:isoamyl alcohol (49:1). The samples were inverted vigorously by hand. Suspensions were centrifuged for 20 min at 10,000g after cooling on ice for 15 min. The upper aqueous phase was transferred to RNase-free 1.5 ml tubes and 1 ml isopropanol added to precipitate the RNA. Samples were incubated at −20°C for 2 hours and centrifuged for 20 min at 10,000g. The RNA pellet was dissolved in 0.3 ml Solution D, 0.3 ml isopropanol added, and the resulting aliquots for each sample combined in one tube. The final suspensions were incubated for 30 min at −20°C and centrifuged for 20 min at 10,000g. The RNA pellets were suspended in 50 μl RNase-free water. DNA contamination was removed using the Qiagen DNase Kit according to the manufacturer’s instructions. The same procedure was applied to bacterial cultures grown in vitro except for the homogenization step with the Dispomix Drive homogenizer. The quality and concentration of total RNA isolated from the cecum and in vitro cultured bacteria was assessed by microcapillary electrophoresis using an Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA). All preparations used in this study had an RIN value >7.0.
The MICROBEnrich Kit (Ambion) was used to enrich bacterial mRNAs in total RNA samples from cecums by removing 18S and 26S rRNAs and polyadenylated mRNAs according to the manufacturer’s instructions. To deplete the bacterial rRNA and tRNA in total RNAs from in vivo and in vitro samples, we used a MICROBExpress Kit (Ambion) according to the manufacturer’s instructions.
RNA libraries for sequencing were prepared using TruSeq RNA kits (Illumina, CA, USA) according to the manufacturer’s instructions, but with the following changes. The RNA samples were EtOH precipitated and subsequent protocols (starting from cDNA synthesis in the Illumina provided protocol) were automated using an MBS 1200 pipetting station (Nordiag AB, Sweden). All purification steps and gel-cuts were replaced by the magnetic bead clean-up methods described previously [56].
Quality and base trimming (5 nt from 5’ end and 5 nt from 3’ end of each read) were performed on 100-nt-long paired-end Illumina 2000 Hiseq reads from in vivo and in vitro sample libraries. Trimmed reads for each in vivo library were mapped to the NCBI 16SMicrobial database to determine the bacterial species in the cecal tissue biopsies with 100% identity. Identified bacterial species with available reference genomes (42 annotated bacterial genomes), Y. pseudotuberculosis YPIII, and the pYV plasmid from NCBI were used as reference genomes for mapping. A variety of mapping tests were performed by loosening or strengthening the alignment settings to optimize the filtration of non-Y. pseudotuberculosis YPIII-specific reads with SNP calling using Probabilistic Variant Detection in CLC Genomics Workbench after each mapping attempt. The rRNA and tRNA annotations were removed from the reference genomes prior to RNA-seq in order to avoid bias from the rRNA depletion procedures. The mRNA expression level and RPKMO value was calculated for each gene in in vivo and in vitro samples using annotated NC_010465 and NC_006153 as reference genomes in CLC Genomic Workbench for RNA-seq. Analysis of differentially expressed genes were performed on normalized RPKMO values by CLC Genomic Workbench for RNA-seq. The zero read values for the ORFs in in vivo samples were normalized by adding 0.01 pseudocounts in order to avoid dividing by zero [57–59]. As many of the reads detected for individual ORFs were only present in samples from either early or persistent infection, proper p-values could not be calculated for the in vivo data set. For in vitro data, the ORFs with less than 10 reads in all four replicates and p >0.05 were removed from the analysis to filter out false-positive values. The raw RNA-seq data have been deposited in NCBI’s Gene Expression Omnibus and are accessible through GEO Series accession number GSE55292.
Total pure bacterial RNA isolated from bacterial cultures grown in BHI/LB medium at pH 7.2 at 26°C/37°C (in some experiments at acidic pH or under anaerobic conditions at 37°C) and heterogenous RNA isolated from infected cecums were used as templates for cDNA synthesis with the RevertAid H Minus First Strand cDNA Synthesis Kit (Fermentas). The qPCR reactions were performed in triplicate for each condition using the Quantimix Easy Syg Kit (Biotools) or KAPA SYBR FAST qPCR Kit (Kapabiosystems) and Bio-Rad i5 Light Cycler. After optimization experiments (S7 Fig.) gyrB was selected as internal control to calculate the relative expression of tested genes.
The design and analysis of the microarray (Agilent, 8×15K format) for the transcriptome analysis of Y. pseudotuberculosis YPIII were described previously [33]. YPIII was grown aerobically and anaerobically in four independent cultures at 25°C to the exponential and stationary phase. Bacteria were pelleted, mixed with 0.2 volumes of stop solution (5% water-saturated phenol), and snap frozen in liquid nitrogen. After thawing on ice, total RNA was prepared using the SV Total RNA Purification Kit (Promega) and remaining genomic DNA removed by rDNAse (Macherey-Nagel) digestion as described by the manufacturer. RNA concentration and quality were determined by measuring A260 and A280 with an Agilent 2100 Bioanalyzer using the Nano 6000 kit, and the absence of DNA was excluded by PCR of intergenic regions. Total RNA from the independent cultures was labeled using the ULS™ Fluorescent Labeling Kit for Agilent Arrays (Kreatech) as follows: 1 μg total RNA was used for RNA-labeling with Cy5 (for wt RNA) and Cy3 (for mutant RNA). Non-incorporated Cy5/Cy3 was removed using KREApure purification columns from the ULS™ Fluorescent Labeling Kit as suggested by the manufacturer, and the degree of labeling was analyzed with a Nanodrop (Peqlab). Subsequently, 300 ng Cy5-labelled RNA and 300 ng Cy3-labelled RNA were mixed, fragmented, and hybridized for 17 h at 65°C to custom-made Agilent microarray slides using the Agilent Gene Expression Hybridization Kit as described by the manufacturer. Four replicates were utilized in each experiment. After washing and drying the microarray slide, the data were scanned using an Axon GenePix Personal 4100A scanner and array images captured using the software package GenePix Pro 6.015. The microarray data was processed using the software package R (www.r-project.org) in combination with the “Bioconductor” software framework [60] as described previously [33]. The overall fold-changes of a gene represented by at least three probes are given as median values for all probes. The set of resulting differentially expressed genes (fold-change ≥ 2) was analyzed by the topGO package for Gene Ontology (GO) term enrichment [61]. MIAME compliant array data were deposited in the Gene Expression Omnibus (GEO) database and are available via the following accession numbers: GSE56475 and GSE56475.
Bacteria from overnight cultures were inoculated into LB and grown to exponential phase. A 5 μl aliquot of each culture was spotted on LB (pH 7.4 or pH 5) with 0.25% agar. Plates were incubated at 26°C or 37°C under aerobic or anaerobic conditions for 48 hours. The bioluminescent signal from bacteria on the plates was monitored using a ChemiDoc XRS System (Bio-Rad).
Overnight bacterial cultures were diluted 25-times in LB media and allowed to grow for 2 hours at 26°C. The medium was changed to T3SS-inducing conditions (5 mM EGTA, 20 mM MgCl2, pH 7.4 or 5) and incubated for 4 hours at 37°C under aerobic and anaerobic conditions. The culture supernatant was collected by centrifugation at 4000 rpm for 10 min and filtered using 0.45-μm filters. The secreted proteins were concentrated by TCA precipitation. Samples were loaded onto 12% SDS-PAGE according to the cultures’ OD600 values after 4 hours incubation.
Bacterial cultures grown overnight were diluted 25-times in LB media and allowed to grow for 2 hours at 26°C to OD600 = 0.2. The growth conditions were then changed to T3SS-inducing conditions at 37°C. One milliliter was taken from the bacterial cultures every hour for analysis with atomic force microscopy. Each sample was centrifuged for 4 min at 1500 rpm, washed once with 2 mM MgCl2, and re-suspended in 50–200 μl of the same solution. Ten microliters of each sample was placed on freshly cleaved ruby red mica (Goodfellow Cambridge Ltd, Cambridge), incubated 5 min at room temperature, and blotted dry before being placed into a desiccator for a minimum of 2 hours. Images were collected by a Nanoscope V AFM (Bruker software) using ScanAsyst in air with ScanAsyst cantilevers at a scan rate of approximately 0.9–1 Hz. The final images were flattened and/or plane-fitted in both axes using Bruker software and presented in amplitude (error) mode.
Windows Microsoft Excel 2011 and CLC Genomic Workbench were utilized for statistical tests and linear regression analysis of RNA-seq and infection data. Multiple RNA-seq were compared by the paired t-test on Gaussian data. Bonferroni correction was employed for multiple comparison analysis. Similarities between replicates were determined by Spearman and Pearson’s R value. One-tailed test were conducted to calculate p-values for difference in mutant infections clearance with 2×2 contingency table by Fisher’s exact test.
Kyoto Encyclopedia of Genes an Genomes (KEGG) accession numbers for the genes mentioned in this study are as follow; arcA (ypy:YPK_3606), fnr (ypy:YPK_1944), rovA (ypy:YPK_2381), frdA (ypy:YPK_3813), hdeB (ypy:YPK_1140), uspA (ypy:YPK_0120), napA (ypy:YPK_1387), wrbA (ypy:YPK_2363), motB (ypy:YPK_0802), cheW (ypy:YPK_1750), csrA (ypy:YPK_3372), crp (ypy:YPK_0248), dnaG (ypy:YPK_0635), gyrB (ypy:YPK_0004), mdh (ypy:YPK_3761), rpoC (ypy:YPK_0341), fliC (ypy:YPK_2381), lpp (ypy:YPK_1854), ompF (ypy:YPK_2649), ompA (ypy:YPK_2630), ftn (ypy:YPK_2438), aspA (ypy:YPK_3825), yhbH (ypy:YPK_3353), pal (ypy:YPK_2955), flhC (ypy:YPK_1746), flhD (ypy:YPK_1745), fliA (ypy:YPK_2380), fliE (ypy:YPK_2390), fliK (ypy:YPK_2396), flgL (ypy:YPK_2415), flgH (ypy:YPK_2419), flgG (ypy:YPK_2420), flgB (ypy:YPK_2425), flgA (ypy:YPK_2426), invA (ypy:YPK_2429), uvrY (ypy:YPK_2326), hfq (ypy:YPK_3799), yopB (pYV0055), yopD (pYV0054), yopH (pYV0094), yopE (pYV0025), yopK (pYV0040), yopM (pYV0047), lcrF (pYV0076).
|
10.1371/journal.pntd.0005568 | Detecting the impact of temperature on transmission of Zika, dengue, and chikungunya using mechanistic models | Recent epidemics of Zika, dengue, and chikungunya have heightened the need to understand the seasonal and geographic range of transmission by Aedes aegypti and Ae. albopictus mosquitoes. We use mechanistic transmission models to derive predictions for how the probability and magnitude of transmission for Zika, chikungunya, and dengue change with mean temperature, and we show that these predictions are well matched by human case data. Across all three viruses, models and human case data both show that transmission occurs between 18–34°C with maximal transmission occurring in a range from 26–29°C. Controlling for population size and two socioeconomic factors, temperature-dependent transmission based on our mechanistic model is an important predictor of human transmission occurrence and incidence. Risk maps indicate that tropical and subtropical regions are suitable for extended seasonal or year-round transmission, but transmission in temperate areas is limited to at most three months per year even if vectors are present. Such brief transmission windows limit the likelihood of major epidemics following disease introduction in temperate zones.
| Understanding the drivers of recent Zika, dengue, and chikungunya epidemics is a major public health priority. Temperature may play an important role because it affects virus transmission by mosquitoes, through its effects on mosquito development, survival, reproduction, and biting rates as well as the rate at which mosquitoes acquire and transmit viruses. Here, we measure the impact of temperature on transmission by two of the most common mosquito vector species for these viruses, Aedes aegypti and Ae. albopictus. We integrate data from several laboratory experiments into a mathematical model of temperature-dependent transmission, and find that transmission peaks at 26–29°C and can occur between 18–34°C. Statistically comparing model predictions with recent observed human cases of dengue, chikungunya, and Zika across the Americas suggests an important role for temperature, and supports model predictions. Using the model, we predict that most of the tropics and subtropics are suitable for transmission in many or all months of the year, but that temperate areas like most of the United States are only suitable for transmission for a few months during the summer (even if the mosquito vector is present).
| Epidemics of dengue, chikungunya, and Zika are sweeping through the Americas, and are part of a global public health crisis that places an estimated 3.9 billion people in 120 countries at risk [1]. Dengue virus (DENV) distribution and intensity in the Americas has increased over the last three decades, infecting an estimated 390 million people (96 million clinical) per year [2]. Chikungunya virus (CHIKV) emerged in the Americas in 2013, causing 1.8 million suspected cases from 44 countries and territories (www.paho.org). In the last two years, Zika virus (ZIKV) has spread throughout the Americas, causing 764,414 suspected and confirmed cases, with many more unreported (http://ais.paho.org/phip/viz/ed_zika_cases.asp, as of April 13, 2017). The growing burden of these diseases (including links between Zika infection and both microcephaly and Guillain-Barré syndrome [3]) and potential for spread into new areas creates an urgent need for predictive models that can inform risk assessment and guide interventions such as mosquito control, community outreach, and education.
Predicting transmission of DENV, CHIKV, and ZIKV requires understanding the ecology of the vector species. For these viruses the main vector is Aedes aegypti, a mosquito that prefers and is closely affiliated with humans, while Ae. albopictus, a peri-urban mosquito, is an important secondary vector [4,5]. We expect one of the main drivers of the vector ecology to be the climate, particularly temperature. For that reason, mathematical and geostatistical models that incorporate climate information have been valuable for predicting and responding to Aedes spp. spread and DENV, CHIKV, and ZIKV outbreaks [5–10].
The effects of temperature on ectotherms are largely predictable from fundamental metabolic and ecological processes. Survival, feeding, development, and reproductive rates predictably respond to temperature across a variety of ectotherms, including mosquitoes [11,12]. Because these traits help to determine transmission rates, the effects of temperature on transmission should also be broadly predictable from mechanistic models that incorporate temperature-dependent traits. Here, we introduce a model based on this framework that overcomes several major gaps that currently limit our understanding of climate suitability for transmission. Specifically, we develop models of temperature-dependent transmission for Ae. aegypti and Ae. albopictus that are (a) mechanistic, facilitating extrapolation beyond the current disease distribution, (b) parameterized with biologically accurate unimodal thermal responses for all mosquito and virus traits that drive transmission, and (c) validated against human dengue, chikungunya, and Zika case data across the Americas.
We synthesize available data to characterize the temperature-dependent traits of the mosquitoes and viruses that determine transmission intensity. With these thermal responses, we develop mechanistic temperature-dependent virus transmission models for Ae. aegypti and Ae. albopictus. We then ask whether the predicted effect of temperature on transmission is consistent with patterns of actual human cases over space and time. To do this, we validate the models with DENV, CHIKV, and ZIKV human incidence data at the country scale in the Americas from 2014–2016. To isolate temperature dependence, we also statistically control for population size and two socioeconomic factors that may influence transmission. If temperature fundamentally limits transmission potential, transmission should only occur at actual environmental temperatures that are predicted to be suitable, and conversely, areas with low predicted suitability should have low or zero transmission (i.e., false negative rates should be low). By contrast, low transmission may occur even when temperature suitability is high because other factors like vector control can limit transmission (i.e., the false positive rate should be higher than the false negative rate). Finally, if the simple mechanistic model accurately predicts climate suitability for transmission, then we can use it to map climate-based transmission risk of DENV, CHIKV, ZIKV, and other emerging pathogens transmitted by Ae. aegypti and Ae. albopictus seasonally and geographically.
Data gathered from the literature [9,13–30] revealed that all mosquito traits relevant to transmission—biting rate, egg-to-adult survival and development rate, adult lifespan, and fecundity—respond strongly to temperature and peak between 23°C and 34°C for the two mosquito species (Ae. aegypti in Fig 1 and Ae. albopictus in Fig A in S1 Text). DENV extrinsic incubation and vector competence peak at 35°C [31–37] and 31–32°C [31,32,34,38], respectively, in both mosquitoes—temperatures at which mosquito survival is low, limiting transmission potential (Fig 1, Fig A in S1 Text). Appropriate thermal response data were not available for CHIKV and ZIKV extrinsic incubation and vector competence.
We estimated the posterior distribution of R0(T) and used it to calculate key temperature values that indicate suitability for transmission: the mean and 95% credible intervals (95% CI) on the critical thermal minimum, maximum, and optimum temperature for transmission by the two mosquito species. At constant temperatures, Ae. aegypti transmission peaked at 29.1°C (95% CI: 28.4–29.8°C), and declined to zero below 17.8°C (95% CI: 14.6–21.2°C) and above 34.6°C (95% CI: 34.1–35.6°C) (Fig 2). Ae. albopictus transmission peaked at 26.4°C (95% CI: 25.2–27.4°C) and declined to zero below 16.2°C (95% CI: 13.2–19.9°C) and above 31.6°C (95% CI: 29.4–33.7°C) (Fig 2). Overall, the thermal response curve for Ae. albopictus is shifted towards lower temperatures than Ae. aegypti, so Ae. albopictus transmission is better suited to cooler environments. For a more realistic scenario in which daily temperature ranged over 8°C, the transmission peak, minimum, and maximum were slightly lower for both Ae. aegypti (28.5°C, 13.5°C, 34.2°C, respectively) and Ae. albopictus (26.1°C, 11.9°C, and 28.3°C, respectively). The lower thermal maximum under fluctuating temperatures occurs because we incorporated empirically supported irreversible lethal effects of temperatures that exceed thermal maxima for survival (see Materials and Methods).
The posterior distribution of R0(T) allows us to evaluate uncertainty in key temperature values that define the transmission range, including critical thermal minimum, maximum, and optimum. Uncertainty was higher for the critical thermal minimum for transmission than for the maximum or optimum, and the two mosquito species overlapped most for this outcome (Fig 2, bottom panels). This result occurred because several trait thermal responses increase gradually from low to mid temperatures but decline more steeply at high temperatures (Fig 1), so uncertainty is greatest at low temperatures. Ae. aegypti has a substantially higher optimum and maximum temperature than Ae. albopictus (Fig 2) due to its greater rates of adult survival at high temperatures (see Supplementary Materials for sensitivity analyses).
We used generalized linear models (GLM) to ask whether the predicted relationship between temperature and transmission, R0(T), was consistent with observed human cases of DENV, CHIKV, and ZIKV. Specifically, we assessed whether R0(T) was an important predictor of the probability of autochthonous transmission occurring and of the incidence given that transmission occurred. We also controlled for human population size, virus species, and two socioeconomic factors. (Note that we focused on testing the R0(T) model, rather than on constructing the best possible statistical model of human case data.) To do this, we used the version of the Ae. aegypti R0(T) model that includes 8°C daily temperature range, along with country-scale weekly case reports of DENV, CHIKV, and ZIKV in the Americas and the Caribbean between 2014–2016. We first addressed the fact that countries with larger populations have greater opportunities for (large) epidemics by creating two predictors that incorporate scaled R0(T) and population size. In the models of the probability of autochthonous transmission occurring we used the product of the posterior probability that R0(T) > 0 (which we notate as GR0) and the log of population size (p) to give log(p)*GR0. (Here, and throughout, log denotes the natural logarithm.) In the models of incidence, given that transmission does occur, we used the log of the product of the posterior mean of R0(T) and population size, log(p*R0(T)). To control for several socioeconomic factors that might obscure the impact of temperature, we also included log of gross domestic product (GDP) and log of percent of GDP in tourism (using logs because the predictors were highly skewed, to stabilize variance). These are potential indicators of investment in and/or success of vector control and infrastructure improvements that prevent transmission. By comparing models that included the R0(T) metric alone, socioeconomic factors alone, or both, we tested whether R0(T) was an important predictor of observed transmission occurrence and incidence (see Table D in S2 Text). Note that R0(T) is out of sample for all validation analyses because it is derived and calculated strictly from laboratory data on mosquitoes, and we perform a validation analyses for R0(T) using independent case incidence reports. For this validation step we assessed model adequacy for the transmission data in two ways. First we used the full dataset for case incidence reports to select the best model (Table D in S2 Text) and to determine whether or not our predicted value of relative R0(T) based on laboratory data was included in the model (“within sample” analysis). Second we used a bootstrapping approach where models were fit on subsets of the case incidence data that were randomly sampled and then predictive accuracy of the competing models (Table D in S2 Text) was assessed on left-out data (“out of sample” analysis).
For the probability of autochthonous transmission occurring, the model that included both the R0(T) predictor and socioeconomic predictors had overwhelming support based on Bayesian Information Criterion (BIC; model PA5 relative probability = 1, Table D in S2 Text). Based on deviance explained, the models that included R0(T), with or without the socioeconomic predictors out-performed the model that did not include R0(T) (Table D in S2 Text; Fig 3A, Fig B in S1 Text). In analyses of out-of-sample accuracy, models that included the R0(T) metric (with or without the socioeconomic factors) were surprisingly accurate. They predicted the probability of transmission with 86–91% out-of-sample accuracy for DENV (Table D in S2 Text). For CHIKV and ZIKV, models that included the R0(T) metric or population alone had 66–69% out-of-sample accuracy (Table D in S2 Text). There were no significant differences in out-of-sample accuracy between the top four models, but for both DENV and CHIKV/ZIKV the best model was significantly better than the worst model [see supplementary code in 39 for full results]. The lower out-of-sample accuracy for CHIKV and ZIKV likely reflects the much lower frequency of positive values and the lower total sample size of this dataset. All results were similar for a set of models that separated GR0 from population size, so for simplicity we show the model predictors that combines GR0 and population size here (see Table D in S2 Text and [39] for results of other models). Further, from a biological perspective, the combined model better describes what we know about disease systems: if either the probability of R0(T) being greater than zero is small or population size is very small, transmission is unlikely to occur. Together, these analyses suggest that R0(T) is an important predictor of transmission occurrence, but that CHIKV and ZIKV need further data to better explain the probability of transmission occurrence (Fig 3A, Fig B in S1 Text).
R0(T) was also an important predictor of incidence, given that autochthonous transmission did occur. Within-sample, incidence was best predicted by the model that included both R0(T) and the socioeconomic predictors (model IM5 in Table D in S2 Text) based on BIC (relative probability = 1). The models that included R0(T) out-performed those that did not based on deviance explained (Table D in S2 Text). In out-of-sample validation, the models that included R0(T) explained the magnitude of incidence based on mean absolute percentage error (85–86% accuracy versus 83% accuracy for models that did not include R0(T); Table D in S2 Text), but this difference was not statistically significant. For illustration, we show the simpler model that only contains the R0(T) predictor in the main text (Fig 3B; model IM1 in Table D in S2 Text). Notably, the models that contained R0(T) predicted incidence well for all three viruses, despite the lower incidence of CHIKV and ZIKV.
Although predicted R0(T) correlated with the observed occurrence and magnitude of human incidence for all three viruses, these observed incidence metrics were higher for DENV than for CHIKV and ZIKV. While the reason for this difference is unclear, the most likely explanation is that DENV is much more established in the Americas, so it is more likely to be detected, diagnosed, and reported. Because ZIKV and CHIKV are newly emerging, they may not have fully saturated the region at this early stage.
The ability of the model to explain the probability and magnitude of transmission is notable given the coarse scale of the human incidence versus mean temperature data (i.e., country-scale means), the lack of CHIKV- and ZIKV-specific trait thermal response data to inform the model, the nonlinear relationship between transmission and incidence, and all the well-documented factors other than temperature that influence transmission. Together, these analyses show simple mechanistic models parameterized with laboratory data on mosquitoes and dengue virus are consistent with observed temperature suitability for transmission. Moreover, the similar responses of human incidence of ZIKV, CHIKV, and DENV to temperature suggest that the thermal ecology of their shared mosquito vectors is a key determinant of outbreak location, timing, and intensity.
The validated model can be used to predict where transmission is not excluded (posterior probability that R0(T) > 0, a conservative estimate of transmission risk). Considering the number of months per year at which mean temperatures do not prevent transmission, large areas of tropical and subtropical regions, including Puerto Rico and parts of Florida and Texas, are currently suitable year-round or seasonally (Fig 4). These regions are fundamentally at risk for DENV, CHIKV, ZIKV, and other Aedes arbovirus transmission during a substantial part of the year (Fig 4). Indeed, DENV, CHIKV, and/or ZIKV local transmission has occurred in Texas, Florida, Hawaii, and Puerto Rico (www.cdc.gov). On the other hand, many temperate regions experience temperatures suitable for transmission three months or less per year (Fig 4). Temperature thus limits the potential for the viruses to generate extensive epidemics in temperate areas even where the vectors are present. Moreover, many temperate regions with seasonally suitable temperatures currently lack Ae. aegypti and Ae. albopictus mosquitoes, making vector transmission impossible (Fig 4, black line). The posterior distribution of R0(T) also allows us to map months of risk with different degrees of uncertainty (e.g., 97.5%, 50%, and 2.5% posterior probability that that R0 > 0), ranging from the most to least conservative (Fig D in S1 Text).
Temperature is an important driver of—and limitation on—vector transmission, so accurately describing the temperature range and optimum for transmission of DENV, CHIKV, and ZIKV is critical for predicting their geographic and seasonal patterns of spread [12,41]. We directly estimated the temperature–transmission relationship using mechanistic transmission models for each mosquito species (Fig 2). These models are built using empirical estimates of the (unimodal) effects of temperature on mosquito and pathogen traits that drive transmission, including survival, development, reproduction, and biting rates (Fig 1, Fig A in S1 Text). Because these trait thermal responses are unimodal across the majority of ectotherm taxa and traits, and because the traits combine nonlinearly to drive transmission, the emergent relationship between temperature and transmission is difficult to infer directly from field data or from individual trait responses. Here, we present a model of temperature-dependent DENV, CHIKV, and ZIKV transmission that advances on previous models because it is mechanistic, fitted from experimental trait data (Fig 1, Fig A in S1 Text), and validated against independent human case data at a broad geographic scale (Fig 3).
Mechanistic understanding is valuable for extrapolating beyond the current spatial and temporal range of transmission (Fig 4), as compared to environmental niche models, for example [5,42,43]. Of the six previous mechanistic temperature-dependent models of DENV, CHIKV, or ZIKV transmission by Ae. aegypti and Ae. albopictus that we were able to reproduce, three had similar thermal optima [7,44,45] while the other three had dramatically higher optima (3–6°C) [9,46] (Fig E in S1 Text). Two of the models were very similar to ours [44,45]; of the remaining four models, two predicted much greater suitability for transmission at low temperatures [46] and all four predicted greater suitability at high temperatures [7,9,46] (Fig E in S1 Text). Only one of these previous models was (like ours) statistically validated against independent data not used to estimate model parameters, and its predictions were very similar to those of our model [44]. Other mechanistic and environmental niche models could not be directly compared with ours [5,10,41–43], either because fully reproducible equations, parameters, and/or code were not provided or because their predicted marginal effects of temperature were not displayed. Visually, our maps are similar to maps based on a previous model of Ae. aegypti and Ae. albopictus persistence suitability indices [41]. Recent environmental niche models of Zika distribution have shown similar but more constrained predicted distributions of environmental suitability, in part because these models include not just temperature suitability but also further environmental, socioeconomic, and demographic constraints [5,42,43,47].
Even though the thermal response data are imperfect—for example, CHIKV and ZIKV thermal response data are missing—and the human case data are reported at a coarse spatial scale, the validation analyses suggest that R0(T) is an important predictor of both the probability of transmission occurring and the magnitude of incidence for DENV, CHIKV, and ZIKV. This has several key implications. First, temperature-dependent transmission is pervasive enough to be detected at a coarse spatial scale. Second, dynamics of the mosquito predict transmission for a suite of Ae. aegypti-transmitted viruses, without additional virus-specific information. Third, climate and socio-economic factors combine to shape variation in incidence across countries. Finally, these simple predictors explain a substantial proportion of the variance in both the probability and intensity of transmission.
Predicting arbovirus transmission at a higher spatial resolution and precision will require more detailed information on factors like the exposure and susceptibility of human populations, environmental variation (e.g., oviposition habitat availability, seasonal and daily temperature variation), and socioeconomic factors. However, as a first step our mechanistic model provides valuable insight because it makes broad predictions about suitable environmental conditions for transmission, it is mechanistic and grounded in experimental trait data, it is validated against independent human case data, and its predictions are applicable across three different viruses. Using these thermal response models as a scaffold, additional drivers could be incorporated to obtain more precise and specific predictions about transmission dynamics, which could in turn be used for public health and vector control applications. For this purpose, all code and data used in the models are available on Figshare [39].
The socio-ecological conditions that enabled CHIKV, ZIKV, and DENV to become the three most important emerging vector-borne diseases in the Americas make the emergence of additional Aedes-transmitted viruses likely (potentially including Mayaro, Rift Valley fever, yellow fever, Uganda S, or Ross River viruses). Efforts to extrapolate and to map temperature suitability (Fig 4) will be critical for improving management of both ongoing and future emerging epidemics. Mechanistic models like the one presented here are useful for extrapolating the potential geographic range of transmission beyond the current envelope of environmental conditions in which transmission occurs (e.g., under climate change and for newly invading pathogens). Accurately estimating temperature-driven transmission risk in both highly suitable and marginal regions is critical for predicting and responding to future outbreaks of these and other Aedes-transmitted viruses.
We constructed temperature-dependent models of transmission using a previously developed R0 framework. We modeled transmission rate as the basic reproduction rate, R0—the number of secondary infections that would originate from a single infected individual introduced to a fully susceptible population. In previous work on malaria, we adapted a commonly used expression for R0 for vector transmission to include the temperature-sensitive traits that drive mosquito population density [12]:
R0(T)=(a(T)2b(T)c(T)e−μ(T)/PDR(T)EFD(T)pEA(T)MDR(T)Nrμ(T)3)1/2
(1)
Here, (T) indicates that the trait is a function of temperature, T; a is the per-mosquito biting rate, b is the proportion of infectious bites that infect susceptible humans, c is the proportion of bites on infected humans that infect previously uninfected mosquitoes (i.e., b*c = vector competence), μ is the adult mosquito mortality rate (lifespan, lf = 1/μ), PDR is the parasite development rate (i.e., the inverse of the extrinsic incubation period, the time required between a mosquito biting an infected host and becoming infectious), EFD is the number of eggs produced per female mosquito per day, pEA is the mosquito egg-to-adult survival probability, MDR is the mosquito immature development rate (i.e., the inverse of the egg-to-adult development time), N is the density of humans, and r is the human recovery rate. For each temperature-sensitive trait in each mosquito species, we fit either symmetric (Quadratic, -c(T–T0)(T–Tm)) or asymmetric (Brière, cT(T–T0)(Tm−T)1/2) unimodal thermal response models to the available empirical data [48]. In both functions, T0 and Tm are respectively the minimum and maximum temperature for transmission, and c is a positive rate constant.
We consider a normalized version of the R0 equation such that it is rescaled to range from zero to one with the value of one occurring at the unimodal peak. Although absolute values of R0 that are used to determine when transmission is stable depend on additional factors not captured in our model, the minimum and maximum temperatures for which R0 > 0 map exactly onto our normalized equations, allowing us to accurately calculate whether or not transmission should be possible at all. Empirical estimates of absolute values of R0 are difficult to obtain in any case, but it is much easier to determine whether transmission is occurring and for how long. While different model formulations for predicting R0 versus temperature can produce results with different magnitudes and potentially different overall shapes [49], the temperatures for which R0 is above or below zero (or one) are mostly model independent. For instance, two competing models differ only by whether or not the formula in Eq (1) is squared, but the square of a number (e.g., an absolute R0 value) greater than one is always greater than one, and the square of a number less than one is always less than one. Therefore, the threshold temperatures at which absolute R0 > 0 or absolute R0 > 1 will be exactly the same for either choice of formula (Fig F in S1 Text). Similarly, because different expressions for R0, including the square of Eq (1), map monotonically onto our function, they will produce identical estimates for the temperatures at which transmission declines to zero and peaks (Fig F in S1 Text). Consequently, our use of relative R0 adequately describes the nonlinear relationship between mosquito and virus traits and transmission.
We fit the trait thermal responses in Eq (1) based on an exhaustive search of published laboratory studies that fulfilled the criterion of measuring a trait at three or more constant temperatures, ideally capturing both the rise and the fall of each unimodal curve (Tables S1-S2). Constant-temperature laboratory conditions are required to isolate the direct effect of temperature from confounding factors in the field and to provide a baseline for estimating the effects of temperature variation through rate summation [50]. We attempted to obtain raw data from each study, but if they were not available we collected data by hand from tables or digitized data from figures using WebPlotDigitizer [51]. We obtained raw data from Delatte [19] and Alto [21] for the Ae. albopictus egg-to-adult survival probability (pEA), mosquito development rate (MDR), gonotrophic cycle duration (GCD, which we assumed was equal to the inverse of the biting rate) and total fecundity (TFD) (Table D in S2 Text). Data did not meet the inclusion criterion for CHIKV or ZIKV vector competence (b, c) or extrinsic incubation period (EIP) in either Ae. albopictus or Ae. aegypti. Instead, we used DENV EIP and vector competence data, combined with sensitivity analyses.
Following Johnson et al. [52], we fit a thermal response for each trait using Bayesian models. We first fit Bayesian models for each trait thermal response using uninformative priors (T0 ~ Uniform (0, 24), Tm ~ Uniform (25, 45), c ~ Gamma (1, 10) for Brière and c ~ Gamma (1, 1) for Quadratic fits) chosen to restrict each parameter to its biologically realistic range (i.e., T0 < Tm and we assumed that temperatures below 0°C and above 45°C were lethal). Any negative values for all thermal response functions were truncated at zero, and thermal responses for probabilities (pEA, b, and c) were also truncated at one. We modeled the observed data as arising from a normal distribution with the mean predicted by the thermal response function calculated at the observed temperature, and the precision τ, (τ = 1/σ), distributed as τ ~ Gamma (0.0001, 00001). We fit the models using Markov Chain Monte Carlo (MCMC) sampling in JAGS, using the R [53] package rjags [54]. For each thermal response, we ran five MCMC chains with a 5000-step burn-in and saved the subsequent 5000 steps. We thinned the posterior samples by saving every fifth sample and used the samples to calculate R0 from 15–40°C, producing a posterior distribution of R0 versus temperature. We summarized the relationship between temperature and each trait or overall R0 by calculating the mean and 95% highest posterior density interval (HPD interval; a type of credible interval that includes the smallest continuous range containing 95% of the probability, as implemented in the coda package [55]) for each curve across temperatures.
We fit a second set of models for each mosquito species that used informative priors to reduce uncertainty in R0 versus temperature and in the trait thermal responses. In these models, we used Gamma-distributed priors for each parameter T0, Tm, c, and τ fit from an additional ‘prior’ dataset of Aedes spp. trait data that did not meet the inclusion criteria for the primary dataset (Table C in S2 Text). We found that these initial informative priors could have an overly strong influence on the posteriors, in some cases drawing the posterior distributions well away from the primary dataset, which was better controlled and met the inclusion criteria. We accounted for our lower confidence in this data set by increasing the variance in the informative priors, by multiplying all hyperparameters (i.e., the parameters of the Gamma distributions of priors for T0, Tm, and c) by a constant k to produce a distribution with the same mean but 1/k times larger variance. We chose the value of k based on our relative confidence in the prior versus main data. Thus we chose k = 0.5 for b, c, and PDR and k = 0.01 for lf. This is the main model presented in the text (Fig 2). It is comparable to some but not all previous mechanistic models for Ae. aegypti and Ae. albopictus transmission (Fig E in S1 Text). Results of our main model, fit with informative priors, did not vary substantially from the model fit with uninformative priors (Figs G-H in S1 Text).
Because organisms do not typically experience constant temperature environments in nature, we incorporated the effects of temperature variation on transmission by calculating a daily average R0 assuming a daily temperature range of 8°C, across the range of mean temperatures. This range is consistent with daily temperature variation in tropical and subtropical environments but lower than in most temperate environments. At each mean temperature, we used a Parton-Logan model to generate hourly temperatures and calculate each temperature-sensitive trait on an hourly basis [56]. We assumed an irreversible high-temperature threshold above which mosquitoes die and transmission is impossible [57,58]. We set this threshold based on hourly temperatures exceeding the critical thermal maximum (Tm in Tables A-B in S1 Text) for egg-to-adult survival or adult longevity by any amount for five hours or by 3°C for one hour. We averaged each trait over 24 hours to obtain a daily average trait value, which we used to calculate relative R0 across a range of mean temperatures. We used this model in the validation against human cases (Fig 3) and the risk map (Fig 4).
To validate the model, we used data on human cases of DENV, CHIKV, and ZIKV at the country scale and mean temperature during the transmission window. Using statistical models (as described below), we estimated the effects of predicted R0(T) on the probability of local transmission and the magnitude of incidence, controlling for population size and several socioeconomic factors. We downloaded and manually entered Pan American Health Organization (PAHO) weekly case reports for DENV and CHIKV for all countries in the Americas (North, Central, and South America and the Caribbean Islands) from week 1 of 2014 to week 8 of 2015 for CHIKV and from week 52 of 2013 to week 47 of 2015 for DENV (www.paho.org). ZIKV weekly case reports for reporting districts (e.g., provinces) within Colombia, Mexico, El Salvador, and the US Virgin Islands were available from the CDC Epidemic Prediction Initiative (https://github.com/cdcepi/) from November 28, 2015 to April 2, 2016. We aggregated the ZIKV data into country-level weekly case reports to match the spatial resolution of the DENV, CHIKV, and covariate data.
We matched the DENV, CHIKV, and ZIKV incidence data with temperature using daily temperature data from METAR stations in each country, averaged at the country level by epidemic week. A previous study found a six-week lagged relationship between temperature and oviposition for Aedes aegypti in Ecuador [40]. Assuming that the subsequent transmission, disease development, medical care-seeking, and case reporting in humans takes an additional four weeks, we assumed a priori a ten-week lag between temperature and incidence (i.e., mean temperature for the week that is ten weeks prior to each case report). METAR stations are internationally standardized weather reporting stations that report hourly temperature and precipitation measures. Outlier weather stations were excluded if they reported a daily maximum temperature below 5°C or a daily minimum temperature above 40°C during the study period, extremes that would certainly eliminate the potential for transmission in a local area. Because case data are reported at the country level, we needed a collection of weather stations in each country that accurately represent weather conditions in the areas where transmission occurs, excluding extreme areas where transmission is unlikely. For the study period of October 1, 2013 through April 30, 2016, we downloaded daily temperature data for each station from Weather Underground using the weatherData package in R [59]. We removed all data from Chile because it spans so much latitude and the terrain is so diverse that its country-level mean is unlikely to be very representative of the temperature where an outbreak occurred.
We accessed available data on projected 2016 gross domestic product (GDP) for countries of interest via the International Monetary Fund’s World Economic Outlook Database (http://www.imf.org/external/ns/cs.aspx?id=28). The direct and total contributions of tourism to GDP in 2016 were compiled from World Travel and Tourism Council economic impact reports (http://www.wttc.org/research/economic-research/economic-impact-analysis/country-reports/#undefined). We retrieved population size data for 2013–2015 from the United Nations Population Division (https://esa.un.org/unpd/wpp/Download/Standard/Population/) and averaged them across the three years for each country. Throughout the analyses below, unless otherwise specified, we used the natural log of the population size and of GDP as our predictors. We have two reasons for this choice. The first is that, intuitively, the relative order of magnitude of the population/GDP is more important in determining observed outbreak sizes or probabilities than their absolute sizes. Second, population sizes and GDPs across countries tend to exhibit clumped patterns with a few outliers that are much larger than the others. From a statistical perspective, using the un-transformed populations (or GDPs) results in those few large/rich countries having very high leverage in the analysis, and thus potentially skewing the results. Taking a log of the population better balances these predictors and is the standard accepted approach when using these kinds of predictors in regression models.
To validate the R0(T) model while controlling for population and socio-economic factors, we used generalized linear models (GLMs) on the weekly case count data. Importantly, we focused on testing whether the case counts were consistent with the transmission–temperature relationship predicted from our model, rather than on maximizing the variation explained in the statistical model. We are more specifically interested in understanding autochthonous transmission (i.e., locally acquired, not just imported cases). We set country-level thresholds for the number of cases defining autochthonous transmission for our three diseases separately, based on current transmission understanding: seven cases of CHIKV, 70 cases of DENV, and three cases of ZIKV. We derived these thresholds in the following way. First, we looked for data on outbreaks of travel related cases in countries that are not expected to experience any local transmission. For instance, in 2014 Canada experienced 320 confirmed, travel-related cases of chikungunya (http://www.phac-aspc.gc.ca/publicat/ccdr-rmtc/15vol41/dr-rm41-01/rapid-eng.php), equivalent to an average of more than six cases per week. Thus, to be conservative in our estimates, we set the threshold of transmission as seven cases/week for CHIKV. The reported weekly cases of DENV transmission in our study sample are considerably higher than for CHIKV (mean DENV incidence was nearly 100 times higher mean CHIKV incidence). We chose a moderately high threshold of 70 cases in a week (i.e., 10 times higher than the CHIKV threshold based on Canadian cases) to reflect higher overall incidence and increased potential for travel related cases. We examined the sensitivity of the results to choice of threshold by varying it from 25 to 100, and we found qualitatively similar results for all thresholds that we tested. As ZIKV is not as well established as either CHIKV or DENV at this time, smaller numbers of cases may indicate autochthonous transmission. Consequently, we chose a threshold of three cases for ZIKV (approximately half the CHIKV threshold). Further, the results were fairly sensitive to the ZIKV threshold as many locations have small numbers of cases. Since higher thresholds exclude a very large proportion of available case data making analysis impossible, we used the slightly less conservative threshold of three cases for autochthonous transmission of ZIKV. The resulting data consisted of zeros for no transmission and positive case counts when transmission is presumed to be occurring. To model these data, we used a hurdle model that first uses logistic regression on the presence/absence of local transmission data to understand the factors correlated with local transmission occurring or not (PA analysis). Then we modeled the log of incidence (number of new cases per reporting week) for positive values with a gamma generalized linear models (incidence analysis).
We were interested in understanding whether R0(T) was an important predictor of human transmission occurrence and incidence, after controlling for potentially confounding factors like population size and socioeconomic conditions. To do this, we fit a series of models with different subsets of predictors that included R0(T) and population size, the socioeconomic variables, or both (see Table D in S2 Text for full models). To control for human population size, we created new metrics based on R0(T) and population size to use for validation against the PAHO incidence data. We define GR0, which is the posterior probability that R0(T) > 0. We use log(p)*GR0, where p is the population size, as the relevant R0-based predictor for the PA analysis. For the incidence analysis, we instead use log(p*R0(T)) as the predictor. In all cases log refers to the natural logarithm. For simplicity, we refer to these as the R0(T) metrics hereafter and in the Results.
In both the PA and incidence analyses, we first used the full data sets to examine which of the candidate models best described the data. Randomized quantile residuals indicated that the logistic and gamma GLM models were performing adequately. We compared the approximate model probabilities, calculated from the BIC scores, as well as the proportion of deviance explained (D2) from each model. Next we examined the performance of the models in predicting out of sample, for both PA and incidence analyses. To do this we created 1000 random partitions, where 90% of the data were used to train the model and 10% were used for testing. In the PA analyses we classified each partition based on presence/absence, with separate classification thresholds for DENV versus CHIKV/ZIKV as these grouping had much different probabilities of occurrence. We assessed the performance of the model for the PA analysis based on the mean misclassification rate. In the incidence analyses we assessed the model performance based on the predictive mean absolute percentage error (MAPE). Since differences in prediction success between the models in both the PA and incidence analyses were not statistically significant, we present the simpler models that only include the R0(T) metrics in the main text (Fig 3) and the models that additionally include socioeconomic covariates in the Supplementary Information (Figs B-C in S1 Text). We plotted the model predictions as a function of the R0(T) metrics together with the observed data for the PA and incidence analyses using the R package visreg [60].
The residuals of the incidence model exhibit “inverse trumpeting,” in which residual variation is larger at low than high predicted incidence (Fig I in S1 Text). This occurs in part because we forced the model to go through the origin, i.e., no transmission when R0(T) or the population size is equal to zero. However, the data did sometimes show transmission where we did not expect it, potentially because of imported cases, errors in reporting, or small pockets of transmission suitability in countries or times that are otherwise unsuitable on average. More local-scale case reporting that separates autochthonous from travel-associated cases would be needed to tease apart the source of this error.
Using the validated model, we were interested in where the temperature was suitable for Ae. aegypti and/or Ae. albopictus transmission for some or all of the year to predict the potential geographic range of outbreaks in the Americas. We visualized the minimum, median, and maximum extent of transmission based on probability of occurrence thresholds from the R0 models for both mosquitoes. We calculated the number of consecutive months in which the posterior probability of R0 > 0 exceeds a threshold of 0.025, 0.5, or 0.975 for both mosquito species, representing the maximum, median, and minimum likely ranges, respectively. The minimum range is shown in Fig 4 and all three ranges are overlaid in Fig D in S1 Text. This analysis indicates the predicted seasonality of temperature suitability for transmission geographically, but does not indicate its magnitude. To generate the maps, we cropped monthly mean temperature rasters from 1950–2000 for all twelve months (Worldclim; www.worldclim.org/) to the Americas (R, raster package, crop function) and assigned cells values of one or zero depending on whether the probability that R0 > 0 exceeded the threshold at the temperatures in those cells. We then synthesized the monthly grids into a single raster that reflected the maximum number of consecutive months where cell values equaled one. The resulting rasters were plotted in ArcGIS 10.3, overlaying the three cutoffs (Fig D in S1 Text). We employed this process for both mosquito species.
|
10.1371/journal.ppat.1007397 | The RNA helicase DDX3X is an essential mediator of innate antimicrobial immunity | DExD/H box RNA helicases, such as the RIG-I-like receptors (RLR), are important components of the innate immune system. Here we demonstrate a pivotal and sex-specific role for the heterosomal isoforms of the DEAD box RNA helicase DDX3 in the immune system. Mice lacking DDX3X during hematopoiesis showed an altered leukocyte composition in bone marrow and spleen and a striking inability to combat infection with Listeria monocytogenes. Alterations in innate immune responses resulted from decreased effector cell availability and function as well as a sex-dependent impairment of cytokine synthesis. Thus, our data provide further in vivo evidence for an essential contribution of a non-RLR DExD/H RNA helicase to innate immunity and suggest it may contribute to sex-related differences in resistance to microbes and resilience to inflammatory disease.
| The establishment of innate immunity to pathogens requires cells to sense microbial molecules and to initiate a de novo transcription-based antimicrobial response. With the identification of Rig I and Mda5, two RNA helicases were shown to serve as pivotal receptors of viral RNA. Subsequently, a considerable number of RNA helicases were proposed to function as sensors or signal transducers for both microbial RNA and DNA. X-chromosome-encoded RNA helicase DDX3X was discovered as an interactor of the S/T kinase TBK1 which regulates the production of type I Interferons (IFN-I). However, the importance of DDX3X for innate immunity in an organismic context remained elusive. Here we describe and analyze mice lacking DDX3X in hematopoietic cells. We show contributions of DDX3X to hematopoiesis and a striking loss in resistance against Listeria monocytogenes. Our data reveal that DDX3X is critically involved in enhancing the expression of numerous antimicrobial genes. Consistently, production of important cytokines such as IL12 or IFNγ is reduced. Furthermore, DDX3X-deficient macrophages show reduced ability to restrict L. monocytogenes growth. Owing to partial redundancy with its close Y-chromosomal homologue, DDX3Y, the observed effects differ between mouse sexes. Thus, DDX3X may contribute to sex differences in immunity to pathogens and inflammatory disease.
| Upon infection, germline-encoded pattern recognition receptors (PRRs) located on the surface of cells, in endosomal compartments and throughout the cytosol initiate an array of signaling cascades that culminate in the production of type I interferons (IFN-I), pro-inflammatory cytokines and chemokines. These cytokines establish an inflammatory response and an antimicrobial state restraining the spread of the infectious agent.
The discovery of Rig-I-like receptors (RLR) as sensors of viral RNA sparked considerable interest in the role of other DExD/H RNA helicases as innate modulators of antimicrobial immune responses [1,2]. DExD/H helicases not belonging with typical RLR also contribute to innate immunity in experimental animals, as recently demonstrated for DDX41 which acts in dendritic cells to limit retroviral growth [3]. We and others have identified the RNA helicase DDX3X as a regulator of IFN-I transcription in cells infected with viruses or with the intracellular bacterial pathogen Listeria monocytogenes [4,5]. DDX3X belongs to the DEAD-box RNA helicase superfamily 2 [6] that has widespread functions in RNA metabolism, including transcription, RNA processing, splicing, decay and translation [7,8]. Moreover, DDX3X is implicated in cellular processes such as apoptosis, cell cycle regulation and tumorigenesis [9]. Deletion of DDX3X in all embryonic tissues causes the death of male embryos at an early postimplantation stage. By contrast, male embryos with epiblast-restricted DDX3X deletion die around E11.5 with widespread occurrence of apoptotic cells and expression of DNA damage markers [10]. This is most likely a direct consequence of a disturbed cell cycle in embryonic tissue lacking DDX3X. This view is further supported by a study investigating the role of DDX3X in early mouse development using siRNA-mediated knockdown [11].
A homologue of DDX3X called DDX3Y is encoded by the non-recombining region of the Y-chromosome. DDX3X and DDX3Y share around 90% homology. While DDX3X is ubiquitously expressed, DDX3Y protein expression was originally thought to be confined to the male germline [12]. More recent proteomic databases list DDX3Y in cells of the immune system, including T-cells, B-cells and NK-cells. Their high degree of similarity supports the idea that DDX3X and DDX3Y are functionally redundant [13]. The heterosomal origin of the DDX3 isoforms suggests they may contribute to sex-related differences in immunity to microbes, the ability to resolve inflammation and the propensity to develop autoinflammatory syndromes [14–16].
Several studies point to an ambiguous role of DDX3X in viral infections. On the one hand, it may promote replication of viruses like HIV or HCV [17–22]. On the other hand, DDX3X stimulates the production of antiviral IFN-I [23,24]. Antimicrobial pathways leading to IFN-I synthesis converge at two related S/T kinases, TBK1 and IKKε. DDX3X interacts with TBK1 and serves as its substrate [4]. It also interacts with IKKε (Schröder et al, 2008; Gu et al, 2013), the other non-canonical IKK kinase responsible for phosphorylation and activation of the interferon regulatory factors (IRF) 3 and 7 that control Ifnb gene transcription. In addition to its function downstream of TBK1/IKKε, DDX3X reportedly associates with the adaptor protein MAVS which localizes upstream of TBK1 and IKKε and supports signal transduction by the two DExH helicases RIG-I and MDA-5. In this context DDX3X was shown to sense viral RNA and to supplement the function of RIG-I and MDA-5 in the early phase of infection [25]. Strengthening its importance for antiviral immunity, a recent report lists DDX3X among genes conferring intrinsic antiviral immunity to stem cells [26].
Like viruses, the Gram-positive bacterium Listeria monocytogenes (Lm) is a potent inducer of IFN-I [27]. After phagocytosis by macrophages, Lm escapes to the cytosol via the secretion of listeriolysin O (LLO) that disrupts the phagosomal membrane [28,29]. The innate response to Lm is characterized by intracellular effector mechanisms as well as the secretion of many pro-inflammatory chemokines and cytokines, among them the IFN-I. Cytoplasmic nucleic acids derived from Lm were shown to trigger strong induction of the IFN-I genes. Lm DNA promotes IFN-I expression through the DNA sensors cGAS and IFI16, the adaptor molecule STING, as well as TBK1 kinase and its downstream targets IRF3 and IRF7 [30,31]. In addition to DNA, Lm RNA was implicated in the induction of type I IFNs via the cytosolic receptor RIG-I [32]. Knock-down of DDX3X phenocopies the silencing of TBK1 on Ifnb gene induction after Lm infection [4], emphasizing the general importance of DDX3X for the TBK-IRF3 axis. Despite the fact that IFN-I are crucial for protective antiviral responses, their impact on Lm infection appears to be detrimental for the host as evidenced by the observation that mice lacking the receptor for type I IFNs (IFNAR) are more resistant to parenteral infection with Lm [33–35]. Unlike IFN-I, the type II interferon (IFNγ) is strongly associated with protective innate immunity against Lm [36,37]. Rapid production of IFNγ is essential and has been attributed to innate lymphocyte responses that include natural killer cells [38]. The resistance against Listeria provided by IFNγ is strongly associated with its role as a macrophage-activating cytokine [39].
Here we report a crucial role of DDX3X in the innate immune responses of cells and mice. We show that besides its role in the regulation of the TBK1-IRF3 axis, DDX3X controls the NFκB signaling pathway and has a profound impact on inflammatory cytokine production. DDX3Y, either alone or together with additional Y-chromosomal genes, partially compensates for the loss of the Ddx3x gene, as homozygous female cells and mice show more severe loss-of-function phenotypes. Mice lacking DDX3X in the hematopoietic system show alterations of bone marrow and splenic cell populations and are highly susceptible to Lm infection. Our data thus demonstrate a vital role of the sex-specific DDX3 isoforms in innate immunity to Listeria.
To investigate DDX3X activity in the immune system we introduced loxP sites flanking exon 2 of the DDX3X locus into the mouse genome (Ddx3xfl/fl; Fig 1A). Deletion of DDX3X in bone marrow-derived macrophages (BMDM) by means of tamoxifen (4-OHT) -inducible Cre recombinase resulted in complete loss of DDX3X protein (Fig 1B). Complete Cre-mediated deletion of DDX3X in mice failed to produce viable offspring, consistent with impaired blastocyst formation and early embryonic lethality reported by others [10,11]. Vav-iCre -mediated deletion of DDX3X in the hematopoietic system allowed the development of male mice (Ddx3xfl/y Vav-iCre). In contrast, homozygous female offspring (Ddx3xfl/fl Vav-iCre) was not obtained. This emphasizes an important role of DDX3 isoforms in hematopoiesis and suggests that the Y chromosome contains genes that compensate for DDX3X deficiency to the point of ensuring survival and an absence of overt phenotypic abnormalities of unchallenged animals. The Y-chromosomal DDX3X homologue DDX3Y is an obvious candidate for this rescue, either alone or in combination with other Y-chromosomal genes. Consistent with an overlapping spectrum of activities, DDX3Y enhanced Ifnb gene expression in an almost identical manner to DDX3X when introduced into Ddx3x-deficient MEFs by transfection (Fig 1C) and knockdown of DDX3Y decreased poly (dA:dT)-stimulated IFNβ expression, similar to the tamoxifen-mediated knockout of DDX3X in fibroblasts derived from Ddx3xfl/y-CreERT2 mice (Fig 1D). Interestingly, both DDX3X and DDX3Y enhanced the activity of a constitutively active IRF7 variant, IRF7-M15 [40], in fibroblasts lacking both TBK1 and IKKε kinases (Fig 1E). This suggests that TBK1/IKKε-mediated phosphorylation is dispensable for the ability of DDX3X/Y to enhance IFNβ synthesis, or that fibroblasts express some constitutive DDX3X/Y kinase activity. Of further note, DDX3X was able to enhance the activity of an NFκB reporter gene (Fig 1F), suggesting its impact on innate immune responses may extend beyond the IRF3/7 pathway.
Based on the role of DDX3X in IFN-I synthesis we investigated whether loss of the helicase causes a defect in innate immunity to virus. In male fibroblasts, derived from Ddx3xfl/y-CreERT2 mice and rendered DDX3X-deficient by treatment with tamoxifen, we observed reduced synthesis of IFNα as well as IFNβ mRNA following infection with vesicular stomatitis virus (VSV; Fig 2A and 2B). In accordance with this, VSV replicated to higher numbers in DDX3X-ablated cells compared to controls that had not been treated with tamoxifen (4OHT; Fig 2C). Surprisingly however, Ddx3xfl/y Vav-iCre mice infected i.v. with VSV showed only a marginal reduction of viral clearance, in spite of the fact that hematopoietic cells contribute to IFN synthesis when i.v. injection is chosen as infection route (Fig 2D; [41,42]). This result suggests that any defect in IFN-I synthesis caused by deletion of DDX3X in hematopoietic cells can be compensated and that VSV infection does not reveal a major defect in the DDX3X-deficient innate immune system.
To further characterize the role of DDX3X in immune responses, we subjected Ddx3xfl/y Vav-iCre mice to intraperitoneal (i.p.) infection with the intracellular bacterial pathogen Listeria monocytogenes (Lm). Compared to Ddx3xfl/y control animals, Ddx3xfl/y Vav-iCre mice were highly susceptible to Lm. All knockout animals succumbed to infection within 6 days, whereas most control animals survived the observation period (Fig 2E). In line with this, the bacterial burdens in spleens and livers of Ddx3xfl/y Vav-iCre animals were strongly increased three days after i.p. infection (Fig 2F). IFN-I deficiency increases the innate resistance of mice against Lm [33–35]. Therefore, pro-survival effects of reduced IFN-I synthesis in Ddx3xfl/y Vav-iCre animals are clearly overwhelmed by immune defects that weaken host resistance.
To address the consequences of DDX3X deficiency and to seek an explanation for reduced resistance to Lm, we determined the composition of mature hematopoiesis-derived cell populations in bone marrow and spleen. In the bone marrow total cell numbers were slightly reduced due to a rather selective loss of bone marrow B lymphocytes (Fig 3A). We found splenic composition to be more dramatically changed. Besides B220+ B cells the numbers of CD3+CD4+ and CD3+CD8+ T cells, CD3+CD1d-tetramer+ NKT cells and CD3-NK1.1+ NK cells were significantly reduced in Ddx3xfl/y Vav-iCre mice. Interestingly, cells of myeloid origin were not generally affected. Numbers of total CD11c+ dendritic cells, CD11b+Ly6G+Ly6Clo neutrophils and CD11b+F4/80+ macrophages unaltered and only CD11b+Ly6Chi monocytes showed a slight reduction (Fig 3B).
To assess whether reduced lymphoid cell numbers were due to alterations in hematopoiesis we counted the numbers of lineage (Gr1, Mac1, Ter119, B220, CD3)—Sca1+ cKit+ (LSK), lineage- Sca1- cKit+ (LK), as well as lineage- Sca1/cKitdouble-dim CD127+ common lymphoid progenitor (CLP) cells and found all three populations to be significantly reduced (Fig 3C), indicating a role for DDX3X in early hematopoietic development. In line with this observation, further analysis of bone marrow B cell development showed a slight reduction in the numbers of pre-pro- and early pro B cells, as well as a particularly strong reduction of progenitors from the small pre- B cell stage onwards, i.e. following the proliferation of large pre B cells (Fig 4A). Surprisingly, the development of NK cell precursors in the bone marrow seemed unaffected by loss of DDX3X, with lineage (Gr1, Ter119, B220, CD3)- CD122+ NK1.1- Dx5- NK progenitors (NKp), lineage- CD122+ NK1.1+ immature NK cells (iNK) and lineage- CD122+ NK1.1+ Dx5+ mature NK cells (mNK) all present at normal numbers, indicating that the reduction in CLP cells can be compensated for during NK cell development (Fig 4B).
Since the numbers of splenic T and NKT cells were also dramatically affected in Ddx3xfl/y Vav-iCre mice we more closely examined development of these lineages in the thymus. Interestingly, the number of thymic CD4/CD8 double-positive (DP) and CD8 single-positive (SP) thymocytes was increased in absence of DDX3X, while CD4 SP thymocytes seemed unaffected, suggesting that a developmental block is not responsible for the reduced numbers of mature T cells observed in the periphery (Fig 4C). In contrast, NKT cell development in the thymus was more dramatically affected with CD1d-tetramer+ CD24- stage 1, CD1d-tetramer+ CD24- CD44+ stage 2 and CD1d-tetramer+ CD24- CD44+ NK1.1+ stage 3 NKT cells all strongly reduced in numbers (Fig 4D).
In summary, our analysis of hematopoiesis in Ddx3xfl/y Vav-iCre mice indicated a differential impact on distinct lineages. While the reduced numbers of peripheral B and NKT cells can be at least partially explained by developmental alterations, the lack of splenic T and NK cells does not seem to result from an obvious developmental block.
Given that DDX3X-deficient embryos show increased apoptosis we investigated whether cell death might contribute to the reduction of DDX3X-deficient splenic leukocytes. As shown in Fig 5, the absence of DDX3X caused elevated rates of cell death in B220+ B cells, CD3+ T cells and CD3- NK1.1+ NK cells, and to a lesser extent in CD11b+ myeloid cells. These results suggest that DDX3X, besides its variable roles in the development of different immune cell lineages, also plays a role in the maintenance of immune cells in the periphery, with the most profound effects observed in lymphocytes and splenic NK cells.
To determine whether Lm infection exacerbates defects in hematopoiesis, splenic cell populations were analysed in infected mice. 24 hours of infection increased the number of total splenocytes in both control and Ddx3xfl/y Vav-iCre mice (first panels of Figs 3A and 6A, Table 1), mainly due to recruitment of Ly6Chi inflammatory monocytes (approx.13-fold versus 12-fold mean increase, respectively), suggesting no obvious recruitment defects caused by the absence of DDX3X. By and large, the cell populations suffering from DDX3X deficiency were the same as in uninfected animals with B220+ B cells, CD3+ CD8+ T cells, CD3+ CD1d-tetramer+ NKT cells and CD3- NK1.1+ NK cells being particularly compromised. By comparison, DDX3X-deficient myeloid cells showed a rather mild decrease that did not reach statistical significance (Fig 6A). Table 1 summarizes the infection-induced relative changes in splenic leukocyte composition. While some differences in the recruitment and/or proliferation of different leukocyte populations was observed, this display of the data emphasizes that most changes in leukocyte numbers precede infection rather than being result of the innate response to Lm.
In our animal experiments, the primary site of Lm infection is the peritoneal cavity. We therefore analysed cell recruitment to this anatomical location. The data summarized in Fig 6B show that unlike the spleen hematopoietic DDX3X deficiency per se has little effect on immune cells residing in the peritoneum, with only the number of Ly6Chi monocytes slightly reduced in PBS treated control animals. Upon infection with Lm we observed a rise in peritoneal exudate cell (PEC) numbers owing primarily to recruitment of Ly6CloLy6G+ neutrophils, Ly6Chi monocytes and CD11b+F4/80+ macrophages. While neutrophil recruitment seemed unaffected, the number of inflammatory monocytes was significantly decreased in Ddx3xfl/y Vav-iCre animals and we observed a tendency for lower macrophage numbers. These results indicate a requirement for DDX3X in the generation, maintenance or recruitment of inflammatory monocytes particularly in the peritoneal cavity.
To address potential protective immune mechanisms under DDX3X control we determined the levels of several serum cytokines and chemokines in mice infected with Lm. Most of the factors we assessed were not significantly different between Ddx3xfl/y Vav-iCre mice and control animals. At 24 hours after infection, serum levels of Il-6, IL-17, IL-10, TNFα and IL1β were unchanged and while the levels of Il-6, IL-10, TNFα and IL1β increased in Ddx3xfl/y Vav-iCre relative to control animals at 72h, the difference only reached significance in the case of IL-10. In addition, we observed a small but significant reduction in IL-4 levels at both time points as well as reduced CCL5/RANTES concentrations at 24 hours (S1 Fig). The most striking difference was the reduction of serum IFNγ and IL-12 in Ddx3xfl/y Vav-iCre mice at 24 hours after infection. At 72hrs IFNγ levels of Ddx3xfl/y Vav-iCreand control mice were similar, either due to a recovery of the former, or to more rapid subsiding of IFNγ production in the latter animals (Fig 7A). The results suggest an inhibited IL12-IFNγ axis in the early stage of Lm infection of Ddx3xfl/y Vav-iCre mice.
Upon L. monocytogenes infection, early production of IFNγ by NK-cells, NKT cells and CD8+ T-cells is crucial for the activation of macrophage effector functions and subsequent bacterial clearance [43–45]. Early IFNγ production requires IL-12, as IL-12 depletion leads to abrogated IFNγ levels and reduced resistance to infection [36,38,46,47]. To address the contribution of these cell populations to the differences seen in IFNγ levels 24 hours post-infection, flow-cytometric analysis of IFNγ production was performed. We found that NK cells were an important source of IFNγ in wt animals at this early time point, because the fraction of producer cells was the largest among the investigated cell types (Fig 7B). Numbers of NK cells are much lower in Ddx3xfl/y Vav-iCre mice and of the few remaining cells a smaller fraction produced IFNγ, possibly as a consequence of reduced IL-12 amounts. The fraction of IFNγ-producing splenic iNKT cells was low and not significantly affected by genotype, while IFNγ– producing CD8+ T cells were below the limit of detection (Fig 7B). These data suggest that NK cells make an important contribution to serum IFNγ in the early phase of Lm infection, and their absence in Ddx3xfl/y Vav-iCre mice is largely responsible for reduced serum levels of IFNγ at the 24 hour time point.
To verify our assumption that NK cells are the primary producers of IFNγ at 24 hours after Lm infection, we depleted NK cells from Ddx3xfl/y and Ddx3xfl/y Vav-iCre mice using an anti-NK1.1 antibody. 72 hours after intraperitoneal administration about 98% of NK cells were depleted, while splenic CD1d-tet+ iNKT cells remained unaffected under these conditions. The NK-depleted mice were then infected with Lm and 24 hours later serum was collected. Depletion of NK-cells strongly reduced serum IFNγ from about 104pg/ml to less than 103pg/ml after NK depletion (leftmost panel in Fig 7A and 7C). NK depletion also abolished the differences between wild type controls and Ddx3xfl/y Vav-iCre mice (Fig 7C). In line with earlier findings [48] [49], depletion of NK cells did not reduce the bacterial burden in infected organs of wt mice (Fig 7D). Therefore, although our data emphasize the contribution of NK cells to serum IFNγ production during the early, innate immune response against Lm, these results suggest that residual serum IFNγ in NK-depleted mice, or local production by other cells, suffices for macrophage activation in the early phase of Lm infection. Our data also support our notion that NK cells are the primary cell type responsible for the lack of sufficient early serum IFNγ production in mice lacking hematopoietic DDX3X.
Macrophages are essential effector cells against Lm. Reduced proinflammatory activity and/or cell-autonomous antimicrobial states in absence of DDX3X might thus explain the reduced innate immunity of Ddx3xfl/y Vav-iCre mice. To study antimicrobial responses of both male and female sexes we generated bone marrow derived macrophages (BMDMs) from Ddx3xfl/fl CreERT2 and Ddx3xfl/y CreERT2 mice. In vitro deletion of DDX3X by tamoxifen treatment allowed us to address the effect of Ddx3x deletion in the presence or absence of the Ddx3y gene. Tamoxifen treatment after 5 days of culturing bone marrow cells in CSF-1-containing medium did not impair the terminal differentiation of macrophages from Ddx3xfl/fl CreERT2 female and Ddx3xfl/y;CreERT2 male mice. However, the fully differentiated, female Ddx3x-/- cells showed a reduced life span and were used no later than 3 days after the termination of tamoxifen treatment. To determine the impact of DDX3X on cell-autonomous clearance of Lm, BMDMs of male (Fig 8A, left panel) and female mice (Fig 8A, right panel) were infected with Lm at an MOI of 10 and colony forming unit (CFU) assays were performed to monitor bacterial growth. DDX3X-deficient cells from both sexes contained higher bacterial loads compared to the wild-type cells 8 hours after infection. Activation of DDX3X-deficient female cells with IFNγ resulted in increased killing of intracellular Lm. The relative drop in bacterial loads were similar to wt (2.3 versus 2.7-fold) but in absolute numbers about twice as many bacteria persisted in DDX3X-deficient macrophages (Fig 8A, left panel). The data suggest that IFNγ responsiveness per se is unaffected, but that the remaining difference in bacterial numbers may result from a higher initial burden compared to wt cells.
The consequences of DDX3X deficiency on the expression of genes contributing to the innate antimicrobial response of macrophages were examined by RNA-Seq analysis of untreated or Lm-infected cells. Genes differentially expressed upon Lm-infection in either male or female cells (S1 Table) are enriched for GO Immune System terms, and the extent to which DDX3X deficiency altered the transcriptome depended on the sex chromosomes of the cells (Fig 8B, heatmap and upper right panel). This suggests that one or more Y-chromosomal genes partially compensate for the loss of DDX3X at the level of transcript synthesis. The data demonstrate control of the macrophage transcriptome by DDX3 isoforms and, in line with the mouse phenotypes, an incomplete redundancy between DDX3X and DDX3Y and/or additional genes present on the Y chromosome.
wt mice of female sex show higher susceptibility to Lm infection and immune-related genes may show sex-dependent differences in expression [50]. In line with this, we observed a considerable variation in the differentially expressed genes in macrophages of infected wt male and wt female mice (heatmap of Fig 8B). Comparison of gene sets differentially expressed upon Lm infection at the same significance threshold in the four experimental conditions revealed that a set of 311 genes responded with altered expression in wt males, but not in wt females, whereas only 7 genes responded to infection in females, but not in males (Fig 8B, Venn diagram). 230 of the 311 male-specific genes and all of the 7 female-specific genes were unresponsive to infection in macrophages of the same sex lacking DDX3X. The data suggest a contribution of DDX3X to sex-specific gene expression in response to infection with Lm.
Validation of critical antimicrobial genes affected by the lack of the Ddx3x gene in RNA-Seq experiments by qPCR is shown in Fig 8C. Cytokines with a strong impact on innate resistance to Lm, such as IL-1, IL-6, IL-12, TNFα as well as chemokines were controlled by DDX3X. Further consistent with the RNA-Seq results, the reduction in expression for most of the examined genes was more pronounced in female (lower panel) than in male macrophages (upper panel). The broad effect of DDX3X deficiency on cytokine production is consistent with the impact of the protein on both IRF3/7 and NFκB pathways (Fig 1C, 1E and 1F). To determine whether the DDX3X effect on these genes is at the level of transcription rather than transcript processing, primary transcript synthesis of a number of mRNAs was examined at 2hrs or 4hrs after infection in female macrophages. Consistent with our earlier observation that DDX3X increases Ifnb gene expression at the level of transcription [4], most primary transcripts were reduced, the Tnfa transcript being a noteworthy exception (Fig 8D). DDX3X-deficiency also blunted responses to the defined pathogen-associated molecular patterns (PAMPs) poly (I:C), poly (dA:dT) and LPS (Fig 9A). With exception of the Il6 gene, the impact of DDX3X deficiency on signalling by the respective pattern recognition receptors was more pronounced in female cells (upper versus lower panels). These results show that DDX3X affects both primary transcription as well as transcript processing of genes targeted by microbial sensor molecules. Finally, to assess the production of Interferon stimulated genes (ISGs) in absence of DDX3X we treated DDX3X-ablated and control BMDMs with IFNs. As shown in Fig 9B, the induction of Mx1, Oas2, Ifit3 and Isg15 mRNAs was not significantly altered between genotypes upon stimulation with IFNβ (left panels). Similarly, induction of Irf1 mRNA after addition of IFNγ was comparable between the genotypes (rightmost panel). These data indicate that DDX3X-deficient macrophages respond normally to exogenous sources of IFN.
In this report we show that DDX3X deficiency in hematopoietic cells results in a striking susceptibility against the bacterial pathogen Listeria monocytogenes (Lm). The lack of effect on VSV-infected mice suggests that ablation of DDX3X in the hematopoietic system does not result in a complete breakdown of innate immunity. Immune responses against Lm commence with an early innate phase including macrophages, inflammatory monocytes, NK-cells, DCs, CD4+, CD8+ and γδT cells that together constrain the infection (Andersson et al., 1998; Lee et al., 2013; Pamer, 2004b; Shi et al., 2011; Unanue, 1997). Protective immunity requires the synthesis of several proinflammatory cytokines such as TNF, IL6 and IL1, whose spatial and temporal appearance is orchestrated by the immune system. The role of IFNγ in the context of Listeriosis is well-established, as evidenced by the increased susceptibility of mice deficient in either IFNγ [37] or the IFNγ receptor (Huang et al., 1993; Lee et al., 2013). Here we identify three factors contributing to the selective susceptibility of mice lacking hematopoietic DDX3X against Lm. First, the selective defects in hematopoiesis reduce the ability to establish the inflammatory milieu needed for innate defence against Lm. Among myeloid cells, inflammatory monocytes are the only cell type strongly dependent on DDX3X and this cell type is indispensable for restricting Lm growth [51]. Second, DDX3X deficiency impairs antimicrobial gene expression in macrophages, a major effector cell against Lm. The absence of the helicase causes a widespread reduction of genes providing innate immunity, amongst which are proinflammatory cytokines, but also intracellular effectors such as iNOS. The findings reported here and by others [5] show that DDX3X enhances the IRF3/7 as well as the NFκB pathways. This provides an explanation for its strong impact on infection-induced genes and poses the question whether other pathways conveying immunity are influenced by DDX3X as well. The alteration of macrophage gene expression has two important consequences, the lack of cell-autonomous defence and, once more, a reduced ability to establish an inflammatory milieu. Third, DDX3X is needed for a fully active IL12-IFNγ axis. Our data support the notion that this results from a defect of IL12 production by macrophages and possibly other cell types, and from a decrease of IFNγ-producing cells.
In spite of being described as a substrate of the TBK1 and IKKε kinases we observed that DDX3X enhances the induction of the Ifnb gene also in the absence of TBK1/IKKε when a constitutively active phosphomimetic mutant of IRF7 is provided. This demonstrates that DDX3X retains the ability to enhance IRF-mediated transcription in absence of TBK1/IKKε. Therefore, activating phosphorylation in this situation is either dispensable or performed by other kinases, for example, the conventional IKK. The latter interpretation is in line with the ability of DDX3X to enhance NFκB-dependent transcription, but requires experimental confirmation.
Ddx3x is essential for development, as its deletion leads to prenatal lethality in mice, where DDX3X is necessary for both embryonic and extraembryonic tissue development [10]. Both DDX3X and DDX3Y are among essential genes in some, but not all investigated human leukemic cell lines [52]. Hematopoietic loss of DDX3X via Vav-Cre mediated deletion resulted in viable male but not female mice, emphasizing the pivotal role of DDX3X in hematopoietic development. The Y-chromosomal homologue DDX3Y shares 90% (mouse) or 92% (human) identity at the protein level. The high degree of homology manifested in its redundant functions in enhancing Ifnb gene expression. Therefore, it is tempting to speculate that DDX3Y alone is responsible for the hematopoietic rescue provided by the Y-chromosome. However, this has not been directly tested and a participation of additional Y-chromosomal genes cannot be ruled out. Sofar we have not examined the hematopoietic defect underlying the death of female Ddx3xfl/y Vav-iCre mice. A failure to produce erythrocytes would seem a likely explanation.
Analysis of hematopoiesis-derived cells in bone marrow and spleen revealed three degrees of susceptibility to the loss of DDX3X. Particularly B cells were impeded in their development, leading to reduced numbers in both organs. NK cells on the other hand developed normally in the bone marrow, but showed reduced presence in the spleen. This defect may result either from reduced recruitment to the spleen, or a difference in survival conditions between splenic and bone marrow NK cells. Finally, mature myeloid cells such as macrophages, neutrophils, or CD11c+ DC showed little dependence on DDX3X in either organ. One likely explanation for reduced hematopoiesis is an increase in apoptosis in DDX3X-deficient cells, particularly B cells and NK cells. Based on studies in mouse embryos this may result from a p53-mediated response to DNA damage and reflect an activity of DDX3X in cell cycle regulation and/or DNA damage repair (Bol et al, 2015; Sun et al, 2013; Li et al, 2014; Chen et al, 2016). In keeping with this notion a preliminary proteomic screen in macrophages identified multiple DDX3X interactors with potential impact on DNA damage and repair pathways. Among these factors are the DBHS family members SFPQ, NONO and PSPC1 [53]. The idea that the increase in apoptosis found in DDX3X-deficient cells is linked to proliferation is supported by cell types found to be particularly sensitive, such as the small preB stage in B cell development, or inflammatory monocytes that, unlike resident splenic macrophages, showed increased apoptosis when lacking DDX3X.
Given the role of DDX3X in the IFN-I synthesis of infected cells we were surprised to find innate immunity to VSV unimpaired. Hematopoietic cells, particularly pDC, are thought to be essential IFN producers during systemic VSV infection [41,42]. One possibility to explain our results is that pDC may not require DDX3X for IFN-I production. Alternatively, serum levels of IFN-I may not accurately reflect the local need to clear infection. This hypothesis is consistent with our findings in Lm-infected mice: NK cells were responsible for high serum IFNγ amounts, but their depletion still left enough IFNγ production for unimpaired innate immunity. An alternative explanation is therefore that our findings with VSV reflect a greater than hitherto suspected need for IFN-I from nonhematopoietic cells. We show that induction of antiviral genes by IFN-I is unimpaired in DDX3X-deficient cells. This further suggests that any IFN-I production defect in Ddx3xfl/y Vav-iCre mice does not reduce the levels of the cytokines below the threshold needed for the antiviral state.
The susceptibility of Ddx3xfl/y Vav-iCre mice to Lm shows that DDX3Y alone is not generally sufficient to compensate for the lack of DDX3X in the immune system. Due to the lethality of DDX3X deficiency in female mice, we cannot assess the relative contribution of DDX3Y to the innate immune response against Lm in vivo. Specifically, we cannot address a potential contribution of the DDX3 isoforms to the increased susceptibility of female C57BL/6 mice noted by others [50,54]. Bone-marrow derived macrophages from Ddx3xfl/fl CreERT2 mice, however, allowed us to compare expression levels of innate immune genes induced upon Lm infection of cells representing both sexes. This approach again revealed an incomplete rescue by DDX3Y. It correlated well with the strongly altered gene expression of infected macrophages in vitro. Since macrophages are ubiquitous primary targets of Lm in infected tissues they are major players in shaping the local cytokine milieu and, concomitantly, the activation status of surrounding cells. Consistent with this notion macrophages lacking one or both DDX3 isoforms showed a significant impairment in limiting the growth of intracellular Lm. Activation by IFNγ increased their antibacterial activity, but not to the same level observed in control cells. Thus, reduced cell-autonomous immunity as a consequence of less activating stimuli in the extracellular environment is likely to cause an impairment of mice to reduce the bacterial burden by innate mechanisms. The data in S1 Fig show that the broad effect observed upon DDX3X deficiency particularly in female macrophages did not translate into a similarly widespread impact on serum cytokine levels in infected mice. On the one hand this may result from the fact that male mice had to be used for infection experiments, hence from the weaker impact of DDX3X deficiency in cells expressing DDX3Y. Alternatively, macrophages may not be responsible for building up high levels of serum cytokines.
In summary, loss of DDX3X impacts on the innate immune system both through its role in hematopoiesis and its regulation of the innate response to Listeria infections. It alters lymphoid organ cellularity and the cytokine cocktail needed for cell recruitment and functional activation of innate immune cells.
Animal experiments were carried out at the University of Veterinary Medicine Vienna and have been approved by the institutional ethics and animal welfare committee and the national authority (Austrian Federal Ministry of education, Science and Research) according to §§26ff of Animal Experiments Act (Tierversuchsgesetz TVG 2012, BGBl. I Nr 114/2012) under the permission license numbers BMWF 68.205/0032-WF/II/3b/2014 and BMWFW-68.205/0212-WF/V/3b/2016. Animal husbandry and experimentation was performed under the Austrian national law and the ethics committees of the University of Veterinary Medicine Vienna and according to the guidelines of FELASA which match those of ARRIVE.
Mouse embryonic fibroblast (MEFs) deficient in either TBK1/IKKε (kindly provided by S. Akira, Osaka University, Osaka, Japan, or in DDX3X (derived from Ddx3xfl/y CreERT2 or Ddx3xfl/fl CreERT2 embryos and immortalized by the 3T3 method), as well as HEK293 cells (American Type Culture Collection CRL-3216) were cultured in DMEM (Sigma, distributed by Sigma-Aldrich Handels GmbH, Vienna, Austria) supplemented with 10% FBS (Sigma, distributed by Sigma-Aldrich Handels GmbH, Vienna, Austria) and with penicillin and streptomycin (Sigma, distributed by Sigma-Aldrich Handels GmbH, Vienna, Austria). Deletion of DDX3X was carried out by Tamoxifen treatment (4-OHT; Sigma, distributed by Sigma-Aldrich Handels GmbH, Vienna, Austria) for 48 hours at a final concentration of 500nM.
The plasmids pIE-NHA-hDDX3X, pIE-NHA-hMAVS, pCS2-N-Myc-mTBK1 and pEF-HA-mIRF3 were described previously [4]. The expression vector pIE-NHA-DDX3Y was generated based on the RefSeq ID: NM_012008.2 using the EcoRI and BspeI restriction sites. FLAG-IRF7-M15 [40] was obtained from Isabelle Marié (New York University, NY, USA). The NFκB-reporter plasmid was kindly provided by Ann J. Richmond (Vanderbilt University School of Medicine, Nashville, Tennessee, USA) [55].
Anti-GAPDH (Clone ABS16; Millipore GesmbH, Vienna, Austria) was used in 1:3000 dilution; anti-DDX3X (Clone A300-474A; Bethyl Laboratories Inc., distributed by Sanova Pharma, Vienna, Austria) was used in 1:1000 dilution; anti-HA.11 Epitope Tag (Clone MMS-101P-200; Cambridge Bioscience, Cambridge, UK) was used in 1:1000 dilution; anti-Myc (Clone 9B11; CST, distributed by New England Biolabs, Frankfurt am Main, Germany) was used in 1:1000 dilution; anti-Flag (anti-ESC; Clone A190-101A; Bethyl Laboratories Inc., distributed by Sanova Pharma, Vienna, Austria) was used in 1:1000 dilution.
Viability stains were carried out according to manufacturer's protocol (Fixable Viability Dyes; formerly ebioscience, now Thermo Fisher Scientific, distributed by Fisher Scientific GmbH, Vienna, Austria). Cells were pre-incubated with TruStain fcX (formerly ebioscience, now Thermo Fisher Scientific, distributed by Fisher Scientific Austria GmbH, Vienna, Austria) and stained with appropriate surface antibodies in PBS 2% FCS. For intracellular cytokine stainings, cells were fixed in 2% PFA and permeabilized using a saponin-containing buffer (BD Perm/Wash; Becton Dickinson Austria GmbH, Schwechat, Austria). Anti-CD11b (Clone M1/70), anti-CD3e (Clone 145-2C11), anti-CD8a (Clone 53–6.7), anti-B220 (Clone RA3-6B2), anti-Ly6G (Clone 18A), anti-CD11c (Clone HL3), anti-CD49b (Clone Dx5) anti-CD44 (Clone IM7), anti-Gr1 (Clone RB6-8C5), anti-Ter119 (Clone TER119) and anti-CD122 (Clone 5H4) antibodies were purchased from BD Bioscience (Becton Dickinson Austria GmbH, Schwechat, Austria) and used according to manufacturer's instructions. Anti-CD4 (Clone GK1.5), anti-NK1.1 (Clone PK136), anti-NKp46 (Clone 29A1.4), anti-IFNγ (Clone XMG1.2), anti-F4/80 (Clone BM8), anti-Ly6C (Clone HK1.4), anti-BP1 (Clone 6C3), anti-CD43 (Clone ebioR2/60), anti-IgM (Clone eB121-15F9), anti-CD24 (Clone M1/69), anti-cKit (Clone 2B8), anti-Sca1 (Clone D7) and anti-CD127 (Clone A7R34) were purchased from Thermo Fisher Scientific (formerly ebioscience, now distributed by Fisher Scientific Austria GmbH, Vienna, Austria) and used according to manufacturer's instructions. PBS-57 loaded and PE-conjugated CD1d tetramers were obtained from NIH Tetramer Core facility (NIH Tetramer Core Facility at Emory University, Atlanta, US) and used according to manufacturer's instructions. Fluorochrome labelled Annexin V and 7-AAD were purchased from Thermo Fisher Scientific (formerly ebioscience, now distributed by Fisher Scientific Austria GmbH, Vienna, Austria) and used according to manufacturer's instructions. Data was acquired on BD FACSAria or BD LSRFortessa and analyzed using FlowJo Software.
MEFs were transiently transfected in 6-well plates (3,5x 105 cells/well) using Turbofect (Thermo Fisher Scientific, distributed by Fisher Scientific Austria GmbH, Vienna, Austria) according to manufacturer’s instructions. DNA amount across the samples was equalized by transfecting an empty vector. Bone-marrow derived macrophages were obtained from Ddx3xfl/fl CreERT2 and Ddx3xfl/y CreERT2 mice via flushing the tibia and femur. Macrophages were differentiated in DMEM (10% FBS, pen/strep) containing recombinant M-CSF (a kind gift from L. Ziegler-Heitbrock, Helmholtz Center, Munich, Ger). On day 5, medium was changed to starvation medium (DMEM, 2% FBS, pen/strep, M-CSF) and half of the cells was treated with 4-OHT (Sigma, distributed by Sigma-Aldrich Handels GmbH, Vienna, Austria) for 48 hours at a final concentration of 500nM to induce deletion of Ddx3x. After 48 hours medium was replaced (DMEM, 10% FBS, pen/strep, M-CSF). On day 8–9 cells were seeded for experiments. BMDMs were stimulated with 30 μg/ml poly (I:C) (formerly Amersham Biosciences, now GE Healthcare Services Europe, distributed by Fischer Scientific Austria GmbH, Vienna, Austria) or 10 μg/ml poly (dA:dT) (Sigma, distributed by Sigma-Aldrich Handels GmbH, Vienna, Austria) using Hyperfect (Qiagen, Vienna, Austria) and Polyfect (Qiagen, Vienna, Austria) respectively according to the manufacturer’s instructions. LPS (Escherichia coli 055:B5; Sigma, distributed by Sigma-Aldrich Handels GmbH, Vienna, Austria) was added dropwise to a final concentration of 100ng/ml. Where indicated, cells were treated with IFNβ (PBL, distributed by Enzo Life Sciences, Lörrach, Germany) at a final concentration of 250U/ml or IFNγ (formerly ebioscience, now Thermo Fisher Scientific, distributed by Fisher Scientific Austria GmbH, Vienna, Austria) at a final concentration of 5ng/ml. L. monocytogenes strain LO28 was grown for 16 hours in brain heart infusion medium (Becton Dickinson Austria GmbH, Schwechat, Austria). BMDMs (seeded the day before in antibiotic free DMEM supplemented with 10% FBS) were infected at a multiplicity of infection (MOI) of 10 for 1 h at 37°C. After this, cells were washed with PBS and complete DMEM containing 50 μg/ml gentamicin (MP Biomedicals, Santa Ana, US) was added to kill extracellular bacteria. After another 1 h, BMDMs were washed with PBS again, and medium was replaced with DMEM containing 10 μg/ml gentamicin and left for another 2 hours. After stimulation, BMDMs were washed two times with PBS and used for RNA isolation.
MEFs were transfected with either 50pmol DDY3Y siRNA or with 50pmol non-targeting siRNA (Dharmacon, distributed by THP Medical Products GmbH, Vienna) using Lipofectamine RNAiMAX (Thermo Fisher Scientific, distributed by Fisher Scientific Austria GmbH, Vienna, Austria) under antibiotic-free conditions. After 48 hours, medium was exchanged and the cells transfected for 4 hours with 10μg/ml poly (dA:dT) (Sigma, distributed by Sigma-Aldrich Handels GmbH, Vienna, Austria) using Polyfect (Qiagen, Vienna, Austria) according to the manufacturer’s instructions. Subsequently, cells were washed once with PBS and used for RNA isolation.
Floxed Ddx3x (Ddx3xfl/fl or DDX3Xfl/y) mice were generated via homologous recombination in ES cells. Specifically, loxP sites flanking exon 2 were introduced by gene targeting. Ddx3xfl/fl CreERT2, Ddx3xfl/y CreERT2 and Ddx3xfl/y Vav-iCre mice on a C57BL/6 genetic background were housed under specific pathogen-free conditions according to FELASA guidelines. NK cell depletion of Ddx3xfl/y, Ddx3xfl/y Vav-iCre and C57BL/6 mice was carried out by injecting the anti-Nk1.1 antibody (150 μg/mouse) 3 days prior to infection. Depletion efficiency was confirmed by flow cytometry of blood and splenic leukocytes. For infection experiments an overnight culture of Listeria monocytogenes was recultured in BHI medium to late logarithmic phase, pelleted and diluted in PBS. The concentration of L. monocytogenes was quantified by optical density measurements at 600 nm. The infectious dose was controlled by plating serial dilutions on BHI agar plates and counting the colonies. For infection, bacteria were diluted in PBS and 200 μl were injected into the peritoneum of 8- to 12-week-old mice. For the infection with VSV, mice were infected i.v. with 100 μl of 1x106 plaque forming unit (pfu) of VSV. The progress of the disease was monitored every 2–4 h during the “day phase” (7 a.m. to 7 p.m.) or both during the “day” and the “night phase” depending on the condition of the animals. In survival or terminal stage experiments, humane endpoint by cervical dislocation was conducted if death of the animals was expected within next few hours.
MEFs were infected with VSV at an MOI of 0,1 under FCS-free conditions. After 1 hour of infection, the medium was exchanged to remove unbound virus. After the indicated periods of time, the supernatants were collected. On the next day, L-929 cells (American Type Culture Collection CCL-1) cells were infected by a dilution series prepared from the collected supernatants in DMEM. After 1 hour of infection, the medium was replaced with DMEM containing 0,5% low melting point (LMP) agarose (Thermo Fisher Scientific, distributed by Fisher Scientific Austria GmbH, Vienna, Austria). After 24 hours, crystal violet solution (Sigma, distributed by Sigma-Aldrich Handels GmbH, Vienna, Austria) was added onto the agarose to stain the cells. Subsequently, plaques were counted and the viral titer was determined.
Bone-marrow-derived macrophages were infected with L. monocytogenes strain LO28 as described above. Additionally, BMDMs were stimulated with the combination of heat-killed Listeria (MOI: 50) and IFNβ (250U/mL) for 4hours. RNA was extracted from bone-marrow-derived macrophages using the NucleoSpin RNA II kit (Macherey-Nagel, distributed by VWR International GmbH, Vienna, Austria). RNA samples were poly(A) selected, libraries were prepared using TruSeq RNA Sample Preparation Kit (Illumina Inc., San Diego, US) as per manufacturers instruction, and sequenced (50 bp single-end read) on an Illumina HiSeq 2000. Sequencing data have been deposited to NCBI Gene Expression Omnibus and are accessible through GEO series accession number GSE86591.50-bp single-end Illumina mRNA sequencing reads were aligned to the mm10 reference genome using STAR (v 2.5.0a), and gene-level read counts were obtained using htseq-count (HTSeq v 0.6.1p1) [56]. Heat map of log 2-fold gene expression changes between Listeria-treated and untreated samples were generated for genes differentially expressed upon treatment in WT samples of either sex (padj <0.01, abs(logFC)>1; DeSeq2 v1.18.1). Functional enrichment analysis of the differentially expressed genes using DAVID v6.8- top 10 enriched GO terms ranked by p-value for the category GOTERM_BP_3 are shown [57].
The assay was performed a recently described [39]. Briefly, 5x104 BMDMs were seeded into 96-well plates in DMEM (10% FBS, no antibiotics) and stimulated with IFNγ as indicated. On the next day, an overnight culture of L. monocytogenes strain LO28 was used to infect BMDMs at an MOI of 10. After uptake of the bacteria medium containing gentamicin was added to kill extracellular bacteria. BMDMs were washed two times with PBS followed by lysis in dH2O at the indicated time points. Serial dilutions of the lysates were plated onto brain heart infusion plates followed by incubation for one day at 37°C. Intracellular bacteria were quantified by counting the number of colonies. Data shown are representatives of at least 3 independent experiments.
Proteins were isolated during RNA isolation with NucleoSpin RNA kit (Macherey-Nagel, distributed by VWR International GmbH, Vienna, Austria). In brief, after adjusting RNA binding conditions with 70% ethanol, lysate were centrifuged. At this step nucleic acids are bound to the silica membrane of the column, whereas proteins flow through. After collecting the flow-through containing the proteins, equal amounts of Protein Precipitator (Macherey-Nagel, distributed by VWR International GmbH, Vienna, Austria) was added and samples were left at room temperature for 10 minutes. After this, samples were centrifuged at 11.000g for 15 minutes and supernatants were discarded. Precipitated proteins were washed with 50% ethanol and centrifuged again. Supernatants were discarded, pellets were dryed and 30 μl of 1x Laemmli sample buffer (62,6 mM Tris pH:6,8, 10% Glycerol, 2% SDS, 7,1% β-mercaptoethanol, bromophenol blue) was added to them. The samples were then subjected to western blot analysis as described above. After primary antibody binding the blots were probed with fluorescence-labelled secondary antibodies (formerly Invitrogen, now Thermo Fisher Scientific, distributed by Fisher Scientific Austria GmbH, Vienna, Austria) at a dilution of 1:15000 and detected by the Odyssey infrared imaging system (LI-COR, Lincoln, NE).
For cytokine analysis blood was taken by cardiac puncture and serum was isolated via centrifugation at 10.000g for 1 minute. Cytokine concentrations were determined using the FlowCytomix system (formerly ebioscience, now Thermo Fisher Scientific, distributed by Fisher Scientific Austria GmbH, Vienna, Austria). In some experiments, IFNγ concentrations were measured by ELISA (formerly ebioscience, now Thermo Fisher Scientific, distributed by Fisher Scientific Austria GmbH, Vienna, Austria).
Total RNA was extracted from mouse embryonic fibroblasts and bone-marrow-derived macrophages using the NucleoSpin RNA II kit (Macherey-Nagel, distributed by VWR International GmbH, Vienna, Austria). The cDNAs was prepared using Oligo (dT18) Primer and the RevertAid Reverse Transcriptase (Thermo Fisher Scientific, distributed by Fisher Scientific Austria GmbH, Vienna, Austria). Real-time qPCR experiments were run on an Eppendorf Mastercycler to amplify the Gapdh (housekeeping gene), using SybrGreen (Promega, Mannheim, Germany). Real-time qPCR assays targeting mRNA of Gapdh, Ifnb, Il-1b, Il-6, Il-12p40, Nos2, Tnfα, Irf1, Mx1, Isg15, Ifit3 and Oas2 and pre-mRNA of Il-1b, Il-6, Nos2, Tnfα, Ccl5 and Cxcl10 were performed with the following forward (f) and reverse (r) primers: mRNA: Gapdh-f: 5’-CATGGCCTTCCGTGTTCCTA-3’; Gapdh-r: 5’-GCGGCACGTCAGATCCA-3’; Ifnb-f: 5’-TCAGAATGAGTGGTGGTTGC-3’; Ifnb-r: 5’-GACCTTTCAAATGCAGTAGATTCA-3’; Il-1b-f: 5’-AGATGAAGGGCTGCTTCCAAA-3’; Il-1b-r: 5’-AATGGGAACGTCACACACCA-3’;, Il-6-f: 5’-CTGCAAGAGACTTCCATCCAG-3’, Il-6-r: 5’-AGTGGTATAGACAGGTCTGTTGG-3’; Il-12p40-f: 5’-TGGTTTGCCATCGTTTTGCTG-3’; Il-12p40-r: 5’-ACAGGTGAGGTTCACTGTTTCT-3’; Nos2-f: 5’- GAGCAACTACTGCTGGTGGT-3’; Nos2-r: 5’- CGATGTCATGAGCAAAGGCG-3’, Tnfα-f: 5’- CAAAATTCGAGTGACAAGCCTG-3’; Tnfα-r: 5’- GAGATCCATGCCGTTGCC-3’; Irf1-f:5'-CCG AAG ACC TTA TGA AGC TCT TTG-3' Irf1-r: 5'-GCA AGT ATC CCT TGC CAT CG-3'; Mx1-f: 5'-GAC TAC CAC TGA GAT GAC CCA GC-3', Mx1-r: 5'-ATT TCC TCC CCA AAT GTT TTC A-3'; Isg15-f: 5'-ATG GCC TGG GAC CTA AAG-3'; Isg15-r: 5'-TTA GGC ACA CTG GTC CCC-3'; Ifit3-f: 5'-CCT CGC AGC CCT GGA GTG TT-3'; Ifit3-r: 5'-TGC GTT GCC TCC CAA ACC CC-3'; Oas2-f: 5'-AAACCTCACACCCAACGAAAA-3'; Oas2-r: 5'- CCACCCTTAGCCACTTCCT-3'. pre-mRNA: Il-1b-f: 5’-AAGATGAAGGTGAGACTCTGAG-3’; Il-1b-r: 5’-CTTGGTGTGTGGCTGTGGTA-3’; Il-6-f: 5’-AATGGAGTTGTTAGGCATGGG-3’; Il-6-r: 5’-TGTAAATCTTTTACCTAAAGGAGGA-3’; Nos2-f: 5’-TCCTTTAAAGAGTAAGTCTGGCTT-3’; Nos2-r: 5’-CAGGACTCAGCAGTGACCT-3’; Tnfα-f: 5’-AGGGATGAGGTGAGTGTCTG-3’; Tnfα-r: 5’-ACGTGTGAACACACTTGTTCGT-3’; Ccl5-f: 5’-GCCTCACCATGTAAGTCGAG-3’; Ccl5-r: 5’-CACAGAAAAGTTCCTCAGAGGA-3’; Cxcl10-f: 5’-TGGGACTCAAGGTAAGGGAC-3’; Cxcl10-r: 5’-CTTTCTTCCCTTCTTCGTTCCT-3’.
Luciferase activity was measured with the Dual Luciferase Reporter Assay System (Promega, Mannheim, Germany). In short, the cell lysates were mixed with the substrate for the Firefly luciferase, then luciferase activity was determined with the luminometer. After this, the stop solution ending the first enzymatic reaction and containing the substrate for the Renilla luciferase was added and the samples were measured again. Quantification was performed normalizing the Firefly luciferase activity to the Renilla luciferase activity. Relative induction levels were derived comparing the samples to an empty control vector.
Bacterial loads in spleens and livers were compared using unpaired t-test. The lines represent mean with the standard error of the mean (SEM). Bacterial loads in in vitro CFU assays were compared with the unpaired t-test and bars on the graph represent the mean values with standard deviation (SD). Significances of survival curves after either Lm or VSV infections were calculated using the Mantel-cox-test. Serum cytokine levels were compared using the unpaired t-test where the lines represent means with the standard error of the mean. The mRNA expression data are represented with the mean values with standard deviation (SD). The differences in mRNA expression data were compared using the paired t-test. All statistical analyses were performed using the GraphPad Prism (GraphPad) software. Asterisks denote statistical significance as follows: ns, P > 0.05; *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001.
|
10.1371/journal.pntd.0002752 | Trichuris suis and Oesophagostomum dentatum Show Different Sensitivity and Accumulation of Fenbendazole, Albendazole and Levamisole In Vitro | The single-dose benzimidazoles used against Trichuris trichiura infections in humans are not satisfactory. Likewise, the benzimidazole, fenbendazole, has varied efficacy against Trichuris suis whereas Oesophagostomum dentatum is highly sensitive to the drug. The reasons for low treatment efficacy of Trichuris spp. infections are not known.
We studied the effect of fenbendazole, albendazole and levamisole on the motility of T. suis and O. dentatum and measured concentrations of the parent drug compounds and metabolites of the benzimidazoles within worms in vitro. The motility and concentrations of drug compounds within worms were compared between species and the maximum specific binding capacity (Bmax) of T. suis and O. dentatum towards the benzimidazoles was estimated. Comparisons of drug uptake in living and killed worms were made for both species.
The motility of T. suis was generally less decreased than the motility of O. dentatum when incubated in benzimidazoles, but was more decreased when incubated in levamisole. The Bmax were significantly lower for T. suis (106.6, and 612.7 pmol/mg dry worm tissue) than O. dentatum (395.2, 958.1 pmol/mg dry worm tissue) when incubated for 72 hours in fenbendazole and albendazole respectively. The total drug concentrations (pmol/mg dry worm tissue) were significantly lower within T. suis than O. dentatum whether killed or alive when incubated in all tested drugs (except in living worms exposed to fenbendazole). Relatively high proportions of the anthelmintic inactive metabolite fenbendazole sulphone was measured within T. suis (6–17.2%) as compared to O. dentatum (0.8–0.9%).
The general lower sensitivity of T. suis towards BZs in vitro seems to be related to a lower drug uptake. Furthermore, the relatively high occurrence of fenbendazole sulphone suggests a higher detoxifying capacity of T. suis as compared to O. dentatum.
| The human whipworm Trichuris trichiura is together with the roundworm Ascaris lumbricoides and the hookworms Ancylostoma duodenale and Necator Americanus the most common intestinal worms worldwide. Together they place more than 5 billion people at risk of infection. The current global control strategy against these worms is regular administration of anthelmintic drugs, mostly albendazole and mebendazole, both belonging to the drug-class benzimidazoles. Both drugs have a low effect against T. trichiura infections, but the reasons for this are not known. We evaluated the in vitro effect of two benzimidazoles; i.e., albendazole, fenbendazole, and another type of anthelmintic, levamisole, on the whipworm (T. suis) and the nodular worm (Oesophagostomum dentatum) of the pig. Oesophagostomum dentatum is highly sensitive towards benzimidazoles in comparison to T. suis. We measured and compared the drug uptake in both species in both living and killed worms. Our results suggest that the reason for the difference in sensitivity is due to a lower drug uptake into T. suis as compared to O. dentatum. Furthermore, T. suis was able to metabolise fenbendazole into an inactive metabolite to a much larger extent than O. dentatum, suggesting a higher detoxifying capacity of T. suis as compared to O. dentatum.
| The whipworm Trichuris trichiura has been estimated to infect 600 million people worldwide resulting in an estimated 1.6–6.4 million disability adjusted life-years lost globally [1]. The current control strategy against T. trichiura and other soil-transmitted helminths (STHs) is administration of single-dose anthelmintic drugs [1], [2]. The benzimidazoles (BZs) i.e. albendazole (ALB) and mebendazole (MBD) are widely used in large-scale control programs where they are administered regularly, at a dosage of 400 mg (ALB) or 500 mg (MBD) [2]. However, the efficacy of single-dose BZ against T. trichiura is not satisfactory. A meta-analysis of 20 randomized, placebo-controlled trials reported an average cure rate (CR) of 28% for ALB (400 mg) and 36% for MBD (500 mg) [3]. Other randomized controlled trials have reported similar low CR and egg reduction rates (ERR) ranging from 31.5–40.3% (CR) and 9.8–54.0% (ERR) for ALB and 22.9–66.7% (CR) and 18.8–81.0% (ERR) for MBD [4]–[7].
The use of the T. muris-mouse model for estimating drug efficacy on T. trichiura is well established [8]–[11]. Trichuris suis is regarded a different but closely related species to T. trichiura [12], [13], hence, T. suis can be considered a valid model for T. trichiura. Another BZ, fenbendazole (FBZ) has shown poor efficacy against T. suis infection in pigs when administered as a single-dose [14], therefore the T. suis-pig model and FBZ may be considered an interesting alternative for studying low treatment efficacy of Trichuris spp. In one controlled trial an oral dose as high as 15 mg/kg, three times the recommended dose of 5 mg/kg for other pig nematodes, was required to obtain a worm count reduction (WCR) of 96.7% [14]. In another controlled study the same oral dose resulted in only a 65.1% reduction in worm burden and a dose of 30 mg/kg resulted in an efficacy of 96.6% [15]. Multiple doses of FBZ (3 mg/kg per day for 3 consecutive days) have shown varied efficacy against T. suis in controlled tests ranging from 66% [16] to 99.8% [14], [17] in WCR. The current recommendation for treatment of T. suis infections in pigs with FBZ is either a single dose of 25 mg/kg, or a long-term treatment where the recommended therapeutic dose is distributed over 7 days [18], [19].
Another nematode of the pig is the nodular worm, Oesophagostomum dentatum which in the adult stage, opposed to T. suis, is highly sensitive to FBZ. An oral dose level as low as 0.25 mg/kg has shown an efficacy of 99.9% and doses of 1, 2.5 and 3.5 mg/kg FBZ have resulted in efficacies of 100% in controlled tests based on worm counts [20], [21]. Trichuris suis and O. dentatum both inhabit the lower part of the intestine namely the caecum and the colon [22]–[25], but in their adult stage, their microhabitat varies significantly. The thin anterior part of T. suis is embedded in the mucosa creating a tunnel-like construction of epithelial cells whereas the thicker posterior part of the body is protruding freely into the lumen [26]. In contrast to T. suis, the adult stage of O. dentatum is not attached to the mucosa but roams freely in the intestinal lumen [27], [28].
Levamisole (LEV), belonging to another class of anthelmintics, the imidazothiazoles, was introduced in 1968 [29] and has like BZs been used against parasitic infections in both animals and humans. In order for BZs and imidothiazoles to exert their pharmacological effect, they need to reach their specific receptors within the target parasites i.e. BZs bind to beta-tubulin [30] and the imidazothiazoles to acetylcholine-gated channels [29], [31]. Passive diffusion through the external surface has been proposed as the main pathway of BZs (i.e. FBZ, oxfendazole (OXF) and triclabendazole sulphoxide (TCBZSO)) in the three main classes of helminth parasites represented by: Moniezia benedeni (cestode), Fasciola hepatica (trematode) and Ascaris suum (nematode) [32]. The uptake of LEV has likewise been demonstrated to occur via a transcuticular mechanism in A. suum, but was observed to take place in four distinct stages, thus suggesting a non-passive up-take mechanism [33]. Once inside an organism, drugs are generally being metabolised. However, our knowledge of the metabolism of anthelmintics in helminths is very limited, although drug metabolising enzymes are well described in mammals and serve as an efficient defense mechanism against potential harmful substances. In brief drugs are (if not excreted unchanged) biotransformed by unique enzymes into more polar compounds that are easier to excrete by the organism in metabolic reactions named phase I-III. In mammals the major phase I reaction is oxidation catalysed by cytochrome P450 superfamily (CYPs) [34]. For many years attempts to detect CYPs in parasitic nematodes were unsuccessful [35] but with the discovery of 75 predicted CYP genes in the free-living nematode Caenorhabditis elegans as well as genomic and transcriptomic-based predictions of proteins produced by helminths, the knowledge has improved [36]. The ability of parasitic helminths to metabolise anthelmintics may serve as an advantageous defence mechanism. Previously, the first step of phase I oxidation of ALB into albendazole sulphoxide (ALBSO) (sulphoxidation) has been reported for F. hepatica, M. expansa, A. suum [37], Dicrocoelium dendriticum [38] and Haemonchus contortus [39]. This metabolite has a lower pharmacological activity than the parent compound [40] and lower effect on nematode motility [41]. The second step of ALB oxidation (sulphonation) into albendazole sulphone (ALBSO2) was reported for D. dendriticum [38]. A similar sulphonation process has been reported for F. hepatica exposed to triclabendazole sulphoxide (TCBZSO) in vitro [42]. To the best of our knowledge no studies has been conducted on the metabolism of FBZ within parasitic nematodes. Comparative in vitro studies of the oxidative metabolism of FBZ by hepatic microsomal fractions from a variety of vertebrate species showed that all species readily produced the sulphoxide metabolite ( = oxfendazole, OXF) and the sulphone metabolite fenbendazole sulphone (FBZSO2) [43]. Oxfendazole is a widely used anthelmintic whereas FBZSO2, similar to ALBSO2, are considered pharmacological inactive [40], [44].
We find the different sensitivity of T. suis and O. dentatum to FBZ in vivo highly interesting because these two species are located in the same compartment of the intestine and thus theoretically exposed to similar concentrations of drugs. We speculate that the difference in sensitivity may be related to differences in uptake and/or metabolism of the drug inside the worms. We hypothesized that the reason for a low or variable treatment efficacy of T. suis infections may be due to a lower drug uptake and/or a higher drug metabolism of T. suis in comparison to O. dentatum. The aim of this study was therefore to examine the motility of T. suis and O. dentatum adult worms in vitro when exposed to FBZ, ALB and LEV and to assess whether these drugs accumulate in the same concentrations within the two species.
Fenbendazole, ALB and LEV were purchased from Sigma-Aldrich (Schnelldorf, Germany), and stock solutions of the drugs (100.000 µM) were prepared in 100% dimethylsulfoxid (DMSO) (Sigma-Aldrich, Schnelldorf, Germany) and stored at 5°C until use within 1 week.
Fourteen pigs were purchased and acclimatized for 1 week prior to experimental infection. The animals had free access to water and were fed restrictively, according to national feeding requirements. For the FBZ in vitro assay, six pigs were orally infected by stomach tube with 2,000 embryonated T. suis eggs (kindly provided by Parasite Technologies A/S, Hørsholm, DK) and two pigs with 5,000 L3 O. dentatum larvae (CEP-strain). The CEP-strain was originally isolated from a farm with no prior use of anthelmintics according to the owner [45], and was later characterized as FBZ susceptible [21]. The T. suis isolate has been used in an in vivo study where experimentally infected pigs were exposed to repeated administration of FBZ (i.e. 5 mg/kg given orally on three consecutive days). Worm count reductions of 51.5 and 98.5% were obtained 24 hours after single and triple dose treatments, respectively; therefore, this isolate was considered FBZ susceptible. For the ALB and LEV in vitro assay, 3 pigs were infected with 5,000 embryonated T. suis eggs and 3 pigs with 4,000 L3 O. dentatum larvae (same strains as above). Due to practicalities the experimental infections for ALB and LEV were performed after the FBZ assay. Patency of infections was confirmed by faecal egg count (EPG) using the modified McMaster technique [46].
The current study was approved by the Experimental Animal Unit, University of Copenhagen, (Denmark) based on national regulations from the Danish Animal Experiments Inspectorate (permission no. 2010/561-1914, C5).
For the FBZ in vitro assay, the O. dentatum infected pigs were euthanized at day 40 post infection (p.i.) and the T. suis infected pigs at day 63 p.i. For the ALB and LEV in vitro assay the O. dentatum and the T. suis infected pigs were euthanized at day 28 and 49 days p.i., respectively. Adult O. dentatum were isolated from the intestinal content according to Slotved et al. [47] and adult T. suis were collected from the intestine by manual plucking. Both parasite species were washed following a common washing procedure which consisted of 4 consecutive washing steps (each 15 min. in 39°C Hanks Balanced Salt Solution (HBSS)) followed by 4 consecutive washing steps (each 60 min. in 39°C RPMI-1640 medium). Both the HBSS and RPMI-1640 media were supplemented with 1% (v/v) amphotericin B-penicillin-streptomycin solution (10,000 U/ml penicillin, 10,000 µg/ml streptomycin, 25 µg/ml amphotericin B) and 0.5% (v/v) gentamicin (10 mg/ml) (All media, antibiotics and anti-mycotic were purchased from Life Technologies, Naerum, DK).
Since FBZ concentrations above 30 µM precipitated during incubation, we tested the following concentrations of FBZ and ALB: 0.01, 0.1, 1, 10 and 30 µM.
Final concentrations of LEV included 0.01, 0.1, 1, 10 and 200 µM. All dilutions contained DMSO (2% v/v) and were made in RPMI-1640 medium supplemented with antibiotics and fungicide as described for the washing procedure. Thirty worms of each species selected at random were placed in a large petri dish (Th. Geyer, Roskilde, DK) containing 40 ml of each of the dilutions described above. Each concentration was tested in triplicate, thus for each drug and each concentration a total of 90 worms were used. Worms incubated in RPMI-1640 with DMSO 2% (v/v) without anthelmintics served as controls. All worms were incubated at 39°C (5% CO2, 21% O2, 90% relative humidity) for 24 or 72 hours. In the motility assay, 21 worms (i.e. 7 worms from each petri dish) of both species were scored by stereomicroscope at 6.3× magnification according to motility grades specific for each species. The motility of T. suis was graded as follows: 3: normal motility (movement of the whole body), 2: low motility (slower movement of the whole body), 1: very low motility (movement of the anterior part only), 0: no movements. The motility of O. dentatum was graded as follows: 3: normal motility (swimming), 2: low motility (slow swimming or jerking movements), 1: very low motility (only movement of the anterior tip of the body), 0: no movements. All motility measurements were blinded except for worms incubated in FBZ, due to lack of resources.
In order to compare the accumulation of drugs in living and killed worms, a number of worms obtained after the common washing procedure was killed by freezing (liquid nitrogen for 1 min.) and thawed at 5°C. Thirty living and 30 killed worms of each species were then incubated for 24 hours in FBZ, ALB or LEV at a final concentration of 10 µM in RPMI-1640 medium with DMSO (2% v/v) using the same conditions as described above. All incubations were performed in triplicates.
After motility measurements and the 24 hour incubation period of living and killed T. suis and O. dentatum, all worms were carefully rinsed in 50 ml HBSS for a maximum of 30 sec. The in vitro assay with FBZ was conducted first, and since the drug concentration within worms was unknown, all worms from each incubation concentration were pooled into one sample to ensure a detectable drug level. Subsequently, triplicates were made for worms incubated in each of the five concentrations of ALB and LEV. After rinsing, worms were transferred to pre-weighed Eppendorf vials, frozen in liquid nitrogen and kept at −20°C until HPLC-analysis.
Vials with worms were thawed and dried under phosphorous pentoxide until constant weight. Each vial with dried worm (10–50 mg) was mixed with 200 µl 0,05M phosphate buffer (pH 7.4) with internal standard (see below). After gentle homogenization with a plastic pestle another 200 µl buffer was added and the homogenization repeated before addition of 400 µl 6M guanidine HCl. The sample was vortexed for 1 minute and left at 20°C for 15 minutes before centrifugation at 8000× g for 10 minutes. The supernatant was transferred to a clean tube and an additional 400 µl of 6M guanidine HCl was added to the sample residue. The procedure was repeated and the two supernatants were pooled and loaded on an activated cartridge (Oasis HLB, 60 mg, 3 mL). The cartridge was activated with 2 mL methanol (100%) followed by 2 mL of water. The loaded cartridge was washed with 2 mL 5% methanol and dried under vacuum for 1 minute, before eluting the analyte with 2 mL methanol. The eluate was evaporated under air at 37°C and the residuum was dissolved in 100 µL 50% methanol and centrifuged at 8000× g before 50 µL were injected into the HPLC-system. Standards in phosphate buffer and guanidine HCl were run in parallel. Concentration of analyte in worms was expressed as µg per g dry worm.
The HPLC system was equipped with an autosampler, 2 HPLC pumps, and a UV detector. HPLC conditions for FBZ, ABZ and LEV are described below:
All motility scores were normalized into percentages relative to controls within species. For each drug the effect of all factors (species, time and log_concentration) and biological meaningful interactions between the factors were tested for statistical significance (P<0.05) using Analysis of Covariance (ANCOVA) with variance heterogeneity using SAS version 9.3 and JMP version 8 (SAS Institute, Cary, North Carolina). Due to significant effects of time, the effect of drug concentrations in the media on the relative motility of the two species was then calculated for 24 and 72 hours separately. Variance heterogeneity was used since the variances between the species were different. Total drug concentrations (parent compound and its metabolites) in living and killed worms of each species were compared using Student's t-test with variance heterogeneity (JMP version 8). Drug concentrations in worms exposed to 5 concentrations of FBZ and ALB were compared using the model ‘One site fit total and nonspecific binding’ (GraphPad Prism 5, GraphPad Software, San Diego, California) which calculates the parameter estimates Kd and Bmax by the following equation: Y = Bmax*X/(Kd+X)+NS*X+background. X and Y are drug concentrations in media and worms, respectively. Kd is the concentration of a ligand which is needed in order to achieve half-maximum binding at equilibrium. Bmax is the maximum specific binding, thus giving the maximum binding capacity of an object or organism. NS is the slope of non-specific binding. Background and NS was constrained to 0 since no binding was observed when measuring the negative controls. The difference of Kd and Bmax between the species was evaluated on a significance level of α = 0.05. Drug concentrations in worms exposed to LEV were compared using Student's t-test (JMP version 8) because only the two highest concentrations yielded detectable levels within the worms. Thus, concentration difference between and within species was evaluated when worms were exposed to 10 and 200 µM LEV respectively. For each drug, all data sets were tested for normality.
The relative motility of T. suis and O. dentatum after exposure to FBZ, ALB and LEV for 24 and 72 hours are presented in Fig. 1. No significant difference in motility between species was observed with increasing concentration over time for FBZ, ALB or LEV (species*time*log_concentration). The motility of T. suis was found to be less affected by time (24 vs. 72 h) than O. dentatum when exposed to FBZ (P = 0.015) and ALB (P<0.0001), but not LEV (species*time). The motility of T. suis was significantly less affected than that of O. dentatum after 24 hours incubation in FBZ (P = 0.003) but not 72 hours (P = 0.73) (species*log_concentration). Although the interaction was not significant after 72 hours, the motility of T. suis was still significantly less affected than the motility of O. dentatum (P<0.0001) (species) and the increasing concentration of FBZ resulted in a significant motility decrease for both species (P = 0.012) (log_concentration). When exposed to increasing concentrations of ALB, the motility of T. suis was less affected than O. dentatum after both 24 hours (P = 0.003) and 72 hours (P<0.0001) (species*log_conc). The opposite was observed for increasing concentrations of LEV where the motility of T. suis was reduced more than O. dentatum after 24 (P<0.007) and 72 hours (P<0.007) (species*log_conc).
The mean concentrations of the parent compounds FBZ, ALB and LEV and the metabolites of FBZ (OXF, FBZSO2) and ALB (ALBSO, ALBSO2) in living and killed worms after incubation in 10 µM of the drug for 24 hours are shown in Fig. 2. In general, the total drug concentrations within both living and killed worm species varied according to type of drug (Fig. 2a, 2b, 2c), with ALB and its metabolite ALBSO occurring at the highest concentration level followed by FBZ and its metabolites and LEV. When incubated in ALB and LEV, the total drug concentrations were found to be significantly lower in T. suis than O. dentatum and this was observed for both living (ALB: P = 0.02, LEV: P = 0.02) and killed (ALB: P = 0.002, LEV: P = 0.008) worms. In both living and dead worms, the total concentration of FBZ and its metabolites was found to be lower in T. suis than O. dentatum. For the dead worms, the difference was significant (P = 0.004) but did not reach significance for living worms (131.1±17.1 pmol/mg dry worm tissue vs. 155.8±33.3 pmol/mg dry worm tissue for T. suis and O. dentatum, respectively).
For O. dentatum the concentration of drug was higher in killed worms as compared to living worms for all three anthelmintics, and the difference was found to be significant when incubated in FBZ (P = 0.006) and ALB (P = 0.011). For T. suis no difference between the living and the killed was observed when incubated in FBZ, whereas the anthelmintic concentration was significantly higher within killed worms when incubated in ALB (P = 0.009) and significantly lower when incubated in LEV (P<0.001). The mean concentrations of OXF in living and killed worms, respectively, were found to be 3.4 and 3.5 pmol/mg dry worm tissue for T. suis and 2.6 and 14.4 pmol/mg dry worm tissue for O. dentatum. The pharmacological inactive metabolite FBZSO2 (mean: 12.7 pmol/mg dry worm tissue) was only observed in living T. suis and amounted 9.7% of the total anthelmintic concentration measured within the worms. The mean concentrations of ALBSO in living and killed worms were 93.8 and 71.9 pmol/mg dry worm tissue, respectively, for T. suis and 133.8 and 124.4 pmol/mg dry worm tissue for O. dentatum. Only trace amount of ALBSO2 (4.71 pmol/mg dry worm tissue) were measured in killed O. dentatum.
The concentration of FBZ and ALB inside living T. suis and O. dentatum after incubation in 0.01, 0.1, 1, 10 and 30 µM of FBZ and ALB for 24 and 72 hours is shown in Fig. 3. The Kd and Bmax values for each species at 24 and 72 hours are given in Table 1. For both anthelmintic drugs no significant difference in the Kd – values were observed between the species neither after 24 or 72 hours of incubation. The Bmax – values were similar for the two species after 24 hours exposure to both BZs, but after 72 hours incubation, these were significantly lower for T. suis than O. dentatum when exposed to FBZ (P<0.0001) and ALB (P = 0.033). The concentrations of LEV found within the worms after exposure to 0.01, 0.1, 1, 10 and 200 µM LEV for 24 and 72 hours were only above the detection limit when exposed to the two highest concentrations (Fig. 4). The concentrations of LEV found within the worms were significantly lower in T. suis than O. dentatum when incubated in 10 and 200 µM for 24 hours (P = 0.01, P = 0.0009). When incubated in 200 µM for 72 hours the concentration of LEV was higher in T. suis (452.5 ng/mg dried worm tissue) than in O. dentatum (187.9 ng/mg dried worm tissue) (P<0.0001). The concentration of LEV within T. suis thus increased significantly with incubation time (P<0.0001) when incubated in 200 µM LEV, whereas the concentration was lower after 72 hours than 24 hours incubation within O. dentatum (P = 0.02).
The concentrations of the metabolites OXF, FBZSO2 and ALBSO measured within living T. suis and O. dentatum are given in Fig. 5. The concentrations of OXF and FBZSO2 within the two worm species were much lower than ALBSO (Fig. 5). Incubation concentrations below 0.1 µM of FBZ and ALB did not result in detectable levels of metabolites. The concentration of OXF within T. suis did not show a concentration or time dependent increase (3.2–5.4 pmol/mg dry worm tissue and 3.8–5.4 pmol/mg dry worm tissue after incubation periods of 24 and 72 hours, respectively) whereas a clear time dependent increase was observed for O. dentatum (5.4–7.9 pmol/mg dry worm tissue and 14.2–15.6 pmol/mg dry worm tissue after 24 and 72 hours, respectively). After 24 hours incubation the inactive metabolite FBZSO2 was only detected in T. suis. Results were inconsistent and are thus not given. After 72 hours incubation, FBZSO2 was detected within T. suis at an incubation concentration as low as 0.1 µM FBZ whereas FBZSO2 only appeared in O. dentatum when incubated in 10 and 30 µM. After 72 hours a concentration dependent formation of FBZSO2 (0.9–17.5 pmol/mg dry worm tissue) was measured within T. suis where it represented between 6–17.2% of the total drug concentration whereas in O. dentatum it only constituted 0.8–0.9%. For both species, the formation of FBZSO2 appeared to be both time- and concentration-dependent as consistent results only were obtained after 72 hours incubation. The ALBSO metabolite showed a clear tendency to reach a higher concentration within O. dentatum than T. suis when incubated for both 24 and 72 hours. The formation of ALBSO within the worms appeared to be both time- and concentration-dependent at incubation concentrations ranging from 0.1 µM to 30 µM. Incubation in 30 µM ALB resulted in ALBSO concentrations equal to or below the concentrations formed when incubated in 10 µM. The metabolite ALBSO2 was not detected within any of the two species. The metabolites OXF and ALBSO showed a clear tendency to reach a higher concentration level within O. dentatum than T. suis when incubated for both 24 and 72 hours, but in relation to the total drug concentration, the average proportion of the metabolites were approximately the same (OXF: T. suis; 4% at 24 hours and 3.6% at 72 hours; O. dentatum: 5.6% and 4%, ALBSO: T. suis; 11.1% and 13.8%, O. dentatum; 15% and 12.2%).
In the present work, we have combined worm motility with concentration measurements of drug-uptake and drug metabolism in two nematode species that inhabit the same part of the large intestine, but differ significantly in their intestinal microhabitat. Our results show that the motility of T. suis was less affected than the motility of O. dentatum when exposed to FBZ for 24 hours and ALB for 72 hours, thus indicating a lower sensitivity of T. suis as compared to O. dentatum towards these compounds. The maximum binding capacity of FBZ and ALB was significantly lower for T. suis than O. dentatum after 72 hours incubation and the total drug concentrations were significantly lower in living and killed T. suis as compared to O. dentatum when incubated in ALB. When living and killed worms were incubated in FBZ, only killed T. suis contained a significantly lower drug concentration than O. dentatum. However, collectively these results suggest T. suis to have a lower uptake of FBZ and ALB than O. dentatum. Furthermore, a relatively higher concentration of FBZSO2 was measured in T. suis than O. dentatum, thus suggesting a higher metabolism of FBZ (or OXF) into FBZSO2 in T. suis. Fenbendazole sulphone is considered anthelmintic inactive due to weak ovicidal activity and lack of inhibition of mammalian tubulin polymerization [44]. The equivalent sulphone metabolite of ALB, ALBSO2, has not only shown complete loss of activity in both egg hatch inhibition assays and inhibition of mammalian tubulin polymerization but also decreased binding affinity to nematode tubulin [40]. Whether the latter also applies for FBZSO2 is not known but due to lack of polymerization inhibition, low ovicidal activity and assumed decreased binding affinity to nematode tubulin, FBZSO2 will in the following be considered “inactive”. However, caution must be taken. Due to uncertainty of detection levels within worms in the first trial, triplicates were not made for T. suis and O. dentatum incubated at different drug levels of FBZ (i.e. 0.01–30 µM). Although triplicates were not obtained, concentration agreement was found within the living worms incubated in 10 µM FBZ in the assay of living and killed worms. Furthermore, the formation of FBZSO2 showed a dose dependent formation.
We found that the motility of T. suis as compared to O. dentatum was less affected by increasing concentrations of FBZ and ALB. A low sensitivity to high concentrations of ALB has also been described for T. muris where doses up to 200 µg/ml (equivalent to 754 µM) of ALB were tested against adult and L3 stages of T. muris in vitro [9]. This dose level, which is approximately 25 times higher than the highest concentration used in our study (30 µM) did not reduce the motility of T. muris by 50% (IC50) after an incubation period of 72 hours. In contrast to T. suis, O. dentatum was found to be more sensitive to increasing concentrations of FBZ and ALB when incubated for 24 and 72 hours respectively. The high sensitivity towards increasing concentrations of ALB and FBZ has also been reported by Petersen et al. [41] who found that a concentration of 0.1 µM was able to inhibit migration of O. dentatum through a mesh by 61% for ALB and 69% for FBZ. An increase in concentration to only 3.16 µM increased the inhibition of migration to 75.3% for ALB and 76.2% for FBZ. The high sensitivity towards increasing concentrations of ALB and FBZ reported by Petersen et al. [41], is in agreement with our results in vitro, but more importantly, it is also in concordance with the high efficacy of FBZ against O. dentatum reported in vivo [20], [21]. Likewise, low sensitivity of T. muris towards ALB in vitro has also been shown to correlate with low treatment efficacy in vivo [9]. Trichuris suis was more sensitive towards increasing concentrations of LEV than O. dentatum. At the highest dose (200 µM) no movement of T. suis was observed neither after 24 or 72 hours incubation. A high sensitivity towards LEV has also been observed for T. muris in vitro (IC50 = 33.1 µg/ml equivalent to 68.5 µM) and in vivo where the worm burden was reduced by 95.9% with a single oral dose of LEV (200 mg/kg) in mice [9]. In pigs, the efficacy of a single oral dose of LEV (7.5–8 mg/kg) has shown varying efficacy on T. suis ranging from 26% [16] to 100% [48], [49].
In the in vitro assay with living and killed worms we found that the total concentrations of anthelmintic drugs were lower in T. suis than O. dentatum (Fig. 2). This applied to all three anthelmintics tested, although the difference was not found to be significant when living parasites were incubated in FBZ (Fig. 2). Incubation in increasing concentrations of FBZ and ALB, ranging from 0.01 to 30 µM for 72 hours revealed similar Kd values for T. suis and O. dentatum which suggests that approximately the same concentrations of FBZ and ALB are needed for both species in order to achieve binding of half of the binding sites at equilibrium. The Bmax values were significantly lower for T. suis than O. dentatum suggesting that T. suis has a significantly lower binding capacity of FBZ and ALB than O. dentatum (Fig. 3, Table 1) which is in accordance with lower effect of these two anthelmintics on motility. The Bmax values measured in O. dentatum were higher after 72 hours than 24 hours incubation. The accumulation of FBZ and ALB may be due to a lower secretion capacity of O. dentatum, in comparison to T. suis, which is supported by the formation of FBZSO2 in T. suis. The concentration of LEV within living worms were below the detection level of the HPLC analysis when incubated in 0.01, 0.1, and 1 µM, but interestingly the concentration of LEV within T. suis was more than two times higher than in O. dentatum when incubated in 200 µM LEV for 72 hours, which was translated into an absence of motor activity in the motility assay.
In the in vitro assay of living and killed worms we found that only living T. suis were able to metabolize FBZ, or possibly OXF, to the inactive metabolite FBZSO2 (Fig. 2), amounting 9.7% of the total anthelmintic concentration measured within the worms. When incubating the worms in increasing concentrations of FBZ for 24 hours we obtained inconsistent results for FBZSO2 (i.e. FBZSO2 was only detected in T. suis, and only when incubated in 1 µM FBZ) (data not shown). After 72 hours a concentration dependent formation of FBZSO2 was measured within T. suis where it represented between 6–17.2% of the total drug concentration whereas in O. dentatum it only constituted 0.8–0.9%. In relation to the maximum binding of FBZ, we measured a significantly lower value for T. suis than O. dentatum (Fig. 3 and Table 1). We therefore suggest that the poor effect of FBZ on T. suis may be related to a lower drug uptake and/or a higher detoxifying capacity of this species, however, some care should be taken with the latter. Albendazole and FBZ are able to undergo spontaneous oxidation to their corresponding derivatives ALBSO and OXF when mixed with DMSO [50]. The average proportions of the metabolites OXF and ALBSO were approximately the same within T. suis and O. dentatum when incubated in increasing concentrations of ALB and FBZ. Furthermore, these metabolites occurred in killed worms of both species and even trace amounts of ALBZSO2 were detected in killed O. dentatum. Therefore these findings indicate that OXF and ALBSO were formed by spontaneous oxidation, and that the formation of FBZSO2 observed in T. suis may be related to the presence and further transformation of OXF. As FBZSO2 were not detected in any of the killed worms or in living O. dentatum when incubated in 10 µM FBZ for 24 hours, it is most likely that the relative high concentrations of FBZSO2 measured in T. suis were not formed by spontaneous oxidation, but by T. suis itself. A trace amount of ALBSO2 (4.71 pmol/mg dry worm tissue) was measured in killed O. dentatum when incubated for 24 hours in 10 µM ALB but was not detected in any of the two species when incubated in increasing concentrations of ALB or in dead T. suis. Therefore it is most likely that occurrence of this compound is a detection uncertainty, which needs to be confirmed in future studies.
The above mentioned findings raise the following questions: a) why is the total drug concentrations of BZs generally lower in T. suis than O. dentatum? b) Why is the difference between concentration of anthelmintic within living and killed worms more pronounced for O. dentatum than T. suis? Considering the first question, possible entry routes of anthelmintic drugs into parasitic nematodes are oral ingestion or passive or active transport across the cuticle. In a study performed by Ho et al. [51], transport across the cuticle was demonstrated to be the main route of entry of lipophilic compounds (hydrocortisone and p-nitrophenol) into the nematode A. suum [51]. This route was confirmed by Mottier et al. [32] who also suggested that as a general rule helminths uptake BZs by passive diffusion [32]. Since previous work indicated that passive diffusion across the cuticle is the main route of uptake of lipophilic anthelmintics, and a transcuticular route also has been shown for the water soluble anthelmintic LEV [33], we therefore assumed that this also was the case for T. suis and O. dentatum. Oral ingestion of anthelmintic was controlled in the present study by killing the worms, but the concentration of all three anthelmintics was lower in T. suis than O. dentatum whether killed or alive, with the exception of living worms exposed to FBZ (Fig. 2). Furthermore, the binding capacity of T. suis was significantly lower than the binding capacity of O. dentatum when exposed to both FBZ (P<0.0001) and ALB (P = 0.033). The average proportions of the metabolites OXF and ALBSO were approximately the same for both species, whereas concentration levels above 5 pmol/mg dry tissue of FBZSO2 were only detected in T. suis. We therefore speculate that the lower total drug concentration of BZs measured both in living (i.e. Bmax values after 72 hours incubation in ALB and FBZ) and killed T. suis may be due to structural differences in the cuticle or different lipid contents. Considering the second question regarding the different concentration of anthelmintic within living and killed worms, Mottier et al. [32] found that the concentration of FBZ was lower within living A. suum as compared to killed worms. These findings correspond to our observation for O. dentatum exposed to all three anthelmintics, although the difference was not significant when the worms were incubated in LEV (P = 0.09). For T. suis, a significantly lower concentration within living worms in relation to the killed, was only observed when exposed to ALB. The rate of drug diffusion across the cuticle of A. suum and other nematodes is restricted by the lipid barrier in the hypodermis, the pKa of the drug, the pH of the aqueous environment within the cuticula and the negatively charged aqueous filled pores within the collagen matrix [52]. Mottier et al. [32] suggested that the lower concentration within living worms is related to the acidic environment at the nematode surface that is created by excretion of acidic organic metabolites from the worms [53]. Benzimidazoles are weak bases [54] and may therefore largely exist in their ionized form in the acidic environment at the nematode surface. The ionized form is not readily diffusible through the lipid layer of the cuticle therefore a smaller amount of BZs may enter the living parasites compared to the killed. This mechanism may be the reason why we observed a lower concentration of anthelmintic in living O. dentatum, and to a lesser extent in living T. suis, compared to the killed specimens.Nevertheless, damage of the cuticle due to freezing and a subsequent increase in permeability or possibly higher drug concentrations trapped in the cuticle of killed worms cannot be ruled out. Furthermore, inactivation of possible ATP-dependent efflux pumps i.e. the ATP-binding cassette (ABC) transporter P-glycoprotein (Pgp) [34], [55] may also contribute to the increased drug concentration observed within the killed worms. Interestingly, we did not observe the same for T. suis when exposed to FBZ and LEV which further supports our hypothesis that the lower drug concentration measured within this species is also related to a lower drug uptake.
An answer to the intriguing question for low to varied treatment efficacy of T. trichiura infections in humans has been sought from a variety of angles. The majority of these has taken an empiric approach by evaluating the effect of different treatment strategies in clinical trials such as: a) comparing the efficacy of single-dose BZs treatment (i.e. ALB (400 mg) and MBD (500 mg)) with the efficacy of combination therapy (i.e. BZs in combination with LEV (40 or 80 mg), ivermectin (200 µg/kg) or diethylcarbamazine (150 mg) [4], [5], b) comparing the efficacy of single-doses with triple-doses of ALB and MBD [6] or c) comparing the efficacy of single and double doses of ALB and MBD given alone or in combination [56]. In the above-mentioned clinical trials the highest CR (70.7%) was obtained using 3×500 mg MBD given over 3 consecutive days [6]. Empiric approaches have also been performed using T. muris as a model where the effect of single-drugs (i.e. monepantel, ALB, LEV, pyrantel pamoate and oxantel pamoate) and drug combinations between ALB, LEV, MBD, pyrantel pamoate, oxantel pamoate and ivermectin (IVM) have been assessed in both in vitro assays and in vivo studies [9], [57], [58]. Albendazole, given as a single-drug, showed poor effect in vivo (600 mg/kg) and low efficacy in vitro (50–200 µg/ml) [9], whereas the combinations of ALB-MBD, MBD-IVM, MBD-LEV and oxantel pamoate-MBD revealed a strong synergistic effect suggesting combination therapy as a future possibility [57]. Yet other approaches have been used in order to find explanations for low to mediocre treatment efficacy of BZs against Trichuris spp. infections. Specific variants of the beta-tubulin gene (i.e. single nucleotide polymorphisms (SNPs) in codon 167, 198 and 200) have been reported to convey BZ-resistance in parasitic nematodes of veterinary importance [59]–[63] and SNPs in codon 200 have been identified in T. trichiura obtained from a human population expected to be unexposed to BZs [64]. Furthermore, there is evidence demonstrating a higher frequency of the resistant genotype in codon 200 (TAC/TAC) in eggs of T. trichiura isolated from human populations in Haiti and Kenya after treatment with ALB [65], indicating that anthelmintic resistance may be involved in the low to mediocre treatment efficacy of BZs reported for this genus. However, such SNPs were not found in other Trichuris spp. [66], and not systematically in human populations [67].
The present work represents yet another approach to address the intriguing question for low to varied treatment efficacy of T. trichiura infections in humans. Based on worm motility, concentration of anthelmintic drugs and their metabolites within the worms and the difference in binding capacity of FBZ and ALB, we suggest that the lower sensitivity of T. suis towards these drugs in vitro is, in comparison to O. dentatum, due to a lower drug uptake. Furthermore, our data indicate that T. suis is able to transform FBZ or OXF into the inactive metabolite FBZSO2. Whether the drug uptake of T. suis in vitro mirrors the drug uptake in vivo is still unresolved. In the host, Trichuris spp. are attached to the mucosa with the anterior part which may give the worms a mechanical advantage in relation to anthelmintic treatment (they do not easily get detached even when temporarily deprived for energy or paralysed). Furthermore, such attachment mayserve as a protective barrier of the anterior part against active drugs in the intestinal lumen and instead render the worms more exposed to less potent anthelmintic metabolites in the blood. However, the posterior part is largely exposed to drugs in the lumen. We do not know whether the majority of the drug acting on Trichuris spp. comes from the intestinal lumen or whether it arrives via the blood supplying the intestine or both, but by using T. suis as a model we have shown that the varied and low drug efficacy against Trichuris spp. in animals and humans may be related to low drug-uptake in the worms.
|
10.1371/journal.pntd.0000129 | Interspecific Hybridization Yields Strategy for South Pacific Filariasis Vector Elimination | Lymphatic filariasis (LF) is a leading cause of disability in South Pacific regions, where >96% of the 1.7 million population are at risk of LF infection. As part of current global campaign, mass drug administration (MDA) has effectively reduced lymphatic filiariasis prevalence, but mosquito vector biology can complicate the MDA strategy. In some regions, there is evidence that the goal of LF elimination cannot be attained via MDA alone. Obligate vector mosquitoes provide additional targets for breaking the LF transmission cycle, but existing methods are ineffective for controlling the primary vector throughout much of the South Pacific, Aedes polynesiensis.
Here we demonstrate that interspecific hybridization and introgression results in an A. polynesiensis strain (‘CP’ strain) that is stably infected with the endosymbiotic Wolbachia bacteria from Aedes riversi. The CP strain is bi-directionally incompatible with naturally infected mosquitoes, resulting in female sterility. Laboratory assays demonstrate that CP males are equally competitive, resulting in population elimination when CP males are introduced into wild type A. polynesiensis populations.
The findings demonstrate strategy feasibility and encourage field tests of the vector elimination strategy as a supplement to ongoing MDA efforts.
| Lymphatic filariasis (LF) is a global health problem, with over 120 million people affected annually. The current LF elimination program is focused on administering anti-filarial drugs to the entire at-risk population via annual mass drug administration (MDA). While the MDA program is proving effective in many areas, other areas may require augmentative measures such as vector control. An example of the latter is provided by some regions of the South Pacific where Aedes polynesiensis is the primary vector. Here, we describe a novel vector control approach based upon naturally occurring Wolbachia bacterial infections. Wolbachia are endosymbiotic intracellular bacteria that cause a form of sterility known as cytoplasmic incompatibility. We show that introgression crosses with mosquitoes that are infected with a different Wolbachia type results in an A. polynesiensis strain (designated ‘CP’) that is incompatible with naturally infected mosquitoes. No difference in mating competitiveness is observed between CP males and wild type males in laboratory assays. The results support continued development of the strategy as a tool to improve public health.
| Lymphatic filariasis (LF) is a global health problem, with over 120 million infected individuals and an estimated one billion people at risk of infection [1]. The current LF elimination campaign is premised upon the lack of a non-human reservoir for Wuchereria bancrofti and is enabled by recent advances in diagnostic tools and treatment as well as the donation of microfilaricidal drugs [1]–[3]. In the absence of appropriate macrofilaricidal prophylactic or therapeutic treatments, the current strategy focuses on interruption of LF transmission via Mass Drug Administration (MDA): treatment of the entire ‘at risk’ population with microfilaricidal compounds to suppress microfilariae levels below that required to sustain transmission. The MDA strategy calls for drug treatment to continue annually over a period exceeding the ∼5 year lifespan of adult worms, [4] with a goal of global LF elimination by 2020.
The efficacy of the MDA strategy is compromised in some regions by the biology of the insect vectors. A notable example is provided in endemic areas within the South Pacific, where the diurnal subperiodic form of W. bancrofti is transmitted by A. polynesiensis. A. polynesiensis displays a pattern of negative density dependent transmission, such that this mosquito is a more efficient vector in low-level microfilaraemics, such as that which occurs with MDA strategies [5],[6]. The stabilizing impact of negative density dependent transmission is hypothesized as a contributor to an inability to eliminate LF in French Polynesia despite decades of ongoing MDA [5],[7]. Since mosquitoes are an obligate host for W. bancrofti, anti-vector interventions are recognized as a supplemental method to break the LF transmission cycle, leading to recommendations for the integration of vector control with MDA in areas where A. polynesiensis is the primary vector [3]. Unfortunately, effective control of A. polynesiensis has never been accomplished, due to problems including the inaccessibility of A. polynesiensis breeding sites and geography of Pacific island nations, which complicate ongoing vector control efforts [3],[8]. In contrast, Pacific island geography can simplify and facilitate an area-wide elimination approach, by subdividing vector mosquitoes into discrete populations with limited immigration.
Area-wide elimination programs targeting mosquito populations have been attempted previously, with mixed results [9]. A notable success was reported in a field trial in which a Culex quinquefasciatus population was eliminated via repeated, inundative releases of male mosquitoes infected with an incompatible Wolbachia pipientis bacteria [10]. Wolbachia are obligate intracellular bacteria that are maternally inherited in insects and other invertebrates [11]. In mosquitoes, Wolbachia cause a form of sterility known as cytoplasmic incompatibility (CI), which results in karyogamy failure and arrested embryonic development. In populations where individuals are infected with different Wolbachia types, bi-directional CI can occur: sterility results in both cross directions between mates infected with different Wolbachia types. Models predict that in natural populations, sterility resulting from bi-directional CI is a transient event, since one infection will predominate and replace the other cytotype [12]. In the Wolbachia-based vector control strategy however, female sterility is artificially sustained by repeated, inundative releases of incompatible males analogous to traditional sterile insect technique (SIT) [13], resulting in mosquito population decrease and elimination. It is emphasized that the released male mosquitoes do not blood feed, vector disease or transmit Wolbachia. Further development and expansion of the Wolbachia-based suppression approach was not subsequently pursued due to strategy complications including immigration of mated females and variable CI patterns observed in Culex populations in different geographic areas. Furthermore, the application was viewed as specialized to Culex, since bi-directional CI was not observed in additional vector species. The latter problem has recently been addressed with a demonstrated ability to artificially generate new Wolbachia infection types in mosquitoes [14].
Problems with insecticidal approaches (e.g., resistance and non-target effects) have led to a renaissance of interest in genetic control of disease vectors, using newly-developed transgenic approaches such as the Repressible Dominant Lethal (RIDL) technology [15]. With the new reality of potential transgenic insect releases, considerable thought is being devoted to addressing the requirements for field implementation of an approach employing transgenic mosquitoes. It is recognized that experience is lacking that demonstrates the efficacy and safety of the transgenic mosquito approach, and a critical question relates to the social acceptance toward the release of transgenic mosquitoes without a demonstrated benefit (i.e., an epidemiologically significant impact on transmission of mosquito borne disease) to offset real and perceived risks. Here we describe progress toward a non-transgenic approach for primary vector elimination within an endemic area, which can allow the subsequent evaluation for an epidemiological impact on disease transmission.
Mosquito strains and details of maintenance and experimental crosses are as previously described [16]: AR = Aedes riversi; ART = Aposymbiotic AR strain (Wolbachia removed via tetracycline treatment); AP = Aedes polynesiensis; APT = Aposymbiotic AP. Each generation of introgression (Figure 1A) was established with >100 individuals of each sex.
For the CP male competitiveness assay, 25 cages were established, each with ten virgin A. polynesiensis females (<2 days post eclosion) and 20 males (<3 days post eclosion). Five days after introduction, females were blood fed, isolated and allowed to oviposit. The egg hatch assay consisted of allowing one week for egg maturation, submerging eggs for three days, and then observing eggs from individual females using a dissection scope. Females producing fewer than ten eggs were excluded from the data set. To confirm female insemination, spermathecae were checked for females producing broods with low egg hatch.
DNA was extracted from individual mosquitoes using DNeasy kits (Qiagen, Valencia, CA) following manufacturers instructions. Six-microsatellite primer pairs were used to amplify loci using PCR conditions as previously described [17]. Left primers were fluorescent labelled with different WellRED dye colors (Integrated DNA technologies, Coralville, IA). Fragments sizes were measured using a CEQ 2000 sequencer (Beckman Coulter, Fullerton, CA) according to manufacturer's instructions. Allele frequencies and genotypic dis-equilibrium were calculated, and Fisher's exact tests [18] were performed using GENEPOP version 3.4. All calculations were performed using the Markov chain method with demorization set to 1000, 100 batches, and 1000 iterations per batch.
Adults were homogenized in 100 µl of buffer containing 10 mM Tris-HCL, 1 mM EDTA, and 50 mM NaCl, at pH 8.2 using a Mini-beadbeater (BioSpec Products, Inc., Bartlesville, OK). After homogenization, samples were boiled for 5 min and centrifuged at 16,000g for 5 min. One µl of supernatant was used for each PCR reaction. PCR conditions were as described previously [19]. Infection type of CP, AP, and AR was determined using PCR primers specific for A type Wolbachia (136F and 691R) or B type Wolbachia (81F and 522R) [20]. To assess maternal inheritance rates, CP females were mated with CP males, blood fed, isolated, and allowed to oviposit. CP females and their progeny were examined via PCR using the 81F and 691R primers [20] using the above-described methods.
To analyze male mating competitiveness a Chi-square goodness of fit test was performed to compare observed and expected numbers of hatching broods for the replicate cages of the varying ratios of CP:AP males. To analyze population suppression in replicate cages, multiple Mann-Whitney tests, with sequential Bonferroni correction were conducted to compare egg hatch rates between cages with varying ratios of CP:AP males. Compatible female egg hatch data from all replicate cages was subjected to a Kruskal-Wallis test.
Prior studies demonstrate that A. polynesiensis and A. riversi, two closely related members of the Aedes (Stegomyia) scutellaris complex are naturally infected with differing Wolbachia infection types (A and B clades, respectively) [16], and that removal of the Wolbachia infection results in egg hatch, which does not occur in interspecific crosses of naturally-infected individuals. Here, we report that hybrids resulting from crosses of uninfected (aposymbiotic) A. polynesiensis and A. riversi are viable and fertile (Table 1). Hybrid fertility allows a strategy in which the A. riversi Wolbachia type is introgressed into the A. polynesiensis genotype, resulting in the ‘CP’ strain (Figure 1A). As shown in Figure 1B, PCR confirms that Wolbachia in the CP strain is B-type Wolbachia, as predicted.
To examine introgression of the CP strain, allelic distributions of six-microsatellite loci were compared between the AP, AR and CP strains. All loci investigated are polymorphic between strains, and four loci are polymorphic within strains. The test for genotypic dis-equilibrium across all pairs was not significant (p>0.2), suggesting loci are not linked. The CP and AP strains were observed to share a similar distribution of alleles across all loci (Fisher's exact test; P>0.3). In contrast, allele distributions of CP and AP are significantly different from AR (Fisher's exact test, P<0.0001), with only two alleles commonly shared by all three strains at two loci (Loci 1 and 3) (Figure 2). Thus, the results are consistent with the hypothesized CP introgression with the AP genotype.
Crosses demonstrate strong bi-directional incompatibility between CP and naturally infected A. polynesiensis, with no egg hatch resulting from >1,800 eggs examined in crosses of AP females and CP males (Table 1). Progeny resulting from crosses of aposymbiotic CPT males (Table 1) demonstrates that the observed sterility in CP crosses is due to the Wolbachia infection.
Experiments show a high level (>99% fidelity) of maternal transmission of Wolbachia from CP females to both sons and daughters. Subsequent to oviposition, CP females were confirmed to be infected using PCR with Wolbachia-specific primers. Approximately ten daughters and ten sons from each of ten infected CP females were PCR tested, and all were observed to be Wolbachia infected (n = 210). The high maternal inheritance rate is similar to that observed in naturally-infected A. polynesiensis and a related mosquito: Aedes albopictus [19],[21].
Strong bi-directional CI and high maternal transmission support investigation of a vector control strategy in which CP male releases are used to suppress A. polynesiensis populations. To examine the strategy: virgin AP females were introduced into cages with varying ratios of CP:AP males (Figure 3). Following mating, females were isolated and the egg hatch rate was examined. As shown in Figure 3, the egg hatch was observed to significantly decrease from 75% to 0% egg hatch, inversely related to the frequency of incompatible CP males. The experimental design also permits an assessment of CP male competitiveness relative to AP males. As illustrated in Figure 3, the number of observed compatibly mated females (i.e., producing hatching broods) did not differ from predictions that assume equal male competitiveness (Chi Square; P>0.1). The data shown in Figure 3 also support the hypothesis that females utilize sperm from one male. A comparison of egg hatch rates resulting from compatibly-mated females did not differ significantly between treatments (Kruskal-Wallis, P>0.3). If females were to utilize sperm from multiple males, then a lower egg hatch rate would be expected in treatments with a mixture of CP and AP males. The results support the feasibility of a CI-based A. polynesiensis suppression strategy and encourage additional experiments to assess the strategy under more natural conditions (e.g., field cages).
The inability of prior MDA efforts to eliminate LF transmission from some Pacific regions represents a potential weakness in the current global campaign. Therefore, an ability to reduce or eliminate the required mosquito vector populations would provide a useful augmentative tool for blocking LF transmission. A. polynesiensis populations provide a logical target, given their broad geographic range in the South Pacific and ability to vector filariasis in low-level microfilaraemics (i.e., ‘limitation’ LF transmission) [5],[7]. Unfortunately, existing vector control tools have proven unsuccessful against this mosquito species.
The results presented here support the feasibility of a species-specific approach in which inundative releases of bi-directionally incompatible males induce sterility in A. polynesiensis females, resulting in vector population elimination. The geography of the South Pacific is ideal for the proposed A. polynesiensis suppression strategy, since the A. polynesiensis population is subdivided into islands with limited immigration [22],[23]. The natural subdivision of A. polynesiensis into isolated populations will facilitate a sequential elimination approach, in which transient entomological teams focus effort on one island and then progress to a subsequent island. Following elimination, a reporting system would be deployed, monitoring for A. polynesiensis reintroduction and reestablishment.
The proposed strategy would be integrated with the existing MDA strategy, to be deployed in areas where LF elimination is complicated by A. polynesiensis biology. It is emphasized that the primary goal of breaking the LF transmission cycle does not require the permanent eradication of A. polynesiensis. Instead, a transient elimination of A. polynesiensis will suffice, as long as the period of vector elimination extends beyond the lifespan of adult W. bancrofti in the reservoir human population. However, A. polynesiensis eradication would be desirable from the broader public health perspective that A. polynesiensis is a biting nuisance and serves as a vector during periodic dengue epidemics. The successful demonstration that releases of bi-directionally incompatible males can impact LF transmission will encourage an extension of the strategy to a broader geographic range and to additional vector species (e.g., Culex spp. or other aedine LF vectors in the South Pacific) using the previously demonstrated ability to artificially generate novel Wolbachia infections in medically important mosquitoes [14].
A concern relates to downstream logistical aspects associated with the subsequent ‘scale up’ required for suppression of larger populations. Due to bi-directional CI, females that are unintentionally released are incompatible with wild type males. With the reduction of the population size due to CI-induced sterility, there is an increasing probability that accidental female releases will permit the establishment of the new infection type, resulting in population replacement instead of population elimination [12]. Thus, strategy success requires releases to consist of males only. While a variety of mechanical sex separation tools for mosquitoes have been developed, available devices are not sufficiently accurate. Therefore, to ensure male-only releases, early field trials will be on a relatively small scale that allows visual verification of mechanically-separated males, similar to prior trials [10]. In the event that visual inspection is not cost effective for subsequent large-scale releases, deployment of the proposed approach in larger areas may require additional technology to improve cost efficacy, such as genetic sexing [24]. An additional possibility is premised upon the observation that female mosquitoes are typically more susceptible to radiation relative to males [25] and would treat release individuals with low levels of radiation to render unintentionally released females impotent (i.e., sterile or of negligible fitness). The risk of compatible matings between released CP males and A. riversi females in the field is not a concern, since populations of A. riversi occur in Japan, [26] remote from South Pacific islands that are proposed for field releases.
The non-transgenic, species-specific, Wolbachia-based elimination strategy proposed here provides a logical segue toward transgenic approaches (i.e., RIDL [15]), which may yield improved efficacy and/or cost. Furthermore, the social palatability of transgenic mosquito releases can be increased via an approach that is integrated with Wolbachia-induced CI. Specifically, if released transgenic males are cytoplasmically incompatible with the targeted mosquito population, the released transgene has a reduced probability of establishing in the field.
Future efforts must define the vectorial competency of CP females relative to wild-type females. In the event that the CP strain is observed to be refractory to LF transmission, replacement of the naturally-susceptible wild-type population with a refractory CP population may be a desirable outcome. Based upon the results of prior vector competency studies examining hybridizations between members of the Aedes scutellaris complex, the prospect of CP displaying reduced vectorial competency is not a remote possibility. Notably: prior cross experiments demonstrate that the Wuchereria refractoriness phenotype is dominant, [27],[28] prior hybridization experiments provide evidence for cytoplasmic inheritance of susceptibility [29] and AR is not recognized as a disease vector. Since periodic dengue epidemics occur within the A. polynesiensis range, the competency of CP to transmit dengue and additional pathogens (e.g., chikungunya) would need to be assessed prior to implementing a population replacement program. The latter complication would be less of a concern with an elimination strategy, since following successful intervention, neither CP or A. polynesiensis would occur in the field. |
10.1371/journal.pcbi.1004025 | Evolution and Phenotypic Selection of Cancer Stem Cells | Cells of different organs at different ages have an intrinsic set of kinetics that dictates their behavior. Transformation into cancer cells will inherit these kinetics that determine initial cell and tumor population progression dynamics. Subject to genetic mutation and epigenetic alterations, cancer cell kinetics can change, and favorable alterations that increase cellular fitness will manifest themselves and accelerate tumor progression. We set out to investigate the emerging intratumoral heterogeneity and to determine the evolutionary trajectories of the combination of cell-intrinsic kinetics that yield aggressive tumor growth. We develop a cellular automaton model that tracks the temporal evolution of the malignant subpopulation of so-called cancer stem cells(CSC), as these cells are exclusively able to initiate and sustain tumors. We explore orthogonal cell traits, including cell migration to facilitate invasion, spontaneous cell death due to genetic drift after accumulation of irreversible deleterious mutations, symmetric cancer stem cell division that increases the cancer stem cell pool, and telomere length and erosion as a mitotic counter for inherited non-stem cancer cell proliferation potential. Our study suggests that cell proliferation potential is the strongest modulator of tumor growth. Early increase in proliferation potential yields larger populations of non-stem cancer cells(CC) that compete with CSC and thus inhibit CSC division while a reduction in proliferation potential loosens such inhibition and facilitates frequent CSC division. The sub-population of cancer stem cells in itself becomes highly heterogeneous dictating population level dynamics that vary from long-term dormancy to aggressive progression. Our study suggests that the clonal diversity that is captured in single tumor biopsy samples represents only a small proportion of the total number of phenotypes.
| We present an in silico computational model of tumor growth and evolution according to the cancer stem cell hypothesis. Inheritable traits of cells may be genetically or epigenetically altered, and traits that confer increased fitness to the cell will be selected for on the population level. Phenotypic evolution yields aggressive tumors with large heterogeneity, prompting the notion that the cancer stem cell population per se is highly heterogeneous. Within aggressive tumors cancer stem cells with low tumorigenic potential may be isolated. Simulations of our model suggest that the cells harvested in core needle biopsies represent less than 10% of the phenotypic heterogeneity of the total tumor population. Dependent on the cells captured in the sample, xenografted tumors may exhibit aggressive growth or long-term dormancy—dynamics that may suggest opposing treatment approaches for the same tumor when translated into clinical decision-making.
| Human organs and tissues are comprised of cells that have evolved to maintain functionality and integrity. Cells of different organs and ages have different traits, such as migration rate, turnover time, proliferation potential and lifespan until senescence. While fetal diploid cell strains in culture demonstrate a large number of divisions before mitotic arrest and culture degeneration (50 ± 10 population doublings, termed the “Hayflick Limit” [1–3]), hematopoietic progenitors may undergo about 20–30 divisions [4], and colonic crypt progenitors complete only four to six divisions before getting washed off at the top of the crypt [3]. Repopulation of the tissue is assured by tissue stem cells that sit on top of a cellular hierarchy [5,6]. In a physiological setting, stem cells are predominantly non-mitotic to prevent malignant transformation [7] and only enter the mitotic cycle when tissue repopulation is required [8–10]. The transit-amplifying offspring of a stem cell may undergo multiple divisions to produce a population of cells that differentiate into tissue-specific cells with determined function and lifespan. The potencies of tissue stem cells and transit-amplifying cells vary with tissue type and age. Transformation may occur at any time in all tissue compartments, but the ability of transformed cells to initiate and sustain pathologic tumor growth requires certain kinetic properties including longevity, migration potential, self-renewal and differentiation capacity. These traits are comparable to physiologic stem cells, and cancer cells with such properties have been termed cancer stem cells causing a long and active discussion about the cell of origin of tumor [5,11,12]. Intestinal cancer may be initiated by a transformed stem cell [13], but a transformed progenitor cell with acquired stem-like traits is more likely in myeloid malignancies and NF1- and PDGF-driven glioblastoma [14,15]. The set of kinetics in the initial cancer stem cell, however, is initially close to that of the untransformed cell. The kinetics will be inherited by the descendent cells yielding tumor population dynamics ranging from microscopic dormancy to aggressive tumor growth [16]. It is conceivable that the variation in kinetics of cells in different positions of the tissue hierarchy and at different ages gives rise to many trait combinations unfavorable for progression [17]. Inferior cancer cell trait combinations could, at least in part, explain the increasing observations of pathologic but non-advancing lesions [18,19]. Trait fitness, mutation and evolution may augment our understanding why tumors may be more prevalent in certain organs at a specific age range [20].
Regardless of kinetics at time of transformation, cells are subject to mutations, which enables evolution of and selection for more aggressive traits. For example, during mitosis the daughter cells inherit telomeres, the non-coding replicative protective ends of the DNA [21]. Telomeres get shortened during mitosis [22], which offers a quantitative visualization of the remaining cellular proliferation potential [22,23]. Abnormally increased or decreased telomerase activity [24] in cancer stem cells lengthens or shortens telomeric DNA that defines the number of cell divisions for non-stem cancer cell progeny [25,26].
We first compare growth of the tumor without (control) and with trait mutations. We then perform detailed analysis of the phenotypes that emerged in the smallest, an average-sized, and the largest tumor in the simulated time frame. Using the phenotypic structure of the largest tumor we investigate the population heterogeneity that is represented in needle biopsies.
We initialize tumor growth simulations with one CSC with the initial traits probability of symmetric division ps = 0.05, proliferation capacity ρ = 10, migration potential μ = 15 and probability of spontaneous death α = 0.01, which has previously been shown to simulate fast tumor growth [16,27]. Motivation for and variation of the discussed parameters α, μ, ρmax and ps and their impact on tumor progression has been discussed in detail elsewhere [16,27,28]. We allow for mutation during symmetric CSC division with a probability of 50% (pmut = 0.5). We simulate tumor growth for t = 730 days, i.e. 2 years, without (control) and with trait evolution. With α = 0.01 and ps = 0.05, stochastic death of the initial CSC before symmetric division and thus regression of the total tumor population is expected and observed in about 20% of the simulations (n = 23). Simulations with successful tumor growth (n = 77) reveal a widespread of tumor sizes after t = 730 days with mutation (standard deviation s.d.>60% of the mean) compared to non-mutating control (s.d.<20% of the mean). Initial growth kinetics (t<200 days) are comparable between both groups, dictated by the chosen initial traits vector. Early tumor growth follows self-metastatic progression as previously described [27,29]. A favorable impact of trait mutations on tumor growth can be seen in significantly steeper growth after selection shapes the population (Fig. 1A). Tumors subject to mutations contain on average approximately 220% more cells with almost a 17-fold increase in CSCs compared to control tumors after t = 730 days. The fastest-growing evolving tumor, however, is more than 6 times larger with 170 times more CSCs than the largest non-evolving control tumor (Fig. 1B). The enrichment in CSC yields a more compact, uniform tumor expansion compared to self-metastatic progression. Evolution, however, may also impair tumor progression. Early mutations that yield unfavorable CSCs kinetics dwarf tumor growth. In the least favorable case, the smallest tumor is only a third the size of the smallest non-evolving control tumor and less than a quarter of the control cancer stem cells (Fig. 1).
We now investigate the CSC trait vectors that dominate the smallest, an average-sized and the largest evolved tumor to identify favorable and unfavorable evolutionary trajectories. Fig. 1C-E shows the temporal evolution of the most common trait vectors and parameter values for the 10 dominant vectors at the end of the simulation (t = 730). Only 28 phenotypes have evolved in the smallest tumor after 730 days. Such small pool of available phenotypes indicates that early unfavorable mutations blocked the ability of the tumor to develop a wide spectrum of phenotypes. The largest tumor, in contrast, evolved more than 11,000 distinct phenotypes. The 10 most common phenotypes reveal that the biggest evolutionary change manifested itself in decreased proliferation capacity ρmax. Indeed, the most common phenotypes in the smallest tumor have increased value of proliferation capacity suggesting competition between CSCs and CCs and long population level dormancy periods as previously described [16,30]. The important role of CC proliferation potential in tumor is further emphasized in detailed analysis of the individual trait evolution and resulting CSC numbers (Fig. 2A). An early decrease in proliferation potential ρmax allows for increased cancer stem cell proliferation and subsequent trait mutation events. An increase in symmetric division probability ps and a decrease in cell death α, and a later increase in cell migration follow the initial decrease in ρmax. An early increase in proliferation potential ρmax inhibits cancer stem cell pool expansion and diminishes further trait evolution (Fig. 2B). Correlation analysis further supports the pivotal role of CC proliferation potential in determining total tumor size. Symmetric division probability ps and cell migration μ correlate positively with tumor size. A weak but significant positive correlation is also observed with spontaneous cell death α (Fig. 2C). This lends support to previous theoretical observations that increased cell death counteracts cancer stem cell confinement [16,31] and promotes self-metastatic tumor expansion [27].
Evolutionary phenotypic heterogeneity leads to unpredictable spatiotemporal distribution of cells with varying aggressiveness. Whilst some tumor regions can be dominated by an ancestral phenotype, more recently evolved clones could be spatially limited to small pockets albeit large fitness. However, similar phenotypes may also develop along independent evolutionary trajectories at different spatial locations throughout the tumor. Cells with phenotypes close to the initial trait vector are contained in the tumor core. More recently evolved clones with increasing aggressiveness are located at the tumor periphery where they dominate outgrowth in circular manner (Fig. 3). Throughout the tumor, the probabilities of symmetric CSC division as well as migration rate have a monotonically increasing average value and follow a bell shaped distribution at t = 730 days. Proliferation potential and cell death distributions are skewed to the left approaching zero (Fig. 3).
We simulate in silico biopsy on the largest evolved tumor to investigate phenotypic diversity in the biopsy sample. A 26 Gauge biopsy needle with a 0.46 mm diameter is simulated to penetrate the tumor from the boundary to the center of mass. All cells (~100,000) in the needle trajectory are collected in a single harvest and their phenotypic traits averaged. Biopsies are collected at different angles in 5 degrees intervals yielding 72 samples. The average value for different phenotypic traits can vary orders of magnitude dependent on biopsy angle (Fig. 4A), suggesting spatially localized evolution and diverse phenotypic dispersal. Whilst the fraction of captured phenotypes (ratio of unique phenotypes in a sample to the total number of unique phenotypes in the whole tumor) varies by one order of magnitude, the fraction of cancer stem cells in the collected samples (ratio of CSC count to the total cell count within the collected sample) vary by as much as two orders of magnitude ranging from 0.7% to 10% (Fig. 4B). Multiple equally distant biopsies within a 90-degrees tumor quadrant increase the average fraction of captured phenotypes from 9% for one biopsy to 25% for 4 or more biopsies (Fig. 4C). Single collected biopsy samples can be divided into multiple subpopulations (10 subpopulations with approx. 10,000 cells each in our settings) and reseeded to obtain sample growth dynamics. Biopsy-specific phenotypic diversity yields tumor populations with grossly different growth dynamics from rapid and persistent growth (for biopsy taken at 135°) to initial decay and long-term dormancy (240°; Fig. 4D).
It becomes increasingly appreciated that tumors are heterogeneous populations of cells with different traits and fates. An intrinsic difference in tumor initiation and propagation potential led to a classification of cancer stem cells (CSC) and non-stem cancer cells (CC) in the majority of hematologic and solid tumors [12,32–34]. The transformed cells that lead to the initial CSC inherit kinetic properties of their somatic cell of origin, which may be significantly varying at different ages and between organs. Different cell kinetics result in different overall tumor population dynamics ranging from population level dormancy to aggressive growth and invasion [17]. Agent-based models are well suited to simulate individual cell dynamics and estimate tumor progression dependencies on single cell kinetics. We have identified a set of orthogonal cell kinetics that include cell migration rate, proliferation potential, spontaneous cell death, and symmetric cancer stem cell division. In a cancer stem cell-driven tumor, proliferation potential and spontaneous cell death have been shown to non-monotonically modulate overall tumor progression [16,30,31] while increased cell migration and symmetric CSC division always lead to accelerated tumor growth [27,35]. The complex interplay of these participating dynamics, however, requires a more elegant investigation of which combinations of simultaneously evolving kinetics yield the most aggressive tumor clones. While parameter evolution yields predominantly fast growing tumors, early unfavorable mutation events may ultimately push tumor dynamics into dwarfed tumor growth or long-term dormancy. Our study suggests that cell proliferation potential is the strongest modulator of tumor growth. Early increase in proliferation potential yields larger populations of CC that compete with CSC and thus inhibit CSC division [16,30]. Conversely, a reduction in proliferation potential decreases intratumoral competition and enables accelerated CSC pool expansion and further evolution. Interestingly, short telomeres indicative of short-lived cancer cells with limited proliferation capacity have indeed been observed in malignant tumor population [26,36], lending further support that limiting the number of CC progeny promotes parental CSC expansion and tumor growth. Phenotypes that eventually dominate the tumor developed relatively late indicating that the chosen initial parameter values were suboptimal for tumor progression despite previous observations of relatively fast growth [17]. Successfully evolving tumors exhibit a heterogeneous distribution of phenotypes with different sub-clones dominating local expansion. This confirms clinical observations of intratumoral heterogeneity and branched evolution with spatially distinct genetic profiles [37]. The spatial heterogeneity of cell traits also leads to a large variation in CSC fraction in biopsy samples, offering yet another angle to the ongoing discussion about the proportion of tumorigenic subpopulations in tumors [38–41]. Furthermore, our work suggests that the so-called sub-population of cancer stem cells is in itself heterogeneous. Whilst some CSC will initiate immediate growth, other CSC will form microscopic tumors that may remain dormant for prolonged periods of time. Indeed, a variety of individual human derived malignant cancer cell lines can each be carefully divided into sub-clones that form fast-growing tumors or stable disease for many months before initiating rapid growth [42–44].
The model presented here is a complex but simple approach for measuring continuous intratumoral CSC evolution. For computational convenience we have limited our study to a two-dimensional model. A previous analysis in a similar cell interaction model has shown that two-dimensional cellular automata qualitatively mimic three-dimensional tumor growth dynamics [45]. The presented results assumed a mutation rate of 50%; other mutation rates yield qualitatively similar results on a different time scale with smaller mutation rates yielding slower growing tumors (S1 Fig). To first understand the trajectories of intratumoral evolution, we limited the study to stem and non-stem cancer cells and ignored tumor environmental factors that will exert external selection pressure on the tumor population and shape parameter fitness values accordingly [46,47]. Future developments of this model may include local tumor-environmental interactions or globally-informed modulations by the host immune system. To further increase complexity, bi-directional plasticity of phenotypes via CSC differentiation and CC de-differentiation may be considered [48–51], which will contribute to increased biological realism of tumor evolution dynamics.
We develop a theoretical framework to investigate tumor progression in response to cancer stem cell evolution. We explore orthogonal cell traits including cell migration to facilitate invasion [52], spontaneous cell death due to genetic drift after accumulation of irreversible deleterious mutations [53,54], symmetric cancer stem cell division that increases the cancer stem cell pool [55,56], and telomere length and erosion as a mitotic counter for inherited non-stem cancer cell proliferation potential [22,57]. We use an agent-based model to simulate the dynamics of single cells and observe cell-cell interactions and population-level tumor formation [16,27]. The model is realized as an asynchronous cellular automaton in which cell events are stochastically driven. A cell, either cancer stem cell (CSC) or non-stem cancer cell (CC), occupies a single grid point of (10μm)2 on a two-dimensional square lattice. Each CSC is characterized by its specific trait vector [ps, ρmax, μ, α] denoting probability of symmetric division, proliferation capacity, migration potential and spontaneous death probability, respectively. According to the cancer stem cell hypothesis, CSCs have unlimited proliferation potential and thus their proliferative capacity ρmax does not exhaust. At each division CSCs produce either another CSC with probability ps (symmetric division) or a CC with probability 1-ps (asymmetric division). CCs that are direct offspring of a CSC inherit the initial proliferation capacity ρ that decreases with each cell division (Fig. 5A). At ρ = 0, CCs die and are removed from the simulation. At each proliferation attempt, cells may undergo spontaneous death with probability α and then be removed from the system. Both tumor subpopulations are equipped with migration potential μ representing number of potential cell displacements into neighboring lattice sites per day. We assume that cells need adjacent space for migration and proliferation, and cells that are completely surrounded by other cells (eight on a two-dimensional lattice) become quiescent (Fig. 5B). In unsaturated environments, cells proliferate and migrate into vacant adjacent space at random. To avoid artifacts caused by computational domain boundaries we introduce a dynamically growing domain [58].
Motivated by the immortal strand hypothesis [59] and non-random DNA strand co-segregation [60] we model mutation events in the malignant subpopulation of CSCs during symmetric division (Fig. 5C). We ignore mutation of traits in CCs as those will be lost from the total population [61]. Traits are marginally mutated allowing for continuous evolution and large phenotypic variety, assuring that the phenotype pool is not limited to a number of trait combinations from which a phenotype is randomly drawn [35,62,63]. We assume that a single mutation affects at most one trait and induces a stochastic positive or negative unit change of the trait parameter value, i.e. ps±0.01, ρmax±1, μ±1, or α±0.001. The trait to be mutated is chosen at random from a discrete uniform distribution. The modified trait vector is inherited by both CSCs and then further propagated to their respective CC populations. If a trait becomes negative the cell is considered unviable and removed from the simulation. Fig. 5D summarizes the simulation process for the presented model.
Simulation time is advanced at discrete time intervals Δt = 1/24 day (i.e., 1 hour), that is 24 simulation steps equal one day. At each simulation time step, cells are considered in random order and the behavior of each cell is updated. Cell proliferation, migration and death are random events with the respective probabilities scaled to simulation time. Cell proliferation and migration are temporally mutually exclusive events, and cell death only occurs when cell actively attempts to proliferate. We assume that cells proliferate on average once per day (proliferation probability pd = 1×Δt), migrate with probability (1-pd)pm and die with probability pdα. Let pm = μ×Δt, where the parameter μ denote motility of cancer cells. Due to the stochastic nature of the model we perform at least 100 independent simulations for each discussed case and report average values and standard deviations.
|
10.1371/journal.pcbi.1003032 | Steered Molecular Dynamics Simulations of a Type IV Pilus Probe Initial Stages of a Force-Induced Conformational Transition | Type IV pili are long, protein filaments built from a repeating subunit that protrudes from the surface of a wide variety of infectious bacteria. They are implicated in a vast array of functions, ranging from bacterial motility to microcolony formation to infection. One of the most well-studied type IV filaments is the gonococcal type IV pilus (GC-T4P) from Neisseria gonorrhoeae, the causative agent of gonorrhea. Cryo-electron microscopy has been used to construct a model of this filament, offering insights into the structure of type IV pili. In addition, experiments have demonstrated that GC-T4P can withstand very large tension forces, and transition to a force-induced conformation. However, the details of force-generation, and the atomic-level characteristics of the force-induced conformation, are unknown. Here, steered molecular dynamics (SMD) simulation was used to exert a force in silico on an 18 subunit segment of GC-T4P to address questions regarding the nature of the interactions that lead to the extraordinary strength of bacterial pili. SMD simulations revealed that the buried pilin α1 domains maintain hydrophobic contacts with one another within the core of the filament, leading to GC-T4P's structural stability. At the filament surface, gaps between pilin globular head domains in both the native and pulled states provide water accessible routes between the external environment and the interior of the filament, allowing water to access the pilin α1 domains as reported for VC-T4P in deuterium exchange experiments. Results were also compared to the experimentally observed force-induced conformation. In particular, an exposed amino acid sequence in the experimentally stretched filament was also found to become exposed during the SMD simulations, suggesting that initial stages of the force induced transition are well captured. Furthermore, a second sequence was shown to be initially hidden in the native filament and became exposed upon stretching.
| There are a large number of infectious bacteria that can be harmful to humans. Some bacterial infections are facilitated by long, tether-like filaments called type IV pili which extend from the surface of bacterial cells and attach to the surface of host cells. Type IV pilus filaments can grow to be many micrometers in length (bacterial cells themselves, on average, are only a couple of micrometers in length and half a micrometer in diameter), and can exert very large forces (up to 100,000 times the bodyweight of the bacteria). Because they extend from the surface of the cell, type IV pili are very good candidates for drug targeting. Computer simulation was used to exert forces on a segment of one of these filaments, in an effort to mimic the effects of tension that would be experienced by the pilus upon binding during infection. Regions of the filament that become exposed to the external environment in the pulled state were determined, in an attempt to identify amino acid sequences that could act as targets for drug design.
| Type IV pili (T4P), long (lengths at the micron scale) filamentous proteins composed of pilin subunits, are associated with a variety of bacteria, and emanate from the surface of the bacterial cell [1], [2]. T4P have been known as virulence factors for a long time as they are borne by many pathogens [1], [3]. They are of paramount importance in mediating attachment between bacteria and other surfaces, and perform a wide variety of functions for the bacterial cell including adhesion, motility, micro-colony formation, infection, and are implicated in immune escape [1], [3]. While other pili such as Type 1 or Type P pili provide function such as adhesion by their presence on the cell surface, T4P are also dynamic [4], [5]. T4P undergo cycles of elongation and retraction as pilin subunits are either added to or removed from the filament in a mechanism that is still poorly understood [1], [2]. When retracting, a single gonococcal (GC)-T4P filament can exert a force greater than 100 pN [6], [7]. The ability of GC-T4P to form bundles of 8–10 individual filaments has been observed, and these bundles can exert forces in the nanonewton range [8]. These are the highest recorded forces generated by bacteria (equivalent of 100,000 times the bacterial bodyweight).
Because of their involvement in surface attachment, GC-T4P filaments often find themselves under tension. The biological role of force in the interaction with host cells has been demonstrated to activate various mechanical signaling pathways in epithelial cells [9]. In addition, the physical forces exerted by the bacteria elicited dramatic rearrangements of the cell cortex [10], [11]. However, the mechanisms at play to go from force generation to biological function have yet to be established. Recent experimental evidence points to the impact of tension on the structure of T4P filaments. Specifically, experiments have shown conformational rearrangements of GC-T4P filaments expose buried amino acid sequences to the environment [12]. It is of interest to determine all of the regions exposed to the environment under tension for understanding the extraordinary plasticity of GC-T4P filaments. In addition, by uncovering what regions of the pilus filament become exposed under strain, more effective drugs, acting as inhibitors to T4P binding, could potentially be engineered [13], [14].
Among the many type IV pilins, the GC-pilin subunit, PilE [15], and the Pseudomonas aeruginosa subunit, PilA [16] have received the most attention, and their structures exemplify the canonical shape of type IV pilin: a globular head attached to a hydrophobic extended α-helix. In these two cases, the full-length subunits were crystallized. The N-terminal half of the helix (α1-N domain) protrudes from the protein, while the other half (α1-C domain) interacts with an anti-parallel four to five stranded β-sheet globular head domain. α1-N is almost completely hydrophobic except for a single charged residue, Glu5, which is conserved in nearly all type IV pilins, with only one exception: an aspartate is found at position 4 in the subunit PilS of S. enterica [1]. It has been speculated that Glu5 (and Asp4 in PilS), may serve to neutralize the electrostatic nature of the core of the filament by compensating for the positively charged N-terminus of α1-N [15], [17].
The globular head domain of the subunits lines the surface of the filament, and is therefore thought to be involved in its interactions with the environment [1]. The globular heads exhibit features relevant to pilus function. The αβ-loop possesses two post-translational modifications in GC-pilin, glycosylation of Ser63 and phosphorylation of Ser68 [1], [18], [19], which may protect epitopes from immune response and change the surface chemistry of the pilus and have been recently shown to play a role in the dispersal of the bacteria [18], [20], [21]. The D-region includes a hyper-variable loop, named as such because of the high variability of its amino acid sequence from one bacterial strain to another, which has been suggested to contribute to immune system evasion and persistent infection [22]–[24]. Additionally, along the filament surface, grooves between the globular heads of adjacent pilins are lined with positively charged residues in some locations, which may help to facilitate GC-T4P binding to DNA [15], [17].
A model for the GC-T4P filament assembly has been constructed by fitting the x-ray structure of GC-pilin (158 residues) into a cryo-EM map of a segment of GC-T4P filament at 12.5 Å resolution [17]. The cryo-EM reconstruction helped to shed light on its structural characteristics. Pilin subunits wind along the central filament axis, with approximately 3.6 pilin subunits per turn [17]. Subunits are arranged following symmetries along the filament axis (right-handed 1-start, left-handed 3-start and right-handed 4-start helices) [1], [2], [17]. These symmetries represent the various ways to divide the filament into progressions of pilin subunits that wind helically around the central filament axis. The right-handed 1-start helix symmetry describes positions of all pilin subunits in the filament using the smallest axial rise. The three left-handed 3-start helices of pilin subunits connect subunits n, n+3, n+6, etc, while the four right-handed 4-start helices of pilin subunits connect subunits n, n+4, n+8, etc, (see Figure 1A). The cryo-EM model also exhibited the presence of a channel of variable width (6–11 Å) through the filament core [17]. The α1-N domains of the subunits are thought to contribute to the strength of the filament due to their extensive hydrophobic interaction network in the core of the structure. Interaction between the N-terminal helices consists of about 75% of the total hydrophobic buried surface area of every pilin subunit [17]. While providing strength, the α1 helices are also expected to be flexible, since they possess glycine and proline that induce kinks and flexibility in α-helices. For example, Pro22 and Gly42 contribute to the S-like shape of α1 and are conserved amongst the Type IVa pilin [1]. In PilA from Pseudomonas aeruginosa, the conserved Pro22 residue leads to a kink in α1-N [16]. When over-expressed, PulG can form pilus filaments that have similar function to T4P [25]. If Pro22 is mutated in the pseudopilus PulG, pilus formation in K. oxytoca is significantly decreased, implying that the flexibility induced by Pro22 may be critically important for filament assembly [26]. The kink in α1-N has also been observed in recent crystal structures for the D. nodosus and F. tularensis pilins, with different crystallization procedures for the F. tularensis subunit implying that kinking in α1-N is natural [27].
Given the wealth of molecular data available on T4P and the importance of force in their biological roles, the powers of in silico methods were used to better understand the role of tension on T4P structure. MD simulations have been previously used to study filaments such as in the case of the actin filaments or microtubules [28], [29]. Complementing biophysical single protein pulling experiments, the effects of the application of external forces to proteins can also be studied using steered molecular dynamics (SMD) simulations [30], where external forces are applied to specific atoms in the biological system. Such SMD studies have been carried out on the adhesion protein FimH, a component of the related Type 1 pili [31], [32]. As T4P are known to sustain a considerable amount of force, understanding at the molecular level how T4P filaments respond under strain can provide insights into their function.
Therefore SMD simulations of GC-T4P using the 18 subunits long cryo-EM reconstruction were carried out to probe the dynamics of GC-T4P under tension, and to gain insights about the response of GC-T4P to external force at an atomistic level of detail. Even though the simulations are based on a 3-bundle model obtained from low-resolution cryo-EM experiments, predictions from simulation in agreement with experimental data would prove the possibility of such of model for the GC-T4P. The current study represents the first SMD simulation of a full pilus filament model, which would help contribute to the growing understanding of the wide variety of biological filaments found in nature. The aim of this computational study is to capture only the initial rearrangements of the filament coming under tension, as a full extension would be computationally prohibitive, in order to identify the strongest and weakest points of the filament structure. Structural changes in the GC-T4P filament, interactions between inter-subunit interfaces and residues that become exposed to the filament's external environment under tension, are discussed.
The pdb coordinates for GC-T4P were obtained from the Protein Data Bank entry 2HIL [17]. It consists of 18 individual pilin subunits, each subunit being 158 residues in length (Figure 1B,C). This system was placed in a water box using the VMD [33] plug-in, Solvate (using the TIP3P force field for the water model), and waters within 2.4 Å of the protein were removed. The water box dimensions were ∼100 Å×100 Å×350 Å. The system was brought to electrostatic neutrality using the VMD Autoionize plug-in to add 47 Na+ ions and 29 Cl− ions. Finally, additional water molecules within 2.7 Å of the alpha-helical core of the T4P system were removed to reduce the number of waters initially present in the filament core which is not expected to be filled with water [17]. This led to 287,272 atoms in the final system.
The package NAMD [34] was used with the CHARMM27 force fields [35] to carry out all simulations. Minimization was accomplished in two segments. First, 500 steps of minimization were carried out in which only the protein atoms were harmonically constrained, followed by 500 steps with all atoms in the system unconstrained. Subsequently to minimization, the system was equilibrated at a constant temperature of 310 K and a constant pressure of 1 atm with all atoms unconstrained for 500 ps. Constant temperature and pressure were maintained through the use of Langevin dynamics [36], with a Langevin damping coefficient of 5 ps−1 and a Langevin piston period of 0.1 ps, and periodic boundary conditions were used. The minimized and equilibrated structure served as the starting point for all simulations. A free simulation with no applied forces was carried out for an additional 20 ns with constant pressure and constant temperature maintained using the same parameters as in the 500 ps equilibration.
SMD simulations were carried out at constant velocity using the approach implemented in NAMD [34]. The spring constant was 500 pN/Å, and pulling velocities of 10 Å/ns, 5 Å/ns, 2.5 Å/ns, and 1 Å/ns were used. The four SMD simulations are referred to as T4P-v10, T4P-v5, T4P-v2.5, and T4P-v1 according to their pulling velocities. The SMD force was applied directly to the atoms in the top four subunits of T4P (Figure 1B) and the atoms in the first 30 residues of the four bottom-most subunits (Figure 1B) were fixed. Subunits that were not directly pulled on during the simulations are referred as the ‘bulk’ subunits. In all SMD simulations, pulling was stopped when the filament was stretched to the edge of its periodic box, which maintained a separation of approximately 30 Å between the filament and its periodic image along the z-dimension.
The length of a pilin subunit was defined as the distance between the center of mass of the alpha-carbon atoms for residues 1–3 and residues 51–53 for that subunit. The length of the complete GC-T4P filament was defined as the distance between the center of mass of the alpha-carbon atoms of residues 1–30 in each of the first four subunits (fixed selection) to the center of mass of the alpha-carbon atoms of residues 51 to 53 in the last four subunits (Figure 1B, subunits p7, p8, p9, p10) of the ‘bulk’ (i.e., excluding the four pulled subunits). Length extensions were defined as the difference between the GC-T4P length (or pilin subunit length) and its initial value.
The separations between globular heads were calculated as the distance between the center of mass of two subunit head domains. The distance between pilin subunits along the left-handed 3-start helices (subunit n, n+3, …) and the right-handed 4-start helices (subunit n, n+4, …) was calculated by finding the change in the z-coordinate of the center of mass of pairs of subunits (Figure 1A). The z-coordinate was used as the z-axis is approximately the central GC-T4P filament axis.
To measure bending angles in the N-terminal half of α1, each α1 helix was divided into three segments: residues 1–13, residues 15–21 and residues 23–53. The angle with Gly14 at its vertex was measured by defining a line of best fit for backbone atoms of residues 1–13 and residues 15–21. The angle made by these lines will be called θG. Similarly for Pro22, the angle, θP, was measured by calculating the angle made by the line of best fit for backbone atoms of residues 15–21 and residues 23–53, which have Pro22 as their vertex. 0 degree corresponds to a straight angle. A schematic of these angles can be viewed in Figure S1.
Contacts were identified as existing between any two residues, which had any atoms coming within 3.3 Å of one another. The number of contacts was then monitored over time. Only the number of contacts based on proximity were tracked, and not their type. The contacts were monitored separately between the α1-domain interfaces, and for globular head interfaces.
For quantities that can be measured for each of the pilin subunits (for example, the tail extension, or the angles θG and θP), data is presented as an average over the “bulk” subunits as defined above and also pictured in Figure 1B. For quantities which are measured for pairs of subunits (such as the separation between p3–p7, which is analogous by symmetry to the separation between p2–p6, p6–p10, etc.), an average over the value for all of the similar pairs is shown. For contacts between interfaces, a representative subunit from the filament, subunit p5 (Figure 1B) was chosen, with data from the T4P-v1 simulation mainly presented. The average number of contacts with all neighboring subunits whose interface involved the α1-domain were calculated. Similar calculations were performed for the contacts involving the globular head interfaces. Additional data for subunits p3, p4 and p6 are presented in the Supplemental Data, as well as data from simulations carried out at different pulling velocities.
The SASA for 5 amino acid long patches (5-mers) was calculated for each of ten frames over a period of 0.75 ns at the end of the T4P-v1 simulation that corresponds to an overall extension of the filament of 5%. The same criteria of 5% extension was also used to choose the frames over which the SASA calculation was performed for the other three SMD simulations. Similarly, SASA were calculated for the cryo-EM structure and over a period of 0.1 ns at the end of the free simulation. The final reported SASA values and their standard deviations were calculated by averaging over all ten frames, and then over the 10 “bulk” subunits (Figure 1B). SASA calculations were performed in VMD [33].
Polyclonal rabbit antibodies were raised against two regions of the pilin primary sequence around the regions that were thought to behave like SM1 in the molecular simulations (Genscript, Inc). Antibody #1 was raised against residues 94–108 (SSGVNNEIKGKKLSL) and antibody #2 was raised against residues 109–120 (WARRENGSVKWF). Those antibodies were further purified against bands of denatured pilins [37].
Pili were purified as previously published [8]. 50 µL of purified pili in 50 mM CHES buffer (∼100 µg/mL) were either added to 50 µL of 50 mM CHES buffer or to 50 µL of 2X Laemmli buffer. The first solution was a solution of T4P filament, after 5 minutes boiling the second was a solution of pilin subunits (denatured pili). Dot blots of 2 µL of either solution were blotted twice on nitrocellulose membranes. The membranes were blocked with 5% dry milk in TBS for one hour, then incubated overnight with either antibody #1 or antibody #2 (1/1,000 dilution), washed 3 times with TBST, incubated for one hour with goat anti-rabbit HRP secondary antibodies (1/5,000 dilution) and revealed using ECL reagents.
Either unstretched or stretched (transitioned) T4P purified from Neisseria gonorrhoeae were obtain in a modified molecular combing technique [12]. Briefly pili sheared from Neisseria gonorrhoeae MS11 were first unspecifically labeled with carboxytetramethylrhodamine (TAMRA), a red fluorophore. They were then let to interact with clean coverslips for 15 minutes at the bottom of a 6 well plate (2 ml of the solution per well/coverslip). They were then either dried by removing excess liquid with a lint free Kimwipes tissue while maintaining the coverslip to obtain stretched samples or let as is to obtain unstretched samples. All wells were then fixed with 4% formaldehyde and subsequently processed for immunostaining.
A free simulation and four SMD simulations of the GC-T4P filament were carried out. The simulations were started from the equilibrated structure as described in the Methods section. As observations across all pulling velocities were similar, results for the T4P-v1 simulation are mainly described. The pulling velocity applied to this system is 1 Å/ns, which is the slowest that is computationally achievable in a reasonable amount of time, even though it is still several orders of magnitude faster than experimental speed [30].
To further characterize changes in the GC-T4P filament, subunit-subunit interfaces for four pilin subunits in the ‘bulk’ of the filament (subunits p3, p4, p5 and p6, see Figure 1B) were studied. Changes seen for ‘bulk’ subunits during the SMD simulations are expected to be more representative of what would occur in the GC-T4P filament in in vitro pulling experiments. Results for contacts for subunit p5 are presented here, while representative results for p3, p4 and p6 and for p5 at the pulling simulation T4P-v2.5 are presented in the Supplemental Data. Subunits p5 and p6 share 1-start, 3-start and 4-start interfaces only with other ‘bulk’ subunits; they do not share any interface with either fixed or pulled subunits. Subunits p3 and p4 share interfaces with fixed subunits as well as with ‘bulk’ subunits.
The interactions in the core of the GC-T4P filament originate from the packing of the α1 domains against one another, and are thought to contribute to the incredible strength of bacterial pili. Each subunit's α1 domain has been proposed to make contact with the α1 domain of six other subunits by participating in three sets of ‘three-helix bundles’ based on the filament model describe by Craig et al. [17]. The SMD simulations demonstrate that such a model could allow for filament extension and underscore the strength of these non-covalent, and in many cases hydrophobic contacts between the α1 domains (Figure S7). Even as the filament and individual subunits extended in length (Figure 2 and 5), the contacts between α1 interfaces remained well conserved (Figure 7, S2 and S4), which have been suggested to provide stability to the filament [17], though such contacts might not be required for assembly, as it has been demonstrated that globular domains of the type IV pilins can assemble into fibers in vitro under certain sets of conditions [39]. These SMD simulations were performed to capture the initial steps of the elongation process, as elongations observed in experiments would be computationally prohibitive. Experimentally observed elongations would require an entirely new packing of the α1 domains to be realized. Coarse-grained simulations could provide insights to the nature of the packing in this extended conformation by probing numbers of contacts and residues in contact in the extended conformation.
One specific interaction, an inter-subunit Glu5 oxygen-Phe1 nitrogen hydrogen bond, is thought to be formed in order to neutralize charge in the filament core and increase hydrophobicity [17]. Additionally, in the crystal structure for the full PAK pilin filament a close contact in between the Phe1 backbone nitrogen and the Glu5 side-chain oxygen within a single subunit (intra-subunit interaction) was observed [16]. For the ‘bulk’ subunits inter-subunit hydrogen bonding between Glu5 and Phe1 was found to occur more frequently in the T4P-v1 simulation compared to the free simulation (Table 1), which could imply that the interaction also plays a role in maintaining stability in the core as the filament comes under tension.
Mutation of Pro22 in the pseudopilin PulG leads to a significant decrease in pilus formation in K. oxytoca, suggesting that the flexibility of α1-N around Pro22 may be critically important for pilus assembly [26]. Kinking of α1-N has also been observed in two more recent pilin crystal structures, suggesting that this bend is natural [27]. Additionally, it has been proposed that flexibility of the α1-N domain could lead to more efficient packing of pilin α1 helices within the filament core [16]. Fluctuations of angles θG and θP observed in the free simulation demonstrate the natural flexibility of α1, which could account for the effects observed upon assembly.
The observed elongations of the ‘bulk’ subunits (Figure 5) may represent initial stages of the transition to the force-induced conformation of GC-T4P that was recently observed experimentally [12]. The more extensive straightening of angles θG and θP in the T4P-v1 simulation may imply that the filament becomes less flexible as it is stretched, and that eventually all subunits in the pulled conformation become straightened.
In the experimentally determined stretched structure [12], the diameter of the filament decreases by 40% and its overall length increases by a factor of 3. In order to reproduce experimental data of this nature, the filament would need to be simulated for a much longer time and in a much larger water box along the z dimension (the filament axis), which would require simulations beyond the allowed time scale of all-atom MD. This computational study was designed to capture the initial rearrangements of the filament coming under tension, in order to identify the strongest and weakest point of the filament structure. While the actual extension of the filament involves a large increase in the axial rise per subunit, neither this feature nor a decrease in filament diameter (Figure S3) was captured by our simulations. However, elongations of α1 upon straightening of θG and θP could represent features in the initial stage of the transition towards the elongated GC-T4P conformation. Furthermore, the thinning of the filament observed experimentally may be the result of significant rearrangements of the pilin subunits that occur at timescales that these simulations cannot access.
The longer extensions of the filament observed experimentally would also require more extensive rearrangements of the pilin subunits than seen in the simulations. Conformational rearrangements between the globular head and the α1 domains of individual subunits are unlikely to produce longer extension; rather the slipping of α1 of one subunit along the α1 of adjacent subunits could produce such extension. In this case, neutralization of the charge of the Glu5 side-chain by hydrogen bonding to a residue on another subunit might become a concern, since of the first 23 residues of α1, only Glu5 is hydrophilic. However, residues 24 to 53 on α1 include side-chains available for hydrogen bonding or salt-bridge formation [17]. The α1 domain of a subunit could potentially slip out of its proposed 3-bundle interactions and translate up along the filament axis until Glu5 is able to interact with one of the adjacent subunit hydrophilic residues. To test this hypothesis, coarse-grained simulations would need to be carried out to study the filament at longer timescales. To further verify the models that would be obtained from such a computational approach, successful modeling of stretched pili based on cryo-EM would be useful, but such results are difficult to obtain.
In contrast to α1, the globular heads of the pilin subunits are considerably more free to move. Water exchange between the external environment and the GC-T4P core can occur in spaces between the globular heads in the free simulation (Figure 6A), as well as through the larger gaps formed between globular heads due to the application of pulling forces. Waters that enter through the surface gaps can proceed to interact with the buried pilin α1-domains, which are water accessible even in the free simulation (Figure 6B).
In the functionally related, but structurally different Vibrio cholerae type IVb pilus (VC-T4P), deuterium exchange experiments demonstrated that the D-region (in the globular head domain) was significantly exposed, and hence could not be buried by interaction with the αβ-loop (connecting the α1 domain to the β-sheets), suggesting that the αβ-loop had to interact with another region of an adjacent subunit [40]. A recent cryo-EM study of VC-T4P shed further light onto the differences between GC-T4P and VC-T4P, including that VC-T4P packing is not as tight as the packing of subunits in GC-T4P, and that in VC-T4P a segment of the pilin α1 domains are exposed through gaps along the filament surface [41]. In the GC-T4P model, both the D-region and the αβ-loop are already well-exposed to the environment [17], [38], though some polar interactions are present [17]. Fluctuations observed in the free simulation can further diminish contacts between globular heads, which include the contacts between the D-region and the αβ-loop in GC-T4P, even when the filament is not under tension (see Figure 7, S2 and S4) as observed experimentally for VC-T4P. Reduction of contacts at these interfaces supports that the globular heads are not packed too tightly against one another, which would potentially limit filament flexibility [2], [17].
Experimentally, it has been observed that T4P can bundle, creating larger filaments able to exert greater force [8]. Pilus bundle formation might be occurring by initial binding of one T4P filament to a surface, which would result in its extension under tension, followed by the association of additional filaments to the initial one [8]. However, the mechanism by which subsequent filaments associate to the first filament is unknown. The increased space between globular heads observed in the SMD simulations, demonstrated both by increases in head-head distances (Figure 2) and changes in the 3-start and 4-start inter-subunit axial distances (Figure 3 and 4), potentially provide locations along the filament surface that adjacent filaments could ‘dock’ into, in turn promoting the creation of the experimentally observed T4P bundles.
Finally, the increased spacing between globular heads produced along the filament surface in the SMD simulations (Figure 2B,C for T4P-v1) also leads to the exposure of the EYYLN sequence (Figure 8, 9, S5 and Movie S1) and S1 (SSGVNNEIKG). The interest of predicting these regions of exposure lies in the possibility of understanding the plasticity of GC-T4P filaments and to potentially developing drugs that target T4P functions during infection. As these sequences were exposed further in the SMD simulations compared to the free simulation, it is most likely a direct consequence of the forces applied to the system. Exposure of EYYLN is consistent with the experimental result in [12] which showed EYYLN could bind with an antibody in its force-transitioned conformation, but not in the absence of tension forces. Exposure of SSGVN under force was demonstrated experimentally following prediction from our simulation.
Exposure of EYYLN and SSGVN in the SMD simulations suggests that a model based on the 3-helix bundle can capture conformational changes in the T4P filament that have been previously observed in vitro [12] or demonstrated in this study. Because simulations of the experimentally observed elongation would be computationally prohibitive, here only the initial changes were probed. In vivo, filaments are dynamic, constantly alternating between retraction and elongation phases while releasing some of the force they are subjected to. Therefore, our simulations also suggest that EYYLN and SSGVN might become accessible early on under physiological conditions.
Conformational rearrangements of the GC-T4P filament under tension were studied utilizing MD simulations starting from the GC-T4P structure determined from a cryo-EM map and the crystal structure of a single GC pilin subunit. These studies were carried out in an effort to better understand the dynamics of the GC-T4P filament, its response to application of external forces and to probe initial stages of the transition between the relaxed and the tension-induced conformation.
Even though SMD simulations are based on 3-helix bundle model derived from low-resolution cryo-EM experiments, exposure of the sequences EYYLN and SSGVN, consistent with in-vitro experiments [12, present study], were observed. Therefore, such 3-helix bundle model could represent the actual structure of the filament. Simulations based on such a model reveal that the strength of the GC-T4P filament comes from the interactions between the α1 domains [17], as during elongation the contacts between these domains were well maintained. Contacts between subunit head domains decreased, creating additional gaps along the surface that could be related to filament bundling. These gaps lead to exposure of regions, which are hidden when not stretched, for potential drug targeting.
This work shows that SMD simulations can be used to narrow down the range of potential binding sites for drug therapy targeting bacterial filaments as the SSGVN was predicted as a possible site and confirmed experimentally. Finally, GC-T4P shares with the T4P from Neisseria meningitidis the presence of multiple post-translational modifications. As the functional importance of certain of these modifications is being discovered [21], simulations including the known modifications could shed more light on the function of the biological systems.
|
10.1371/journal.pgen.1005457 | Runx1 Transcription Factor Is Required for Myoblasts Proliferation during Muscle Regeneration | Following myonecrosis, muscle satellite cells proliferate, differentiate and fuse, creating new myofibers. The Runx1 transcription factor is not expressed in naïve developing muscle or in adult muscle tissue. However, it is highly expressed in muscles exposed to myopathic damage yet, the role of Runx1 in muscle regeneration is completely unknown. Our study of Runx1 function in the muscle’s response to myonecrosis reveals that this transcription factor is activated and cooperates with the MyoD and AP-1/c-Jun transcription factors to drive the transcription program of muscle regeneration. Mice lacking dystrophin and muscle Runx1 (mdx-/Runx1f/f), exhibit impaired muscle regeneration leading to age-dependent muscle waste, gradual decrease in motor capabilities and a shortened lifespan. Runx1-deficient primary myoblasts are arrested at cell cycle G1 and consequently differentiate. Such premature differentiation disrupts the myoblasts’ normal proliferation/differentiation balance, reduces the number and size of regenerating myofibers and impairs muscle regeneration. Our combined Runx1-dependent gene expression, ChIP-seq, ATAC-seq and histone H3K4me1/H3K27ac modification analyses revealed a subset of Runx1-regulated genes that are co-occupied by MyoD and c-Jun in mdx-/Runx1f/f muscle. The data provide unique insights into the transcriptional program driving muscle regeneration and implicate Runx1 as an important participant in the pathology of muscle wasting diseases.
| In response to muscle injury, the muscle initiates a repair process that calls for the proliferation of muscle stem cells, which differentiate and fuse to create the myofibers that regenerate the tissue. Maintaining the balance between myoblast proliferation and differentiation is crucial for proper regeneration, with disruption leading to impaired regeneration characteristic of muscle-wasting diseases. Our study highlights the important role the Runx1 transcription factor plays in muscle regeneration and in regulating the balance between muscle stem cell proliferation and differentiation. While not expressed in healthy muscle tissue, Runx1 level significantly increases in response to various types of muscle damage. This aligns with our finding that mice lacking Runx1 in their muscles suffer from impaired muscle regeneration. Their muscles contained a significantly low number of regenerating myofibers, which were also relatively smaller in size, resulting in loss of muscle mass and motor capabilities. Our results indicate that Runx1 regulates muscle regeneration by preventing premature differentiation of proliferating myoblasts, thereby facilitating the buildup of the myoblast pool required for proper regeneration. Through genome-wide gene-expression analysis we identify a set of Runx1-regulated genes responsible for muscle regeneration thereby implicating Runx1 in the pathology of muscle wasting diseases such as Duchenne muscular dystrophy.
| Striated muscles are highly organized structure composed of bundles of multinucleated myofibers. Each myofiber harbors peripheral nuclei and highly-organized myofibrils, granting the muscle its contractile force [1]. Muscle satellite cells (SC) comprise 2–5% of adult muscle cells [2]. Located at the myofiber periphery, SC are quiescent, myoblast-committed cells that serve as the muscle’s “stem cell” reservoir. Muscles subjected to regeneration-inducing damage, such as trauma or muscle dystrophy, use this reservoir to create new muscle fibers. Muscle regeneration involves the sequential induction of muscle-specific transcription factors (TFs), including the myogenic regulatory factors (MRFs) Myf5, Myod, Myog and Mrf4. Proliferating myoblasts express Myf5 and Myod, whereas Myog is induced at the onset of differentiation and drives myoblast terminal differentiation [3]. Yet, the role of Runx1 TF in muscle regeneration remains to be determined.
Runx1 is a member of the RUNX family of TFs, which regulate cell lineage determination in several developmental pathways [4]. While Runx1 is not detected in naïve embryonic developing muscle [5,6] or in adult muscle tissue [7], it is highly expressed in muscles exposed to myopathic damage. RUNX1 expression was found to be significantly increased in samples of muscle dystrophies, including mouse models of Duchenne muscular dystrophy (DMD) [8] and amyotrophic lateral sclerosis (ALS) [9], myopathy patients (including EDMD, DMD, AQM [10]) and in cardiotoxin (CTX)-treated muscle [11]. Genome-wide ChIP-seq analysis using C2C12 cells revealed enrichment of RUNX and AP-1 motifs at MyoD-bound regions [12]. Runx and AP-1 motifs were also enriched in C2C12 cell MyoD-bound enhancers [13], and several genomic loci co-occupied by MyoD and AP-1 factor c-Jun also bound Runx1 [13]. Based on these findings in C2C12 cells, it was suggested that Runx1, MyoD and c-Jun assemble on the same regulatory regions, to promote myoblasts differentiation. However, other experiments involving myoblastic or transformed cell lines led to conflicting conclusions regarding the role of Runx1 in myoblasts. Inhibition of Runx1 activity in C2C12 either directly or by knockdown of its obligatory cofactor Cbf-β or led to enhanced differentiation [14]. On the other hand, similar enhanced differentiation was observed upon forced expression of Runx1 in rhabdomyosarcoma cells [15]. These data suggested that Runx1 could function as both repressor or activator of myoblast differentiation.
To investigate the function of Runx1 in muscle regeneration in a direct in vivo approach, we first generated mice lacking muscle Runx1 (Runx1f/f). Using these mice we found that Runx1 is switched on in response to muscle damage and participates in muscle regeneration by preventing premature myoblasts differentiation. Moreover, when crossed onto the DMD mouse model (mdx mice), the Runx1-deficient mdx mice (mdx/Runx1f/f) encountered defects in muscle mass and muscle strength that are not part of the mdx phenotype thereby highlighting the involvement of Runx1 in muscle regeneration. At the cellular level mdx-/Runx1f/f mice showed impaired myoblast proliferation that impeded muscle regeneration and contributed to the severity of muscle deterioration. Genome-wide analyses of Runx1f/f primary myoblasts (PM) revealed that PM Runx1 cooperates with MyoD and c-Jun to transcriptionally regulate a subset of genes that prevent premature myoblast differentiation. These data add unique insight on the transcriptional program driving muscle regeneration and implicate Runx1 as an important participant in the pathology of muscle-wasting diseases.
As noted above, Runx1 RNA expression was reported previously in various types of human muscle diseases including ALS and DMD and their respective mouse models tg-mSOD1 and mdx. Immunohistochemistry (IHC) analysis of gastrocnemius muscles by anti Runx1 antibodies (Ab) revealed no signal in untreated wild-type (WT) muscle (Fig 1A) and in developing muscle (S1A Fig), whereas it was readily detected in tg-mSOD1 muscles (Fig 1B) and in denervated muscles (see S2 Fig). Significantly, Runx1 was also readily detected in nuclei of regenerating CTX-treated or mdx muscles (Fig 1C and 1D). This observation suggests that Runx1 participates in muscle regeneration, an interpretation further supported by the presence of a cell population that co-expressed Runx1 and the SC-expressing TF Pax7 (Fig 1E), indicating that Runx1 is expressed in SC during regeneration. Finally, Runx1 expression was also observed in cultured PM (Figs 1F and S1A). Of note, all the Runx1+ cells in these PM cultures also expressed Pax7 (S1B Fig). Thus, muscle Runx1, which is not expressed during development or in resting WT muscle, is activated in response to either neuronal-mediated muscle damage, or myonecrosis.
To elucidate Runx1 function during muscle regeneration, we first created mice lacking muscle Runx1 by crossing Runx1LoxP/LoxP (Runx1L/L) mice [16] onto transgenic Myf5::Cre mice that express Cre in early muscle development and regeneration [17] (S2A Fig, left panel, Runx1f/f). As Runx1 expression was previously reported to be elevated in denervated muscle [7], we determined the levels of muscle specific Runx1 mRNA (S2B and S2C Fig) and protein (S2D Fig) in denervated muscle and thymus of Runx1f/f mice compared to WT Runx1L/L mice. Runx1 RNA and protein levels were elevated in the denervated WT muscle, yet its levels did not change upon denervation of the Runx1f/f muscle (S2B–S2D Fig). No significant differences were observed in thymi of Runx1L/L or Runx1f/f mice (S2B–S2D Fig). Of note, while Myf5::Cre is active from early stages of muscle development, loss of Runx1 is actually confined to fibers responding to muscle damage. Indeed, body weight and myofiber size of Runx1f/f and Runx1L/L littermate mice were similar (S2E and S2F Fig). The muscle specific Runx1-deficient mice (Runx1f/f) were then crossed onto an mdx mice to generate mdx mice lacking muscle Runx1 (S2A Fig, right panel, mdx/Runx1f/f). Muscle specific ablation of Runx1 was verified in affected muscles of mdx/Runx1f/f compared to mdx/Runx1L/L mice (S2G Fig). The mdx/Runx1f/f mice represent a useful model for investigating the role of Runx1 in muscle regeneration in vivo. As mdx mice lack dystrophin expression the mice undergo recurrent cycles of muscle necrosis and regeneration. However, in contrast to human DMD patients who encounter muscle waste and paralysis at early childhood and die during the second or third decade of their lives [18], mdx mice exhibit extensive muscle regeneration, resulting in no loss of muscle mass, and have a normal life span (reviewed in [19]). Analysis of various litters showed that body weight of neonate and juvenile mdx/Runx1f/f are comparable to those of mdx/Runx1L/L littermates. However, starting at the age of 2 months, mdx/Runx1f/f mice did not gain weight, unlike their mdx/Runx1L/L littermates (Fig 2A). As a result mdx/Runx1f/f mice became underweighted compared to Runx1L/L, Runx1f/f or mdx/Runx1L/L mice. In addition, mdx/Runx1f/f mice were notably smaller and thinner when reaching maturity (6–7 months) (Fig 2B). To evaluate the mechanism underlying this weight differences between mdx/Runx1L/L and mdx/Runx1f/f mice, we monitored their relative lean weight. Compared to mdx/Runx1L/L mice, the mdx/Runx1f/f mice have lower lean weight starting at 4 month of age and throughout the recording period (Fig 2C), suggesting that the mdx/Runx1f/f mice bear loss of muscle mass. (Fig 2C). This loss of muscle mass is consistent with the possibility that Runx1 plays a role in mdx related muscle regeneration.
Because loss of Runx1 seemed to affect mdx related muscle regeneration we assessed whether lack of muscle specific Runx1 will affect the life span of mdx/Runx1f/f mice. Kaplan-Meyer survival analysis revealed that life expectancy of Runx1f/f and mdx/Runx1L/L mice was similar to that of WT mice, whereas mdx/Runx1f/f exhibit a significantly (p = 4e-291, χ2 test) shorter life span with deaths occurring as early as at 12 weeks of age, with a median survival age of 28.5 weeks (Fig 2D). Histological analysis of diaphragm muscles of mice at the median survival age revealed extreme muscle deterioration, with extensive fibrosis and a pronounced decrease in diaphragm size (Fig 2E). We therefore postulate that the likely cause of death was respiratory failure. This profound reduction in life span underscores the contribution of Runx1 to the muscle pathology observed in mdx/Runx1f/f mice.
To characterize the muscular dystrophy of the mdx/Runx1f/f mice, we compared its muscle strength to that of mdx/Runx1L/L strain. It was previously reported [20] that mdx mice exhibits a transient decline in muscle strength at juvenile stages, which dramatically improves in mice older than 2 months. We therefore monitored muscle strength by recording treadmill performance of mice from the age of 2 to 7 months. Follow-up post-hoc comparisons (Bonferroni corrected for multiple comparisons) revealed no significant differences between WT, Runx1f/f and mdx/Runx1L/L mice at all time points. Conversely, mdx/Runx1f/f mice reached exhaustion significantly faster (p<0.01) than WT, Runx1f/f or mdx/Runx1L/L mice (Fig 2F). We further evaluated muscle performance by the grip strength test, which measures the maximal force a mouse can apply when gripping a rod with its forelimbs. Again, mdx/Runx1f/f mice exhibited a significant reduction (~50%, p = 2e-5, student t-test) in muscle strength compared to mdx/Runx1L/L mice (Fig 2G, left), regardless of their muscle mass (Fig 2G, right). These results indicate that the impaired muscle performance of mdx/Runx1f/f mice is due not only to shear loss of muscle mass, but also due to a reduction in capabilities of the remainder muscle tissue. Together, the complementary outcome of these muscle strength experiments demonstrated the importance of muscle Runx1 to mdx related muscle regeneration.
We next investigated whether the muscle wasting encountered by mdx/Runx1f/f mice involves a decrease in the number (i.e. regeneration defect) or size (i.e. enhanced atrophy/inability to produce proper hypertrophy) of myofibers. Analysis of soleus muscles of 8-week-old mdx/Runx1f/f mice revealed a significant reduction in the number of total myofibers compared to mdx or mdx/Runx1L/L mice (Fig 3A and 3B). Importantly, the amount of centrally nucleated myofibers, a hallmark of regenerative muscle tissue, was significantly decreased in soleus muscle of mdx/Runx1f/f mice both in terms of absolute numbers (70.2±20.58 vs. 344.38±25.36, p = 5.5e-8, unpaired student t-test) and percentage of total fibers (18.75±3.98% vs. 49.72±3.18%, p = 5.6e-6, unpaired student t-test) (Fig 3B). Interestingly, a similar reduction in centrally nucleated myofibers was also observed in CTX-treated muscle of Runx1f/f mice (S3A and S3B Fig). This data supports the possibility that loss of Runx1 leads to a decrease in muscle regeneration in mdx/Runx1f/f mice.
To evaluate whether the profound muscle waste in the mdx/Runx1f/f mice could be attributed to the ability of Runx1 to attenuate muscle atrophy, as previously observed in denervated muscle [21], we determined the total myofiber size, i.e., the average cross-sectional area (CSA), in the soleus and gastrocnemius muscles. We found no significant reduction in myofiber CSA in either muscle type of mdx/Runx1f/f mice as compared to mdx/Runx1L/L mice (Fig 3C and 3D). This finding indicates that a Runx1 function other than its role in muscle atrophy must be the underlying cause for the striking muscle waste in mdx/Runx1f/f mice. Indeed, when the regenerating myofibers were recorded separately by counting the centrally nucleated myofibers, a significant CSA reduction was noted in mdx/Runx1f/f mice muscles compared to those of mdx/Runx1L/L mice (Fig 3C and 3D). Moreover, quantitative analysis of CSA distribution revealed a significant increase of small myofibers fraction in the mdx/Runx1f/f muscles, which was more pronounced in the centrally nucleated myofiber subset (Fig 3E and 3F). The decrease in the CSA of centrally nucleated myofibers and the change in CSA distribution indicate that regenerating myofibers in mdx/Runx1f/f mice were formed by fusion of a smaller number of myoblasts, conceivably due to decreased myoblast proliferation in mdx/Runx1f/f muscles. Interestingly, a similar reduction in CSA of centrally nucleated myofibers was found in the CTX- treated muscles of Runx1f/f compared to Runx1L/L mice (S3A and S3C Fig). To directly address whether the muscle regeneration deficit of mdx/Runx1f/f and Runx1f/f mice was due to impaired myoblast proliferation, we recorded cell proliferation by BrdU staining. A significant decrease in the number of BrdU+ cells was observed within the damaged muscle of mdx/Runx1f/f compared to mdx/Runx1L/L muscle (S3D and S3E Fig). The reduction in proliferating myoblasts was also manifested in decreased number of Pax7+ cells in regenerating muscle of mdx/Runx1f/f mice compared to mdx/Runx1L/L mice (S3F and S3G Fig). We then examined SC proliferation by co-staining muscles of mdx/Runx1L/L and mdx/Runx1f/f mice with anti-Pax7 and anti-Ki67 Ab (S3H and S3I Fig). Significantly, marked reduction in the number of double positive Pax7+/Ki67+ cells was noted in muscle of mdx/Runx1f/f mice compared to mdx/Runx1L/L mice (S3I Fig, left panel). Moreover, the percentage of proliferating Pax7+ cells within the total SC population was also markedly reduced in muscles of mdx/Runx1f/f mice compared to mdx/Runx1L/L mice. Together, the complementary results obtained using anti- BrdU, Pax7 and Ki67 Ab demonstrate a reduced proliferation capacity of SC in regenerating muscle of mdx/Runx1f/ mice compared to mdx/Runx1L/L mice. Similar phenotype was observed in CTX-treated muscles of Runx1f/f mice (S3J and S3K Fig). Collectively, these findings suggest that Runx1 promotes muscle regeneration-associated myoblast proliferation and loss of Runx1 in mdx/Runx1f/f or Runx1f/f impairs muscle regeneration causing marked muscle wasting in the mdx/Runx1f/f mice.
The impaired muscle regeneration seen in both CTX-treated Runx1f/f and mdx/Runx1f/f mice is compatible with the notion that loss of Runx1 in SC-derived myoblasts leads to proliferation defects. We directly examined this possibility by culturing PM from Runx1f/f and Runx1L/L littermates under proliferation-inducing conditions. Runx1f/f PM proliferation was attenuated as indicated by the significantly longer doubling time compared to Runx1L/L PM (Fig 4A). This prolonged doubling time resulted from Runx1f/f PM arrest in the G1 phase (Fig 4B–4D). Prolonged doubling time resulted from impaired cell cycle progression was also observed in adenovirus-Cre-GFP (Adeno-Cre)-infected Runx1L/L PM (S4A–S4C Fig), underscoring the finding that lack of Runx1 is the cause for this phenotype. We then used PI staining to determine whether the cell cycle progression impairment was associated with cell death. Comparing, No reduction in Runx1f/f PM viability compared to Runx1L/L PM was noted (S4D and S4E Fig), indicating that loss of Runx1 did not induce myoblasts death.
We also evaluated the role of Runx1 in PM differentiation by analyzing the expression of myosin heavy chain (MHC), a myofiber differentiation marker using immunofluorescence (IF). Compared to Runx1L/L, the Runx1f/f cultures contained a significantly higher number of MHC-positive, multinucleated myofibers (Fig 4E). This Runx1-/--dependent phenotype was further characterized by counting the number of fusion events in proliferating Runx1L/L and comparing it to that measured in Runx1f/f PM cultures (Fig 4F). Runx1f/f myoblasts displayed a significantly lower number of mononuclear cells and a two-fold increase in the amount of multinucleated myofibers (34.28±6.2% vs. 17.13±2.8%; p = 0.023). Similar results were obtained with cultured Runx1L/L PM infected with Ad-Cre-GFP (S4F Fig). Together, the attenuated proliferation and spontaneous differentiation of Runx1f/f PM, suggest that Runx1 participates in myoblast cell-fate decision to proliferate or differentiate and when lost the normal proliferation/differentiation balance is disturbed.
Because Runx1 affects the PM proliferation/differentiation balance, we questioned whether ectopic Runx1 expression inhibits myoblast differentiation. Cultured PM were infected with either Runx1-expressing Ad-Runx1-GFP or Ad-GFP viral constructs. IF analysis revealed fewer multinucleated MHC-expressing myotubes in the Ad-Runx1-GFP infected culture compared to those of Ad-GFP (Fig 4G and 4H), indicating that ectopic expression of Runx1 causes delayed differentiation. This result correlates with the reciprocal effect of Runx1-/-, which manifested in enhanced PM differentiation (Figs 4E, 4F and S4D). To further characterize this ectopic Runx1-induced delayed differentiation phenotype, we analyzed expression of Myog and Mef2c TFs and the sarcomeric genes Myomesin, Troponin T, Myh2 and Myh8, which are induced during myoblast differentiation. RT-qPCR analysis revealed reduced expression of these genes in ectopically expressing Runx1-differentiating PM (Fig 4I and 4J), supporting the observation that Runx1 expression delays myoblast differentiation.
The high expression level of Runx1 in proliferating PM (Fig 1F), prompted us to conduct a complementary assessment of its levels during PM differentiation. Western blot analysis indicated that Runx1 levels decrease during differentiation (S5A and S5B Fig) and that addition of the proteasome inhibitor Bortezomib attenuated this decline (S5C and S5D Fig). In contrast, Runx1 RNA levels did not significantly change during myoblast differentiation, as determined by RT-qPCR (S5E Fig). We therefore conclude that Runx1 is actively degraded in differentiating myoblasts and that this breakdown facilitates myoblast differentiation.
We next investigated the Runx1-mediated transcriptional program involved in the early stages of muscle regeneration. A genome-wide analysis of cultured PM was perform following the strategy described in Fig 5A. First, we identified Runx1-responsive genes by comparing gene expression profiles of Runx1L/L and Runx1f/f PM. Runx1-responsive genes were defined using an FDR q-value <0.1 and >1.5-fold change as the significant threshold. The analysis revealed 636 differentially expressed genes, of which 478 were upregulated and 158 were downregulated in Runx1f/f PM (Fig 5B and S1 Table). This Runx1-responsive gene-subset contains genes known to play a role in the myoblast proliferation/differentiation balance including, the MRFs Myf5, MyoG and Mef2c (Fig 5B). We also noted a change in the expression levels of muscle structural genes, including Myomesin (Myom2) and Troponin (Tnnt2) isoforms that compose the sarcomere and are activated in differentiated myoblasts (Fig 5B). Other genes that were found to be differentially expressed included the signaling pathway-related genes Dlk1 of the Delta-Notch and the Igfbp2 of the Igf-1/PI3K/Akt pathways, and Cdkn1c a cell-cycle regulator encoding the cyclin-dependent kinase-inhibitor p57Kip2 protein (Fig 5B). Of note, in contrast to the considerable role played by Cbf-β in C2C12 myoblastic cell-line differentiation [14], no significant change was noted in Cbf-β levels in Runx1f/f PM (see S1 Table). These results suggest that in PM, Runx1 regulates the expression of MRFs, sarcomeric genes and cell cycle-control genes, thereby promoting myoblast proliferation and attenuating their differentiation.
Runx1-responsive gene analysis represents changes in genes that are either direct or indirect targets of Runx1. To identify genes directly regulated by Runx1 in PM we conducted Runx1 ChIP-seq using proliferating PM, which enable us to single- out Runx1-responsive genes that are bound by Runx1 (Fig 5A). Runx1 occupancy pattern displayed enrichment at promoter regions, defined as 1 kb upstream and downstream from an annotated transcription start sites (TSS). However, most Runx1-bound regions were distal to annotated TSS; 42% were located within 10–100 kb from TSS, and 15% were found in “gene deserts” (>100 kb from any TSS) (Fig 5C). All in all, most Runx1-bound regions (85% of Runx1 chip-seq peaks) were located within 100 kb from known TSS (Fig 5C) with more than 94% of the peaks located up to 200 kb from known TSS. A similar Runx1 occupancy pattern was observed in differentiating megakaryocytes [22] and in hematopoietic progenitors [23]. In C2C12 myoblasts, the median enhancer-TSS distance was defined as ~53kb [13]. We then identified the Runx1-regulated genes by cross-analyzing the Runx1-responsive gene dataset with the ChIP-Seq results. Out of the 636 Runx1-responsive genes, 83% contained Runx1-bound regions within 200kb from their TSS; these 531 genes were considered as Runx1-regulated genes (Fig 5D, S2 Table). A partial list of Runx1-regulated muscle-relevant genes is presented in S3 Table. Collectively, the gene expression and ChIP-seq analyses indicate that during myoblasts proliferation, Runx1 regulates muscle-specific genes that encode MRFs and structural proteins and that it may serve as a component of the Igf-1/PI3K/Akt and Delta-Notch pathways.
De-novo TF motif analysis of the Runx1-bound regions revealed a significant enrichment of the canonical RUNX motif (p = 4.3e-935) as well as MyoD and AP-1 TF motifs (Fig 5E). In fact, over 95% of Runx1-bound regions contained at least one RUNX motif. Previous studies have found cooperation between AP-1 and Runx1 in proliferating megakaryocytes [22], an enrichment between of Runx and AP-1 motifs in C2C12 cells [12] and enrichment of c-Jun and Runx1 that are recruited by MyoD to several muscle specific enhancers in C2C12 cells [13]. These findings prompted us to analyze the proliferating PM ChIP-seq data for TF module enrichment, defined as TF motifs within 50bp spanning the bound RUNX motif. Analysis revealed Runx-Runx, Runx-MyoD and Runx-AP1 to be among the most enriched modules (Fig 5F). The preponderance of the two latter modules further supports the possibility of cooperation between Runx1, MyoD and AP-1 TFs in driving the transcription program that regulates PM proliferation/differentiation balance. Analysis of Runx1-bound regions using the GREAT program [24], which predicts meaningful biological functions from the landscape of TF-bound regions, indicated enrichment for many skeletal muscle-related terms and relevant signaling pathways (S4 Table). The enrichment of MyoD and c-Jun motifs, as well as Runx-MyoD and Runx1-Ap1 modules, in Runx1 ChIP-seq data suggests that these TFs cooperate during muscle regeneration.
As noted above, unbiased de novo motif-finding analysis of Runx1-bound regions in PM revealed a significant enrichment of MyoD and AP-1 motifs as well as Runx-MyoD and Runx-Ap1 modules. To obtain a better understanding of the Runx1-mediated transcriptional program and derive the signature of Runx1 in proliferating PM we characterized the regulatory regions bound by the TFs Runx1, MyoD and c-Jun (RMJ). We started by performing independent MyoD- ChIP-seq using proliferating PM (Fig 5A). While MyoD binding was enriched at the promoter regions, it was more abundant at TSS distal regions (S6A Fig), as was also observed by Cao et al in C2C12 cells [12]. Motif-finding analysis of MyoD-bound regions revealed the MyoD motif to be the most enriched followed by the AP1 and Mef2a motifs (S6B Fig). ChIP-seq data analysis revealed a significant overlap between Runx1-occupied regions and those bound by MyoD (S6C Fig). Specifically, 46% of Runx1-bound regions were co-occupied by MyoD in PM (p<1e-4, bootstrap test) (S6C Fig). This co-occupancy of Runx1 and MyoD suggests a genome-wide cooperation of the two TFs in PM.
To further define the TF myoblast regulatory regions, we examined the genome-wide binding pattern of c-Jun in PM. Since c-Jun was implicated as a negative regulator of differentiation in the myoblastic C2C12 cell line [25,26,27] and was shown to co-bind with Runx1 and MyoD at genomic loci in these cells [13], we first examined its expression in differentiating PM. Interestingly, c-Jun mRNA and protein levels gradually decreased during myoblast differentiation (S7A and S7B Fig), reminiscent of the Runx1 decay noted before (S5 Fig). This finding corroborates the possibility that c-Jun and Runx1 cooperate during myoblast proliferation, prompting us to perform a c-Jun ChIP-seq in proliferating PM. Motif analysis revealed that c-Jun-bound regions are highly enriched for AP-1, RUNX and MyoD motifs (S7C Fig). Moreover, 47% of the c-Jun-occupied regions are co-bound by Runx1 (S7D Fig), and a substantial number of peaks were bound by all three TF (Fig 5G, p<1e-4, bootstrap test). We further characterized the Runx1-cJun co-occupied regions by conducting c-Jun ChIP-seq using PM lacking Runx1 (Runx1f/f PM). Interestingly, qPCR analysis revealed a pronounced reduction in bound c-Jun at several Runx1f/f PM loci compared to WT PM loci (S7E Fig). Of note, the observed decrease in c-Jun binding upon loss of Runx1 was not due to a reduction in c-Jun protein levels (S7F Fig). These findings support the notion that Runx1 plays a role in recruiting c-Jun to at least a portion of their co-bound sites.
Cross-analysis of the RMJ-bound genomic regions with the Runx1-responsive gene subset yielded a significant (2e-16, hypergeometric test) list of 408 genes highly enriched for muscle-related GO terms (S5 Table), designated RMJ-regulated genes (Fig 5A). Importantly, this gene subset includes a preponderance of Runx1-repressed genes (S8A Fig), along with genes involved in myoblast proliferation and/or differentiation (S8B Fig). To further characterize the RMJ-regulated gene subset, we analyzed the epigenomic status of Runx1- and RMJ-bound regions in PM (Fig 5A). First, we performed ChIP-seq of H3K4 monomethylation (H3K4me1) and H3K27 acetylation (H3K27ac). These two histone modifications are known to mark active enhancer loci (reviewed in [28]). Analysis revealed that over 70% of Runx1-bound and 90% of RMJ-bound regions overlap with the histone marked enhancer subset (S9A and S9B Fig), underscoring the notion in PM that the three TFs occupy a subset of myoblast active enhancers that form the core of Runx1-mediated regulatory network. A more stringent analysis that enabled the identification of nucleosome-free open chromatin was achieved using the recently developed assay of ATAC-seq [29]. This evaluation showed that ~ 25% and ~40% of Runx1- and RMJ-bound regions, respectively, have a nucleosome-free structure (S9C and S9D Fig). We then cross-analyzed the combined RMJ-bound ChIP-seq data of histone-marked enhancers and ATAC-seq open chromatin regions with the 636 Runx1-responsive expressed genes (see Fig 5B). This analysis yielded a subset of 229 high-confidence Runx1-regulated genes (p<1e-15, Monte Carlo FDR) that responded to the loss of Runx1 and had RMJ-bound to adjacent nucleosome-free histone marked enhancers. This subset includes a number of major muscle regulatory and structural genes, including Myog, Myh2, Tnnt1 and Myom2, the signal transduction-related genes Dlk1, Hey1 (Delta/Notch pathway) and Igfbp2, Igfbp3 and prkcd (Igf-1/AKT/mTor pathway) (Fig 5H). The finding of H3K4me1 and H3K27Ac, which mark active enhancers, at Runx1 bound loci that mediate gene silencing was puzzling. These Runx1 bound loci could represent poised enhancers that are activated upon differentiation. To test this possibility, we performed H3K27me3 and H3K4me1 qChIP that when occurred together mark poised enhancers [30,31]. While all examined loci were enriched for H3K4me1 (S9E Fig), in most of them the level of H3K27me3 in differentiated WT PM decreased (S9F Fig). Comparison of these loci in proliferating Runx1f/f PM to WT PM revealed similar pattern in some, but not all the loci (S9F Fig). This finding might reflect the heterogeneity of Runx1f/f PM cultures that contain both proliferating and differentiating myoblasts (See Fig 4A–4D).
To further evaluate the participation of Runx1 in regulation of the high-confidence Runx1-regulated genes we examined four RMJ-bound active enhancers regions located at the vicinity of Myog, Tnnt1, Myh8 and Myom2 gene loci. These four Runx1-responsive genes play key roles in muscle development and regeneration [3]. The four genomic regions were cloned into pTK-Luc reporter construct and evaluated by ectopic expression. Following co-transfection with Runx1 expression vector into HEK293 cell line, the four regions conferred Runx1-dependent repression of the basal promoter activity (S10A Fig). This Runx1-dependet repression was abrogated by mutations in the RUNX binding site of Myog and Tnnt1 constructs (S10B Fig). The intact and mutated reporter constructs were also transfected into PM cultures normally expressing Runx1 (Fig 2F). In comparison to the mutated construct, activity of the intact construct was repressed (S10C Fig), presumably by the endogenous PM Runx1 binding to the intact RUNX motif.
Taken together, the histone-marked-enhancer and ATAC-seq-open-chromatin regions, the RMJ ChIP-seq and the differential gene expression results have stringently identified, at high confidence, a subset of MyoD-bound Runx1-regulated genes. These data suggest that in proliferating PM Runx1 cooperates with c-Jun to repress MyoD-activated genes that drive myoblast differentiation, and thereby participates in maintaining a proper proliferation/differentiation balance. In the absence of Runx1, this delicate equilibrium is disrupted resulting in impaired muscle regeneration.
To gain insight into Runx1 activity in muscle regeneration in vivo, we analyzed the transcriptional program of Runx1 in affected mdx muscles. Gene expression profiles in mdx/Runx1L/L and mdx/Runx1f/f muscles were determined by RNA sequencing (RNA-seq). Runx1-responsive genes were defined using an FDR q-value <0.05 and >1.7-fold change as the significant threshold. Analysis revealed 1432 differentially expressed genes, with 1240 and 192 genes were up- and down- regulated, respectively, in mdx/Runx1f/f soleus muscles (Fig 6A and S6 Table). Among the mdx Runx1-responsive genes were muscle-related genes such as Myog, sarcomeric structural genes (Myh7, Myh4, Myh3 Mybph, Tnnt2) and signal transduction pathway-related genes such as Igf1, Igfbp4, Igfbp6 (Igf-1/PI3K/Akt pathway) Dlk1 (Delta- Notch pathway) and Mstn (Myostatin- of the Tgf β family). This list corresponds with the Runx1-responsive gene subsets found in PM.
To identify genes directly regulated by Runx1 in DMD-induced muscle regeneration, we cross-analyzed the mdx/Runx1f/f gene expression data with the Runx1f/f PM RMJ-regulated gene subset. This analysis yielded 62 genes (Fig 6B and S7 Table), representing a high-confidence Runx1-regulated gene subset in mdx/Runx1f/f soleus muscle. Fig 6C–6E depicts examples of ChIP-seq and RNA-seq readouts of three high-confidence Runx1-regulated genes Myog, Dlk1 and Mstn all known to participate in myoblast proliferation/differentiation balance. The expression of these genes was verified by RT-qPCR in Runx1f/f PM, mdx muscles and PM overexpressing Runx1 (Fig 6C–6E right panels). In summary, using differential gene expression acquired in the mdx mice combined with the PM-derived expression and ChIP-seq data we were able to identify a subset of high-confidence Runx1-regulated genes participating in myoblast proliferation/differentiation balance. In mdx mice lacking Runx1, normal regeneration is impaired, leading to the adverse muscle waste phenotype of mdx/Runx1f/f mice.
Muscle regeneration following injury is mediated by the activation, proliferation and differentiation of adult SCs [2]. Maintaining the balance between myoblast proliferation and differentiation is crucial for proper muscle regeneration heighted by the fact that insufficient proliferation causes a reduction in myoblast pool leading to incomplete reconstitution of muscle mass. Indeed, disruption of the proliferation/differentiation equilibrium results in an impaired regeneration phenotype characteristic of muscle-wasting diseases [32,33,34].
In this study we perform an in-depth investigation of the function of Runx1 in muscle regeneration and its role in regulating myoblast proliferation/differentiation balance. While Runx1 is not expressed in normal healthy muscle, its expression is highly induced by different types of muscle damage. Muscle-specific ablation of Runx1 in mdx/Runx1f/f mice impairs muscle regeneration in vivo. This diminished regeneration causes a decrease in the number and size of regenerating myofibers, leading to loss of muscle mass and motor capabilities. Similarly, CTX-induced muscle damage in muscle-specific Runx1-deficient Runx1f/f mice results in decreased myoblast proliferation relative to Runx1L/L mice. Consequently, the number of regenerative fibers and their size in CTX-treated Runx1f/f mice are reduced compared to Runx1L/L mice. At the cellular level, loss of Runx1 delays PM proliferation by affecting their cell cycle: Runx1f/f PM linger in the G1 phase and consequently, spontaneously differentiate. Interestingly, differentiation of WT PM is associated with gradual Runx1 degradation, suggesting that Runx1 tapering plays a role in the progression of myoblast regeneration. The finding that forced expression of Runx1 reduces PM capacity to differentiate, supports this notion. These results indicate that Runx1 prevents premature differentiation of proliferating myoblasts, thereby facilitating the buildup of the myoblast pool required for proper regeneration. Upon induction of differentiation, Runx1 is degraded allowing myoblasts to differentiate (Fig 7).
Having shown the pivotal role of Runx1 in regulating the balance of myoblast proliferation/differentiation, we used cultured PM to derive a Runx1 genome-wide occupancy pattern and identify its regulated genes during the early stages of muscle regeneration. Sequence analysis of Runx1-occupied regions revealed enrichment for the RUNX, MyoD and AP-1/c-Jun motifs. This finding corresponds with the observation that in the C2C12 myoblastic cell line, Runx1, MyoD and c-Jun co-bind to the same genomic loci [13] and supports the possibility that in PM the three TFs cooperate to prevent premature myoblast differentiation. ChIP-seq of Runx1-, MyoD- and c-Jun-occupied regions revealed Runx1-responsive genes bound by RMJ to be highly enriched for genes involved in myogenesis. These findings underscore the notion that Runx1 cooperates with MyoD and c-Jun to attenuate myoblast differentiation. Specifically, it suggests that during early regeneration RMJ cooperate to activate PM proliferation genes and repress genes that drive myoblast differentiation, thereby affecting the proliferation/differentiation balance. This could occur through repression by Runx1 and c-Jun of MyoD pro-differentiation target genes. Following Runx1 and c-Jun degradation, repression is relieved, allowing MyoD-mediated differentiation to proceed [12].
We then derived differential gene expression of mdx/Runx1L/L and mdx/Runx1f/f muscles and cross-analyzed this data with the RMJ-regulated gene subset of Runx1f/f PM. This analysis singled out a small subset of 62 genes, which we defined as in vivo high-confidence Runx1-regulated genes. This subset included several groups of genes known to affect myoblast proliferation/ differentiation balance providing clues regarding the mechanisms underlying the function of Runx1 in muscle regeneration. For example the muscle-related TFs, Myog and Mustn1. Myog (encoding Myogenin) is a myoblast differentiation-promoting MRF [3]. Thus, its repression by Runx1 would prevent premature myoblast differentiation. Mustn1 (Mustang, Musculoskeletal Embryonic Nuclear Protein) encodes a nuclear protein highly expressed in adult regenerating muscle [35]; its knockdown in C2C12 cells inhibits cell differentiation [36]. Therefore, Mustn1 repression by Runx1 would again prevent myoblasts differentiation. Another interesting group of high-confidence in vivo Runx1-regulated genes is the signal transduction pathways-related genes, including members of the Delta-Notch, Igf/Akt/mTor and Tgf-β pathways. The Delta/Notch pathway is activated by the Delta like 1 (Dll1), which upon binding to the Notch receptor, induces an anti-differentiation signal by upregulating Hey1 and MyoD, which in turn, prevent the expression of pro-myogenic genes [37]. Interestingly, in Runx1 PM and mdx muscles, we found that Delta- like homolog 1 (Dlk1), a putative Delta- Notch antagonist [38], to be significantly upregulated. As it was previously shown that Dlk1 inhibits proliferation in avian [39] and mouse [40,41] myoblasts it is tempting to speculate that Runx1 participates in the myoblast Delta-Notch signaling pathway by repressing the antagonist Dlk1 thereby promoting Delta-mediated myoblast proliferation. In case of the Igf-1/Akt/mTor pathway, which regulates muscle hypertrophy [42] and SC proliferation and differentiation [43], we found the two isoforms of Igf-1 downstream mediator protein kinase C (PKCβ and PKCδ) among the in vivo high-confidence gene subset. PKCδ (Prkcd) specific inhibition delays differentiation of C2C12 cells and primary human myoblasts [44]. In the rat myoblastic cell line H9c2, PKCδ is activated during differentiation, and its knockdown results in reduced myoblast differentiation [45]. Runx1 could regulate the pro-differentiation branch of the Igf-1 signal by repressing PKC isoforms, especially PKCδ. Finally, we found that Myostatin (Mstn, Gdf-8), a member of the Tgf-β family, is repressed by Runx1. Expressed in muscle Mstn serves as a negative regulator of muscle mass [46] and as attenuator of myoblast [47] and SC proliferation [48]. Ablation of Mstn improves muscle regeneration [49], and has been proved beneficial in mdx mice [50]. Therefore, Runx1 repression of Mstn in dystrophy-induced muscle regeneration could promote myoblasts proliferation.
The data we obtained from both in vivo and in vitro systems show Runx1 function during muscle regeneration to promote myoblast proliferation by repressing myoblast differentiation-inducing genes. Its activity in regenerating muscle is therefore required for the production of the critical amount of myoblasts needed for proper restoration of muscle mass (Fig 7A). Runx1 expression is confined to the PM proliferation stage, which mimics the first stages of muscle regeneration. Thus, it is conceivable that its affect is manifested during the first days post myonecrosis. Loss of Runx1 activity leads to premature myoblast differentiation, resulting in the diminution of the myoblasts pool and subsequent impaired regeneration (Fig 7B). Of potential relationship, Runx1 promotes the proliferation of adult stem cells in other tissues. For example Runx1 promotes adult hair follicle stem cell proliferation thereby increase the cell pool size prior to terminal differentiation [51]. In mesenchymal stem cells, RUNX1 is induced upon activation by an TGF-β signal and drives progenitor cells proliferation and restricts terminal differentiation into myofibroblast [52].
Prior work addressing Runx1 function using the C2C12 cell line or rhabdomyosaracoma myoblasts resulted in conflicting conclusions. While the C2C12 cell data [14] supported an anti- differentiation function of Runx1 the human rhabdomyosarcoma cell data [15,53] indicated that Runx1 promotes myoblast differentiation and that Runx1, MyoD and c-Jun cooperate to induce this differentiation [13,54]. However, in the PM cultures describe here the protein level of both Runx1 and c-Jun decreased at the onset of differentiation, rendering a potential role in later stages of differentiation unlikely. Additionally, c-Jun [26] and another AP-1 family member, Fra-2 [25] were found to repress myoblast differentiation. All in all, this discrepancy may have resulted from intrinsic differences between in vivo mouse models and cultured PM stem cells, and transformed/ immortalized myoblastic cell lines.
In summary, our findings support the conclusion that in response to injury, muscle Runx1 is switched on and cooperates with MyoD and c-Jun in order to regulate a muscle regeneration transcription program that involves changing the proliferation/differentiation balance by repressing genes that participate in myoblast differentiation. These data add unique insights into the transcriptional program driving muscle regeneration and implicate Runx1 as a potential participant in the pathology of muscle wasting diseases.
The experiments were conducted in strict accordance with the recommendations of the US National Institutes of Health Guide for the Care and Use of Laboratory Animals. The protocols were approved by the Committee on the Ethics of Animal Experiments of the Weizmann Institute of Science (Permit Number: 01190113–2, 12720814–3). All surgery was performed under Ketamine/Xylazine anesthesia, and all efforts were made to minimize suffering.
Runx1f/f mice were generated by crossing Myf5::Cre mice [17] onto Runx1L/L C57bl/6 mice [16,55]. mdx/Runx1f/f mice were generated by crossing Runx1f/f mice onto mdx mice [56]. Transgenic SOD1 mutant mice (B6.Cg-Tg (SOD1*G93A)dl were obtained from Jackson Laboratory, USA and bred on C57Bl/6. Genotypes were determined by PCR of tail tissue. Mice weight was monitored once a month. Kaplan- Meyer survival curve was calculated using the PRISM© software. For body composition measurements, we used EcoMRI- 100H analyzer (Echo medical systems, USA). Mice body composition was measured monthly, as an average of three separate measurements for each mouse. For muscle damage experiments, 0.75 μg CTX (Sigma-Aldrich, Israel) in 50 μl of sterile phosphate-buffered saline was injected into the right gastrocnemius muscle of 8 weeks old mice. To record cell proliferation in vivo, 10 mg/ml BrdU (150 μl/30 g mouse) was injected intraperitoneally, gastrocnemius muscles were harvested 24h (in mdx mice) or 2h (in CTX treated mice) post BrdU injection and subjected to BrdU IHC using anti BrdU antibody (Ab) (#MCA2060, Serotec, UK).
PM cultures were established as previously described [57], following isolation of SC from gastrocnemius muscles of 2–3 week old C57bl/6 Runx1L/L or Runx1f/f mice. Muscles were treated with collagenase type I (C-0130, Sigma-Aldrich, Israel) for 3h and isolated myofibers were then seeded in proliferation media (BIOAMF-2, Biological Industries, Israel), in GHR Matrigel (BD Biosciences, USA) coated plates for 3 days, to facilitate SC delamination. PM were enriched by three stages of pre-plating. Cells were grown at 37°C, 5% CO2 on GHR Matrigel in proliferation medium, which was replaced daily. For differentiation assay, cells were grown as above to 75–80% confluency and then induced to differentiate by serum starvation in differentiation media (DM): DMEM containing 4% horse serum (Gibco, UK) and 0.04U/mL human Insulin. DM was replaced after 48h when needed. For analysis of immunostained myoblasts, 2x104 cells were seeded and grown for 16 h on a Lab-Tek 8-well chamber slide, pretreated with Matrigel. For viral infection, cells were exposed for 24 hours to 6.5×107 virus particles/ml of either Ad5CMV-eGFP, Ad5CMVCre-eGFP (both from Gene Transfer Vector Core, University of Iowa USA) or Adeno-Runx1 constructed in house using the AdEasy system [58]. For determination of average cell doubling time, 105 primary SC/sample were plated, grown for 48h in proliferation medium and counted at the end point. For cell cycle analysis, SCs were grown under proliferation conditions for 48h, until reaching 70% confluency. Myoblasts were then fixed using cold ethanol, stained with propidium iodide (PI), and analyzed by FACS. For measurement of cell death, PM cultures were collected and washed twice with PBS and FACS analysis was performed promptly following addition of PI. As positive control of PI staining, WT PM were permeablized by incubation at 65°C for 2 min followed by mixing with untreated PM at a 1:3 ratio.
IHC of mouse tissue paraffin sections and of satellite cell cultures were performed as previously described [57]. Primary antibodies (Abs) used included mouse anti-MHC (MH-20, Developmental Studies Hybridoma Bank, USA at 1:5 dilution), mouse anti-Pax7 (Developmental Studies Hybridoma Bank, USA at 1:1000 dilution), rabbit anti-MyoD (sc-304, Santa Cruz Biotechnology, USA at 1:100 dilution), rat anti BrdU (MCA2060, Serotec, USA, 1:100), rabbit anti-Ki67 (275R, Cell Marque, 1:200) and our in-house affinity purified rabbit anti-Runx1 (at 1:100 dilution) [5]. HRP based IHC was performed using MOM kit (PK-2200, Vector, USA) according to the manufacturer’s instructions. Immunofluorescence analysis was performed using Cy2-, Cy3-, or Cy5-conjugated secondary Abs (Jackson ImmunoResearch, USA), at a dilution of 1:200–1:500. Images were acquired using a Zeiss LSM510 confocal microscope. For recording BrdU+ cells or regenerating myofibers, we subjected relevant sections to either anti BrdU IHC or H&E staining, respectively. The stained sections were photographed using a Nikon E800 light microscope, coded and manually counted by an unbiased estimator. For determining the average size of myofibers, sections were stained with H&E and the CSA of 400–500 fibers was measured by an unbiased estimator using the “count” procedure of ImagePro+ software. For recording Pax7+/Ki67+ cells, sections were reacted with anti-Pax7 and anti-Ki67 Ab and analyzed using the Zeiss LSM780 confocal microscope. Number of Pax7+, Ki67+ and Pax7+/Ki67+ cells was determined using the Fiji software (ImageJ 1.47v, NIH, USA). For fusion index-determination assay, 2x104 primary myoblasts were transferred to chamber slides and grown in either Bio-AMF2 or DM, as indicated. Cultures were coded and stained for MHC and DAPI. Single, double and multinucleated cells were counted by an unbiased estimator using 4 biological repeats per experiment, comprising 12 different fields per repeat.
Treadmill assay was performed by monthly training on a treadmill (Panlab Mouse 5-Lane Treadmill; model#: 760309; HARVARD APPARATUS, USA) over a period of 8 months, starting at the age of 2 months. Mice ran on the treadmill at 20 degrees uphill, starting at a speed of 10 meters/min. After 10 minutes, the speed was increased gradually to a final speed of 20 meters/min. The mice then ran for an additional 10 minutes at this speed. Performance was determined by comparing running time till exhaustion (defined as stepping off the running lane 5 times with less than 0.5 sec. intervals). Performance of each mouse was recorded at three consecutive days. Differences in treadmill performance at the ages of 2–9 months were assessed by one factor ANOVA (analysis of variance) for Gene (the four genotype groups) in each time point. The analyses were performed using IBM® SPSS® Statistics version 20.0. For grip strength assay, we use the TSE grip strength meter (#303500, TSE systems, Germany). 4 months old mice grip strength was monitored at three consecutive days, 5 times each day (15 measurements per mouse).
cDNA was synthesized by superscript II RT kit (#18064–022, Invitrogen, USA) using 1μg of purified RNA and analyzed by qPCR using light cycler 480 (Roche, US). The following Taqman gene expression assays (Applied Biosystems, USA) were used to quantify RNA level: Mm01213404_m1 and Mm0123405_for Runx1, Mm0044614_m1 for Myog, Mm01340842_m1 for Mef2c, Mm00500665_m1 for Myom2, Mm00449089_m1 for Tnnt1, Mm01332564_m1 for Myh2, Mm01329494_m1 for Myh8 and Mm00446973_m1 for Tbp1, used as an internal calibrator. Other genes were quantified using miScript SYBR green PCR kit (#218073, Qiagen, Germany). The primers used are detailed in S5 Table. Each qPCR experiment consisted of three biological repeats each using two cDNAs independently prepared. Statistics were performed using the Excel based REST software.
Nuclear protein extracts were obtained following collection and sonication of cultured SCs as previously described [59]. WB was performed using our in house anti-Runx1 (1:5000) as described [5]. Primary Abs used included rabbit anti- c-Jun (sc-1694, 1:1000), rabbit anti-Emerin (sc-15378, 1:104) (Santa Cruz Biotechnology, USA) and mouse anti-GAPDH (MAB374, Chemicon, USA, 1:1000). Secondary Abs used were either anti-rabbit HRP or anti-mouse HRP (Jackson ImmunoResearch, USA). Quantification of WB protein bands was conducted using the Image Quant LAS4000 (GE) device and endogenous Image Quant TL software.
RNA was isolated by PerfectPure RNA tissue kit (# 2302410, 5 PRIME, Germany) according to the manufacturer's instructions. Purified RNA was reverse-transcribed, amplified, and labeled with Affymetrix GeneChip whole transcript sense target labeling kit. Labeled cDNA was analyzed using Affymetrix Mouse Gene 1.0 ST microarrays, according to the manufacturer's instructions. Microarray data were analyzed using Partek Genomic Suite software. CEL files (containing raw expression measurements) were imported and data was preprocessed and normalized using the Robust Multichip Average (RMA) algorithm [60]. To identify differentially expressed genes ANOVA was applied and genes fold-changes were calculated.
For RNA-seq analysis RNA was isolated from 2 months old mice Soleus muscle extracts using the PerfectPure RNA tissue kit, as mentioned above. Illumina TruSeq® RNA Sample Preparation v2 was used according to manufacturer's instructions. Indexed samples were sequenced in a Illumina HiSeq 2500 machine in a single read mode. The obtained reads, 50 bp long, were mapped to the mm9 mouse genome assembly using TopHat2 [61]. Version 2.0.12.0.10 with default options. Expression at the gene level was quantified by applying HTSeq (version 0.6.1) [62], and using the known genes from UCSC in gtf format as annotation. Differential expression was calculated utilizing the DESeq2 software (version 1.2.10) [63].
ChIP was performed essentially as described [22]. Briefly, cross-linked chromatin from approximately 1.2x108 freshly isolated primary WT myoblasts was prepared and fragmented to an average size of approximately 200 bp by 35 cycles of sonication (30 seconds each) in 15-ml tubes using the Bioruptor UCD-200 sonicator (Diagenode, USA). The following Abs were used for immunoprecipitation of fragmented chromatin: 170μl of in house anti-Runx1; 24μg of mouse anti-MyoD (sc-32758, Santa Cruz Biotechnology, USA); 24μg of rabbit anti c-Jun (sc-1694, Santa Cruz Biotechnology, USA). Rabbit pre-immune serum or mouse IgG (278–010, Ancell), were used as control for Runx1 or MyoD and c-Jun ChIP-seq, respectively. DNA was purified using QIAquick spin columns (QIAGEN) and sequencing performed using Illumina HiSeq 2500. Two biological repeats were conducted and separately sequenced for each ChIP-seq experiment. For ChIP-seq analysis, the reads were aligned to the mouse genome (mm9) allowing one mismatch and using the Bowtie aligner [64]. Reads with a unique best alignment were retained for further processing. Immunoprecipitated samples were compared against the negative control to find binding sites using the MACS software with default parameters [65].
ChIP products of Runx1L/L and Runx1f/f PM were purified as described for ChIP-seq using 3x107 freshly isolated PM per reaction. ChIP products and input DNA were diluted to the same DNA concentration and subjected to qPCR using SYBR-green (miScript #218073, Qiagen, Germany). Each experiment consisted of three biological repeats, and input DNA served as control. Statistics were performed using the Excel based REST software.
ATAC was performed as previously described [29]. Briefly, 5x104 freshly isolated PM were harvested, and underwent the recommended transposition protocol without the lysis stage. The resulting transposed DNA was enhanced using 12 cycles of PCR, as described. The resulting libraries were sequenced using Illumina HiSeq 2500. For ATAC-seq analysis, we used similar parameters as for the ChIP-seq (see above).
For dual Luciferase assay, Runx1 bound genomic DNA fragments related to the Myog, Tnnt1, Myh8 and Myom2 genes were generated by PCR using primers listed in S8 Table. RUNX binding site in the Myog and Tnnt1 regulatory elements was mutated by overlap PCR using primers indicated in S8 Table. Intact and mutated genomic elements were cloned into the Renilla Luciferase expression vector pTK-Luc, upstream to the TK promoter, using HindIII and BamHI restriction sites. HEK293 cells in 24-well plates were co-transfected using Lipofectamine 2000 according to the manufacturer protocol (#11668–027, Invitrogen) with 1μg of the reporter vector, 1μg of expression vector (empty pcDNA3.1 or pcDNA3.1-Runx1) and 0.01μg of pGL4.13 vector carrying firefly Luciferase as internal transfection control. PM were co-transfected in 24- well plates with 1μg of the reporter vector and 0.01μg of pGL4.13, using the Nepa21 electroporation system (Nepagene) at the following settings: Poring phase of 2 pulses of 225V for 2.5ms with a 50ms interval, followed by a transfer phase of 5 pulses of 30V for 50ms with a 50ms interval. Firefly and Renilla Luciferase activities were measured 24 h after transfection using a dual luciferase assay kit (Promega).
Ingenuity Pathway Analysis tool (https://apps.ingenuity.com/) was used for GO annotation of Runx1-regulated genes and GREAT software [24] was used for Chip-seq peak GO analysis. MEME-ChIP suit (http://meme.nbcr.net/meme4_6_1/cgi-bin/meme-chip.cgi/), was used for de-novo motif finding in ChIP-seq TF-bound regions with default parameters and Genomatix Genome Analyzer RegionMiner tool (http://www.genomatix.de/solutions/genomatix-genome-analyzer.html) was used for deriving overrepresented TF modules. All microarray, ChIP-seq and ATAC-seq data are available in the GEO public database under the SuperSeries accession number GSE56131.
|
10.1371/journal.ppat.1005653 | Inhibition of Nuclear Transport of NF-ĸB p65 by the Salmonella Type III Secretion System Effector SpvD | Salmonella enterica replicates in macrophages through the action of effector proteins translocated across the vacuolar membrane by a type III secretion system (T3SS). Here we show that the SPI-2 T3SS effector SpvD suppresses proinflammatory immune responses. SpvD prevented activation of an NF-ĸB-dependent promoter and caused nuclear accumulation of importin-α, which is required for nuclear import of p65. SpvD interacted specifically with the exportin Xpo2, which mediates nuclear-cytoplasmic recycling of importins. We propose that interaction between SpvD and Xpo2 disrupts the normal recycling of importin-α from the nucleus, leading to a defect in nuclear translocation of p65 and inhibition of activation of NF-ĸB regulated promoters. SpvD down-regulated pro-inflammatory responses and contributed to systemic growth of bacteria in mice. This work shows that a bacterial pathogen can manipulate host cell immune responses by interfering with the nuclear transport machinery.
| Salmonella Typhimurium replicates in macrophages through the action of effector proteins translocated into host cells by a type III secretion system (T3SS). We show that the T3SS effector SpvD targets the NF-ĸB pathway by interfering with nuclear translocation of p65. SpvD interacts with the exportin Xpo2. Perturbation of Xpo2 disrupts recycling of importin-α from the nucleus, leading to abrogation of p65 nuclear translocation. These data show that a bacterial pathogen manipulates host cell immune responses by interfering with nuclear transport machinery.
| The NF-ĸB signalling pathway has a central role in the host response to infection by microbial pathogens, by stimulating innate and acquired host immune responses. Under normal physiological conditions, transcription factors of the NF-ĸB family such as p65 remain inactive in the cytoplasm through their interaction with inhibitors, the IĸBα proteins, which mask the nuclear localisation signal (NLS) of transcription factors. Following engagement of extracellular bacterial LPS by Toll-Like Receptor 4 (TLR4) or tumour necrosis factor (TNF) by the TNF receptor (TFNR), different pathways lead to phosphorylation and proteasomal degradation of IĸBα, allowing the NF-ĸB subunits to bind the adaptor protein importin-α (KPNA). This complex then interacts with one of up to 20 importin-β family members to enable nuclear transport through the nuclear pore complex. Within the nucleus, RanGTP binds to importin-β, dissociating the import complex and releasing the NF-ĸB subunits to initiate transcription of their target genes. Importin-β complexed with RanGTP is recycled to the cytoplasm while export of KPNA follows its interaction with the β-karyopherin exportin-2 (Xpo2, also called CAS) and RanGTP. Finally, cytoplasmic Ran GTPase activating protein (RanGAP) stimulates the Ran GTPase, generating RanGDP, which dissociates from the importins and thereby releases them for another import cycle [1,2].
Many pathogens, including Salmonella enterica, have acquired mechanisms that interfere with NF-ĸB signalling. Numerous components of the NF-ĸB signalling pathway are targeted by pathogen-mediated post-translational modifications (PTMs) that result in attenuation of the NF-ĸB dependent responses [3–9].
Salmonella has two type III secretion systems (T3SS) encoded within the Salmonella pathogenicity islands (SPIs) 1 and 2 that deliver virulence effector proteins into the host cell. The SPI-1 T3SS effectors are translocated across epithelial cell plasma membranes and mediate bacterial invasion and intestinal inflammation [10], while the SPI-2 T3SS translocates approximately 30 different effectors across the vacuolar membrane. Some of these maintain vacuolar membrane integrity and enable bacterial growth [11,12]. A few have been shown to interfere with host cell inflammatory responses. For example, SpvC has phosphothreonine lyase activity on MAPKs [13,14], SspH1 (translocated both by SPI-1 and SPI-2 T3SSs [15]) binds to the kinase PKN1 [16], which in turn regulates NF-ĸB and JNK signalling and AvrA (also translocated both by SPI-1 and SPI-2 T3SSs [17] inhibits NF-ĸB pathway in epithelial cells [18] via the JNK pathway [19] and tight junction stabilization [20]. PipA, GogA and GtgA redundantly target components of the NF-ĸB signaling pathway to inhibit transcriptional responses leading to inflammation [21].
Here, we used macrophages lacking TLR4 to reveal that the SPI-2 T3SS effector SpvD suppressed production of pro-inflammatory cytokines. We found that SpvD interfered with the NF-ĸB signalling pathway by preventing nuclear accumulation of p65. This was associated with nuclear accumulation of importin-α family members that are required for nuclear import of p65. Interestingly, SpvD interacted specifically with exportin Xpo2, which mediates nuclear-cytoplasmic recycling of importins. Together, this work reveals that a bacterial pathogen prevents host cell immune responses by interfering with nuclear transport machinery.
We previously reported increases of more than 10-fold in mRNA for 26 genes in wild-type mouse bone marrow-derived macrophages (BMM) infected for 10 h by S. Typhimurium [22]. Many of these are likely to be due to the predominant effect of LPS on pro-inflammatory signalling pathways, since LPS and Salmonella cells induced similar changes in RAW 264.7 macrophages gene expression at 4 h post-challenge [23]. To analyse the effect of the SPI-2 T3SS in the context of TLR4-induced responses, we selected three genes (tnf-α, il1-β and serpinB2) that are induced by LPS [23,24]. Their mRNA levels were measured in BMM from wild-type and TLR4 knock-out (TLR4-/-) mice, infected with wild-type or ΔssaV (SPI-2 T3SS null mutant) S. Typhimurium for 10 h to allow sufficient time for SPI-2 T3SS effector translocation and activity, while minimising the difference in overall intracellular growth rate between wild-type and ΔssaV mutant bacteria. The fold-increase in intracellular bacterial numbers (calculated from the ratio of the cfu at 10 h compared to 2 h post-uptake) was 2.20 ± 0.34 for the wild-type and 0.89 ± 0.09 for ΔssaV mutant bacteria. Following infection of wild-type BMM for 10 h, the RNA levels of tnf-α, il1-β and serpinB2 (quantified by real-time PCR (qRT-PCR)) increased by between 40 and 400 fold, and no significant differences were observed between BMM infected with wild-type or ΔssaV mutant bacteria (Fig 1A and 1B, S1 Fig). Similar increases were detected following exposure of wild-type BMM to LPS, suggesting that TLR4 signalling accounted for the majority of these changes (Fig 1A and 1B, S1 Fig). However, TLR4-/- BMM infected with the ΔssaV mutant had statistically significant higher levels of tnf-α and il1-β mRNA transcripts than BMM infected with wild-type bacteria (Fig 1A and 1B). No difference was observed for serpinB2 mRNA transcripts levels (S1 Fig), indicating that transcription of not all the LPS-responsive genes was affected by the SPI2-T3SS. Levels of secreted TNF-α from TLR4-/- BMM infected with SPI-2 null mutants (ΔssaV or ΔsseB) were also significantly greater than from TLR4-/- BMM infected with wild-type bacteria or a complemented SPI-2 null mutant (ΔsseB, psseB) (Fig 1C). The level of secreted Il1-β was below the threshold of detection at 10 h post-bacterial uptake but was also reduced in a SPI-2 T3SS-dependent manner at 24 h post-bacterial uptake (Fig 1D). Together these results provide evidence that the SPI-2 T3SS suppresses pro-inflammatory cytokine mRNA levels and protein secretion in TLR4-/- macrophages. The failure to detect such differences using wild-type BMM presumably reflects the potent effect of LPS on pro-inflammatory signalling pathways in this experimental system.
To identify SPI-2 T3SS effector(s) involved in down-regulation of pro-inflammatory immune responses, several SPI-2 T3SS mutant strains (representing single mutants lacking individual effectors) [25] were screened for levels of secreted TNF-α from TLR4-/- BMM at 10 h post-uptake (S1 Table). Of these, 2 strains (ΔspvC and ΔspvD) induced significantly more cytokine secretion from infected BMM compared to BMM infected by the wild-type strain (Fig 1E and 1F). SpvC and SpvD are encoded by the spvRD operon (composed of spvA, B, C and D, and regulated by spvR [26] on the Salmonella virulence plasmid (pSLT)) and SpvC is a phosphothreonine lyase that interferes with MAPK signalling [14]. BMM infected with deletion mutants carrying the low copy number plasmid pACYC184 containing the corresponding wild-type allele of the gene (ΔspvC, pspvC or ΔspvD, pspvD) produced significantly less TNF-α and Il1-β than BMM infected with the single mutant strains (Fig 1E and 1F). To analyse if the effects of SpvD are confined to macrophages, Hela cells were infected with Salmonella strains for 8, 12 or 24 h and levels of IL-8 in supernatants were quantified by ELISA. Infection with wild-type bacteria induced IL-8 production but no increased production was detected in cells infected with the ΔspvD mutant or the ΔspvC mutant (as shown in [14]) (S2 Fig). However, cells infected with the ΔspvD, pspvD or ΔspvC, pspvC strains showed significantly less IL-8 production compared with the wild-type strain at all time-points tested (P ≤ 0.05 for all time points). Therefore, the effects of SpvD occur both in macrophages and another cell type that does not normally respond to LPS.
SPI-2 T3SS-dependent secretion of SpvD into minimal medium and its translocation into host cells (using a CyaA fusion protein) were shown previously [29]; however its function is unknown and its contribution to virulence of S. Typhimurium in mice is not clear [27–30]. To confirm SPI-2 T3SS-dependent secretion and translocation of SpvD, strains producing double-HA tagged SpvD from pSLT were used for an in vitro secretion assay and for macrophage infection. Secretion of SpvD-2HA from bacteria grown in minimal medium was investigated following shift of ambient pH from 5.0 to 7.2, which stimulates effector secretion from the SPI-2 T3SS [31]. A similar amount of SpvD-2HA was present in lysates from wild-type and ΔssaV mutant bacteria (S3A Fig). SpvD-2HA was also detected in the secreted fraction from wild-type bacteria but not from the ΔssaV mutant. To investigate SpvD translocation into host cells, RAW macrophages were infected for 22 h with Salmonella strains producing SpvD-2HA. Cells were then fixed and labelled to detect Salmonella and the HA tag. Translocated SpvD-2HA was not detected within infected macrophages by immunofluorescence microscopy under these conditions. However when cells were incubated for 2 h prior to fixation with the proteasome inhibitor MG132, weak but distinct and reproducible cytoplasmic labelling of SpvD-2HA was detected in the cytoplasm of the majority of cells infected with wild-type but not ΔssaV mutant Salmonella (S3B Fig). These results suggest that relatively small amounts of SpvD are translocated into the host cell cytoplasm.
Since SpvC inhibits MAP kinase signalling [14] we tested if SpvD affects this pathway in cells expressing a luciferase reporter gene under the control of MAPK-dependent AP-1 binding sequences. HEK 293 cells were co-transfected with vectors expressing myc-SpvD or myc-SpvC (as positive control) or myc vector alone (empty pRK5 vector), along with vector expressing the luciferase reporter under the control of AP-1 promoter and vector constitutively expressing the Renilla luciferase, as an internal control for normalisation of data. Cells were exposed to PMA for 6 h to stimulate MAP kinase pathways. Then cells were lysed and bioluminescence was measured to determine the fold-difference of MAP kinase activation in relation to non-transfected cells. Immunoblots of cell lysates following SDS-PAGE, using anti-myc and anti-tubulin antibodies, confirmed that similar numbers of cells expressing similar levels of effectors were analysed. Ectopic expression of SpvC resulted in a strong inhibition of luciferase activity, while transfection of empty vector or of vector producing SpvD did not inhibit luciferase activity (Fig 2A), suggesting that unlike SpvC, SpvD does not inhibit the MAP kinase signalling pathway.
Some T3SS effectors inhibit pro-inflammatory cytokine production and secretion by interfering with proteins of the NF-ĸB pathway [6]. To determine the potential effect of SpvD on nuclear translocation of the NF-ĸB transcription factor p65, confocal microscopy was used to quantify the amounts of nuclear p65 in TLR4-/- BMM infected with different Salmonella strains at 10 h post-bacterial uptake. Cells were immunolabelled with anti-Salmonella and anti-p65 antibodies and nuclei were stained with the DNA dye, DRAQ5 (Fig 2B). Three-dimensional image projections were acquired randomly and nuclear p65 intensity was quantified using Volocity software. The intensity of p65 in the nuclei of infected TLR4-/- BMM was significantly increased when SpvD was absent (ΔspvD, ΔspvRD (corresponding to ΔspvR-spvA-spvB-spvC-spvD mutant strain) or ΔspvRD,pspvC strains) and was restored to wild-type levels by mutants carrying pspvD (ΔspvD,pspvD or ΔspvRD,pspvD strains) (Fig 2C), indicating that of spv-encoded proteins, only SpvD prevents nuclear accumulation of p65. NleC, a T3SS effector of enteropathogenic E. coli is a protease that cleaves p65 directly [32–35]. Therefore, we analysed total cell levels of p65 in TLR4-/- BMM infected with different Salmonella strains at 10 h post-uptake. No significant differences were detected (S4A Fig). To determine if the influence of SpvD on p65 localisation was cell-type specific, its effects on p65 and its localisation were analysed during infection of HeLa epithelial cells. The amount of p65 in total cell lysates and in nuclei of HeLa cells infected with Salmonella strains was quantified by immunoblot after cell fractionation. No differences were detected in total cell extracts, but the amount of p65 in the nuclei of cells infected with the ΔspvD mutant was consistently increased compared to wild-type infected cells (S4B Fig). The amount of nuclear p65 was restored to wild-type levels in cells infected with the ΔspvD mutant carrying a plasmid containing spvD (S4B Fig). In addition, total levels of p65 in HeLa cells transfected with pRK5myc-SpvD, pRK5myc-SpvC (used as a negative control) or pRK5myc-NleC (used as a positive control) were analysed by FACS after labelling of the cells with anti-p65 and anti-myc antibodies. Levels of p65 were determined in myc-positive cells and normalised to non-transfected cells. Degradation of p65 occurred only in cells expressing NleC (S4C and S4D Fig). These results indicate that SpvD reduces nuclear translocation of p65 in both mouse macrophages and human epithelial cells, but does not affect the overall cellular pool of p65. The amounts of two other immune related transcription factors (p50 and STAT2) in total cell lysates and in nuclei of HeLa cells infected with Salmonella strains were quantified by immunoblot after cell fractionation. No differences were detected in total cell extracts, but the amount of p50 in the nuclei of cells infected with the ΔspvD mutant was consistently increased compared to wild-type -infected cells (S4B Fig). The amount of nuclear p50 was restored to wild-type levels in cells infected with the ΔspvD mutant carrying a plasmid containing spvD. No differences in STAT2 levels in nuclei or total cell lysates were detected. These results indicate that SpvD specifically targets the import of NF-ĸB -associated proteins.
The predicted product of spvD is a hydrophilic protein of 216 amino acids [36] and visual alignment of a region of SpvD with several prokaryotic phosphothreonine lyases [37] revealed a region in SpvD (between amino acids 181 to 213) that is conserved among these enzymes (S4E Fig). This area contains an invariant Lys residue (K185) and mutation of this residue in SpvC results in a loss of phosphothreonine lyase activity [38]. As a further test of the effect of SpvD on the NF-ĸB signalling pathway, and to investigate the importance of K185, HEK 293 cells were co-transfected with a reporter plasmid encoding luciferase under the control of NF-ĸB promoter and a plasmid encoding either myc alone (empty pRK5 vector), myc-SpvC (as a negative control), myc-SpvD or myc-SpvDK185A (SpvD carrying a lysine-to-alanine substitution at residue 185). Luciferase activity was assayed 8 h after TNF-α stimulation and normalised to non-transfected and unstimulated cells. Non-transfected cells underwent an approximately 20-fold increase in activation following stimulation with TNF-α when compared to resting cells. Transfection of 500 ng of empty vector or vector expressing SpvC did not prevent NF-ĸB activation, whereas expression of SpvD or SpvDK185A reduced activation by approximately 65% (Fig 2D). Inhibition of NF-ĸB activation by SpvD was observed after transfection of 50 ng of DNA and increased with increasing amounts of DNA. Together, these results indicate that SpvD is sufficient to block the activation of an NF-ĸB -regulated promoter and that K185 is not required for SpvD activity.
To investigate if SpvD interferes with IĸBα degradation induced upon stimulation with TNF-α, HEK 293 cells were transfected with plasmids encoding either myc alone (empty vector), myc-SpvD or myc-YopP (a T3SS effector of Yersinia enterolitica that inhibits IĸBα degradation [39–41], used here as a positive control), then stimulated with TNF-α for 5, 10 or 20 min to induce IĸBα degradation. Proteins in cell extracts were then analyzed by SDS-PAGE and immunoblotting. As expected, degradation of IĸBα was inhibited in cells expressing myc-YopP (Fig 2E). However, IĸBα was degraded in cells transfected with the empty vector or with the plasmid encoding myc-SpvD (Fig 2E). Because only approximately 60% of the cell population was efficiently transfected (as measured by microscopic examination), a potential effect of SpvD on IĸBα degradation might be difficult to detect when analysing the whole cell population by immunoblotting. Therefore, IĸBα degradation following TNF-α stimulation was also analysed by flow cytometry after labelling of the cells with anti- IĸBα and anti-myc antibodies. Levels of IĸBα were quantified in myc-positive cells and normalised to non-stimulated cells. TNF-α -induced IĸBα degradation was only prevented in cells expressing YopP (S4F Fig), indicating that SpvD does not inhibit IĸBα degradation and that it must prevent nuclear accumulation of p65 by interfering with the NF-ĸB pathway after IĸBα degradation.
Translocation of p65 into the nucleus requires proteins of the karyopherin (KPNA) family [42,43], also called importin-α. To determine the potential effect of SpvD on KPNA1 and KPNA3, two importin-α proteins known to interact with p65 [43–45], confocal microscopy was used to analyse their localisation in HeLa cells. Since attempts to detect endogenous KPNA1 or KPNA3 using commercial antibodies failed, HeLa cells were co-transfected with plasmids encoding either KPNA1-FLAG or KPNA3-FLAG and plasmids encoding myc-SpvD or myc-SpvC (as a control). Cells were immunolabelled with anti-FLAG, anti-myc and anti-lamin antibodies. Consistent with the localisation of SpvD following its translocation from bacteria (S3B Fig), SpvD was found in the host cell cytoplasm after transfection (Fig 3A). In addition to nuclei containing diffuse labelling of KPNA1-FLAG or KPNA3-FLAG, in some nuclei, the labelling was predominantly in the region of the nuclear envelope, where it colocalised partially with lamin (Fig 3A). This lamina-associated labelling was observed in all transfected cell populations, but its frequency was enhanced 2.7-fold by KPNA1 (P<0.001) and 1.6-fold by KPNA3 (P<0.05) when SpvD was co-expressed (Fig 3B). In addition, the enhanced frequency of lamina-associated labelling was not observed when SpvC was co-expressed (Fig 3B), indicating that it is specific to SpvD. To determine if the reorganisation of KPNA1 induced by myc-SpvD was localised inside or outside the nucleus, differential membrane permeabilisation was done using either Triton X-100 (under conditions in which it permeabilises both the plasma and nuclear membranes) or Saponin (at a concentration where it permeabilises only the plasma membrane). SpvD-induced accumulation of KPNA1 at the nuclear envelope was only observed when cells were permeabilised with Triton X-100, indicating that this importin accumulates within the nucleus (Fig 3C).
To analyse the effect of SpvD on importin-α localisation during infection, HeLa cells transfected with plasmids encoding KPNA1-FLAG or KPNA3-FLAG were infected for 14 h with Salmonella strains. Cells were immunolabelled with anti-Salmonella and anti-FLAG antibodies and nuclei stained with DRAQ5. Deletion of SpvD led to a significant decrease of the percentage of cells with nuclear lamina-associated KPNA1 (Fig 3D and 3E) or KPNA3 (S5 Fig). Cells infected with the deletion mutant carrying a plasmid-borne wild-type allele of spvD induced a greater proportion of these structures than the deletion mutant (Fig 3E; p<0.01). These structures might be an artefact of overexpression of KPNA1 and KPNA3 but do suggest an SpvD-dependent accumulation of KPNA1 and KPNA3 in the nucleus. To analyse the effect of SpvD on the localisation of endogenous KPNA1, the amount of endogenous KPNA1 in nuclei of HeLa cells infected with Salmonella strains was quantified by immunoblotting after cell fractionation. A significant decrease in the amount of endogenous KPNA1 in the nuclei of cells infected with the ΔspvD mutant was detected, compared to wild-type Salmonella-infected cells (Fig 3F and 3G). The amount of nuclear KPNA1 was restored to that of the wild-type level in cells infected with the ΔspvD mutant carrying a plasmid containing spvD (Fig 3F and 3G). Therefore, in addition to inhibiting nuclear accumulation of p65, SpvD also causes nuclear accumulation of KPNA1.
Recycling of KPNAs from the nucleus to the cytoplasm is mediated by Xpo2 (also named CAS or CSE1) [46]. To confirm that depletion of Xpo2 inhibits recycling of KPNAs, confocal microscopy was used to analyse the localisation of KPNA1-FLAG in HeLa cells depleted of Xpo2 (by siRNA of Xpo2, S6A Fig) and then transfected with a plasmid encoding KPNA1-FLAG. Cells were immunolabelled with anti-Xpo2 and anti-FLAG antibodies and the percentage of cells having detectable cytoplasmic KPNA1 were scored. siRNA of Xpo2 in HeLa cells caused accumulation of KPNA1 in the nucleus (S6B Fig) and a decrease in the percentage of cells having detectable KPNA1 in the cytoplasm (S6C Fig), compared to non-treated cells or cells exposed to a scrambled siRNA. However, the nuclear KPNA1 that was observed in the absence of Xpo2 was not localised at the lamina, as was observed in the presence of SpvD.
Knock-down of Xpo2 in HeLa cells was reported to cause a decrease of p65 translocation after stimulation with TNF-α [47]. To determine the effect of inhibition of Xpo2 expression on p65 localisation in our experiments, confocal microscopy was used to analyse localisation of p65 in HeLa cells depleted of Xpo-2 and stimulated with TNF-α for 45 min. Cells were immunolabelled with anti-Xpo-2 and anti-p65 antibodies and the intensity of p65 within the nucleus was quantified using Volocity software. siRNA of Xpo2 in HeLa cells led to a decrease in levels of nuclear p65 after stimulation with TNF-α compared to control cells (non-treated or scrambled siRNA) (S6D and S6E Fig). Since cytoplasmic SpvD and inhibition of Xpo2 expression both led to nuclear accumulation of KPNA1 and reduced levels of nuclear p65, we hypothesized that SpvD might interact with Xpo2. To investigate this, HEK cells were co-transfected with a plasmid encoding Xpo2-FLAG and a plasmid encoding either myc-SpvD or myc-SpvC. Cells were lysed and proteins immunoprecipitated with anti-myc antibody-conjugated beads. Immunoblot analysis revealed that SpvC failed to interact with Xpo2 but that SpvD and Xpo2 formed a complex (Fig 4A). To determine if SpvD interacts with endogenous Xpo2 in infected cells, RAW macrophages and HeLa cells were infected with Salmonella expressing either SpvD-HA or two other effectors translocated by the SPI-2 T3SS (SpvC-HA or SseL-HA), followed by immunoprecipitation using anti-HA conjugated beads. Input and output samples were analysed by SDS-PAGE and immunoblotting with anti-Xpo2 and anti-HA antibodies. SpvD bound to endogenous Xpo2 following its translocation from intracellular bacteria (Fig 4B and 4C). No binding was observed when cells were infected with Salmonella expressing SpvC-2HA or SseL-2HA despite the fact that these two effectors were evidently more abundant than SpvD-2HA (Fig 4B and 4C). Xpo1 and Xpo5 are additional members of the exportin family, have some structural similarities to Xpo2 and mediate the export of proteins with a nuclear export signal and of RNA respectively [48]. However, when membranes were probed with anti-Xpo1 and anti-Xpo5 antibodies, no interaction between SpvD-2HA and these proteins was detected (Fig 4C). To analyse if the interaction of Xpo2 with SpvD affects the cellular localisation of Xpo2, the amount of Xpo2 in total cell lysates and in cytoplasm of HeLa cells infected with Salmonella strains was quantified by immunoblotting after cell fractionation. No cytoplasmic increase of Xpo2 in wild-type -infected cells was detected (S7A Fig). In addition, HeLa cells were transfected with plasmids encoding myc-SpvD or myc-SpvC (as control) and the amount of Xpo2 in total cells lysates and in cytoplasm was quantified by immunoblotting after cell fractionation (S7B Fig). No cytoplasmic increase of Xpo2 in cells transfected with plasmid encoding myc-SpvD was detected, suggesting that SpvD does not sequester Xpo2 in the cytoplasm.
These results show that SpvD localises to the host cell cytoplasm, inhibits nuclear accumulation of p65, causes nuclear accumulation of KPNA1 and KPNA3, and binds specifically to Xpo2. To determine if nuclear lamina-associated KPNA1 induced by SpvD is dependent on Xpo2, confocal microscopy was used to analyse the localisation of KPNA1-FLAG in HeLa cells depleted of Xpo2 and then transfected with a plasmid encoding KPNA1-FLAG with or without a second plasmid encoding myc-SpvD. Cells were immunolabelled with anti-FLAG and anti-myc antibodies and the number of cells with nuclear lamina-associated KPNA1-FLAG were counted (Fig 4D). When SpvD was present, the frequency of lamina-associated KPNA1 was significantly enhanced by 2.1-fold in non-treated HeLa cells and by 2.0-fold in cells exposed to a scrambled siRNA. However, no statistical difference was observed in HeLa cells depleted of Xpo2. Together, these results indicate that the ability of SpvD to induce lamina-associated KPNA1 is dependent on the presence of Xpo2.
Previous work failed to reveal a clear and consistent role of SpvD in virulence in vivo [27–30]. To reassess the possibility of an effect of SpvD on systemic growth of bacteria, mice were inoculated by intraperitoneal injection of a mixture of equivalent cfu of kanamycin resistant wild-type bacteria and the ΔspvD mutant strain. After 72 h of infection, bacteria were quantified after plating spleen homogenates on LB medium and LB medium containing kanamycin to distinguish the strains. The competitive index (CI) for the ΔspvD mutant was 0.48 ± 0.11 (P≤0.05) when compared to wild-type (Fig 5A). To verify that this attenuation was due to loss of SpvD, a second experiment was conducted in which mice were infected as above but with a mixture of ΔspvD mutant and the same strain carrying spvD allele on the pACYC plasmid. The CI for this combination was 0.47 ± 0.11 (P≤0.005) (Fig 5A), showing that the spvD allele rescues the growth defect of the mutant, thereby confirming that SpvD contributes to systemic virulence in the mice.
To assess the effect of SpvD on LPS-induced genes in mice, RNA levels of serpinB2, tnf-α and il1-β in infected splenic macrophages of C57BL/6 mice inoculated with Salmonella strains were compared by qRT-PCR (Fig 5B–5D). For serpinB2 mRNA transcripts, no significant differences were detected between mice infected with different Salmonella strains (Fig 5B). No significant differences in tnf-α and il1-β were detected between mice infected with the wild-type or ΔspvD mutant bacteria, but macrophages infected with ΔspvD, pspvD bacteria contained fewer tnf-α and il1-β transcripts (Fig 5C and 5D), indicating that SpvD can depress cytokine expression in vivo.
As a basis for identifying SPI-2 T3SS effectors that interfere with host proinflammatory immune signalling, we analysed the effect of the SPI-2 T3SS on mRNA levels of tnf-α and il1-β in wild-type BMM at 10 h post-bacterial uptake. No obvious differences were detected, suggesting that either SPI-2 effectors do not have a significant effect on macrophage mRNA levels or that their activities are masked by the massive changes induced by LPS [23] and possibly other PAMPs. Use of TLR4-/- BMMs provided evidence that the SPI-2 T3SS does suppress pro-inflammatory cytokine production. Among several SPI-2 T3SS effector mutants that were tested subsequently, infection with either ΔspvC or ΔspvD strains led to increased tnf-α and il1-β secretion. SpvD and SpvC-dependent effects also occurred in HeLa cells, showing that their functions are not confined to macrophages. The effect of the ΔspvC mutant was unsurprising, since two previous studies have indicated an anti-inflammatory effect of SpvC, caused by its phosphothreonine lyase activity on ERK1/2, p38 and JNK [14,49].
SpvD was identified in a proteomic study as being secreted by both the SPI-1 and SPI-2 T3SSs and its translocation into J774 macrophage-like cells was confirmed using a CyaA fusion protein assay [29]. Using an epitope-tagged version under the control of its own promoter in the virulence plasmid, we showed that SpvD is secreted into the culture medium in a SPI-2 T3SS-dependent manner. SPI-2 T3SS-dependent translocation of SpvD into the cytoplasm of infected cells was also confirmed by immunofluorescence microscopy, but only when cells had been treated with a proteasome inhibitor. In addition, bacterially translocated SpvD-2HA was barely detectable in the input sample from infected HeLa cells and RAW macrophages, whereas SpvC-2HA and SseL-2HA were clearly more abundant (Fig 4B and 4C). This indicates that SpvD is translocated in relatively small amounts and/or is more prone to degradation compared to other effectors. Despite some amino acid similarity in a region containing the catalytic lysine of SpvC, SpvD did not require the corresponding lysine to suppress TNF-dependent activation of an NF-ĸB reporter. Furthermore, following transfection, SpvC but not SpvD inhibited activation of an AP-1-dependent promoter, showing that these effectors are likely to have distinct biochemical activities and physiological effects.
Several other bacterial effectors have been shown to interfere with NF-ĸB signalling. In most cases they affect the degradation of IĸBα and/or the release of p65 [4–6]. For example, the Shigella flexneri kinase OspG interacts with host ubiquitin-conjugating enzymes to abrogate IĸBα degradation and thus blocks nuclear translocation of p65 [50]. NleC of enteropathogenic and enterohaemorrahgic E. coli is a metalloprotease that degrades both free cytosolic p65 and p65 when complexed with IĸBα [33–35]. Abrogation of p65 translocation has also been described for the effectors NleB and NleE from enteropathogenic E. coli and OspZ from Shigella flexneri [51,52] NleE is an unusual methyltransferase that modifies crucial cysteines in the zinc finger domains in ubiquitin-chain sensory proteins TAB2 and TAB3 and thereby disrupts NF-ĸB signalling [53]. NleB is a N-acetylglucosamine transferase whose action blocks signalling from the TNFR and its associated adapters [54–56]. We found that SpvD prevents nuclear accumulation of p65 by interfering with the NF-ĸB pathway after IĸBα degradation but it did not affect overall cellular levels of p65. SpvD might not be the only effector that inhibits the nuclear translocation of p65 following IĸBα degradation. Indeed, the Bordetella effector BopN [57] localises to the nucleus and appears blocks nuclear translocation of p65 without affecting IĸBα degradation; however, its host target(s) remain to be identified. SpvD might specifically target the import of NF-ĸB -associated proteins as it also prevents nuclear accumulation of NF-ĸB subunit p50 but not of STAT2.
There are 7 different importin-α (KPNA) isoforms and several have the capacity to interact with p65. Fagerlund et al. reported an interaction between p65 and both KPNA4 (α3) and KPNA3 (α4) but not KPNA2 (α1) [43,44], while Cunningham et al. showed that p65 binds to KPNA2 [42], Liang et al. reported that nuclear translocation of p65 relies mainly on KPNA2 [47] and Sun et al. showed that p65 translocation in leucocytes does not involve KPNA4, but rather KPNA1 (α5) [45]. p65 might use different isoforms depending on host cell type, time and strength of pathway activation and other conditions. Furthermore, in vitro binding assays with purified proteins or lysates might not reflect normal physiological interactions. However, our results showing that SpvD affects nuclear accumulation of both p65 and p50 are consistent with its effect on KPNA4 (α3), which mediates nuclear import of both transcription factors [43].
We were unable to distinguish endogenous KPNAs by confocal microscopy using commercial antibodies and therefore cells were transfected with plasmids encoding FLAG-tagged KPNAs and anti-FLAG antibodies were used to detect them by confocal microscopy. Following transfection or bacterial translocation, SpvD stimulated a nuclear lamina-associated accumulation of overexpressed KPNA1 and KPNA3. No effect was observed on KPNA2 or KPNA4 localisation. We are unaware of other bacterial proteins that cause a similar effect, but Hantaan virus nucleocapsid protein N inhibits p65 nuclear translocation by binding directly to KPNAs [58]. We were unable to obtain evidence for an interaction between SpvD and KPNA1, 2, 3 or 4 by co-immunoprecipitation assays. Instead, following transfection or bacterial translocation, SpvD interacted specifically with endogenous Xpo2, which is required for recycling of KPNAs from the nucleus to the cytoplasm [46]. Inhibition of Xpo2 expression led to nuclear accumulation of KPNA1 and also reduced nuclear translocation of p65 ([47] and this work). Despite apparently stable binding between SpvD and Xpo2, Xpo2 localisation was not noticeably affected by SpvD. It is possible that SpvD is an enzyme that modifies Xpo2. However, following immunoprecipitation by SpvD-2HA after infection of macrophages, mass spectrometric analysis of Xpo2 did not reveal any post-translational modifications such as phosphorylation, acetylation or conjugation by ubiquitin and ubiquitin-like proteins. It is possible that such modifications did occur but are relatively transient, labile or did not survive the extraction or ionization processes. Alternatively SpvD might affect Xpo2 function non-enzymatically, or use Xpo2 as a docking site or cofactor to mediate an effect on another component of the importin/exportin machinery. A link between binding of SpvD to Xpo2 and the effect of SpvD on KPNA1 was established by showing that the SpvD-induced accumulation of KPNA1 at the nuclear lamina was dependent on Xpo2. Sequestration of KPNA in the nuclear envelope has been observed in cells treated with histone deacetyltransferase inhibitors [59]. However, this treatment also resulted in accumulation of Xpo2 in nuclear aggregates [59], which are clearly distinct from the effects of SpvD.
Due to the lack of clear and consistent data on the contribution of SpvD to Salmonella virulence [27–30], we reassessed its effect on systemic growth of bacteria in mice. Our evidence indicates that SpvD does contribute to systemic virulence in the mouse. The disparity might be attributable to the different mouse strains used in each study. We did not detect a significant difference in the levels of tnf-α and il1-β mRNA in splenic macrophages of wild-type mice infected with the wild-type or spvD mutant strains. However, the complemented mutant in which SpvD was produced from a plasmid caused a significant reduction of tnf-α and il1-β mRNA levels in splenic macrophages of wild-type mice. It seems likely that the effects of SpvD produced by wild-type bacteria are subtle and can be masked by other immune-modulating bacterial molecules present in these relatively short-term infection assays, as evidenced by the initial experiments using wild-type macrophages.
Together, our evidence suggests that an interaction between SpvD and Xpo2 disrupts the nuclear/cytoplasm cycling of KPNA1 and KPNA3, which interferes with nuclear translocation of p65 and activation of NF-ĸB regulated promoters (Fig 5E). Whatever the precise mechanism, it is remarkable that Salmonella has evolved the means to inhibit the two main pro-inflammatory signalling pathways that are activated following detection of bacterial pathogens through two proteins (SpvC and SpvD) that are encoded by adjacent genes on the same operon (Fig 5E).
The S. Typhimurium strains (wild-type 12023 and its mutant derivatives) and plasmids used in this study are listed in S2 Table. Primers used for construction of all strains or plasmids are listed in S3 Table. Bacteria were grown in Luria Bertani (LB) medium at 37°C with shaking and supplemented with ampicillin (50 μg/ml), kanamycin (50 μg/ml) or chloramphenicol (34 μg/ml) as appropriate. S. Typhimurium mutant strains were constructed using a one-step λ Red recombinase chromosomal inactivation system [60]. Plasmid pKD4 was used as the template to amplify the kanamycin (km) resistance gene and amplification reaction products were transferred into pKD46-containing bacteria expressing λ Red recombinase by electroporation. To excise the km resistance marker, the mutant strains were transformed with pCP20 helper plasmid expressing the FLP recombinase. Virulence plasmid-encoded spvD was double hemagglutinin (HA)-tagged using the procedure described previously [61]. To obtain pACYCspvD-2HA, the double-HA tagged spvD was amplified from the virulence plasmid by PCR and cloned into pACYC184. To obtain pRK5-myc-SpvD, pRK5-myc-SpvC, pRK5-myc-YopP and pRK5-myc-NleC, the genes of interest were amplified by PCR and cloned into pRK5-myc. To obtain pRK5-myc-SpvDK185A, site-directed mutagenesis was performed by inverse PCR using pRK5-myc-SpvD as DNA template. Lysine at residue 185 of SpvD was changed to Alanine using SpvDK185A1 and SpvDK185A2. To create pcDNA3.1Xpo2-FLAG, the pCDNA3.1Xpo5-FLAG construct [62] was digested by NheI and XhoI restriction enzymes to obtain pcDNA3.1 vector backbone. In parallel, total cellular RNA was extracted from HeLa cells using the Qiagen RNeasy mini kit following manufacturer’s instructions. RNA were reverse transcribed to cDNA using the SuperScript II Reverse Transcriptase kit (Invitrogen) and Xpo2, encoded by the gene CSE1L, was amplified with the Expand Long Template PCR system (Roche) from total cDNA using primers CAS NheI and CAS XhoI and cloned into pcDNA3.1. Obtained plasmids were verified by sequencing.
The following antibodies were used for immunofluorescence, FACS and immunoblot analysis: goat anti-Salmonella (CSA-1, Kirkegaard and Perry Laboratories), rabbit anti-DnaK [63], rat anti-HA (3F10, Roche), mouse anti-HA (HA11, Covance), mouse anti-FLAG (M2, Sigma-aldrich), rabbit anti-myc (Cell Signaling), mouse anti-β tubulin (Sigma-aldrich), rabbit polyclonal anti-GAPDH (Abcam), rabbit or goat anti-p65 (Santa Cruz), mouse anti- IĸBα (Cell Signaling), rabbit polyclonal anti-lamin B1 (Abcam), rabbit polyclonal anti-Xpo2 (CSE1L, Abcam), rabbit anti-Xpo1 (CRM1, Santa Cruz), rabbit anti-Xpo5 (Sigma-aldrich), rabbit anti-histone H3 (Abcam), rabbit anti-KPNA1 (Proteintech), mouse anti-p50 (Biolegend) and rabbit anti-STAT2 (Santa Cruz). Alexa Fluor 488-, 555- and 633- conjugated donkey anti-rat, anti-mouse, anti-goat and anti-rabbit antibodies were from Life technologies, UK.
Protein secretion assays were done as described previously [31]. Briefly, bacterial strains were grown overnight in LB broth and sub-cultured in MgM-MES medium pH 5.0 for 4 h to assemble and activate SPI-2 T3SS. Bacterial cells were then collected by centrifugation, re-suspended into MgM-MES at pH 7.2 and incubated at 37°C for 1.5 h.
HeLa (human epithelial cell line) cells, HEK (human embryonic kidney) cells and RAW264.7 macrophages used in this study were obtained from the European Collection of Animal and Cell Cultures (Salisbury, UK) and maintained in Dulbecco’s modified Eagle’s medium (DMEM) (Life technologies) supplemented with 10% foetal calf serum (FCS) (PAA Laboratories or Sigma-Aldrich) at 37°C in 5% CO2. Primary bone-marrow macrophages (BMM) were obtained from C57BL/6 wild-type (Charles River) or TLR4-/- mice (kind gift from Prof Maria Belvisi and Dr Mark Birrell (Respiratory Pharmacology, National Heart and Lung Institute, Imperial College, London)). BMM were grown in RPMI (Life Technologies) supplemented with 10% FCS, 2 mM glutamine, 1 mM sodium pyruvate, 10 mM HEPES, 50 μM β-mercaptoethanol, 100 U/ml penicillin/streptomycin (Sigma-Aldrich), and L929 cell-conditioned medium 20% (vol/vol; National Institute for Medical Research). After 3 days of culture, further fresh complete medium containing L929-cell conditioned medium was added to the growing macrophages. On day 7, cells were washed and seeded in complete medium without antibiotic and incubated for 24 h before bacterial challenge.
HeLa cells were seeded in 24-well plates at a concentration of 2 x 104 cells/well 24 h before transfection with Lipofectamine 2000 (Life technologies) following the manufacturer’s protocol. Unless indicated, 100 ng of DNA were used. Cells were used 24 h after transfection.
The siRNA oligo duplex targeted against Xpo2 was custom synthesized using a previously described targeting sequence [64] (ThermoScientific). The scramble control oligos (ThermoScientific) were designed not to target any human mRNA transcript.
HeLa cells were seeded in 24-well plates at a concentration of 5 x 104 cells/well and grown until they were 70% confluent. siRNA transfection using RNAiMAX (Life technologies) was carried out according to the manufacturer’s protocol with a final concentration of siRNA oligos of 10 nM. A scramble oligo sequence was included in all experiments as a negative control. Cells were used 72 h after siRNA transfection.
HeLa cells and RAW264.7 macrophages were infected with S. Typhimurium strains as described previously [65]. BMM were infected as described previously [66].
HeLa cells were seeded into 24-well plates at a concentration of 2 x 104 cells/well 24 h before stimulation. Unless indicated, cells were stimulated with TNF-α (10 ng/mL) for 45 min at 37°C. For IĸBα degradation analysis, at indicated times after TNF-α stimulation, cells were washed and either lysed for immunoblotting analysis or trypsinised for flow cytometry assays.
All samples were fixed in 3% paraformaldehyde (PFA) and unless indicated, permeabilised in 0.2% Triton-X100 for 6 min. All antibodies were diluted to the appropriate concentrations in PBS containing 10% horse serum. The coverslips were washed twice in PBS, incubated with primary antibodies for 1 h, washed 3 times in PBS, incubated with secondary antibodies for 30 min and stained with the nucleic acid dye, DRAQ5 (Alexis) for 20 min. For specific plasma membrane permeabilisation, all antibodies were diluted in PBS containing 10% horse serum and 0.1% saponin (Fisher) and after fixation with PFA, cells were washed twice in PBS, incubated with primary antibodies for 1 h, washed 3 times in PBS and incubated with secondary antibodies for 30 min. Coverslips were washed and mounted onto glass slides using Mowiol mounting medium. Cells were analysed using either an epi-fluorescence microscope (BX50; Olympus) or a confocal laser-scanning microscope (LSM510 or LSM710; Zeiss GmBH). All scorings were done blindly and at least 100 cells were analysed per coverslip. To determine p65 intensity in the nucleus, confocal three-dimensional Z-stacks were acquired for each sample using a 63x objective with a slice of increment of 0.5 μm. At least 50 infected cells were imaged for each condition. Rendered three-dimensional stacks were analysed with Volocity image analysis software (Perkin Elmer).
HeLa cells were collected from a well in a 24-well plate 3 days after transfection with siRNAs. Infected BMM were collected from a well in a 24-well plate at 10 h post-bacterial uptake. For IĸBα degradation analysis, at indicated times after TNF-α stimulation, cells were washed and collected after trypsinisation. Cells were lysed in 5x sample buffer (0.25 M Tris-Cl pH 6.8, 10% SDS, 10% β-Mercaptoethanol, 10% glycerol, 0.05% Bromophenol Blue). Samples were boiled for 5 min and separated by SDS-PAGE. Samples were then transferred onto PVDF Immobilon-P membranes (Millipore), and Immunoblotting was carried out according to the manufacturer’s instructions. Immunoblots shown in each figure are representative of three independent experiments.
For cell fractionation, 5 x 106 cells were collected, washed twice with cold PBS and lysed in cold hypotonic buffer (10 mM HEPES pH 7.9, 10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA, 1 mM DTT) supplemented with complete protease inhibitor cocktail without EDTA (Roche) on ice for 15 min. NP-40 was added to a final concentration of 0.625% and cells were vortexed vigorously for 10 s. Samples were centrifuged for 30 s at 16000 g and the supernatants were harvested as cytoplasmic fraction. The nuclear pellets were then resuspended in cold hypertonic buffer (20 mM HEPES pH 7.9, 0.4 M NaCl, 1 mM EDTA, 1 mM EGTA, 1 mM DTT) supplemented with complete protease inhibitor cocktail (Roche). The samples were incubated at 4°C for 15 min with agitation. The supernatants were collected after centrifugation for 5 min at 16000 g as nuclear proteins. Cytoplasmic and nuclear proteins were frozen in 5x sample buffer at -20°C until use.
For flow cytometry assays to determine protein levels by immunolabelling, trypsinised cells were washed twice in PBS prior to fixation in 3% PFA for 20 min at room temperature. Cells were then washed twice in PBS and permeabilised in 0.08% Triton X-100 for 10 min. Cells were incubated for 1 h at room temperature with primary antibodies diluted in PBS. Cells were then washed twice and incubated for 30 min with secondary antibodies diluted in PBS. Finally, cells were washed and resuspended in PBS. Flow cytometry was performed on a two-laser, four colour FACSCalibur cytometer (BD Biosciences) using Cell Quest Pro software. For each sample, 45,000–50,000 events were analysed. Collected data were analysed with FlowJo 8.1.1 software (Treestar).
Cells subjected to LPS (1 μg/ml) stimulation for 2 h were used as controls. At 10 h post-uptake, BMM were washed and RNA was isolated using TRIzol according to the manufacturer’s directions (Invitrogen). Contaminating genomic DNA was removed using DNaseI (Qiagen). RNAs (400 ng) were reverse transcribed with Quantiscript Reverse transcriptase (QuantiTect Reverse Transcription kit, Qiagen) for 25 min at 42°C. Quantification of the mRNA levels was done using SensiMix dT kit (Quantace) and specific primers (S3 Table) on Rotor-Gene 3000 (Corbett Research).
For TNF-α and IL1-β quantification, at 10 h or 24 h post-bacterial uptake or after 2 h of LPS (1 μg/ml) stimulation, supernatants from infected BMM were collected, centrifuged and stored at −80°C. For IL-8 quantification, at 8 h, 12 h or 24 h post-bacterial uptake, supernatants from infected HeLa cells were collected, centrifuged and stored at −80°C. The amount of TNF-α, IL1-β or IL-8 released in the culture supernatant was determined by enzyme-linked immunosorbent assay (ELISA; R&D Systems) and cytokine concentrations were assessed according to the manufacturer's instructions.
HEK293 cells were seeded at a density of 5 x 104 cells per well in a 24-well plate 16 h prior to transfection. Cells were transfected for 16 h with 50 ng of luciferase reporter plasmid (AP-1 or NF-ĸB dependent luciferase reporter plasmid), 30 ng of pTK-Renilla luciferase, and 100 ng of expression vectors (myc-effector or myc vector alone). For MAPK reporter assays, cells were then incubated with 25 ng/ml of PMA for 6 h and harvested in 100 μl of passive lysis buffer (Promega). For NF-ĸB reporter assays, cells were then incubated with 10 ng/ml of TNF-α for 8 h and harvested in 100 μl of passive lysis buffer (Promega). Luciferase activity was measured using Dual Luciferase reporter assay system (Promega) and a TD20/20 Luminometer (Turner Designs) and normalised according to Renilla luciferase intensity. The data presented are from at least three independent experiments.
HEK cells were seeded at a density of 2 x 105 cells per well in a 6-well plate 16 h prior to transfection. Cells were transfected with 1 μg of indicated DNA with Lipofectamine 2000. Three wells were used per condition. Cells were lysed 24 h post-transfection in GTPase lysis buffer (50 mM Tris Cl pH 7.4, 150 mM NaCl, 5 mM MgCl2, 0.5% Triton X-100) supplemented with complete protease inhibitor cocktail without EDTA (Roche) on ice for 30 min. Lysates were then centrifuged at 16000 g and the supernatants were harvested and incubated with 25 μl of anti-myc conjugated beads for 2 h at 4°C. Samples were then centrifuged at 1500 g and supernatant was removed. Pellets were washed 3 times with 50 mM Tris Cl pH 7.4. Samples were then separated by SDS-PAGE and analysed by immunoblot.
HeLa cells and RAW macrophages were grown on 150 mm dishes (Corning) seeded at a concentration of 4x106 cells and 8x106 per dish, respectively. Two dishes were used per condition. Cells were then infected with S. Typhimurium strains expressing either SpvD-2HA, SpvC-2HA or SseL-2HA for 18 h. Proteasome inhibitor MG132 (10 μg/ml) was added to the culture medium for 2 h prior to cell lysis. Cells were lysed in GTPase lysis buffer supplemented with complete protease inhibitor cocktail without EDTA (Roche) on ice for 30 min. Lysates were then centrifuged at 16000 g and the supernatants were harvested and incubated with 25 μl of anti-HA beads (Pierce) for 2 h at 4°C. Samples were then placed on a magnetic rack and supernatant was removed. Pellets were washed 3 times with 5 mM Tris Cl pH 7.4, 15 mM NaCl, 0.5 mM MgCl2, 0.05% Triton X-100. Proteins bound to anti-HA beads were eluted using HA-peptide (Sigma-Aldrich) and were then separated by SDS-PAGE and analysed by immunoblotting.
To prepare the inocula, bacteria were first grown overnight in LB broth and then subcultured at a dilution of 1:100 for a further 2 h. Cultures were diluted to a concentration of 2.5 × 104 cfu/ml in physiological saline.
For competitive index (CI) measurements, bacterial cultures were mixed for intra-peritoneal inoculation (0.2 mL per mouse). Viable bacteria in inocula were quantified by dilution and plating onto LB agar plates with appropriate antibiotics to distinguish between strains. Female C57BL/6 mice (Charles River, 6–12 weeks) were sacrificed at 3 days post inoculation. The spleens were removed aseptically and homogenized in distilled water by mechanical disruption. Serial dilutions were plated on LB agar for cfu enumeration. Strains were distinguished by differential counting or replica plating on antibiotic-supplemented plates. For each mouse, the CI was calculated by dividing the output ratio (i.e. strain a versus strain b) divided by the input ratio. The log CI values were used to calculate means and for statistical analyses. A wild-type strain carrying a kanamycin resistance cassette on the chromosome (STM0857, [14]) was used to represent wild-type bacteria in the CI assay. Its CI was 0.97 ± 0.18 (Fig 5A) when compared to wild-type bacteria, indicating that kanamycin resistance cassette does not affect virulence.
For mRNA extraction, mice were infected by intra-peritoneal inoculation with 1 × 104 bacteria and sacrificed 48 h post-inoculation. Spleens were harvested, homogenized and splenic CD11b(+) cells enriched using magnetic beads following manufacturer’s instructions (Milteneyi Biotec). RNA of purified cells was isolated using TRIzol according to the manufacturer’s directions (Invitrogen). mRNA reverse transcription and qRT-PCR were done as described above.
Animals were used in accordance with UK Home Office regulations. The Imperial College Animal Welfare and Ethical Review Body (AWERB) committee approved the project licence for animal research (70/7768). The following people formed the panel: Applicant Scientist, CBS site manager / NACWO, NVS, Peer Scientist and a Lay person.
All results are reported as mean ± Standard Error of the Mean (SEM). Statistical analyses were done using two-tailed unpaired Student’s t-test or ANOVA followed by Bonferonni's multiple comparison test. Differences denoted in the text as significant fall below a p-value of 0.05.
|
10.1371/journal.pgen.1003646 | Genome-Wide Association Mapping in Dogs Enables Identification of the Homeobox Gene, NKX2-8, as a Genetic Component of Neural Tube Defects in Humans | Neural tube defects (NTDs) is a general term for central nervous system malformations secondary to a failure of closure or development of the neural tube. The resulting pathologies may involve the brain, spinal cord and/or vertebral column, in addition to associated structures such as soft tissue or skin. The condition is reported among the more common birth defects in humans, leading to significant infant morbidity and mortality. The etiology remains poorly understood but genetic, nutritional, environmental factors, or a combination of these, are known to play a role in the development of NTDs. The variable conditions associated with NTDs occur naturally in dogs, and have been previously reported in the Weimaraner breed. Taking advantage of the strong linkage-disequilibrium within dog breeds we performed genome-wide association analysis and mapped a genomic region for spinal dysraphism, a presumed NTD, using 4 affected and 96 unaffected Weimaraners. The associated region on canine chromosome 8 (pgenome = 3.0×10−5), after 100,000 permutations, encodes 18 genes, including NKX2-8, a homeobox gene which is expressed in the developing neural tube. Sequencing NKX2-8 in affected Weimaraners revealed a G to AA frameshift mutation within exon 2 of the gene, resulting in a premature stop codon that is predicted to produce a truncated protein. The exons of NKX2-8 were sequenced in human patients with spina bifida and rare variants (rs61755040 and rs10135525) were found to be significantly over-represented (p = 0.036). This is the first documentation of a potential role for NKX2-8 in the etiology of NTDs, made possible by investigating the molecular basis of naturally occurring mutations in dogs.
| Neural tube defects (NTDs) are birth defects resulting from errors in the closure of the neural tube, an embryonic structure which develops into tissues of the central nervous system during pregnancy. NTDs commonly lead to costly lifelong disabilities. They are considered to be caused by a combination of nutritional, inherited and environmental factors, and their interactions. However, an obvious mechanism is currently unknown. Genetic studies in human populations are made difficult by the multifactorial nature of NTDs and because multiple cases within a single family are rare. Animal models are helpful in dissecting the genetics of such complex traits; however existing rodent models do not explain all of the NTD cases in humans. Dogs are excellent biomedical models for humans since they receive comparable medical care, share our home environment, and develop naturally occurring diseases comparable to those in humans. We used a naturally occurring NTD in Weimaraner dogs, termed spinal dysraphism, to identify a mutation in an associated regulatory gene, NKX2-8. Mutations in NKX2-8 were subsequently documented in human patients with a generally similar NTD termed spina bifida. This is the first documented evidence that NKX2-8 has a role in NTDs. It is expected that this discovery will contribute to our understanding of the mechanisms leading to NTDs.
| Neural tube defects (NTDs) is a general term used to describe developmental defects resulting from abnormal closure or development of the neural tube during embryogenesis. The resulting pathologies may involve the brain, spinal cord, and associated structures such as vertebrae, soft tissue or skin [1]. The condition is reported among the more common birth defects in humans (incidence of ∼1 per 1,000 pregnancies worldwide), leading to significant infant morbidity and mortality [2], [3]. Clinically, human NTDs are defined as a severe, “open” form, in which tissues of the nervous system are exposed to the environment or a “closed,” form, characterized by skin-covered lesions [1]. NTDs are the outcome of aberrant primary or secondary neurulation during embryogenesis. During primary neurulation, the neural plate folds on itself and fuses on the midline into the neural tube [4]. The brain and most of the spinal cord are formed by primary neurulation. Flawed primary neurulation typically leads to open forms of NTDs which include anencephaly and spina bifida. Spina bifida (failure of vertebral fusion) usually occurs secondary to failure of closure of the neural tube, and causes a spectrum of physical and developmental disabilities, depending on the magnitude and position of the spinal defect [3]. Spinal dysraphism is an alternative terminology, describing conditions with malformations of structures relating to the midline raphe of the developing spine; generally implying neural tube defects [5]. Secondary neurulation is defined as the formation of the caudal portion of the neural tube from the pluripotent cells of the tail bud and does not require folding as the solid cell mass undergoes cavitation. Secondary neurulation creates most of the sacral and all of the coccygeal tissues of the spinal cord [4]. Pathologic secondary neurulation may result in closed forms of spina bifida, where the nervous tissue fails to separate from the other tissues of the tail bud [6].
The etiology of NTDs remains poorly understood and genetic, nutritional (folate, inositol), and environmental factors, or a combination of these, are known to play a role in the development of NTDs [1], [2]. Genes involved in the complex multistep process of neurulation such as those in the the planar cell polarity (PCP) pathway [4], [7], and genes involved in folate metabolism [8], [9], have been found to contribute to NTDs. Mouse models have led to the identification of over 200 genes with roles in NTDs [10], [11]. In addition to null mutants, many knock-down or compound mutants have been identified, and some mutations, like the kinky tail mouse [12]–[14], cause NTDs in oligogenic combinations, or like the curly tail mouse, display complex inheritance [11], [15], [16].
Similarly, it is presumed that most human NTDs have a multifactorial etiology and arise in a similar fashion to some of the mouse phenotypes [6]. Although multiple genetic variants that increase the risk of developing NTDs have also been recognized for humans [17], the identified genetic variation does not explain the total genetic contribution to the incidence of NTDs observed in human populations [3]. Utilizing conventional approaches such as positional cloning and genetic linkage to identify additional associated variants is hampered by the rarity of families with multiple affected individuals, and because of undersized cohorts leading to suboptimal power for association studies [3], [17].
In contrast to association studies in human populations, the dog is a large animal model that is particularly useful for whole genome association studies due to its unique population structure [18], [19]. Haplotype blocks in LD (linkage disequilibrium, non-random association of alleles at two loci or more) extend across 0.4 to 3.2 megabases [20], [21], and are similar to the extent of LD in inbred strains of mice; simplifying genetic analyses [22], [23]. In addition, the variable conditions associated with NTDs occur naturally in dogs [24], and spina bifida has been reported to occur sporadically in breeds such as the English Bulldog [25], Toy Poodle [26], [27] Collie, Chihuahua, mixed-bred dogs [27] and Samoyed [28]. Furthermore, Weimaraner dogs with a NTD, commonly referred to as “spinal dysraphism” were studied previously [29], [30]. Research colonies of affected Weimaraners have been maintained [31], [32], and breeding experiments included mating of severely affected Weimaraner dogs which resulted in 10/10 affected fetuses [32], [33], suggesting a recessive mode of inheritance. However, other breeding experiments did not support a conclusive mode of transmission [30] and the disorder was speculated to have a complex mode of inheritance in the Weimaraner [29].
Extensive studies by McGrath [30], demonstrated that the heterogenic spinal pathology in Weimaraners includes duplicated, stenotic, or absent central canal, hydromyelia or syringomyelia, chromatolysis and loss of nerve cell bodies in gray matter, disrupted dorsal median septum and ventral median fissure, and gray matter ectopias [30]. In any affected Weimaraner, these histopathological changes may be present in varying degrees within different spinal cord segments, but occur most frequently in the lumbosacral region. Engel and Draper [33] reported that abnormalities in affected Weimaraner prenates were evident in embryos (24 days of gestation) and consisted of failure of the dura mater to separate from the periosteum, absence of the ventral median fissure and fusion of ventral white matter, and disruption of gray matter structure. Central canal diverticula were common and the diameter ratio of gray matter/spinal cord was significantly greater in affected fetuses [33].
In the live Weimaraner, McGrath [30] observed abnormal hair streams along the back, similar to those observed in some of the human patients [34], kinked tails resembling the curly and kinked tail mouse phenotypes [15], [12], and scoliosis of the vertebral column in the lumbar spinal region [30]. Clinical signs include paraparesis and a symmetric “bunny-hopping” or simultaneous use of the pelvic limbs, a bilateral withdrawal reflex; pinching one paw elicits flexion of both hindlimbs, a crouched stance, and deficient proprioception in the pelvic limbs [30]. Although spina bifida was not observed in Weimaraner cases, they share many of the human and mouse NTDs phenotypes which suggests that an evolutionary conserved molecular pathway may be contributing to the pathoetiology of NTDs across these species.
Whereas previously maintained colonies of affected Weimaraners allowed for a thorough description of the phenotype and experimental breeding has confirmed that “spinal dysraphism” is an inherited condition in the Weimaraner breed, no mutation has been identified to date. We used four cases of presumed spinal dysraphism to map a genomic location for the disorder in the breed. A regional homeobox candidate gene with functions in neuronal development, NKX2-8, was sequenced and a frameshift mutation was identified in affected Weimaraners. Human patients with spina bifida had a significant increase in the rate of rare missense mutations within evolutionary conserved residues of NKX2-8.
Four unrelated Weimaraners showing the clinical signs typical of “spinal dysraphism” (Video S1, Figure S1) and 96 unaffected Weimaraners were genotyped using ∼173 k SNP markers. Following quality control, 114,775 SNPs were retained in the genome-wide association study (GWAS) analysis using PLINK, and an associated region on canine chromosome 8 was observed (Figure 1A and 1B). To confirm that this region was not falsely associated due to population substructure, we reviewed the quantile–quantile plots with the associated SNPs on chromosome 8 (λ = 1.03, Figure S2A) and without the SNPs on chromosome 8 (λ = 1.01, Figure S2B). The associated region extended over ∼1.5 Mb. Within this region, a distinct homozygous haplotype was present within the affected dogs (Figure 1C), and was absent in all 96 unaffected dogs.
Of the 18 regional genes (Table S1) NKX2-8 (chr8: 18,156,525–18,157,928) was shown to regulate key steps in spinal accessory motor neuron development in the mouse, and adult mice with targeted disruptive NKX2-8 mutations exhibit abnormal locomotion, including a permanent or intermittent hopping gait. The two exons of NKX2-8 were sequenced in genomic DNA and an alteration of a G to AA was identified within exon 2 in an affected Weimaraner when compared to unaffected Weimaraners and to the Boxer reference genome [21]. Two obligate carriers (parents of affected dogs) and two littermates of affected dogs were heterozygous for the mutation. The three genotypic variants observed within exon 2 of NKX2-8 are shown in Figure 2.
In order to determine the NKX2-8 protein sequence, we acquired the complete cDNA sequence, including the 5′ and 3′ untranslated regions from brain tissue of an unaffected Beagle. Subsequent translation of the exonic sequence of an affected Weimaraner revealed that the identified alteration functions as a frameshift mutation which introduces an amino acid change (A150VfsX1) and a downstream stop codon (Figure 3). No additional mutations were identified within the promoter region or within the exon-intron boundary of NKX2-8 in genomic DNA of affected dogs. 109 additional unrelated unaffected Weimaraners were tested for the presence of the A150fs mutation by direct sequencing and three carrier dogs were identified. The mutation frequency was therefore calculated to be ∼1.4% within the Weimaraner breed. One additional case of clinically affected Weimaraner had no copies of the mutation. Additionally, 496 unaffected dogs, from six breeds reported to be clinically affected by NTDs, and a Chesapeake Bay Retriever diagnosed with myelodysplasia, absent ventral median fissure, hydromyelia, and syringomyelia by histopathology, were tested to determine whether or not this is an allelic mutation. No copies of the mutation were found within non-Weimaraner dogs.
To investigate a potential role for NKX2-8 in cases of NTDs in human patients, 149 unrelated samples from patients with lumbosacral myelomeningocele, (spina bifida), were sequenced. Six missense variants were identified in exon 2 of NKX2-8 within the spina bifida cohort. Five patients (3 females, 2 males), all European Americans were heterozygous for variant rs61755040, which has a reported minor allele frequency (MAF) of 0.0073 in dbSNP. This missense variant, results in an amino acid change of serine to threonine at position 62 of human NKX2-8, within an area of complete evolutionary conservation (Figure 3). The S62T alteration is predicted to be “probably damaging” by PolyPhen. Of the 149 samples, only 19 belonged to African American spina bifida patients. While we did not have adequate sample size to examine this ethnic group statistically, we did find that one female African American was heterozygous for variant rs10135525, which has a reported MAF of 0.0014 in dbSNP. This missense mutation results in an amino acid change of alanine to threonine at position 94, within the evolutionary conserved homeobox functional domain of human NKX2-8 (Figure 3) and is also predicted to be “probably damaging” by PolyPhen. As shown in Figure 3, both variants reside within domains of 100% identity between human (Homo sapiens), dog (Canis lupus familiaris), cat (Felis catus), cow (Bos taurus), bat (Pteropus alecto), wild boar (Sus scrofa), mouse (Mus musculus), tree-shrew (Tupaia chinensis), chicken (Gallus gallus) and zebra fish (Danio rerio).
Using the Exome Variant Server (EVS) data as a control population for spina bifida, we compared missense variants in the European American spina bifida population versus the EVS population. The EVS European American database contains 6 variants in NKX2-8 (nonsense or missense) in a total of 72 variant alleles out of an average of 8,500 alleles sequenced (Table S2). The difference between the frequency of missense variants in spina bifida cases versus controls was significant by one tailed Chi-squared analysis with Yate's correction (p = 0.036).
Using LD mapping in dogs, we identified an A150fs frameshift mutation which segregates within the Weimaraner breed in spinal dysraphism affected dogs and their relatives. The frameshift mutation was absent in 496 dogs from six breeds that were previously reported in the literature as presenting with cases of spina bifida, and in a case of spinal dysraphism in a Chesapeake Bay Retriever dog. Our results suggest that this is a private mutation in Weimaraners which is not shared between breeds. The mutation does not segregate as a benign polymorphism in canine populations, supportive of a causative role for the mutation.
The mutation was found in a homozygous state within spinal dysraphism cases and recessive Mendelian transmission was verified by genotyping two parents (obligate carriers) and two littermates whose samples were available. Additionally, the ∼1.5 Mb genome-wide associated region is comprised of a homozygous haplotype present in the affected dogs which best fit a recessive model of inheritance [35]. Previously, Karlsson et al. demonstrated that a Mendelian recessive trait could be successfully mapped within a single dog breed with fewer than 15 cases and 15 control dogs [18]. We used a genome-wide case-control association study to map spinal dysraphism with merely four cases; providing further evidence for the effectiveness of the dog as a model organism for inherited diseases.
The ∼1.5 Mb associated haplotype contained a tight cluster of associated SNPs and 18 regional candidate genes. Among these genes, NKX2-8 was an appealing candidate since it belongs to a family of vertebrate developmental regulators (homeodomain transcription factors) that are homologues of the Drosophila homeodomain transcription factor, NK2 [36]–[38]. NKX2-8 is expressed in the developing neural tube [39], connecting it with the sub-group of the Nk2 genes which is expressed in the central nervous system [36], [37]. This group includes the Drosophila vnd gene and the vertebrate Nkx2.1 and Nkx2.2 genes [39]. This suggests an early role for the Nk2 gene family in the development of the nervous system before the divergence of deuterostomes. The NKX2-8 protein has an N-terminus conserved homeobox DNA binding domain; involved in the transcriptional regulation of key developmental processes, and a C-terminus conserved NK specific domain with transcriptional activity [40]. The A150fs frameshift mutation identified in Weimaraners introduces an early stop codon and the truncated protein lacks the NK specific domain. This mutation may lead to impaired NKX2-8 function during embryonic development and thus the observed neurospinal pathologies in homozygous Weimaraners. Support for this proposed etiology come from extensive experiments in mice [41].
The function of the murine NKX2-8 homolog, Nkx2-9, was evaluated by targeted disruptive mutations in the related homeodomain transcription factor Nkx2-2 [42], and in Nkx2-9 [43]. The experiments suggested that both proteins play essential and partially redundant roles in the development of distinct neuronal populations in hindbrain and ventral spinal cord [41]. The authors observed impaired floor plates that led to defects in axonal pathfinding of commissural neurons in Nkx2-9 mutants. Intriguingly, adult mice with disruptive mutations in Nkx2-9 exhibit varying degrees of abnormal locomotion, observed predominantly for hindlimbs as continuous hopping with no alternating activity of left and right legs, similar to the Weimaraner phenotype (supplementary video) [41]. In vitro recordings in spinal cord preparations from newborn mutant mice, showed markedly reduced coordination of locomotor-like activity with increased variability in both left-right and flexor-extensor coordination [41]. The authors concluded that disruption of the Nkx2-9 gene results in a strong walking impairment and substantial locomotor deficits, both in vitro and in vivo [41]. Interestingly, only 75–90% of homozygous Nkx2-9 mutants exhibit the hopping gait phenotype, suggesting that there is reduced penetrance in the mouse, or possibly a “leakage” of the null phenotype [44]. While spinal dysraphism in Weimaraners was previously reported as a disorder for which the penetrance is reduced [30], we have not seen evidence for reduced penetrance within the tested group of Weimaraners (n = 210).
A single case of a Weimaraner with clinical signs of ataxia and paraparesis did not share the NKX2-8 A150fs mutation. It is possible that a second mutation exists in Weimaraners, which may account for the previously reported inconsistent transmission [30]. It also cannot be ruled out that in this case the clinical signs are the result of environmental factors such as maternal hyperthermia, nutritional imbalances, medication or abnormal glucose metabolism; factors mentioned in epidemiological studies as leading to congenital spinal defects [1]. In absence of diagnostic imaging or histopathological evidence, only a suggestive diagnosis could be made for spinal dysraphism based on case history, signalment and findings on a neurological examination. The identification of a mutation which segregates in the breed may aid in reaching a diagnosis in live pet dogs. In the future, Weimaraner breeders will be able to select against this mutation through DNA screening of prospective breeding animals. Identification of a mutation leading to a NTD in the dog could be utilized to improve our understanding of NTDs in human patients.
When a cohort of 149 spina bifida patients was tested, we identified 6 cases that had one of two heterozygous missense mutations (rs61755040 or rs10135525) within exon 2 of the NKX2-8 gene. Interestingly, the dog frameshift mutation is also located in exon 2 of the NKX2-8 gene, suggesting that exon 2 might be more susceptible to damaging DNA mutations than exon 1 of NKX2-8. The missense mutations identified in spina bifida patients alter evolutionary conserved amino acid residues and functional consequences are predicted. While these variants are rare (MAF<0.007) in controls, they may explain 4% of the cases within this cohort. Nonetheless, future functional studies are needed in order to confirm that these are harmful mutations which may cause a mutant phenotype under certain conditions. Previous studies that present existing evidence to support a causative role for the variants identified within NKX2-8 include the work performed on VANGL1, one of the genes of the well-studied PCP pathway [45], [46]. Similar to our results; Loop-tail (Lp) mice with NTDs had recessively inherited mutations, but when human patients were sequenced, heterozygous missense mutations were identified within the VANGL1 gene. This suggests that heterozygous missense mutations within genes with critical functions during development play a role in the etiology of NTDs in human patients. Another discovery of heterozygous variants in human patients was made while investigating the FZD6 gene of the PCP pathway [47]. While the inheritance in the mouse model was recessive, the variants discovered within human patients were heterozygous [48], resembling our results.
It is possible that the variants identified in spina bifida patients, have a dominant negative effect. NKX2-8 has a DNA-binding homeobox domain which is shared by a large variety of transcriptional regulators involved in controlling development [49]. The mutations in spina bifida patients were identified in domains of absolute evolutionary conservation; both adjacent to, and within the homeobox domain. Missense mutations within homeobox domains of various genes were studied previously to reveal that in patients with congenital diseases, most missense mutations have dominant effects [49]. Amino acids of homeobox domains play the critical roles of determining the correct structural fold of proteins, regulating DNA-protein interactions, regulating protein-protein interactions and signaling nuclear localization. It was previously proposed that heterozygous mutations within homeobox genes may have a detrimental effect during development due to haploinsufficiency [50], [51].
While it is possible that the mutations identified in NKX2-8 have a dominant negative effect, it was previously hypothesized that NTDs inheritance is multifactorial [1]. A threshold model is used to explain multifactorial contribution to the NTD dichotomous phenotype, where multiple genetic variants interact with environmental factors to cause NTDs [52]. According to the threshold model, the variants contributing to the elevated risk would be present in controls, but a significant higher frequency of these variants is expected in cases [52], [53]. Our results are consistent with this theory; however, further studies are required in order to determine the inheritance pattern of NKX2-8 mutations in NTDs patients.
Using the tractable genome of the dog for association mapping of naturally occurring NTDs, we identified a frameshift mutation in NKX2-8. Additionally, rare missense variants in NKX2-8 were identified in 4% of the cases in a cohort of spina bifida patients. To the best of our knowledge, this is the first documentation of a potential role for NKX2-8 in the development of NTDs. Future functional studies are required in order to provide insights into the mechanisms and etiologies which constitute NTDs in both species.
All research involving human participants was approved by Northwestern University (Chicago) and the University of Iowa institutional review boards (IRBs). Informed consent was obtained and all clinical investigation must have been conducted according to the principles expressed in the Declaration of Helsinki.
DNA samples of domestic dogs (Canis familiaris) owned by private individuals were used in this study. We accepted samples from dogs of all ages and of both sexes. The sample collection protocol was approved by the University of California, Davis Animal Care and Use Committee (protocol #16892).
Weimaraner samples were solicited using advertisements posted in the Weimaraner Club of America (WCA) magazine, on the WCA website, by direct communication with Weimaraner owners and treating veterinarians, and via the Veterinary Information Network (VIN). All of the samples used in this study came from shorthaired American dogs. Presumed (no histopathology) spinal dysraphism cases were brought to our attention by treating clinicians, Weimaraner breeders and owners. Veterinary evaluation of congenital non-progressive neurological abnormalities which consisted of pelvic limb ataxia, paraparesis, and delayed proprioceptive positioning in the pelvic limbs; together with patient signalment and history served to make a suggestive diagnosis. Samples from additional breeds of dogs were obtained from patients of the Veterinary Medical Teaching Hospital at UC Davis.
DNA was extracted from blood samples in EDTA using a commercially available kit (Puregene, Gentra Systems, Minneapolis, MN). Additional DNA isolation from buccal swabs was performed as previously described [54].
DNA samples were genotyped using the Illumina 170K CanineHD BeadChip (Illumina, San Diego, CA). Quality control checks on the canine dataset were performed for individuals and SNPs using GenABEL in the R statistical package [55]. SNPs were excluded if they had a minor allele frequency (MAF)<5%, a genotype call rate <95%, or if they deviated from the Hardy-Weinberg Equilibrium (HWE). A total of 114,775 SNPs passed the quality control check and were available for analysis. The retained SNPs were then used for case-control chi-square statistical analysis by PLINK [56], and Manhattan and quantile-quantile (QQ) plots were generated using GenABEL [55]. We assessed the effect of population stratification by examining the QQ plots for deviation of the p-values from the null hypothesis. We considered a significant genome-wide association if the SNPs p-value was below the 5% Bonferroni-corrected threshold (p≤0.05; −log 10≥1.3). To derive the genome-wide significance thresholds we repeated the GWAS with 100K Max (T) permutations.
Adult beagle total RNA was obtained from Zyagen (San Diego, CA, USA). cDNA was synthesized with the SuperScript III First-Strand Synthesis System for RT-PCR (life technologies, Grand Island, NY 14072, USA). Primers for the complete cDNA of NKX2-8 were designed using the Primer3 program [57]. cDNA PCR products were cloned using the TOPO TA Cloning kit (pCR2.1-TOPO vector) with One Shot TOP10 Chemically Competent E. coli (life technologies, Grand Island, NY 14072, USA). Products were isolated with the Qiaprep Spin Miniprep kit (QIAGEN, Valencia, CA 91355, USA) and sequenced as described below. Nucleotide sequences were translated into amino-acid sequences with Vector NTI software (Applied Biosystems, CA 92008).
Primers for the two exons and for the 3 and 5′ UTRs of NKX2-8 were designed using the Primer3 program [57] (Table S3). The primers, E2F2: 5′ CTGGTAGGCGGGGAAGAG; and E2R2: 5′ GGTTCCAGAACCATCGCTAC, were used to generate PCR products which flank the frameshift mutation within exon 2 of the NKX2-8 gene. PCR was performed using 40 ng of DNA, 1 unit of AccuPrime GC-Rich DNA Polymerase, 5 ul of buffer A (AccuPrime high GC DNA polymerase kit; life technologies, Grand Island, NY 14072, USA), 50 ng forward and reverse primers, in 25 ul reaction volume. Cycle conditions of 3 min at 95°C followed by 35 cycles of 30 s at 95°C, 30 s at 62°C, and 1 min at 72°, with a final extension of 20 min at 72°C were used. Primers (Table S1) were used to generate overlapping sequences to complete the genomic sequence of NKX2-8. The sequence upstream of the gene was missing on the May 2005 CanFAm2.0 genome assembly (viewed using the UCSC genome browser), and was captured using the LongAmp Taq PCR Kit (New England BioLabs Ipswich, MA 01938, USA).
The PCR products were electrophoresed on 1–2% agarose, and cleaned using ExoSAP-IT. Purified PCR products were sequenced using the Big Dye terminator mix on ABI 3500 Genetic Analyzer (Applied Biosystems, CA 92008). Sequences were visualized using Chromas2 (Technelysium, Tewantin, QLD, Australia) and analyzed with Vector NTI software (Applied Biosystems, CA 92008).
Cases comprised a total of 149 patients. Unrelated European American (n = 130), and unrelated African American (n = 19) samples from patients with lumbosacral myelomeningocele (spina bifida). Collected at Children's Memorial Hosptial in Chicago, IL, USA. All cases had open spina bifida (myelomeningocele).
Genomic DNA fragments spanning the two exons of NKX2-8 were amplified by PCR. Purified PCR products were sequenced using Big Dye terminator chemistry (Applied Biosystems) and analyzed on a MegaBACE 1000 (Amersham). Sequence reads derived from both strands were assembled, aligned and analyzed for nucleotide differences using Sequencher (GeneCodes). We assessed the presence of variants in the Exome Variant Server, NHLBI GO Exome Sequencing Project (ESP), Seattle, WA (URL: http://evs.gs.washington.edu/EVS/); data release ESP6500, November 2012. PCR primers and conditions used are shown in Table S4.
NCBI BLASTP [58] was used to compare protein sequence conservation across species. The biological sequence alignment tool, Bio Edit, (Ibis Biosciences, Carlsbad, CA), was used to align protein sequences. The PolyPhen online tool was used to predict the possible impact of missesne mutations [59].
|
10.1371/journal.pntd.0004290 | Systematic Review and Meta-analysis of the Impact of Chemical-Based Mollusciciding for Control of Schistosoma mansoni and S. haematobium Transmission | Programs for schistosomiasis control are advancing worldwide, with many benefits noted in terms of disease reduction. Yet risk of reinfection and recurrent disease remain, even in areas with high treatment coverage. In the search for means to better prevent new Schistosoma infections, attention has returned to an older strategy for transmission control, i.e., chemical mollusciciding, to suppress intermediate host snail species responsible for S. mansoni and S. haematobium transmission. The objective of this systematic review and meta-analysis was to summarize prior experience in molluscicide-based control of Bulinus and Biomphalaria spp. snails, and estimate its impact on local human Schistosoma infection.
The review was registered at inception with PROSPERO (CRD42013006869). Studies were identified by online database searches and hand searches of private archives. Eligible studies included published or unpublished mollusciciding field trials performed before January 2014 involving host snails for S. mansoni or S. haematobium, with a primary focus on the use of niclosamide. Among 63 included papers, there was large variability in terms of molluscicide dosing, and treatment intervals varied from 3–52 weeks depending on location, water source, and type of application. Among 35 studies reporting on prevalence, random effects meta-analysis indicated that, on average, odds of infection were reduced 77% (OR 0.23, CI95% 0.17, 0.31) during the course of mollusciciding, with increased impact if combined with drug therapy, and progressively greater impact over time. In 17 studies reporting local incidence, risk of new infection was reduced 64% (RR 0.36 CI95% 0.25, 0.5), but additional drug treatment did not appear to influence incidence effects.
While there are hurdles to implementing molluscicide control, its impact on local transmission is typically strong, albeit incomplete. Based on past experience, regular focal mollusciciding is likely to contribute significantly to the move toward elimination of schistosomiasis in high risk areas.
| Infection with Schistosoma blood flukes is a leading cause of chronic parasitic disease in at-risk areas of Africa, South America, Asia, and the Philippines. Over past decades, many national programs have implemented regular drug treatment to control or prevent the advanced complications of Schistosoma infection. However, these periodic treatments do not stop transmission of the parasite, which occurs when human sewage contaminates local water bodies and parasite eggs infect intermediate host snails. In this systematic review, we collated past experience of using chemically-mediated snail control for prevention of schistosomiasis. This approach, used in many Schistosoma-affected countries before the advent of the current oral drug regimens, has the potential to significantly reduce transmission if properly applied. Our meta-analysis of 63 studies (performed 1953–1981) catalogued a wide variety of water treatments and schedules employed. Among studies reporting on human infection, we found that snail control reduced local human prevalence and incidence of infection in most, but not all locations. Estimates from the aggregated studies indicate that snail control (alone) typically reduced new infections by 64% and local prevalence declined over a period of years. This decline was accelerated and more profound (84% reduction) if drug treatment was also made available.
| Schistosomiasis, the chronic human disease caused by Schistosoma spp. parasite infections, is a preventable illness that, if left untreated, is associated with long-term undernutrition, anemia, organ scarring and fibrosis, resulting in disabling patient symptoms [1, 2]. Current anti-schistosomiasis chemotherapy programs focus on controlling or preventing morbidity by treating school-age children who typically have the highest levels of Schistosoma infection [3]. However, because pre-school infection [4–6] and recurrent infection during childhood [7, 8] are associated with significant risk for disease, optimal disease prevention can occur only when parasite infection or reinfection can be effectively blocked [9]. By themselves, Preventive Chemotherapy (PCT) campaigns [3] using mass drug administration have not been very successful in limiting transmission in high-risk areas [4, 10–13]. The WHO roadmap’s new focus on 'transmission control, wherever possible' [14] means it is appropriate to re-examine the efficacy of intermediate-host snail control for prevention of human-to-snail-to-human parasite transmission.
Reduction in infected snail numbers at the places where humans come into contact with freshwater could substantially reduce each patient's frequency of exposure to infecting parasite larvae (cercariae), and, hence, reduce the frequency of reinfection. In the 1960s early theoretical modelling [15] suggested that a greater than 90% reduction in snail numbers, in conjunction with population drug treatment, had the potential to extinguish Schistosoma populations from local ecosystems. Chemical molluscicides, including copper sulfate, sodium pentachlorophenate (NaPCP), N-tritylmorpholine (Frescon), and niclosamide (Bayluscide, Bayer 73) were used extensively in the 1950s, 1960s, and 1970s for schistosomiasis control in Africa, South America and Asia [16], but following the introduction of oral drug therapies, molluscicides have not seen as much use in the last 30 years [17]. As present-day programs contemplate integrated strategies for schistosomiasis control, it is important to systematically review the efficacy of molluscicide use in snail suppression and its effectiveness for infection prevention, so that planners can project likely costs and impacts when targeting parasite elimination in at-risk locations.
In the present systematic review and meta-analysis, we compiled the results of field trials of chemical mollusciciding focusing primarily on control of S. mansoni and S. haematobium species and the use of niclosamide molluscicide, now the most commonly used agent for host snail control. The decision to include just two parasite species was derived from the African focus of our sponsor, the Schistosomiasis Consortium for Operational Research and Evaluation (SCORE), and was taken in light of previous publication of a meta-analysis of mollusciciding impacts in China [18, 19]. While we encountered a number of limitations in the available study literature, we found sufficient quantitative evidence that routine mollusciciding can effectively reduce snail numbers in a manner that significantly reduces reinfection or new Schistosoma infection in typical at-risk human populations.
The data used in this project were aggregated, anonymized data from previously published studies; as such, this study does not constitute human subjects research according to U.S. Department of Health and Human Services guidelines (http://www.hhs.gov/ohrp/policy/checklists).
The protocol for this project was developed prospectively by the authors, then registered and published in the International Prospective Register of Systemic Reviews (PROSPERO) online database, http://www.crd.york.ac.uk/prospero/index.asp, number CRD42013006869, on 16 December 2013. Our a priori review question was, “Does chemical mollusciciding effectively reduce snail numbers in a manner to prevent reinfection or new Schistosoma infection in at risk human populations?” focusing primarily on control of S. mansoni and S. haematobium species and the use of niclosamide molluscicide (2-amino ethanol salt of 2', 5'-dichloro-4'-nitro salicylanilide, sold as Bayluscide, Mollutox, and other names). The PRISMA checklist and the PROSPERO protocol for this study are provided as Supporting Information files S1 and S2 Files.
To quantify the effects of repetitive use of chemical mollusciciding, we aimed to include any available published or unpublished reports on its use for control of Bulinus or Biomphalaria species for prevention of S. haematobium or S. mansoni infection. No limits were placed in terms of location or language of the report. However, we did not include studies of S. japonicum or S. mekongi control, which was the topic of a recently published meta-analysis from China [18, 19]. Studies had to include periodic application of chemical compounds to transmission water contact sites or experimental locations, as well as the names of the snail species treated, and treatment doses, frequency, habitat (static vs. flowing water), region, and season of application. Information about local human prevalence and incidence of Schistosoma infection, before and after intervention, was also sought as secondary outcomes for meta-analysis. We aimed to include any studies performed after the development of niclosamide molluscicide compounds, (i.e., after January 1961) to the close of the search phase of the project, 1 January 2014. Historical perspectives, observational studies and prospective trials were eligible for inclusion if they provided the necessary quantitative data.
We identified published studies using PubMed, Google Scholar, Web of Science, SCIELO, African Journals Online, as well as resources such as WHO technical reports and archived files at Case Western Reserve University and the Schistosomiasis Consortium for Operational Research and Evaluation (SCORE). Where published bibliographies of the recovered studies were found to contain promising citations (including grey literature) not included in online searches, these papers were obtained, whenever possible, and screened for inclusion in the meta-analysis.
We examined the available electronic database literature using combination searches of the following terms: 'molluscicide'; 'snail control/prevention'; 'Biomphalaria (Australorbis)’; 'Bulinus'; 'field [trial]' ‘schistosomiasis/prevention and control’ ‘transmission’, and/or 'niclosamide’. Secondary report finding was done by scanning PubMed 'similar articles' feature, and by using the Google- and PubMed-generated listings of papers that cited papers that we found to contain well-conducted snail control intervention trails. As relevant articles were identified, we broadened our search by accessing additional titles through the online databases’ automated ‘related articles’ links. Full titles and abstracts were recovered for the initial screening phase of study selection.
Review of titles and abstracts was performed by two trained reviewers, searching for data content meeting study requirements. The studies found suitable for inclusion—including historical, observational, and prospective studies—were then obtained for full-text review from online or library sources. Where a single report contained data on multiple individual community surveys, each survey was also separately abstracted for inclusion in some of the sub-group comparison analysis. We excluded studies where sufficient details of snail control measures were not reported, or when the data on the community or individual Schistosoma infection levels were not sufficiently detailed to confirm the reported incidence and prevalence of infection following the implementation of niclosamide or other molluscicide treatments. Cases of duplicate publication or extended analysis of previously published data were also excluded. Full listings of included and excluded studies are provided as Supporting Information files S3 and S4 Files.
Included papers were abstracted and their relevant features entered into a purpose-built database created in Microsoft Excel 2013 software (Redmond, WA). These papers were archived by the authors in both paper and electronic (pdf) formats at the Center for Global Health and Diseases, Case Western Reserve University. In addition to full citation information and year of publication, information was collected on the country and region where the study was performed, along with snail genus and species, chemical mollusciciding treatment, whether the study was performed in a research laboratory or in the field, the concentration range of molluscicide delivered, and percentage kill of the observed molluscs. The effective days of molluscicide-mediated snail control were also captured, as well as the beginning and end values for population-wide incidence and prevalence. Data entries were fully verified by the second reviewer before final data analysis was carried out.
Reported data on snail outcomes was quite diverse in terms of delivery, metrics, and timing of interventions and follow ups. Those results were thus summarized only qualitatively. The impact of snail control interventions on local prevalence and incidence of infection among at-risk human populations could be compared across studies, however. Results from each study and sub-study were entered into Comprehensive Meta-Analysis software, v.3 (CMA, Biostat, Englewood, NJ) for calculation of summary estimates of treatment impacts, along with their confidence intervals. Potential modifying factors assessed in preliminary analysis included parasite species, starting infection prevalence, age-group monitored, study era, region, duration of control, type of habitat, and impact of added drug treatments.
Heterogeneity among studies was expected due to differences in habitat, snail and parasite species, and human populations involved [16]. Heterogeneity levels were scored using Higgins’s and Thompson’s I2 statistic [20]. Summary estimates of intervention effects were computed using Der Simonian and Laird random-effects modeling [21] implemented in CMA software. The data sets used for this analysis are provided as Supporting Information files S1 and S2 Datasets. Linear meta-regression by CMA was also used to estimate the impact of duration of control on the odds/risk of Schistosoma infection. Potential publication bias of studies reporting impact on human incidence or prevalence of Schistosoma infections was assessed by visual inspection of funnel plots, and calculation of the Egger test for plot asymmetry. These funnel plots and statistics are provided as Supporting Information file S5 File.
Fig 1 contains a flow chart that details the results of the search and selection strategy for the studies included in this systemic review. Of the 357 listings recovered, 315 were obtained from online database searches, and 42 from bibliographies and archived materials. After titles and abstracts were assessed, 140 reports were selected for full review, ultimately yielding 63 studies (from 40 papers) to be included in the qualitative or quantitative analysis reported here (see Supporting Information files: S3 File ‘Listing of Included Studies’ and S4 File ‘Listing of Excluded Studies’). As detailed in the Methods section, the included studies were not limited in terms of publication date or language. The study focus was limited, however, to the control of Biomphalaria or Bulinus snail species for prevention of S. mansoni and S. haematobium transmission. Studies of S. japonicum and other human and animal schistosome species were not included.
Sixty-three reports provided information on mollusciciding’s impact on snail numbers and/or the duration of its suppressive effects. Of these, one was a graduate student thesis and the remainder were reports published in peer-reviewed journals. Twenty-seven (published) papers reported on the impact of mollusciciding campaigns on Schistosoma spp. prevalence among the local human population, and 12 papers provided estimates of the effects on local incidence of infection among children and/or adults. Whereas most studies reported on the use of niclosamide compounds for snail control, 6 studies used either sodium pentachlorophenate (NaPCP) [22–24] or N-tritylmorpholine (Frescon) [25] as molluscicides. For studies of niclosamide effects, publication dates ranged from 1960 to 2013 (median 1981). Of the 63 studies, 33% were from East Africa (Tanzania, Kenya, Sudan, Ethiopia), 17% were from southern Africa (Zimbabwe and South Africa), 17% were from Brazil or the Caribbean (St. Lucia), 11% were from North Africa (Egypt and Morocco), 11% were from West Africa (Gambia, Ghana, Mali), 6% were from Central Africa (Burundi, Cameroon), and 4% were from Iran.
Among 35 studies reporting impact on human Schistosoma spp. infection prevalence, 14 (40%) focused on prevalence among school age children, 20 (57%) reported results for the general population, and 1 (3%) reported separate results for adults and children. Eleven (31%) of these studies reported the impact of snail control as used alone, 20 programs (57%) employed snail control plus some form of community-based screening and treatment intervention (i.e., only egg-positive persons were treated), 2 studies (6%) used snail control plus a school-based treatment strategy, and 2 (6%) used snail control combined with targeted mass drug administration.
Research performed in the pre-1990s era often did not report quality-related trial details in peer-reviewed publications, and therefore, we did not pursue quality weighting in the present analysis. Many of the researchers who were involved are now deceased or retired, so access to primary data was not available. Most of the reported studies involved a single intervention site, the studies were non-randomized, and for the most part the comparison of intervention effects involved historical and not concurrent comparison data (i.e., they used a one-group, pre-test/post-test study design [26]). It is possible that the selection of study sites favored extremely high or low transmission locations. In the studies where concurrent untreated comparison sites were monitored, non-homogeneity of risk between the treated and untreated areas was often likely (e.g., Tamiem et al. [27]).
Threats to validity in assessing molluscicide-related impact on Schistosoma transmission included secular trends among other interventions, and maturation of environments [23, 28], heterogeneities within landscapes and populations [29], loss to follow up, and the lesser reliability of egg-count diagnostics as prevalence and intensity of Schistosoma infections decline [30–33]. In addition, there was potential unreliability of treatment implementation, and unknown risk of reinfection from population movement within or outside the control areas [34, 35]. The role of diffusion of information about schistosomiasis was also unknown. Observation (Hawthorne) effects were possible during implementation, particularly on plantation estate programs run by employers [36, 37]. However, in our analysis of reporting bias, we did not find evidence of underreporting of negative results or adverse outcomes (see Supporting Information file S5 File for funnel plot and regression analysis).
A wide variety of molluscicide delivery approaches were used in the included studies, depending primarily on the speed of water flow in the snail habitat, the extent of the water area to be treated, and the multiplicity of local human water contact sites in the area targeted for control. More rapidly flowing water bodies were often treated by drip-feed delivery systems that provided a constant dose of molluscicide to a stream or canal over a period of 1–2 days [38–41]. In irrigation schemes, where water flow could be diverted and controlled, some projects utilized impoundments of molluscicide-containing water that could be slowly transferred through the canal system to fully treat the entire irrigation system [42, 43]. Slow moving streams and static and seasonal ponds were often treated with focal spraying of shallows and vegetation at the water’s edge—the primary habitat of the intermediate host Biomphalaria and Bulinus spp. snails. Large lakes presented a special problem for treatment dosing because of rapid molluscicide dispersal by currents and wave action [44]. In one study, plastic sheeting was employed to isolate lake shore areas to retain molluscicide for a sufficient time to effect snail control [45].
Reporting on the lethal impact of molluscicide treatment and duration of its effects varied widely among studies. Early laboratory and field trials established that niclosamide concentrations of 0.1 to 3 mg/L (units equivalent to ‘parts per million’ (ppm), a term often used in the older literature) could kill over 90% of intermediate host snails [46], and that the dose effect was dependent on the duration of exposure—lower concentrations (0.04–0.53 mg/L) were effective if applied for 24h, higher concentrations (0.5–1.2 mg/L) could be lethal if applied for only 6h [47–49]. In the schistosomiasis control campaigns included in our meta-analysis, the estimated concentrations delivered to local water bodies ranged from 0.025 [37, 50] to 10 mg/L [51]. The median and most common dose target in our included studies was 1 mg/L, consonant with the experience summarized by Andrews, et al. in their detailed 1983 review [46]. Immediate impact of mollusciciding was usually assessed at breeding sites 24 h after delivery [52], and then if living snails were still present, reapplication of chemical, sometimes at higher concentrations [53], was frequently used to maximize snail suppression. Where drip feed administration was applied to running streams, significant molluscicide effects on snail numbers were detectable 900 meters [54], 1375 meters [55], 1700 meters [40], even up to 10 km [38] downstream. Snail mortality was assessed in many different ways. Some studies reported live/dead snails at various intervals after treatment, others reported on treatment impact on caged sentinel snails placed in the treated water habitats [38, 54]. Because early reapplication was used in several programs but not clearly detailed, we could not determine a summary estimate of niclosamide efficacy in terms of (dose X exposure time) [46, 47, 49, 54, 56–60] across the reported field studies. Vegetation, wave action, and debris were noted in several studies to be confounding factors affecting the molluscicidal efficacy of chemical applications [24, 44, 45, 61].
Many projects achieved 100% elimination of targeted snails for periods lasting from several weeks to several months. Some projects failed to reach 100% snail elimination following mollusciciding [35, 39, 55, 58, 61–64]. Nevertheless, these sites were able to obtain 88–99% immediate reduction in snail numbers. When snail re-emergence occurred, repopulation times ranged from 2 weeks [62] up to 18 months [65], depending on location and habitat. Serial monitoring for significant snail repopulation at human water contact sites was an important part of implementation, and was most often used to decide the intervals needed for repeated mollusciciding.
Fig 2 indicates the between-treatment mollusciciding intervals reported by 47 studies of S. mansoni and S. haematobium control in different areas of Asia (Iran), Africa, South America (Brazil) and the Caribbean (St. Lucia and Puerto Rico). There was a large range of working between-treatment intervals reported (21 days to 365 days), depending in part on the seasonality of transmission, the type of water treated (flowing, static, or canal), and the desired lethality of the treatment applied (suppression vs. elimination of snails). Decisions regarding treatment intervals were most often based on snail repopulation detected on regular waterside surveys at treated water contact sites. The median interval used in the reported studies was 90 days, with an inter-quartile range (IQR) of 42–90 days.
Thirty-five studies reported in 28 publications [8, 22–24, 27, 28, 34, 35, 39, 41, 42, 66–81] reported on the interval impact of programs that included snail control campaigns on the prevalence of detectable Schistosoma infection among local human populations. Heterogeneity in prevalence outcomes was quite high (I2 = 99.9) among studies. Fig 3 indicates the pre- and post-intervention levels of Schistosoma prevalence for individual studies; the median pre-control prevalence for all studies was 45%, (Range: 5% to 92%, IQR 26% to 58%) while the median post-control prevalence was significantly lower, 17.5% (Range: 0% to 53%, IQR 6% to 30%, (P < 0.001 by Wilcoxon signed rank test)). Supporting Information file S1 Fig shows the forest plot for the prevalence studies. Fig 4 graphs summary estimates of the odds ratio for infection after molluscicide treatment intervention as compared to pre-treatment levels (numeric details for this graph are provided in Supporting Information file S1 Table). Overall, surveyed populations had a significantly reduced odds of infection (OR 0.23, CI95% 0.169, 0.309) following snail control intervention. The impact was less strong where snail control was used alone (OR 0.47, CI95% 0.276, 0.800), and greatest among studies where snail control was combined with community-based screening and treatment programs (OR 0.162, CI95% 0.116, 0.225). There was not a significant difference in terms of impact between S. mansoni- and S. haematobium-endemic locations. Treatment of natural water sites had greater overall impact than treatment of irrigation systems, which may account for the observation that North Africa (primarily irrigation locations) had less improvement than sites elsewhere in Africa, Asia, South America, and the Caribbean. Studies focused only on school age children reported lesser gains in terms of post-intervention prevalence when compared to general population studies, likely reflective of school age groups’ much greater risk for infection/reinfection. Not shown, starting prevalence of infection did not have a clear effect on the size of prevalence reductions obtained during a mollusciciding program (for details, see Supporting Information file S3 Fig, which shows a forest plot of the range of outcomes data (ORs) arranged according to starting prevalence of Schistosoma infection).
Of note, three (9%) of the 35 studies [66, 77, 81], two in Egypt and one in Zimbabwe, did not demonstrate reductions in local Schistosoma prevalence. In addition, another three studies reported less than a five percentage point drop in local human Schistosoma prevalence during their mollusciciding trial period [24, 39, 73]. These three less successful studies were performed in Egypt and in Liberia and involved both S. mansoni and S. haematobium areas. A summary of the implementation, population, and environmental features of these six projects having relatively limited mollusciciding impact is included in Supporting Information file S2 Table. Their individual reports provided several possible explanations for their limited program impact. These included i) having only a short duration of follow-up (i.e., 1 year after mollusciciding implementation) [24]; ii) a lesser impact of supplemental drug treatments on S. mansoni as compared to S. haematobium [39, 77]; iii) inability of the implemented molluscicide program to reduce snail numbers at transmission sites [66, 81]; and iv) incorrect timing of mollusciciding application relative to maximal seasonal transmission [73].
In consideration of the long-term effects of multi-year programs, a meta-regression of mollusciciding effects on prevalence odds ratios vs. time is presented in Fig 5. It suggests progressively greater reductions in infection prevalence as mollusciciding programs extend beyond the first few years of snail control.
Seventeen studies reported in 12 publications [28, 35, 38, 41, 50, 61, 67, 69, 71, 73, 76, 82] reported on the impact of mollusciciding programs on the incidence of new Schistosoma infections before and during control. As for prevalence, above, heterogeneity in incidence outcomes was high (I2 statistic = 93.2) among the reported studies. From our random effects meta-analysis of all 17 studies, the risk of new infection was estimated to be reduced by 64% (relative risk = 0.36, CI95% 0.25, 0.50) in schistosomiasis control programs that included mollusciciding. The forest plot for studies reporting incidence outcomes is provided in Supporting Information file S2 Fig. Yearly incidence dropped from a median 22% (Range: 4% to 78%, IQR 14% to 54%) before intervention to a median 8% (Range: 2% to 80%, IQR 4% to 12%, P = 0.001 for the difference) during the course of these mollusciciding intervention programs. Fig 6 graphs summary estimates of the risk ratio for infection after molluscicide treatment intervention as compared to pre-treatment levels (numeric details are provided in Supporting Information file S3 Table). Of note, reduction of incidence was greater (i.e., RR was lower) for areas with natural water sources as compared to irrigation schemes (RR 0.36 vs. 0.55). There was also no apparent difference in incidence reduction effect when drug treatments were included in the control programs (RR for snail control alone was 0.33 vs. 0.32 for snail control plus community screening and drug treatment). Fig 7 shows the shift in pre- and post- incidence values for individual studies. Fig 8 graphs a meta-regression of the observed impact of mollusciciding on incidence (in terms of log-risk ratio) according to the duration of program implementation, indicating what appears to be an effect on incidence in most locations within 1–3 years.
As noted earlier, most included studies focused on one area using historical control data to assess impact. Nine studies reported in eight publications [22, 23, 27, 28, 38, 61, 71, 76] reported on concurrent infection outcomes in untreated areas near the molluscicide trial site. Table 1 summarizes the observed effects on prevalence and incidence of Schistosoma infection in the treated and untreated zones in each study. Of note, treatment and comparison zones had only one area unit each, and assignment was not randomized.
In general, all molluscicide-treated areas saw greater declines in prevalence or incidence than untreated areas. Remarkably, though, prevalence fell significantly without molluscicide intervention (or population-based drug treatments) over a period 8 years in the Brazil study region [28] and over 13 years in Puerto Rican districts [22, 23], indicating interval changes in local risk for transmission that were unrelated to the snail control intervention. In the Ghana-2101 project, Lyons [76] observed a spontaneous drop in S. haematobium incidence over a 2 year period in an untreated comparison area, which was associated with an unexplained interval disappearance of local bulinid snails. Five study areas saw no change in prevalence in their untreated comparison areas but concurrent reductions in prevalence of 24% to 89% within snail control areas.
This systematic review and meta-analysis summarizes what has been a broad and lengthy experience with the use of mollusciciding for control of S. mansoni and S. haematobium transmission. Results of our analysis suggest that chemical-based snail control, particularly with the compound niclosamide, can effectively reduce local transmission of Schistosoma parasites when delivered at regular interval and under skilled supervision [17]. Direct treatment effects on snails were difficult to summarize, because of the many differences in sampling and reporting used in the included studies. However, where snail reductions were quantified, most programs saw very significant reductions or complete disappearance of local Schistosoma host snails during program implementation. Earlier studies tended to favor broad, intensive mollusciciding in an attempt to eliminate intermediate host Biomphalaria or Bulinus spp. snails. With such approaches, and particularly where water flow could be controlled in irrigation systems and transmission was more seasonal, treatment intervals could be extended to 6–12 months [8, 79, 83]. Later studies, often dealing with natural water bodies and more focal human water contact, tended to favor more frequent focal administration of mollusciciding [13, 41, 43, 62, 66, 84], allowing snails to persist elsewhere outside the main human water contact zones. While not fully explored, several preliminary studies suggested that slow-release strips or pellet formulations [85, 86] or delayed-release molluscicide capsules [87] might better focus the impact of chemical molluscicide and extend its duration of impact, with concomitant cost-savings due to a reduced need for frequent delivery.
Given the cumulative experience of the programs summarized here, it becomes clear that focality of snail habitats, combined with overlap into human water contact zones, represent factors in successful Schistosoma transmission. Not all waterbodies within a control area are suitable for host snails [88], but human movement among water contact sites can strongly facilitate regional persistence of transmission [89, 90]. For these reasons, implementation of focal snail control requires an adequate surveillance component to have an accurate working knowledge of local snail habitat, of human water contact zones, and of the seasonal factors affecting the abundance of snails and the likelihood of transmission [66, 73, 91].
As Shiff [92] points out, the ultimate value of a mollusciciding campaign is measured by its impact on human infection. Where snail control was used alone, early reductions in human incidence and later reductions in human Schistosoma infection prevalence could usually be obtained. When snail control was combined with population screening and selective or mass drug therapy, prevalence was reduced more quickly and incidence diminished. However, transmission was frequently not eliminated. When most successful, programs involving snail control achieved 85–100% reductions in local prevalence of Schistosoma infections [8, 22, 23, 68–71, 75, 78]. However, some control programs appeared to have only minimal impact on local prevalence [39, 66, 73, 77, 81]. While there were some apparent differences in effects by region and by parasite species, the snail species named in the included studies were too diverse to draw meaningful comparisons for prevalence or incidence outcomes stratified at the intermediate snail host species level.
Authors cited the fact that established control is vulnerable to resurgence of snail populations from local refugia [93–95]. The homing characteristics of miracidia for host snails, combined with the homing of cercariae towards human skin and the high degree of asexual multiplication within infected snails, strongly favor persistence of transmission [70]. Given the presence of untreated human individuals within the control area, whether from refusal of drug therapy or in-migration from (or temporary travel to) areas not under Schistosoma transmission control [64, 70, 76, 77, 96], new infections are likely to continue to occur. Habitat changes, growth in human populations, and breakdowns in program performance are other factors that can contribute to limit the impact of snail control programs—whereas Egypt did very well with mollusciciding campaigns in the 1960s [71], by the 1980s their programs were having limited effectiveness [66, 73]. An independent on site WHO review performed in 1985 identified gaps in communication between snail control teams and health personnel, errors in selection of snail sampling sites, inefficiencies in snail testing, and an over-reliance on infrequent area-wide mollusciciding (as opposed to more frequent focal mollusciciding) as contributing causes to poor performance in that era [66].
The evidence summarized in this meta-analysis appears, in general, to favor mollusciciding as an effective method to reduce Schistosoma infections over time, with an additive effect on prevalence where population-based drug control is also given. However, the quality of the reported evidence is limited. The studies included in the analysis were non-randomized interventional trials, often with only historical data used for comparison in assessing the magnitude of snail control outcomes [26]. Where concurrent comparison areas were used, transmission was inconsistent in some areas, suggesting that secular trends or temporal fluctuations were occurring, which means that there is risk of over- or under-estimating the impact of mollusciciding in studies that use only historical controls. As such, and in view of the variability of ecology in Schistosoma transmission settings, the formal scientific evidence for a ‘generalizable’ consistent effect of snail control can be considered only minimally strong at this time. Other limitations in performing our analysis stem from study-to-study differences in snail control implementation, measures of snail impact, monitoring of human population outcomes, and duration of control.
For the reader considering implementation of a snail control program based on niclosamide mollusciciding, the 1983 monograph by Andrews, et al. [46] provides extensive information on the chemistry, biology, and toxicology of niclosamide, as well as its effects on non-targeted plant and animal species. Niclosamide’s environmental impacts have been more recently reviewed by Dawson [97], who concluded that there is minimal risk to humans and the environment, provided its application is appropriately dose-limited, informed, and supervised. It is important to be aware, however, that niclosamide is harmful to fish, amphibia, certain insect larvae, and in higher doses, to aquatic vegetation [46, 97, 98]. Because niclosamide quickly decays over 24 hours, animals that can rapidly move away from an area of application may return in a matter of days [62]. In aggregate, it appears that metered and very focal niclosamide administration at human water contact sites has the potential to provide the greatest impact on Schistosoma transmission with the least impact on local ecosystems.
Overall, the impacts reported in the included studies predominantly lean toward a positive effect of mollusciciding in reducing Schistosoma transmission, with longer duration of control leading to a greater impact. These findings hold promise that the benefits of mollusciciding could be further defined in modern, well-designed comparison trials. A randomized comparison trial of mass drug administration ± snail control is in now progress in Zanzibar as part of a program that is attempting local elimination of S. haematobium [99]. Additional mollusciciding trials for S. mansoni control and elimination are under development. We look forward to the results of these efforts, which are expected to provide valuable evidence to further inform Schistosoma control policy. Based on past experience, regular focal mollusciciding is likely to contribute significantly to the move toward elimination of schistosomiasis in high risk areas.
|
10.1371/journal.pntd.0002777 | Challenges in Dengue Fever in the Elderly: Atypical Presentation and Risk of Severe Dengue and Hospita-Acquired Infection | To better understand dengue fever in the elderly, we compared clinical features, World Health Organization (WHO) dengue classification and outcomes between adult (<60) and elderly (≥60) dengue patients. We explored the impact of co-morbidity and hospital-acquired infection (HAI) on clinical outcomes in the elderly. All patients managed at the Communicable Disease Centre, Singapore, between 2005 and 2008 with positive dengue polymerase chain reaction (PCR) or who fulfilled WHO 1997 or 2009 probable dengue criteria with positive dengue IgM were included.
Of the 6989 cases, 295 (4.4%) were elderly. PCR was positive in 29%. The elderly suffered more severe disease with more dengue haemorrhagic fever (DHF) (29.2% vs. 21.4%) and severe dengue (SD) (20.3% vs. 14.6%) (p<0.05). Classic dengue symptoms were more common in the adult group. The elderly were less likely to fulfill WHO 1997 (93.6% vs. 96.4%) (p = 0.014), but not WHO 2009 probable dengue (75.3% vs. 71.5%). Time to dengue diagnosis was similar. There was no significant difference in the frequency of warning signs between the two groups, but the elderly were more likely to have hepatomegaly (p = 0.006) and malaise/lethargy (p = 0.033) while the adults had significantly more mucosal bleeding (p<0.001). Intensive care admission occurred in 15 and death in three, with no age difference. Notably, the elderly stayed in hospital longer (median 5 vs. 4 days), and suffered more pneumonia (3.8% vs. 0.7%) and urinary infection (1.9% vs. 0.3%) (p = 0.003). Predictors of excess length of stay were age (adjusted odds ratio [aOR] 2.01, 95% confidence interval [CI] 1.37–2.88), critical illness (aOR 5.13, 95%CI 2.59–9.75), HAI (aOR 12.06, 95%CI 7.39–19.9), Charlson score (aOR 6.9, 95%CI 2.02–22.56) and severe dengue (DHF/dengue shock syndrome/SD) (aOR 2.24, 95%CI 1.83–2.74).
Elderly dengue patients present atypically and are at higher risk of DHF, SD and HAI. Aside from dengue severity, age, co-morbidity and HAI were associated with longer hospital stay.
| Dengue is a neglected tropical disease that is increasingly affecting elderly patients; however, there is a paucity of data on clinical presentation and outcomes in this group. The limited data suggests that elderly dengue patients have the highest case-fatality rate but the pathogenesis of mortality in elderly dengue patients remains unclear. To better understand dengue fever in the elderly we compared clinical features, WHO dengue classification and outcomes between adult (<60) and elderly (≥60) dengue patients and explored the impact of co-morbidity and HAI on clinical outcomes in the elderly. We found that diagnosis in the elderly may be challenging due to atypical presentation. Elderly patients have worse outcomes compared with their younger counterparts with increased rates of DHF and SD. Elderly patients have higher rates of HAI placing them at risk of infection-related mortality. Aside from dengue severity, age, co-morbidity and HAI were associated with longer hospital stay. This will place further burden on already stretched hospital systems.
| Dengue is the most significant mosquito-borne virus in humans [1] and is endemic to Singapore. In Asia dengue classically affects children with the majority of cases of dengue hemorrhagic fever (DHF), dengue shock syndrome (DSS), and dengue related mortality observed in this group [2]. However, in Singapore dengue predominantly affects young adults possibly as a result of lowered herd immunity and acquisition outside of the home [3]. However, with aging population there has been an increase in dengue incidence rates in older adults [4], [5], [6]. In Taiwan older adults have the highest reported dengue incidence rate and risk of fatality [7]. Likewise in Singapore elderly patients accounted disproportionately for the majority of dengue deaths [8] highlighting the urgent need for enhanced understanding of dengue in the elderly to improve clinical management and outcome.
The cornerstone of management of dengue patients and prevention of dengue-related mortality is early diagnosis and recognition of clinical syndromes requiring intervention [9]. However diagnosis and appropriate management of elderly dengue patients may be delayed as they present atypically [10]. Fever may be the only symptom and leukopenia occurs less frequently compared with younger adults. Both World Health Organization (WHO) 1997 [11] and 2009 [9] dengue classifications have reduced sensitivity in older adults, because of the absence of classic dengue symptoms, potentially delaying diagnosis [12]. The atypical presentation is likely due to age-related decline in immune function, predominantly affecting cell-mediated and humoral immunity resulting in impaired cytokine response altering disease presentation [13].
Severe hospital-acquired infection (HAI) may contribute to dengue deaths in adults [14], [15]. Concurrent bacteremia was more common in elderly DHF patients (17.4%) compared with non-elderly DHF patients (3.4%) in Taiwan [16], but there was no significant difference in bloodstream infection rates between the fatal and non-fatal cases [16]. The impact of HAI in elderly dengue patients requires further investigation.
The limited data on elderly dengue patients suggests that this group has the highest case-fatality rate [7], [8], [17]. The pathogenesis of mortality in elderly dengue patients remains unclear but co-morbidities may play a role. Seventy-five percent of dengue fatalities in Singapore had co-morbidities [8]. Likewise, in Taiwan fatality from dengue was associated with age above 55 years and pre-existing hypertension, chronic renal impairment or diabetes [18]. In elderly patients with DHF pre-existing pulmonary disease and the development of DSS or acute renal failure (ARF) were associated with mortality in Taiwan [16]. In the 2002 dengue outbreak in Taiwan renal failure was both a risk factor for DHF and mortality, with the degree of renal impairment correlating to risk of mortality [19].
The aim of our study was to compare clinical features, WHO 1997 [11] and 2009 [9] dengue classification and outcomes between adult (<60 years of age) and elderly (≥60 years of age) dengue patients and explore the impact of co-morbidity and HAI on clinical outcomes in the elderly.
The study was approved by the National Healthcare Group Domain Specific Review Board (DSRB/E/2008/00567) with a waiver of informed consent for the collection of anonymized data.
A retrospective study of all adult dengue patients managed at the Communicable Disease Center (CDC), Tan Tock Seng Hospital was conducted between 1 January 2005 and 31 December 2008. Patients were hospitalized if they had suspected DHF or if they met previously published admission criteria [20]. Intensive care unit (ICU) referral criteria included patients with compensated shock (systolic blood pressure [BP] >90 mmHg but narrow pulse pressure <20 mmHg) and those with hypotension (systolic BP <90 mmHg). Patients who were not admitted had daily clinical assessment and full blood count until clinically stable.
All patients were managed using a standardized dengue care path improving consistency of clinical, laboratory, treatment and outcome data. The care path included daily documentation of symptoms (abdominal pain, bleeding, breathlessness and vomiting), examination findings (BP, rash, pleural effusions, ascites) and laboratory parameters (platelet count and hematocrit). The care path provided clear criterion for intravenous fluids and blood products. Medical interventions, diagnosis and patient outcomes (dengue fever, severe dengue [SD], DHF, DSS, severe bleeding, severe organ involvement) were documented daily within the care path. The hospital electronic medical records were used to extract laboratory, microbiological and radiological data. Data extraction was performed by medically trained research assistants.
Inclusion criteria were patients with positive dengue reverse-transcriptase polymerase chain reaction (PCR) [21] or probable dengue (WHO 1997 [11] or WHO 2009 [9]) with positive dengue IgM [22], [23].
An elderly patient referred to one whose age was 60 or greater [24]. A Charlson co-morbidity score was assigned for each patient based on the presence and severity of diseases listed in the index [25] and a Pitt bacteremia score validated against APACHE II score was calculated for each patient as previously described [26].
Leukopenia was defined as total white cell count (WCC) <4×109/L for patients managed during 2005 and as total WCC <3.6×109/L for patients managed from 2006 on-wards (laboratory reference range revised in 2006). Warning signs from the WHO 2009 guideline included: abdominal pain or tenderness, persistent vomiting, mucosal bleeding, clinical fluid accumulation, lethargy, hepatomegaly and rise in hematocrit (≥20%) concurrent with rapid platelet drop to <50×109/L [9]. The diagnosis of DHF based on the WHO 1997 guidelines required the presence of fever, thrombocytopenia (platelet count <100×109/L), bleeding (bleeding from the mucosa, gastrointestinal tract, injection sites or other locations) and plasma leakage (clinical fluid accumulation, hypoproteinemia, increase in hematocrit of ≥20% or decrease in hematocrit of ≥20% after fluid resuscitation) [11]. DSS was diagnosed in patients with either rapid and weak pulse with narrow pulse pressure (<20 mmHg) or hypotension in a patient with DHF [11]. SD defined according to the WHO 2009 guideline required one of the following criterion; severe plasma leakage, severe bleeding or severe organ involvement [9]. Severe plasma leakage was defined as either clinical fluid accumulation or hematocrit change of >20% in combination with at least one of the following; tachycardia (pulse >100/minute), hypotension (systolic BP <90 mmHg) or narrow pulse pressure (<20 mmHg). Severe bleeding was defined as hematemesis, melena, menorrhagia or drop in hemoglobin requiring transfusion of blood products. Severe organ involvement included hepatic injury (alanine aminotransferase >1000 U/L or aspartate aminotransferase >1000 U/L), impaired consciousness or myocarditis [9].
HAI was defined as infection acquired after two days of hospital admission [27]. Urinary tract infection (UTI) was defined as a positive urine culture with clinical features of UTI. Bloodstream infection was defined as a clinically significant bacterium isolated from blood culture. Pneumonia was defined by the presence of new consolidation on chest X-ray with compatible clinical signs and symptoms. Clostridium difficile infection was defined as positive stool Clostridium difficile toxin with diarrhoea.
Outcome variables were categorized into dengue severity and poor clinical outcome. Dengue severity included patients with DHF, DSS and SD. Poor clinical outcome included patients who died, were admitted to the ICU or had excess length of stay (LOS). Excess LOS was defined as hospital admission greater than six days.
The chi-squared test and Fisher's exact test were used to compare univariate associations between categorical variables and the Mann-Whitney U test was used to compare continuous variables. A multiple logistic regression model based on inpatient data was built to ascertain how age was associated with excess LOS. The model adjusted for potential confounders including Pitt bacteremia score, HAI, Charlson co-morbidity score and dengue severity. The Hosmer-Lemeshow goodness-of-fit-test was applied to ensure the model fitted the data appropriately. A change in Pearson chi-squares graph was generated to identify the outlying and influential observations that may have affected the goodness-of-fit. Once identified the outlying and influential observations were tentatively removed and an auxiliary logistic regression model was rebuilt. The results of the two models were then compared in terms of the change in adjusted odds ratio (aOR) and their 95% confidence intervals (C.I.). All statistical analyses were performed using R version 2.15.2 and Stata 12.0 (Stata Corporation, Texas, U.S.A.). All tests were carried out with the 95% C.I. (equivalent to 5% ce level).
During the study period, of the 6989 dengue patients managed at CDC, 295 (4.3%) were elderly. There were 2034 (29%) who were PCR positive and 4955 (71%) who met the WHO criteria for probable dengue and were dengue IgM positive. There were a significantly higher proportion of PCR positive patients in the elderly (115/295, 39.1%) versus the adults (1919/6694, 28.7%).
Clinical, laboratory features and co-morbidities are shown in Table 1. Classical dengue symptoms of headache, rash and aches and pains were more common in the adults at presentation. Mucosal bleeding was significantly more common in the adults (24.2%) versus the elderly (12.5%) (p<0.001). Leukopenia at presentation was significantly more likely in the adults (5105/6694, 76.3%) versus the elderly (188/295, 63.7%) (p<0.001). The elderly were less likely to fulfill WHO 1997 probable dengue (93.6% versus 96.4%, p<0.001) as they were less likely than adults to have headache. In contrast, WHO 2009 probable dengue classification, which does not include headache as a criterion, was similar between the two groups (p = 0.167). Despite the atypical presentation of dengue in elderly patients there was no significant difference in time to dengue diagnosis between the two groups. The majority of patients (96%) were diagnosed on day one of admission with possible selection bias as the cohort was managed by the infectious diseases unit.
The elderly were significantly more likely to have co-morbidities (Table 1) including hypertension, diabetes, chronic renal impairment and chronic obstructive pulmonary disease. As expected a high Charlson co-morbidity score (>3) was significantly more common in elderly patients (p<0.001) reflecting a higher burden of co-morbidities.
Overall there was no significant difference in the frequency of warning signs between the two groups. However when individual warning signs were analyzed the elderly were more likely to have hepatomegaly (p = 0.006) and malaise/lethargy (p = 0.033) while the adults had significantly more mucosal bleeding (p<0.001).
The elderly were more likely to require hospitalization (265/295, 89.8%) versus the adults (5363/6694, 80.1%) (p<0.001), this is expected as the admission criteria from March 2007 included elderly patients with co-morbidities. Clinical outcomes are shown in Table 2. DHF occurred significantly more in the elderly cohort (86/295, 29.2%) versus the adults (1431/6694, 21.4%) (p = 0.002). Likewise, SD was more likely in the elderly (60/295, 20.3%) versus their younger counterparts (975/6694, 14.6%) (p = 0.006). Despite more severe disease in the elderly, ICU admission and death were not significantly different between elderly and adult dengue patients.
LOS was longer in the elderly with a median of five days versus four days in adults. HAI occurred at greater frequency in the elderly (13/295, 4.9%) versus the adults (66/6694, 1.2%). Notably pneumonia and UTI were the most common HAIs. There were no episodes of bloodstream infection in the elderly versus 14 episodes in the adults, but this did not reach significance. Clostridium difficile infection was detected in one adult patient. Older people were more likely to be admitted longer, after adjusting for HAI, Charlson score, Pitt bacteremia score and dengue severity (Table 3).
Dengue is a neglected tropical disease that is increasingly affecting elderly patients. As the dengue epidemic evolves and the population ages dengue in the elderly is likely to be commonplace. Diagnosis in this group may be challenging as the presentation can be atypical. Delayed diagnosis may delay lifesaving interventions. Elderly patients have worse outcomes compared with younger counterparts with increased rates of DHF, SD and dengue-related mortality. Elderly patients have higher rates of HAI placing them at risk of infection-related mortality. Elderly patients have an increased length of hospitalization as a result of severe disease, co-morbidity and HAI. This will place further burden on already stretched hospital systems.
Elderly patients had worse clinical outcomes with significantly higher rates of DHF and SD as previously reported [10], [17]. In our study this did not result in increased mortality unlike in Puerto Rico [17] and Taiwan [7], [18]. The reasons for this are unclear as the elderly cohort suffered more severe disease. The common manifestations of SD in the elderly were severe plasma leakage and severe bleeding. Despite more severe disease in the elderly, they did not experience more warning signs. In adult patients with confirmed dengue no single warning sign was highly sensitive in predicting either DHF or SD [28]. However hepatomegaly, persistent vomiting, hematocrit rise concurrent with rapid platelet drop and clinical fluid accumulation are highly specific for the development of both DHF and SD [28]. In our study hepatomegaly occurred significantly more commonly in the elderly. Vigilance for the above warning signs in the elderly is warranted to ensure rapid provision of potentially lifesaving interventions, while recognising the limitations of warning signs so as not to be falsely reassured in their absence.
DHF characterized by plasma leakage and bleeding is more common in secondary dengue infections [29]. In Singapore, older age is a significant risk factor for past dengue infection [30]. In a seroepidemiologic study of adults, 88.9% aged 55–74 years had evidence of past infection compared with 17.2% of young adults (18–24 years) [30]. Based on this data it is likely that many patients in our elderly cohort had secondary dengue infections increasing their risk of DHF [18].
Higher rates of DHF in the elderly may be the result of co-morbidities. In our cohort, elderly patients were significantly more likely to have hypertension and diabetes, both of which were recognised as risk factors for DHF [31]. A case-control study of DHF and dengue patients in Singapore demonstrated that adult patients with concomitant diabetes and hypertension were at higher risk of DHF, compared with patients without these co-morbidities [31]. The pathogenesis underlying this relationship is unclear but diabetes mellitus results in immune dysfunction [32] in addition to concomitant immunosenescence [13]. Elderly dengue patients with diabetes should be admitted for close monitoring, as close monitoring and early intervention with fluid therapy maybe lifesaving.
Concurrent infection has been reported in patients with dengue fever, including malaria, leptospirosis and Staphylococcus aureus [33], [34], [35]. Acute dengue infection can impair T-cell proliferation in vitro suggesting that dengue may modulate the immune system increasing susceptibility to co-infection [36]. Likewise, age results in immune dysregulation with defects in T and B cell function and impaired cytokine response [13]. Dengue and immunosenescence could explain the increased rates of HAI in elderly patients in our cohort. Median time from illness onset to death in dengue patients in Singapore was 12 days suggesting that HAI may have contributed [8]. Bacteremia was documented in 14.3% of the deaths with a median duration of 6.5 days from admission [8]. Prolonged fever (>5 days) and ARF were independent predictors of concurrent bacteremia in DHF patients in Taiwan [37]. Leukocytosis was more common in patients with dual infection versus controls but did not reach significance in this small cohort [37]. In our cohort neutrophilia was associated with nosocomial infection and excess LOS. Clinicians need to be aware of the potential for bacterial co-infection in elderly patients as they are at higher risk of mortality from severe sepsis versus younger counterparts [13], [38].
There are limitations to our retrospective study. Firstly, HAI may have been under reported as patients could have received empiric treatment without clinical investigation. Secondly, the risk of HAI can be increased by the presence of invasive devices such as intravenous and urinary catheters. The number of patients in this study who had invasive devices prior to the onset of HAI is unknown. Thirdly, the study was conducted at a single medical centre so the severity of illness may be biased by referral pattern. Future studies should include elderly patients managed in the community to enhance our understanding of the factors that affect hospital admission and outcome.
Dengue in the elderly is an emerging phenomena and remains incompletely understood. Dengue should be considered in the differential diagnosis of fever in elderly patients with appropriate epidemiological exposure, and diagnostic testing should be considered as this group presents atypically. Early diagnosis is critical for appropriate monitoring. Elderly patients are at higher risk of DHF and SD, especially those with pre-existing co-morbidities. A lower threshold for hospital admission may be required in this group for close monitoring. Clinicians need to remain vigilant for HAI in elderly dengue patients as these occur at increased frequency and may confer mortality risk.
|
10.1371/journal.pgen.1003412 | Drosophila DJ-1 Decreases Neural Sensitivity to Stress by Negatively Regulating Daxx-Like Protein through dFOXO | DJ-1, a Parkinson's disease (PD)–associated gene, has been shown to protect against oxidative stress in Drosophila. However, the molecular mechanism underlying oxidative stress-induced phenotypes, including apoptosis, locomotive defects, and lethality, in DJ-1-deficient flies is not fully understood. Here we showed that Daxx-like protein (DLP), a Drosophila homologue of the mammalian Death domain-associated protein (Daxx), was upregulated under oxidative stress conditions in the loss-of-function mutants of Drosophila DJ-1β, a Drosophila homologue of DJ-1. DLP overexpression induced apoptosis via the c-Jun N-terminal kinase (JNK)/Drosophila forkhead box subgroup O (dFOXO) pathway, whereas loss of DLP increased resistance to oxidative stress and UV irradiation. Moreover, the oxidative stress-induced phenotypes of DJ-1β mutants were dramatically rescued by DLP deficiency, suggesting that enhanced expression of DLP contributes to the DJ-1β mutant phenotypes. Interestingly, we found that dFOXO was required for the increase in DLP expression in DJ-1β mutants and that dFOXO activity was increased in the heads of DJ-1β mutants. In addition, subcellular localization of DLP appeared to be influenced by DJ-1 expression so that cytosolic DLP was increased in DJ-1β mutants. Similarly, in mammalian cells, Daxx translocation from the nucleus to the cytosol was suppressed by overexpressed DJ-1β under oxidative stress conditions; and, furthermore, targeted expression of DJ-1β to mitochondria efficiently inhibited the Daxx translocation. Taken together, our findings demonstrate that DJ-1β protects flies against oxidative stress- and UV-induced apoptosis by regulating the subcellular localization and gene expression of DLP, thus implying that Daxx-induced apoptosis is involved in the pathogenesis of DJ-1-associated PD.
| DJ-1 is an antioxidant protein that has been implicated in autosomal recessive Parkinson's disease (PD), although the mechanism by which DJ-1 deficiency causes PD remains elusive. Drosophila DJ-1 mutants are highly sensitive to oxidative stress and UV irradiation, and oxidative stress-induced cell death is significantly increased in dopaminergic neurons. In this study, we characterized a Drosophila homologue of death domain-associated protein (Daxx), Daxx-like protein (DLP), as a key player in the process of the oxidative stress-induced cell death in DJ-1 mutants. Upon oxidative stress, DLP expression was increased in the DJ-1 mutants, and locomotive defects and oxidative stress-induced phenotypes including apoptosis and lethality were dramatically rescued by DLP deficiency. More interestingly, we revealed that Drosophila forkhead box subgroup O was required for the increased DLP expression in DJ-1 mutants. Additionally, Drosophila DJ-1 suppressed DLP and Daxx translocation from the nucleus to the cytosol in both fly brain and mammalian cells. Interestingly, targeted expression of Drosophila DJ-1 to mitochondria efficiently inhibited Daxx translocation. Our results show that Drosophila DJ-1 protects dopaminergic neurons from oxidative stresses by regulating the subcellular localization and gene expression of DLP, providing a clue to understanding the molecular mechanism underlying oxidative stress-induced neuronal death in PD.
| Oxidative stress, a state of imbalance between the generation and elimination of reactive oxygen and nitrogen species, has been implicated in a variety of neurodegenerative diseases [1]–[3]. The central nervous system is presumed to be particularly vulnerable to oxidative stress, as it consumes abundant quantities of oxygen and employs nitric oxide as a biological messenger, both of which create reactive species as by-products [1]. Oxidative stress provokes various cytotoxic processes, such as overstimulation of glutamate receptors (excitotoxicity), ER stress, and mitochondrial dysfunction, which lead to apoptosis, the predominant form of cell death in aging-related neurodegenerative diseases [1], [2].
Parkinson's disease (PD) is characterized by typical motor dysfunction and is thought to be caused by the loss of nigrostriatal dopaminergic (DA) neurons that connect the substantia nigra pars compacta (SNpc) to other brain regions [3]–. The death of these neurons has been closely linked to oxidative stress [3]–[5]. Markers of oxidative damage to lipids, proteins and DNA, as well as mitochondrial DNA deletions, which can be caused by oxidative stress, are significantly elevated in postmortem samples of the SNpc of PD patients [4], [5]. The nigrostriatal pathway is sensitive to 6-hydroxydopamine and 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP)/MPP+, which destroy DA neurons via induction of oxidative stress [6]. Moreover, oxidative stress plays an important role in the function of the familial PD-related genes, α-synuclein and parkin [3]. Oxidative damaged α-synuclein is aggregated into Lewy bodies [7], and the function of Parkin is impaired by oxidative modifications [8], [9]. Additionally, various animal models of familial PD show greater damage in response to oxidative stress [10]–[16]. Although these data demonstrate a correlation between PD and oxidative stress, the molecular mechanisms underlying oxidative stress-induced DA neuronal death in PD are not well understood.
Among the genes related to PD, DJ-1 is the most closely associated with oxidative stress [17]. DJ-1 was originally identified as an oncogene that transforms mouse NIH3T3 cells in cooperation with ras [18], and its gene expression is increased in various types of cancer [19]–[21]. Later, DJ-1 was linked to an autosomal-recessive early-onset type of familial PD [22], [23]. The impairment of DJ-1 function sensitizes animal models to oxidative stress [11], [12], [14], [24]–[26]. DJ-1 performs several critical functions in response to oxidative stress via diverse cellular mechanisms [5], [17], [23]. First, DJ-1 functions as an atypical peroxiredoxin-like peroxidase that scavenges peroxides by oxidizing Cys106 [14]. Second, DJ-1 regulates expression of several antioxidant genes [27]–[30] and stabilizes the antioxidant transcriptional master regulator, Nrf2 [29], [31]. Third, DJ-1 inhibits UV- and oxidative stress-induced cell death by suppressing pro-apoptotic factors [32]–[34]. In Drosophila, there are 2 homologues of human DJ-1: DJ-1α and β [11], [12], [24]. DJ-1α is predominantly expressed in the testes, whereas DJ-1β is expressed in most tissues [11], [24], similar to the expression pattern of mammalian DJ-1 [18]. Several previous studies have demonstrated that DJ-1β loss-of-function mutants are acutely sensitive to oxidative stress and prone to locomotive dysfunction, resembling the phenotypes seen in PD [11], [12], [14], [24].
In this study, we identified a Drosophila homologue of death domain-associated protein (Daxx), Daxx-like protein (DLP), as a mediator of Drosophila DJ-1β mutant phenotypes. Daxx, originally identified as a binding partner of the pro-apoptotic receptor Fas (also called CD95) [35], performs a pivotal function in apoptosis [36], [37]. Daxx activates apoptosis signal-regulating kinase 1 (ASK1), which in turn increases c-Jun N-terminal kinase (JNK) activity leading to apoptosis [38], [39]. Daxx is increased in cells upon exposure to hydrogen peroxide and functions as a mediator of oxidative stress-induced apoptosis [33], [40], [41]. In neuronal cells, dominant negative-Daxx blocks Fas-induced cell death [42]. In addition, FADD/caspase-8 cascade-triggered cell death requires the transcriptional activation of Daxx in normal embryonic motor neurons [43]. Furthermore, Daxx has been identified as a potential component of the pathogenesis of neurodegenerative diseases, including PD [33], [44]. For example, Daxx interacts with DJ-1 [33], and MPTP induces translocation of Daxx from the nucleus to the cytoplasm and activates the ASK1 signaling pathway in mouse SNpc [44].
Although previous studies have demonstrated that Drosophila DJ-1β mutants are acutely sensitive to oxidative stress [11], [12], [14], [24], the cellular consequence of DJ-1β deficiency in the oxidative stress response remains unclear. Furthermore, the link between DJ-1 and PD, especially in the context of oxidative stress, has not been thoroughly explored. In this study, we used Drosophila and mammalian cells to investigate the functional interaction between DJ-1 and its downstream target, Daxx/DLP. We also characterized the relationship between loss of DJ-1 and PD-related phenotypes, such as DA neuronal degeneration and locomotive dysfunction, and examined the molecular mechanism of oxidative stress sensitivity in Drosophila DJ-1 mutants.
Drosophila DJ-1β mutants do not have gross morphological defects or loss of DA neurons when raised under standard laboratory conditions (Figure S1) [24]. However, expression level of tyrosine hydroxylase and number of DA neurons were significantly reduced in DJ-1β mutant flies after hydrogen peroxide treatment compared to those in wild-type animals (Figure 1A). The numbers of DA neurons in 3 major clusters of the posterior brain, dorsomedial clusters, dorsolateral clusters 1, and posteriomedial clusters, were significantly decreased by oxidative stress (Figure 1B). Next, we tested the effect of similar oxidative stress conditions on the neurons of larvae. As shown in Figure 1C, oxidative stress-induced cell death was dramatically increased in the brains of DJ-1β mutant larvae compared to that in wild-type controls. This suggests that DJ-1β mediates developmentally universal protection of DA neurons from oxidative stress-induced cell death.
To characterize the protective mechanism of DJ-1β, we used microarrays to compare the gene expression profiles of DJ-1β mutants and wild-type controls under oxidative stress conditions. We identified 143 upregulated and 134 downregulated genes (>1.5-fold changes in expression DJ-1β mutants) in the mRNA extracted from fly heads (Table S1). The role of DLP in the DJ-1-dependent oxidative stress response was first examined among the upregulated genes because Daxx, the mammalian homologue of DLP, has been implicated in oxidative stress-induced apoptosis [40] and identified as a potential component of PD pathogenesis [33], [44]. DLP is a 183.9-kDa protein with approximately 46% similarity to human Daxx in the Daxx-homology region; it is the only Daxx homologue in the Drosophila genome [45]. Interestingly, DLP protein levels were significantly higher in the fly head than in the body (Figure 2A and Figure S2), suggesting that it performs an important function in the brain. Furthermore, DLP mRNA and protein levels were increased by both UV irradiation and oxidative stress (Figure 2B and 2C), implying that DLP is involved in these stress responses, similar to its mammalian counterpart.
To confirm our microarray data, we evaluated DLP expression in the heads of wild-type and DJ-1β mutant flies following treatment with H2O2 using real-time quantitative PCR and western blot analysis. The levels of DLP mRNA and protein did not differ between the DJ-1β mutants and wild-type controls under standard laboratory conditions (Figure 2D and 2E). However, as anticipated, the levels of DLP mRNA and protein were significantly elevated in DJ-1β mutants in comparison to wild-type controls under oxidative stress (Figure 2F and 2G), suggesting that the increasing rate of DLP expression by oxidative stress in DJ-1β mutants is higher than that in wild type. When DJ-1β was overexpressed with a pan-neuronal elav-GAL4 driver, the levels of DLP transcript and protein were significantly reduced (Figure 2H and 2I). Interestingly, mammalian Daxx gene expression was also higher in DJ-1 null cells than in wild-type controls under oxidative stress condition (Figure 2J). These results indicate that DJ-1 functions as a negative regulator of Daxx/DLP gene expression under oxidative stress conditions.
As mammalian DJ-1 inhibits translocation of the nuclear Daxx to the cytosol [33], [46], we examined whether Drosophila DJ-1β also regulates subcellular localization of DLP. First, we fractionated DLP protein from the cytosol and nucleus of wild-type and DJ-1β mutant fly heads (Figure 2K). The proportion of cytosolic DLP relative to nucleic DLP was increased more than 3-fold in DJ-1β fly heads (Figure 2K and 2L). Consistently, immunohistochemical analysis with anti-DLP antibody showed that the cytosolic DLP level was significantly increased in the brain and eye imaginal disc of DJ-1β mutant flies compared to that in wild-type flies (Figure 2M and Figure S3, respectively), which was highly similar in DJ-1 null mouse DA neuroblastoma cells (Figure 2N).
Since we previously showed that DJ-1 is partially localized in mitochondria [24], we examined whether mitochondrial translocation of DJ-1 is important for the cytosolic localization of Daxx by comparing the effect of several forms of DJ-1 that are targeted to various subcellular regions including mitochondria, Golgi, nucleus, and cytoplasmic membrane. Interestingly, the mitochondrial DJ-1β efficiently inhibited translocation of Daxx from the nucleus to the cytosol under oxidative stress conditions, like wild-type or nucleus targeted DJ-1β (Figure 2O and Figure S4). These results suggest that DJ-1 regulates the translocation of Daxx/DLP as well as their gene expression in response to oxidative stress.
To evaluate the role of DLP in the hypersensitivity of DJ-1β mutants to oxidative stress and UV irradiation, we generated and characterized DLP mutants. The mutant EY09290, which harbors a P-element inserted in the 5′ region of the DLP gene, was acquired from the Bloomington Drosophila Stock Center, and deletion mutants were generated via P-element mobilization (Figure 3A). Two deletion alleles were generated and designated DLP1 and DLP2. The genomic deletions were confirmed by PCR (Figure 3B) and DNA sequencing. These alleles have deletions of 1,311 and 1,076 bp, respectively, which removes the first 2 exons, including the translational start site of DLP (Figure 3A). Western blot analysis and RNA in situ hybridization results confirmed that DLP protein and mRNA levels were markedly reduced in DLP1 and DLP2 compared to the levels in wild-type controls (Figure 3C and 3D). In order to exclude the genetic background effect, we generated two other DLP deletion mutants, DLP3 and DLP4, using another P-element line, KG01694 (Figure 3A), and we obtained another DLP mutant, DLPU42, with a different genetic lineage [45]. Therefore, we used five mutant alleles in three different genetic backgrounds to characterize the DLP mutants. Furthermore, Upstream Activation Sequence (UAS)-DLP-RNAi was used for DLP knockdown to confirm the DLP mutant phenotypes.
Because Daxx has been implicated in oxidative stress and UV responses, the role of DLP in oxidative stress-induced lethality was assessed using these DLP mutants. As expected, DLP mutants proved more resistant to H2O2 treatment than wild-type or trans-heterozygotes between DLP revertant (DLPrv) and DLP1 (Figure 3E and Figure S5). Moreover, DLP knockdown in all neurons using UAS-DLP-RNAi conferred increased resistance to oxidative stress, whereas DLP overexpression in the same neurons increased their sensitivity to oxidative stress (Figure 3F). Oxidative stress-induced cell death was also attenuated in the DLP mutant brain (Figure 3G). UV irradiation-induced pupal lethality and apoptosis were also significantly reduced in DLP mutants (Figure 3H and 3I), indicating that DLP is involved in UV-induced stress responses. These results indicate that the neuronal function of DLP is important to the oxidative and UV stress responses at both cellular and organism level.
Daxx was originally identified as a pro-apoptotic gene that induced cell death [35]. However, the role of Daxx and DLP in apoptosis is somewhat controversial [38], [45]. In order to confirm the function of DLP in non-stress-induced apoptosis, we utilized the UAS-GAL4 system to evaluate the effects of DLP overexpression on developing tissues. As EY09290 harbors a P-element with UAS in the 5′ region of DLP and the direction of the P-element is oriented to induce DLP gene expression (Figure 3A), the mutant was employed to study DLP overexpression. The induction of DLP expression was confirmed by RNA in situ hybridization in the wing imaginal discs of the EY09290 line harboring MS1096-GAL4 driver (Figure S6A). Interestingly, DLP overexpression in the developing wing under the control of MS1096-GAL4 reduced organ size in a dose-dependent manner (Figure 4A). When DLP was overexpressed in neurons, reduced survival and defects in locomotive behavior were observed (Figure S6B and S6C), suggesting that increased DLP expression has an adverse effect on neuronal development and function in Drosophila. Indeed, DLP overexpression strongly induced cell death in the imaginal disc (Figure 4B; compared with MS1096/Y and MS1096>DLP×2) without affecting the cell cycle or differentiation (Figure S6D and S6E). Furthermore, the DLP-induced reduction in wing size was suppressed almost completely by co-expression of Drosophila inhibitor of apoptosis protein 1 (DIAP1; Figure 4A), a caspase inhibitor. These results demonstrate that DLP activity induces apoptosis through caspase activation and reduces overall survival, as demonstrated by the reduced survival of DLP-overexpressing flies (Figure S6B).
In mammals, Daxx induces apoptosis by activating the JNK signaling pathway [35], [39]. We attempted to determine whether the JNK signaling pathway is activated by DLP in Drosophila. We initially examined the genetic interaction of DLP with basket (bsk), a Drosophila JNK, and hemipterous (hep), a Drosophila JNK kinase (JNKK). Although overexpression of DLP, bsk, or hep in the wing produced only a slight reduction in wing size, overexpression of either bsk or hep in conjunction with DLP resulted in severely rumpled and shrunken wings (Figure 4C). Consistently, co-expression of hep with DLP strongly induced cell death in the wing imaginal disc (Figure 4B). On the other hand, bsk or hep loss-of-function mutation or co-expression of puckered (puc), a negative regulator of JNK (a JNK phosphatase), suppressed the cell death and wing deformities induced by DLP overexpression (Figure 4B and 4C). Furthermore, co-expression of bsk or hep with DLP resulted in a strong increase of JNK phosphorylation (Figure 4D). These results demonstrate that DLP activates the JNK signaling pathway, similar to mammalian Daxx.
As FOXO is a target of JNK in mammalian systems [47] and acts downstream of JNK signaling in the control of apoptosis [48], we investigated the role of FOXO in DLP-induced apoptosis. We examined the effect of FOXO deficiency on the DLP-induced gain-of-function phenotype and apoptosis. FOXO deficiency (dFOXO21) suppressed the DLP-induced wing (Figure 4E) and cell death (Figure 4B) phenotypes, suggesting that the JNK/dFOXO pathway is downstream of DLP.
It has been shown that JNK activates FOXO4 activity through phosphorylation at Thr447 and Thr451 [47]. Therefore, we assessed whether dFOXO is also phosphorylated by JNK. Amino acid sequence analysis between mammalian FOXOs and dFOXO did not reveal the conserved Thr447 and Thr451 phosphorylation sites in dFOXO. Furthermore, we could not see any evidence to support direct phosphorylation of dFOXO by JNK in in vitro phosphorylation experiments (data not shown). However, interestingly, expression of constitutively active JNKK (hepCA) significantly increased the protein level of dFOXO (Figure 4F), implicating that JNK regulates dFOXO activity by increasing its protein level or stability in Drosophila.
Because DLP expression is regulated by DJ-1β under oxidative stress conditions (Figure 2) and DLP is important for oxidative stress-induced apoptosis and lethality (Figure 3 and Figure S5), we assessed the role of DLP in the oxidative stress-related phenotypes of DJ-1β mutants, specifically their acute sensitivity to oxidative stress and locomotive dysfunction. To accomplish this, we generated DLP and DJ-1β double mutants and asked whether DLP deficiency could rescue various DJ-1β mutant phenotypes. We first evaluated the H2O2 sensitivity of these lines. As shown in Figure 5A, DLP and DJ-1β double mutants displayed survival rates similar to those of wild-type controls, whereas DJ-1β mutants were acutely sensitive to H2O2. However, the survival rates of DLP and DJ-1β double mutants was still lower than those of DLP mutants (compare Figure 5A with Figure 3E), which suggests that the H2O2 sensitivity of DJ-1β is not fully dependent on DLP. Supporting this, gene expression of 18 oxidative stress-related genes (11 up-regulated and 7 down-regulated) was altered in DJ-1β mutants versus wild-type controls (Table S1). Although DLP is not expected to mediate the whole effect of the oxidative stress responses in DJ-1β mutants, DLP deficiency strongly suppressed the oxidative stress-induced cell death observed in DJ-1β mutant brains (Figure 5B). Consistently, oxidative stress-induced DA neuronal death in DJ-1β mutants was almost completely inhibited by DLP deficiency (Figure 5C and Figure S7). These findings indicate that DLP is a key mediator in the oxidative stress-induced neuronal cell death of DJ-1β mutants. Furthermore, DLP deficiency rescued the UV sensitivity (Figure 5D) and locomotive dysfunction (Figure 5E) of DJ-1β mutants. These results strongly suggest that DLP mediates the H2O2-induced oxidative stress responses, UV sensitivity, and locomotive dysfunction of DJ-1β mutants.
Next, we investigated the mechanism by which DJ-1β regulates DLP expression. Previous studies with mammalian DJ-1 suggest that DJ-1 affects oxidative stress-related gene expression by stabilizing Nrf2 [29] or by suppressing p53 transcriptional activity [34]. Alternatively, DJ-1 regulates phosphatidylinositol 3-kinase (PI3K)/Akt signaling [19], [49], [50], which inhibits FOXO. Therefore, we tested whether cap'n'collar C (cncC, Drosophila Nrf2), p53, or dFOXO is involved in the regulation of DLP gene expression by DJ-1β under oxidative stress conditions. As demonstrated in Figure 6A and Figure S8, neither cncC nor p53 affected the DLP levels in wild-type or DJ-1β mutant flies. However, the increased DLP mRNA (Figure 6B) and protein (Figure 6C) levels in DJ-1β mutants were restored to normal levels by dFOXO deficiency, suggesting dFOXO is required for elevation of DLP expression in DJ-1β mutants. Thus, when dFOXO was overexpressed in neurons, DLP expression was elevated more than 2-fold for both mRNA (Figure 6D) and protein (Figure 6E). Since a putative FOXO recognition element (FRE, AAAAACA) is located at 1,041 bp upstream of the transcription start site of the DLP gene (Figure 6F) [51], to examine the positive effect of dFOXO on the DLP gene expression at the transcriptional level, we cloned two different sizes of the DLP promoter region into a firefly reporter plasmid; 0.5 kb and 1.3 kb, respectively. Upon transient co-transfection with a construct expressing the constitutively active form of dFOXO (pMT-dFOXO A3), the construct containing the 1.3-kb fragment of the promoter region, but not the 0.5-kb fragment, exhibited the dFOXO-dependant promoter activity in a dose dependent manner (Figure 6F). To confirm whether this putative FRE site is critical, a mutation that disrupts the dFOXO binding was introduced into the 1.3-kb promoter construct by site-directed mutagenesis. Upon transfection, the construct containing the mutant 1.3-kb promoter region no longer showed dFOXO-dependent promoter activity (Figure 6F), indicating that dFOXO regulates DLP expression through this FRE site.
Additionally, we revealed that phospho-Akt (an activated form of Drosophila Akt (dAkt), a negative regulator of dFOXO) was significantly reduced in DJ-1β mutants (Figure 6G). We also found that gene expression of Drosophila 4E-BP, a target of dFOXO, increased in DJ-1β mutants, although dFOXO gene expression remained unaltered (Figure 6H). Moreover, overexpression of PTEN, a negative regulator of the Akt signaling pathway and, therefore, an activator of dFOXO, elevated the DLP level (Figure 6I). These results consistently indicate that DJ-1β regulates DLP gene expression through the PI3K-Akt-dFOXO pathway.
Finally, we investigated whether DJ-1β could suppress dFOXO-induced apoptosis. As shown in Figure 6J, DJ-1β overexpression strongly suppressed dFOXO-induced eye degeneration and apoptosis (compare ey>dFOXO with ey>dFOXO+DJ-1β).
Recently, several Drosophila PD models, each of which represents a different PD-associated gene mutant, have been developed and characterized [10]–[13], [16], [24], [49], [52]–[54]. Although each exhibits a distinct phenotype, a common feature of all these models is sensitization to oxidative stress. This, along with pathology data from PD patients [55]–[57], strongly indicates a significant role for oxidative stress in the development and progression of PD. Since the DJ-1 mutations have been linked to familial PD and hypersensitivity to toxins that induce oxidative stress [23], we examined a signaling pathway that controls oxidative stress responses by DJ-1 in Drosophila. We demonstrated that DJ-1β inhibits oxidative stress-induced neuronal apoptosis by regulating DLP gene expression and protein subcellular localization, suggesting a causal relationship between DJ-1β mutation and oxidative stress-induced DA neuronal loss in PD.
Our genetic and cellular analyses indicate DLP functions as a pro-apoptotic gene and as a JNK activator in Drosophila, like its mammalian homologue Daxx. Previous studies have shown that Daxx is upregulated in response to oxidative stress and UV irradiation; it also mediates apoptosis in these contexts [40], [41]. Consistent with these reports, our results demonstrate that DLP expression is elevated by H2O2 and UV exposure. Moreover, the apoptosis induced by these insults is reduced dramatically in DLP mutants. In contrast to oxidative stress or UV irradiation, γ-ray irradiation (40 gray)-induced apoptosis was unaffected by DLP deficiency (data not shown). This is consistent with a previous report, which demonstrated that DLP is not associated with radiosensitivity [45]. These findings suggest DLP does not function as a general pro-apoptotic factor, but rather exerts a pro-apoptotic function in response to specific insults, including oxidative stress and UV irradiation. Moreover, the level of DLP in neurons was associated with fly survival rates under oxidative stress conditions. The pan-neuronal overexpression of DLP rendered flies more sensitive to oxidative stress than controls, while knockdown or loss of DLP resulted in resistance to oxidative stress. Therefore, we believe DLP functions as a stress response mediator that generates appropriate cellular responses to oxidative stress. The similarities between the functions of DLP and Daxx suggest that this oxidative stress response pathway is highly conserved from insects to mammals.
Due to the pronounced increase in oxidative damage within DJ-1β mutants [52], we hypothesized that DLP functions as an important mediator of hypersensitivity to oxidative stress in DJ-1β mutants. Indeed, DLP expression and translocation from the nucleus to the cytoplasm increased in DJ-1β mutants, and DLP deficiency almost completely rescued the phenotypes of DJ-1β mutants, including oxidative stress-induced DA neuronal loss. Moreover, overexpression of DJ-1β reduced the level of endogenous DLP. These findings suggest DLP plays an important function in the oxidative stress-related phenotypes of DJ-1β mutant flies and that DJ-1β protects flies against oxidative stress, at least in part, by suppression of DLP expression and cytosolic localization.
These observations raised the question of how DJ-1β negatively regulates DLP expression at the transcriptional level under oxidative stress conditions. Our data indicate that DJ-1β controls DLP gene expression by regulating the activity of dFOXO. Furthermore, DLP harbors a consensus FRE in its promoter region and dFOXO overexpression increased DLP expression in neurons. Previous studies as well as this work identified DJ-1 as a positive regulator of the PI3K/Akt pathway [19], [49], [50], which suppresses the activity of FOXO by phosphorylation [58]. Therefore, it was not surprising to see that loss of DJ-1β function reduced dAkt activity and increased the transcriptional activity of dFOXO (Figure 6G–6H). FOXO activity may be crucial for setting the sensitivity threshold for oxidative stress and determining the appropriate level of stress responses, which ultimately determines whether the cells live or die. From this perspective, the elevated level of dFOXO activity in DJ-1β mutants may render their neurons more sensitive to stress, and thus neurons in mutant animals die more readily than their wild-type counterparts. We also found that dFOXO performs a dual role in DJ-1β mutant flies, as it is required for the upregulation of DLP and is an effector of DLP. Both of these roles increase the DLP-mediated apoptosis in response to oxidative stress in DJ-1β mutants. This suggests that dFOXO is involved in the loss of neurons due to oxidative stress, and possibly, DJ-1 mutation-associated familial PD cases.
In addition to transcriptional regulation, DJ-1 controls DLP translocation from the nucleus to the cytosol. Daxx is translocated to the cytosol under oxidative stress conditions [59] and this translocation is important for its pro-apoptotic function [60]. These studies and our results showed that both mammalian and Drosophila DJ-1 strongly suppress the cytosolic translocation of Daxx/DLP. The molecular mechanism by which DJ-1 suppresses the DLP translocation is elusive. It has been proposed that mammalian DJ-1 directly binds to Daxx and inhibits its translocation. However, we did not observe prominent binding between Drosophila DJ-1 and DLP (data not shown), suggesting Drosophila DJ-1 may regulate the DLP translocation by an alternative mechanism. Interestingly, the mitochondrial targeted DJ-1β efficiently inhibited the Daxx translocation under oxidative stress conditions, suggesting the function of DJ-1 in mitochondria is important for the translocation. Further studies are necessary to understand how DJ-1 inhibits the cytosolic localization of Daxx/DLP under oxidative stress.
Our work with DJ-1β and its downstream effector, DLP, has led us to propose the models illustrated in Figure 7. In a wild-type animal, DJ-1 protects the cells from oxidative stress-induced apoptosis. This protection is a result of activation of the PI3K/Akt pathway that inhibits dFOXO. dFOXO, among its many functions, induces DLP transcription. DLP expression levels have a direct positive correlation with the likelihood of a cell to undergo apoptosis in response to oxidative stress. It is important to note that not only does DJ-1β suppress DLP expression, but DJ-1β also prevents DLP translocation to the cytosol, which may be critical for the pro-apoptotic function of DLP. However, once cells are damaged by oxidative stress and UV irradiation, the DLP protein acts through the JNK pathway to initiate apoptosis. Since the JNK pathway can increase dFOXO activity, DLP expression can be further increased by a hypothetical feed-forward loop of DLP-JNK-dFOXO. In wild-type animals, this pro-apoptotic loop can be negatively regulated by DJ-1β, while in DJ-1β mutant animals, the inability to control this process leads to increased DLP levels and apoptosis. This increased chance of apoptosis may be an important factor in the development of PD.
DJ-1βex54 and p53E4 were previously described [24], [61], and UAS-HA-DJ-1β was generated via microinjection of the corresponding plasmid into w1118 embryos. EY09290, KG01694, basket1 (bsk1), UAS-DIAP1, UAS-dFOXO, UAS-PTEN, elav-GAL4, Glass multimer reporter (GMR)-GAL4, tubulin (tub)-GAL4, eyeless (ey)-GAL4, sevenless (sev)-GAL4, wingless (wg)-lacZ, and engrailed (en)-lacZ were acquired from the Bloomington Drosophila Stock Center (Bloomington, IN, USA). UAS-DLP-RNAi was obtained from the Vienna Drosophila RNAi Center (Vienna, Austria). UAS-basket (bsk) and UAS-hemipterous (hep) were gifts from Dr. M. Mlodzik (EMBL, Germany). UAS-puckerd (puc) and MS1096-GAL4 were generously provided by Dr. M. Peifer (University of North Carolina, Pembroke, NC) and Dr. M. Freeman (MRC Laboratory of Molecular Biology, Cambridge, UK), respectively. UAS-cncC and UAS-cncC-RNAi were gifts from Dr. D. Bohmann (University of Rochester Medical Center). DLPU42 and dFOXO21 were gifts from Dr. I. M. Boros (University of Szeged, Hungary) and Dr. E. Hafen (University of Zurich, Switzerland), respectively. hemipterous1 (hep1) was gift from Dr. S. Noselli (CNRS, France). UAS-hemipterousCA (hepCA, the constitutively active form of Drosophila JNKK) was gift from Dr. K. Mastsumoto (Nagoya University, Japan). All fly strains were maintained at 25°C.
Total RNA was extracted from the heads of hydrogen peroxide-treated wild-type and DJ-1β mutant flies using an RNeasy Mini kit (Qiagen) in accordance with the manufacturer's instructions. Total RNA was used as a probe for microarray analyses. GeneChip Drosophila Genome 2.0 Arrays for Drosophila melanogaster were probed, hybridized, stained, and washed in accordance with the manufacturer's recommendations. Hybridized arrays were scanned using an Affymetrix Command Console, and normalization was conducted using an Affymetrix Expression Console 1.1 (MAS5). These experiments were repeated three times for each sample. We required that the fold change difference between the average of three independent wild-type samples and the average of three independent DJ-1β mutant samples exceed 1.5 (p≤0.005).
To generate DLP mutant flies, we used the EY09290 and KG01694 lines (Bloomington, USA) containing the UAS in the first exon and first intron of the DLP gene, respectively. Two DLP mutants, DLP1 and DLP2, were generated via the imprecise excision of a P-element in the EY09290 line, and DLP3 and DLP4 were obtained from the KG01694 line. Additionally, the revertant line, DLPrv, was also generated via precise excision of the P-element in the EY09290 line. Among the excision lines, the deletion lines were selected via genomic DNA PCR. Genomic DNA was prepared from each independent line, and PCR was conducted with primer pairs to amplify sequences lying upstream and downstream of the P-element insertion site. The deletion sites of the selected lines were determined by sequencing the PCR products. To isogenize the genetic background, DLP1 and DLP2 were backcrossed with w1118 ten times.
A polyclonal antibody against the C-terminus of Drosophila DLP (amino acids 1300–1559) was generated in rabbits via injection of pGEX-fused DLP. The specificity of the DLP antibody was verified by immunoblotting and immunohistochemistry using wild-type and DLP mutant fly tissues (Figure 3C and Figure S2). Cell and fly lysates were prepared in lysis buffer A (150 ml NaCl, 25 mM Tris, 10% Glycerol, 0.1% NP-40, and 1 mM EDTA). For western blotting, the membranes were probed with anti-DLP (1∶1,000 in Tris-buffered saline with Tween 20 (TBST)), anti-lamin (1∶2,000 in TBST; Developmental Studies Hybridoma Bank (DSHB)), anti-β-tubulin (1∶2,000 in TBST; DSHB), anti-phospho-Drosophila Akt (Ser505) (1∶1,000 in TBST; Cell Signaling Technology), anti-Akt (1∶1,000 in TBST; Cell Signaling Technology), anti-dFOXO (1∶2,000 in TBST; Cosmo Bio, Japan), or anti-Actin (1∶2,000 in TBST; DSHB at the University of Iowa) antibodies. Western blot analyses were conducted with standard procedures using horseradish peroxidase-conjugated secondary antibodies (1∶2,000 in TBST; Cell Signaling Technology).
To separate nuclear and cytosolic fractions, 15 fly heads were collected and homogenized. Nuclear and cytosolic fractions from the fly heads were isolated with Nuclear Extract Kit (Active Motif) in accordance with the manufacturer's instructions.
In situ hybridization experiments were conducted using a digoxigenin-labeled RNA probe (Roche Applied Science) in accordance with the manufacturer's instructions. The probe was prepared using the PCR product of DLP. Hybridization was conducted at 55°C, and the RNA hybrids were detected with alkaline phosphatase (AP)-conjugated anti-digoxigenin antibody followed by nitro blue tetrazolium (NBT)/5-bromo-4-chloro-3-indolyl-phosphate (BCIP) staining.
One hundred embryos of each genotype were placed on grape juice agar plates. After incubation for 2 days at 25°C, the number of larvae hatched was counted to determine embryonic lethality. These larvae were then transferred to standard media (cornmeal, yeast, molasses, agar) and aged at 25°C in upright standard plastic shell vials. Larvae were maintained under non-crowded conditions with 20 individuals per vial. The numbers of pupae and enclosed adult flies were counted. Experiments were repeated 3 times with 100 flies per genotype.
UV irradiation experiments were conducted as previously described with some modifications [48]. In brief, 10 mid-aged pupae were collected, and the pupal shells surrounding the heads were surgically removed. The samples were UV irradiated at 10 mJ/cm2 using a UV crosslinker (CL-1000, UV Products). Following irradiation, the pupae were kept in darkness until processing, and the survival rate was determined. Each experiment was repeated more than 5 times. To analyze the effects of UV irradiation on cell death or DLP gene expression in the embryos, 0–3 h embryos were UV-irradiated at a dose of 50 mJ/cm2.
The effects of oxidative stress on the survival of the indicated lines were evaluated by feeding with hydrogen peroxide. Fifty 3-day-old male flies of the indicated lines were starved for 6 h and then transferred to vials containing 1% hydrogen peroxide in 5% sucrose solution. Surviving flies were counted semi-diurnally. We carried out each survival experiment at least 5 times with 50 flies per genotype (n≥250). In order to evaluate gene expression and protein levels under oxidative stress conditions, the flies were fed with 1% hydrogen peroxide for 3 days.
For real-time quantitative PCR, total RNA from the 20 fly heads was isolated with an RNeasy Protect Mini kit (Qiagen). Then, cDNA was synthesized with a Maxime kit (iNtRON Biotechnology), and real-time quantitative PCR was undertaken using SYBR Green PCR Master Mix (Applied Biosystems) according to the manufacturer's recommended protocols. Real-time quantitative PCR was performed using StepOne Real-time PCR system (Applied Biosystems). Quantification was performed using the ‘delta-delta Ct’ method to normalize to Actin transcript levels and to control. Each experiment was repeated at least 5 times (n≥5). The relative level of DLP, d4E-BP or dFOXO mRNA to Actin mRNA was statistically analyzed by Student's t-test.
To determine the deletion sites of the DLP mutant lines, PCR was conducted using the following primer pairs: 5′-ACTGCAAATAGTGAATTAAGGCAAC-3′ and 5′- TGCAACATGGGAAGTCTCTG-3′. To quantify the level of gene expression, real-time quantitative PCR was conducted using the following primer pairs: DLP, 5′-CACATCCCCAGTGGAATCAC-3′ and 5′-TGCCAACATTGATCTGCTTC-3′; Actin, 5′-CACCGGTATCGTTCTGGACT-3′ and 5′-GCGGTGGTGGTGAAAGAGTA-3′; dFOXO, 5′-GCCTGGAGGTGCTCAATAAC-3′ and 5′-GTGGCCAGCGGTATATTGAT-3′; and d4E-BP, 5′-CCATGATCACCAGGAAGGTT-3′ and 5′-GAAAGCCCGCTCGTAGATAA-3′.
For the TUNEL assay, embryos treated with UV were fixed in 4% paraformaldehyde in phosphate-buffered saline (PBS) for 30 min at room temperature. The samples were then washed with PBS and permeabilized by a 2-min incubation in PBS containing proteinase K (10 µg/mL) and 0.1% Triton X-100 on ice. After extensive washing, the samples were incubated for an additional 3 h in TUNEL reaction solution (Roche Applied Science) at 37°C in accordance with the manufacturer's recommendations. After three rinses with PBS, the embryos were incubated with anti-digoxigenin-AP antibody and stained with NBT/BCIP.
Full-length DJ-1β cDNA was cloned into dsRed2-Mito vector (Clontech) to target DJ-1β to mitochondria. dsRed2-Mito vector contains the mitochondrial targeting sequence from subunit VIII of human cytochrome c oxidase at the N-terminus. For the nucleus-targeted DJ-1β, pEF/3×NLS/Myc vector (Invitrogen) was used. To construct the cytoplasmic membrane-targeted DJ-1β, 20-amino acid farnesylation signal from c-Ha-Ras was fused to the C-terminus of DJ-1β resulting in pcDNA3 3×HA DJ-1β-F. For Golgi-targeted DJ-1β, Golgi targeting sequence from β 1,4-galactosyltransferase was fused to the N-terminus of DJ-1β resulting in pcDNA3 Golgi DJ-1β.
HeLa cells were grown in DMEM (Invitrogen) supplemented with 10% fetal bovine serum (Invitrogen) at 37°C in a humidified atmosphere of 5% CO2. WT and DJ-1 null SN4741 cells were established from the substantia nigra region of E13.5 wild-type and DJ-1 knockout mouse embryos, respectively [62]. SN4741 cells were grown in RF medium containing DMEM supplemented with 10% fetal bovine serum, 1% glucose, and L-glutamine (2 mM) at 33°C with 5% CO2. The transfection of expression plasmids was performed using Lipofectamine plus reagent (Invitrogen), or PEI (Polyethylenimine, Sigma) according to the manufacturer's instruction.
For immunocytochemistry, SN4741 or HeLa cells were sub-cultured on 12-well culture plates coated with poly-L-lysine (Sigma). Appropriately treated cells were washed once with PBS and fixed in 2% paraformaldehyde for 15 min, followed by permeabilization with 0.5% Triton X-100 in PBS for 5 min. Then, the cells were washed with 0.1% Triton X-100 in PBS (PBS-T) and incubated in blocking solution (4% BSA and 1% normal goat serum in PBS-T) for 1 h. Primary antibodies were added to the blocking solution and the cells were incubated overnight at 4°C. After washing with PBS-T 3 times, the cells were incubated with appropriate secondary antibodies in blocking solution for 45 min at room temperature. The antibody-labeled cells were washed with PBS-T 6 times and mounted with mounting solution [100 mg/mL 1,4-diazabicyclo[2.2.2]octane (DABCO) in 90% glycerol]. The slides were observed with LSM710 laser-scanning confocal microscope (Carl Zeiss). All immunostaining experiments with HeLa cells were conducted at least 3 times (n = 300). Anti-mouse DJ-1β [24] and anti-rabbit Daxx (Cell Signaling Technology) were used as primary antibodies. MitoTracker Red CMXRos (Invitrogen) was used to visualize mitochondria.
For immunohistochemistry, the wing or eye imaginal discs or adult brains were fixed in 4% paraformaldehyde in PBS at room temperature. The tissues were then washed in PBT (PBS+0.5% Triton X-100) and blocked in PBT with 2% normal goat serum (NGS). The samples were incubated first with rabbit anti-phospho-JNK antibody (1∶200 in PBT containing 2% NGS; Promega) or rabbit anti-phospho-histone H3 antibody (1∶200 in PBT containing 2% NGS; Upstate Biotechnology) or rabbit anti-DLP antibody (1∶200 in PBT containing 2% NGS) or rabbit anti-tyrosine hydroxylase antibody (1∶50 in PBT containing 2% NGS; Pel-Freez Biologicals) and then subsequently incubated with rhodamine-labeled goat anti-rabbit immunoglobulin G secondary antibody (1∶200 in PBT; Sigma-Aldrich).
The UAS-GAL4 system was used to evaluate the phenotypes induced by the overexpression of several target genes, including DLP. The GAL4 gene was placed near a tissue-specific enhancer, allowing for the ectopic expression of the target gene in the desired tissue. GMR-GAL4, MS1096-GAL4, elav-GAL4, tub-GAL4, and ey-GAL4 were used to induce target gene expression in the eye, the whole wing, the nervous system, the whole body, and the eye, respectively.
Acridine orange staining was conducted as previously described [63] with some modifications. The wing or eye imaginal discs of stage L3 larvae were dissected in PBS. In order to characterize the effects of oxidative stress on cell death, we incubated the larval brains for 24 h in Schneider's Drosophila media with 0.1% hydrogen peroxide. The discs or brains were then incubated for 5 min in 1.6×10−6 M acridine orange (Sigma-Aldrich) and briefly rinsed in PBS. The samples were subsequently observed under an Axiophot2 fluorescence microscope (Carl Zeiss).
For X-gal staining, the wing discs were fixed for 4 min in 4% formaldehyde in PBS, washed, and incubated in standard X-gal staining solution (4.9 mM X-gal, 3.1 mM K4Fe(CN)6, 3.1 mM K3Fe(CN)6, 1 mM MgCl2, 150 mM NaCl, 10 mM Na2HPO4, 10 mM NaH2PO4, 0.3% Triton X-100) for 30 min at 37°C before observation.
Drosophila S2 cells were transiently transfected using the Effectene transfection reagent (Qiagen Inc., Valencia, CA) using a standard protocol. After 24 h, CuSO4 (Sigma) was added to a final concentration of 0.6 mM for the optimal expression of the dFOXO A3 construct (a gift from Dr. Oscar Puig, Roche). Dual luciferase assays were performed using the dual luciferase assay system (Promega Corp., Madison, WI). Harvested cells were lysed with luciferase cell lysis buffer. Cell lysates (20 µl out of 30 µl total lysate per sample) were analyzed for firefly luciferase activity by adding 20 µl of firefly reaction buffer. Furthermore, Renilla luciferase activity was measured by adding 20 µl of Renilla reaction buffer. Luminescence was measured from a 96-well plate by using VictorX5 multilabel plate reader (Perkin-Elmer). The data are presented as fold changes relative to the negative control (transfection with the pMT empty vector as an effecter plasmid) normalized to 1.
The climbing assay was conducted as previously described [64], [65] with some modifications. Ten male flies of the indicated lines were transferred into the climbing ability test vial and incubated for 1 h at room temperature for environmental acclimation. After tapping the flies down to the bottom, we counted the number of flies that climbed to the top of the vial within 4 sec. Ten trials were conducted for each group. The experiment was repeated at least 10 times with independently derived transgenic lines. Climbing scores (ratio of the number of flies that climbed to the top to the total number of flies, expressed as a percentage) were obtained for each test group, and the mean climbing score for 10 repeated tests was compared to the scores of the wild-type flies. All climbing assay experiments were conducted at 25°C.
Western blotting data were measured using the Multi gauge V3.1 (Fuji, Japan) software program and converted into ratios of band intensity relative to the controls. Using the non-parametric Wilcoxon signed-rank test or the Kruskal-Wallis test, the data were analyzed to detect any statistical differences between treatments. In particular, when the data analyzed with the Kruskal-Wallis test revealed a statistical difference, the data were arcsine-transformed and subsequently analyzed by ANOVA followed by Tukey's HSD post-hoc analysis. The climbing assay data were arcsine-transformed, and then ANOVA with Tukey's HSD post hoc analyses were conducted to detect any differences in climbing ability between treatments. The Kaplan-Meier estimator and the log-rank test were conducted on the pooled cumulative survival data to determine whether each treatment had any effect on the longevity of individuals using Online Application Survival Analysis Lifespan Assays (http://sbi.postech.ac.kr/oasis) [66].
|
10.1371/journal.pcbi.1004413 | Bulk Genotyping of Biopsies Can Create Spurious Evidence for Hetereogeneity in Mutation Content | When multiple samples are taken from the neoplastic tissues of a single patient, it is natural to compare their mutation content. This is often done by bulk genotyping of whole biopsies, but the chance that a mutation will be detected in bulk genotyping depends on its local frequency in the sample. When the underlying mutation count per cell is equal, homogenous biopsies will have more high-frequency mutations, and thus more detectable mutations, than heterogeneous ones. Using simulations, we show that bulk genotyping of data simulated under a neutral model of somatic evolution generates strong spurious evidence for non-neutrality, because the pattern of tissue growth systematically generates differences in biopsy heterogeneity. Any experiment which compares mutation content across bulk-genotyped biopsies may therefore suggest mutation rate or selection intensity variation even when these forces are absent. We discuss computational and experimental approaches for resolving this problem.
| Researchers who take multiple samples from a cancer or pre-cancer tissue and find that some samples show far more mutations than others are likely to conclude that the high-mutation samples reflect cells with an abnormal mutation or growth rate. We considered the common practice of testing a bulk sample for mutations, which finds only mutations that are common within the sample. Our computer simulations show that even when all cells have identical mutation and growth rates, testing bulk samples frequently leads to spurious detection of rate differences. This can lead to false conclusions about the causes and progress of cancer. We discuss possible solutions involving either genetic testing of single cells or the use of computer algorithms to detect rare mutations within a sample.
| The somatic genotypes of cancerous and pre-cancerous tissues are frequently assayed by taking biopsies containing a substantial number of cells and genotyping each biopsy as a whole (via SNP chip, exome or genome sequencing, or other techniques). For example, in a study of Barrett’s esophagus genotypes were derived from biopsies containing approximately one million epithelial cells (e.g. [1]). We will refer to this type of data collection as bulk-biopsy genotyping. It is generally much less expensive and technically difficult than single-cell genotyping. In this study we examine the use of multiple biopsies from a single tissue or tumor. It is tempting to think that differences in mutation content observed in bulk-biopsy genotyping reflect underlying differences in the number of mutations per cell, which could be informative about the spatial or temporal evolution of the tissue. But is this really true?
The roles of mutations in cancer have often been assessed by analyzing single tumor samples or paired tumor/normal samples from many patients. However, recently it has been recognized that multiple samples from a single patient, separated in time or space, offer additional information. Spatial heterogeneity of clones implies that sampling a single region of a neoplasm may not be representative of the entire neoplasm. The force of natural selection may vary in different parts of a tissue (edge versus center, primary tumor versus metastasis) or over time (early versus late progression, before versus during or after chemotherapy). Multiple samples from a single individual also offer the possibility of phylogenetic analysis to infer relationships among different lineages and reconstruct past events in the history of the tissue. Table 1 shows a sampling of recent studies in which multiple cancer samples per patient were obtained, and phylogenetic methods were either used or could have been used. These studies considered both spatial separation–different parts of a tumor or neoplasm, a tumor and its metastases– and temporal separation– samples taken at different times, such as early and late in progression to cancer, or before and after chemotherapy. They show the potential power of the multiple-sample approach, which we expect will become increasingly important as genotyping costs decrease.
Existing methods do not have the resolution to detect all variants present in a million-cell sample. Variants present in just a few cells will go undetected. Peiffer et al. [17] found minimum frequencies of 33% to 50% for reliable detection of copy-number variants in heterogeneous tumor data using a SNP array. Deep sequencing can detect single nucleotide variants in cancer samples at lower frequencies, down to 1% [18] for 40x sequencing, but the threshhold for reliable detection is much higher: even for an average of 2000x sequencing, single nucleotide variants were reproducibly detected only at >15% allele frequency and indels at >5% allele frequency [19]. Even when low-frequency variants are detected, they are often disregarded as they are difficult to quantify and assign to haplotypes. Thus, the result of bulk-biopsy genotyping is generally a survey of locally high-frequency variants only.
In a well-mixed tissue such as blood, variants which are at high frequency in one sample will generally be at high frequency in all samples. In such tissues, bulk genotyping will miss low-frequency variants, and will thus be biased toward detecting older rather than younger mutations. However, this bias will affect all samples equally and will not tend to produce spurious evidence of non-neutrality.
However, solid tissues are not well-mixed. We will consider the behavior of bulk-biopsy genotyping in a simulated tissue similar to Barrett’s esophagus (BE): a sheet of tissue rolled into a cylinder, with very limited mobility of cell lineages except during initial development. While our simulations are inspired by BE, our conclusions should apply directly to neoplasms in two-dimensional epithelial sheets such as colon, skin, bladder and lung, and conceptually similar effects are also likely in three-dimensional tumors. The key factor is growth with limited mixing.
An increasingly common objective in taking multiple biopsies from a neoplastic tissues is to look for evidence of natural selection or heightened mutation acting on specific clones. This is distinct from standard methods of detecting selection or enhanced mutation via comparison of single samples from many different tumors. A straightforward statistical approach to detecting perturbing forces from multi-sample data would be to infer the evolutionary tree connecting samples from the same individual, and test if that tree conforms to a molecular clock. We simulate this experiment on data which do have a molecular clock, and show that bulk-biopsy genotyping very often leads to the spurious rejection of the clock, and thus to a conclusion of non-neutrality, even when the underlying data are completely neutral.
We emphasize that the bias we observe is not specific to the use of a phylogeny-based molecular clock test, but will influence any formal or informal comparison of apparent mutation content differences among biopsies. For example, if researchers use bulk genotyping to identify a biopsy with an unusually high number of mutations, and conclude that the highly mutant biopsy represents a genetically unstable lineage, they are implicitly assuming that bulk data have a molecular clock in the absence of perturbing forces. As we will show, this is not the case.
We model the BE segment as a 300 x 300 grid of crypts rolled into a cylinder, approximating the size of a typical BE segment. We treat a crypt as the basic replicative unit, since genetic drift is expected to rapidly homogenize the genotype within each crypt. Crypts have a birth rate representing crypt fission. According to Totafurno’s model of the crypt cycle, when stem cells double in number crypt fission is triggered, which results in halving the doubled stem cell population into two new daughter crypts [20]. In this study, we model the crypt fission cycle by allowing a crypt to either eliminate a neighboring crypt or fill in a space lacking crypts. Crypts also have a death rate, a dead crypt leaving an empty space.
The BE segment is thought to expand from the gastro-esophageal junction. An alternative possibility is that cells gain a BE phenotype and spread clonally from squamous duct glands that are situated throughout the esophagus [21]. In both scenarios, BE segments must be rapidly established, since BE segments have not been endoscopically observed in the process of expanding. However, when we simulated the esophagus beginning from a uniform field of non-mutant cells no tree structure arose even after many simulated decades. We do not show results for this case as it is trivially predictable from coalescent theory: our simulations cover approximately 50 generations, and a thoroughly mixed population of size 90,000 will have an average of 1765 distinct ancestors 50 generations ago and will thus appear as approximately 1765 unrelated patches. A less thoroughly mixed sheet of cells will be even patchier. This is not consistent with actual BE data [1] which show mutations shared among biopsies. It is possible that BE arises in situ and is then “overwritten” by an early selective sweep; but if so, this seems little different from the BE segment itself arising by growth from one or a few ancestral crypts. We therefore model the establishment of BE as an expansion from the gastro-esophageal junction.
To simulate BE data we used the agent-based forward simulator of [22]. While this simulator provides for loci whose mutant alleles modify the growth or mutation rates, in the majority of experiments presented here we used a purely neutral model. We simulated 1000 neutral loci for phylogeny inference. Mutations were scored as number of changes from ancestral state; there was no back mutation. We considered neutral mutation rates per locus per crypt per year (μ) of 0.001 and 0.002. Data with the lower rate are fairly sparse, while data with the higher rate are highly polymorphic. We set the probability that a dividing crypt could displace a neighbor at 1. The crypt birth rate was 0.02 and death rate 0.001. We also did simulations with 100 neutral loci and mutation rates of 0.001, 0.002, and 0.004, presented in Supporting Information.
For illustrative purposes we also did a small number of simulations with five potentially selected loci, each having a mutation rate of 10−7 per locus per crypt per year and a twofold selective advantage for the mutant type over the wild type.
The mutation rates given here do not correspond directly to per-cell mutation rates since, when a mutation arises in a crypt, it may be lost rather than fixed. The per-cell mutation rate would be higher by a factor of the mean number of stem cells in the crypt. In any case our mutation rates are chosen in order to give ample mutations for phylogenetic analysis with a limited number of loci. Real data would have fewer mutations per locus, but far more loci. We do not expect this to substantially change the results.
Numerical estimates of BE crypt birth and death rates are not available. Our chosen numbers, which were roughly inspired by values measured for human colon crypts [23], produce an initial spread which is slower than in BE, and a subsequent steady state which probably has faster turnover. However, this is conservative for our conclusions: a faster initial spread and slower subsequent turnover with the same expected amount of mutation would show even greater distortion of the molecular clock. We believe that the details of our parameter choice will not affect our qualitative conclusions as long as the pattern of rapid spread followed by slow turnover is conserved.
We started with a single randomly placed crypt and simulated 20 years of growth. This was generally enough to allow crypts to fill the lower esophagus. A small proportion of simulations resulted in the death of the nascent Barrett’s epithelium; these were discarded. We then randomly chose 10 biopsies which were squares of 10x10 crypts, constrained not to overlap. Rarely, a biopsy was found to contain no live crypts; in such cases the entire simulation was discarded. Additional simulations were run to replace discarded simulations.
These simulation conditions imply a molecular clock, as the mutation rate is the same in all crypts. We tested for presence of a clock in single crypt samples and in biopsies of different sizes using PAUP* 4.0 [24]. For analytic purposes we treated all loci with one or more mutations as one state, and loci with zero mutations as another state. This corresponds to the presence/absence scoring typically used for BE data. To enable use of available phylogenetic software, these states were coded as purine and pyrimidine ambiguity codons. We tested both estimation of the state frequencies from the data using the EMPIRICAL algorithm in PAUP* (results shown in paper), and setting the frequencies equal (results shown in Supporting Information). We performed maximum-likelihood analyses of the recoded data with and without the clock constraint, and assessed the difference in log-likelihoods using a likelihood ratio test [25, 26] with a 5% significance cutoff. When multiple tied trees were produced, we used the first listed tree for analysis.
This use of the likelihood ratio test can be criticized as it assumes that the clocklike and non-clocklike best trees had the same topology [26], which was not always the case. We applied the test to all pairs of trees, even those differing in topology. Our argument is that when the topological difference is trivial (rearrangement across branches of near-zero length) the result of the test will be almost exactly the same as it would for identical topologies; and when the topological difference is non-trivial rejection of the clock is justified even though an exact statistical test is not available.
To measure the influence of biopsy size on detection of rate heterogeneity, we subsampled our biopsies. That is, to produce a 4x4 biopsy we took a 4x4 subsample from the original 10x10 biopsy. To avoid dead crypts, we examined subsamples in turn starting in the upper left and chose the first one in which at least 1 live crypt was found.
To measure the influence of detection threshholds, we used the same sets of simulated biopsies, but varied the cutoff used to establish the biopsy “genotype.” For example, when the cutoff was 30%, we scored a mutation as present if it appeared in 30% or more of the sampled living crypts from the biopsy, and absent otherwise.
In the simulations with 100 loci and μ = 0.001, which had the smallest amount of information per phylogeny, a few cases with large biopsies and stringent cutoffs could not be run. Stringent cutoffs can generate biopsies with no detectable mutations, and having too many such biopsies in a single tree causes failure of the phylogeny analysis. Such runs were discarded. No more than 15/500 runs failed for any combination of conditions; the number of failed runs for each condition are given in the legends to S5 and S6 Tables.
Our simulated data is archived on Dryad at http://dx.doi.org/10.5061/dryad.hf93c.
Our simulations were inspired by Barrett’s esophagus (BE), a neoplastic condition in which the lower esophagus is colonized by a tissue organized into crypts. We treat crypts as the fundamental unit of our simulation, and assume that all spread of genotypes results from reproduction (fission) of crypts which either replace their neighbors or spread into unoccupied areas. The details of the simulator are described in [22].
At the beginning of the simulation each crypt began with an identical genome of 100 or 1000 loci. Mutations in these loci were selectively neutral: they were used solely to infer the relationships among biopsies.
The first striking effect of bulk sampling was seen when the simulation was seeded with a completely filled grid of crypts. At the end of the simulation the tissue consisted of tiny patches of related crypts, each patch unrelated to its neighbors. This reflects the very low gene flow in a static crypt-organized tissue without natural selection. In a tissue of this kind, bulk genotyping would lead to the incorrect conclusion that there are few or no mutations present.
Bulk biopsy sampling of actual BE segments shows abundant mutations [1]. We therefore considered a theory of BE origin in which it spreads from a few crypts. We represented this by seeding the simulation with a single randomly placed crypt. Biopsies sampled from such a tissue did contain genetic variants detectable with bulk genotyping, consistent with actual BE data.
The spatial distribution of mutations in real BE segments is poorly known, as normally only a few biopsies are analyzed per individual. In our simulations we could readily examine the entire pattern, as well as taking simulated biopsies. The simulated BE segments developed a strongly sectored pattern, with small diverse patches of cells near the original seeding area, and larger, more homogeneous patches far from it. Sharp borders between genetically distinct lineages were seen; these borders ran vertically along the simulated esophagus, roughly parallel to the direction of tissue growth. A typical example, captured partway through colonization of the simulated esophagus, is shown in Fig 1.
These patterns reflect the effect of “gene surfing” [27]. Gene surfing is a phenomenon in population genetics, observed when a population is rapidly expanding into a new geographical region but the mobility of individuals is limited. Colonization is therefore driven by a few individuals on the leading edge of the population, and their genotypes will be disproportionately represented in the newly colonized area. Patterns visually similar to our simulations can be seen when two different strains of bacteria are mixed and seeded onto a plate: sectors of pure strains are generated by replication of the few individuals on the colony edge [28] even in the absence of any selective advantage.
Our simulations, seeded with a single crypt, thus produced data that were broadly consistent with observations of actual BE segments. We next asked whether biopsies sampled from these purely neutral simulations would pass tests for neutrality. Based on ten biopsy samples from each of 100 simulated BE segments, we inferred phylogenetic trees and tested whether those trees rejected the molecular clock at the 5% level. We considered biopsies of sizes from 1 (a single crypt) to 10x10 (100 crypts). For biopsies of size greater than 1, we also considered detection cutoffs from 10% (mutations present in 10% or more of crypts were scored) to 100% (only mutations present in all crypts were scored).
If biopsy sampling provided a phylogenetically unbiased sample of mutations occuring in our data, we would expect to see a molecular clock in our inferred trees with any size of biopsy. The proportion of inferences (out of 500) rejecting the molecular clock are shown in Figs 2 and 3 and are presented in table form in S1 and S2 Tables. In these figures, the white color seen at the left-hand edge (single-crypt samples) represents an acceptable clock rejection rate of 5%. (Note that detection cutoff does not affect the results from single crypts, and thus all of the left-hand results represent the same analyses.) All larger biopsy sizes, even 2x2 biopsies with only 4 crypts, rejected the clock at high rates for all conditions studied.
The choice of cutoff had a noticable impact on clock rejection. Cutoffs in the 30%-50% range were better than higher or lower cutoffs; the larger the biopsy, the lower the optimal cutoff. However, no cutoff tested restored the clocklike nature of the underlying data.
Superficially satisfactory results can be obtained by using only 100 neutral loci and inferring frequencies of the mutant and non-mutant states (S3 Fig). However, this apparent improvement merely represents lack of statistical power to detect clock violations, as seen by the dramatic worsening of results with 1000 neutral loci and the same model (Fig 2).
Use of equal frequencies of mutant versus non-mutant states produces higher clock rejection: results are shown for completeness in S1 and S2 Figs for 1000 neutral loci and S4, S5 and S6 Figs for 100 loci.
We show a randomly selected pair of inferred trees from the simulation of S6 Table in Fig 4. The topologies of the two trees differed in ordering of the short bottommost branches. Larger discrepancies were seen in the branch lengths. The single-crypt tree (A) showed some heterogeneity of branch lengths, but it was well within the expected range for a data set of this size, and the clock was not rejected. The 10x10 crypt tree (B) was much more distorted, and rejected the molecular clock. It would be tempting to conclude that biopsy 10, in particular, had a higher mutation rate than biopsy 8; yet they arose from a simulation with perfectly equal rates.
We have presented our results in terms of rejection of the molecular clock in a formal test. However, their significance is not limited to such tests. When we first examined bulk-genotyped BE data for multiple biopsies per patient we saw a striking difference in the mutation content of different biopsies. It was natural to read this as a difference in the underlying mutation rate. After further thought we realized that it could also reflect a difference in the growth rate, since rapidly growing cells will form more uniform samples and therefore appear to have more mutations. Only after performing simulations did we discover that hetereogeneity in apparent mutation content is a general feature of this type of data and should be expected even when neither mutation rate nor growth rate varies. Comparison of Figs 2 and S5 shows that the more informative the data, the stronger this tendency to reject the clock. The clock test formalizes a scientist’s intuition, but both the test and the intuition are liable to error in this case.
We stress that our findings do not challenge the important role non-neutral processes play in the development of cancer. Instead, they warn us against drawing conclusions about non-neutrality that cannot be supported.
Two factors combine to produce this spurious evidence for non-neutrality. Gene surfing causes biopsies taken near the origin of the growing population to be much more heterogeneous than those taken far from the origin (see Fig 1). Bulk-biopsy sampling then translates this difference in diversity into a difference in detectable mutation content: a homogeneous cell sample will have more high-frequency mutations than a heterogeneous one, and bulk sampling detects only high-frequency mutations.
This is most easily understood by considering common ancestry. Consider, as an example, a detection cutoff of 50%. Mutations which reach this cutoff must exist in 50% or more of the cells in the biopsy, and thus, barring convergent evolution, must be inherited from a common ancestor of 50% of the cells. If this common ancestor existed early in the development of the tissue, it likely had relatively few mutations, so few mutations will be shared by its descendants. If it existed more recently, it likely had more mutations (since mutations accumulate over time) and its descendants will have more shared mutations. Cells from a biopsy whose common ancestor is ancient will, individually, have just as many mutations, but a much larger proportion will be at low frequency in the biopsy. Such mutations are difficult to detect with bulk genotyping.
We did not model complicating factors in analysis of bulk data such as differences in ploidy among lineages or typing errors. However, when a bias is present in analysis of clean, error-free genotypic data, there is no reason to believe that better results would come from dirtier data.
Regrettably, no tested frequency cutoff rule was successful in resolving this problem. We suggest three possible approaches.
Once subclones within a biopsy have been detected via approaches (2) or (3), this information needs to be incorporated into the analysis. For analytic methods involving phylogenetics, mixed samples are particularly challenging because when two variants are found at similar frequency in a sample, there is no easy way to determine whether they represent one lineage with two mutations or two lineages with one mutation each [33]. In principle it would be possible for a statistical analysis to sum over these possibilities using an approach analogous to that of [34]. The high computational burden of this approach will have to be compared with the experimental burden of small-sample typing. Alternatively, one could use the minority alleles to estimate biopsy diversity, without attempting to reconstruct minority genotypes.
One potential use of diversity estimations would be as the basis for corrected mutational distances to be used in phylogeny inference. Simulations or heuristics could be used to establish the relationship between observed mutational distance between two biopsies, internal diversity of each biopsy, and the true mutational distance. Distances corrected according to this relationship could then be used in a distance-based phylogeny algorithm to produce trees whose branch lengths more accurately represented the underlying mutational frequencies. We are currently developing such an algorithm.
Accurate inference of branch lengths is important in distinguishing, for example, a mutation rate increase in a specific lineage (presumably due to a mutator mutation or epigenetic change) from a mutation rate increase at a given time across the entire tissue (presumably due to an environmental change, since it manifests in unrelated lineages). Naive tree-drawing based on uncorrected data, as shown by the results in this paper, cannot answer such questions as the branch lengths of its trees are not proportional to time. This is shown dramatically in real BE data, where no separation is seen between data collected from time points many years apart [1]. This situation makes it difficult to draw conclusions about changes in rate over time, though in some cases coherent patterns have been detected [1]. Genetic distances corrected for the bias inherent in bulk-biopsy sampling could allow much more accurate separation of neutral from non-neutral processes in the development of tissues and cancers.
One further positive finding from this study is that the spread of a growing tissue tends to produce a characteristic fan-shaped pattern, as seen in Fig 1. As dense sampling of cancer and pre-cancer tissues becomes more feasible, it will become possible to detect this pattern or deviations from it which may indicate selection. An example is shown in Fig 5, which shows four typical results from a simulation with selected as well as neutral mutations. Note the disruption of the fan pattern by lateral growth of selected clones. The spatial distribution of clones within an expanding tumor or neoplasm may therefore reveal selective processes, as has been explored by [11] in colorectal cancer. In other words, these simulations provide predictions for the nascent field of tumor phylogeography. More work is needed both to detect these patterns and to assess their significance.
|
10.1371/journal.ppat.1004418 | A Critical Role for IL-17RB Signaling in HTLV-1 Tax-Induced NF-κB Activation and T-Cell Transformation | Human T-cell leukemia virus type 1 (HTLV-1) infection is linked to the development of adult T-cell leukemia (ATL) and the neuroinflammatory disease HTLV-1 associated myelopathy/tropical spastic paraparesis (HAM/TSP). The HTLV-1 Tax protein functions as a potent viral oncogene that constitutively activates the NF-κB transcription factor to transform T cells; however, the underlying mechanisms remain obscure. Here, using next-generation RNA sequencing we identified the IL-25 receptor subunit IL-17RB as an aberrantly overexpressed gene in HTLV-1 immortalized T cells. Tax induced the expression of IL-17RB in an IκB kinase (IKK) and NF-κB-dependent manner. Remarkably, Tax activation of the canonical NF-κB pathway in T cells was critically dependent on IL-17RB expression. IL-17RB and IL-25 were required for HTLV-1-induced immortalization of primary T cells, and the constitutive NF-κB activation and survival of HTLV-1 transformed T cells. IL-9 was identified as an important downstream target gene of the IL-17RB pathway that drives the proliferation of HTLV-1 transformed cells. Furthermore, IL-17RB was overexpressed in leukemic cells from a subset of ATL patients and also regulated NF-κB activation in some, but not all, Tax-negative ATL cell lines. Together, our results support a model whereby Tax instigates an IL-17RB-NF-κB feed-forward autocrine loop that is obligatory for HTLV-1 leukemogenesis.
| The retrovirus HTLV-1 is the causative agent of an aggressive lymphoproliferative disorder known as adult T-cell leukemia (ATL). The HTLV-1 Tax regulatory protein constitutively activates the host NF-κB transcription factor to promote T-cell proliferation, survival and cell transformation. However, it remains unknown precisely how Tax persistently activates NF-κB in T cells. In this study, we used next-generation sequencing to identify genes that were differentially expressed upon HTLV-1 infection and immortalization of primary T cells. We found that IL-17RB, the receptor for the IL-25 cytokine, was highly induced in HTLV-1 transformed T cells and was required for NF-κB activation, cell proliferation and survival. Tax induced the expression of IL-17RB and established a positive feedback loop together with IL-25 that triggered persistent NF-κB activation and the upregulation of IL-9 and other genes critical for T-cell proliferation and survival. IL-17RB was also overexpressed in a subset of acute ATL patient specimens and therefore may potentially be targeted by monoclonal antibodies as a novel ATL therapy.
| The retrovirus human T-cell leukemia virus type 1 (HTLV-1) infects between 10–20 million people worldwide [1]. HTLV-1 is the etiological agent of the neuroinflammatory disease HTLV-1-associated myelopathy (HAM/TSP) and adult T-cell leukemia (ATL), a CD4+CD25+ T-cell malignancy [2], [3]. ATL develops in about 5% of HTLV-1-infected individuals after a long latent period spanning 40–60 years [4]. The HTLV-1 genome encodes the Tax protein that exerts pleiotropic roles and is an essential regulator of viral replication and oncogenic cell transformation [5]. Tax modulates the activation of several key signaling pathways and cell cycle proteins to enhance T-cell proliferation and survival. One of the key cellular targets important for transformation by Tax is the NF-κB transcription factor [6].
NF-κB is composed of heterodimeric DNA binding proteins consisting of RelA, c-Rel, RelB, p50 and p52 [7]. In the canonical NF-κB pathway, NF-κB heterodimers are sequestered in the cytoplasm by ankyrin-repeat containing inhibitory proteins including IκBα [8]. A wide variety of stimuli including stress signals, proinflammatory cytokines or virus infection activate the IKK kinase complex consisting of the catalytic subunits IKKα and IKKβ and the regulatory subunit IKKγ (also known as NEMO) [9]. IKKβ phosphorylates IκB proteins to trigger their ubiquitin-dependent degradation thus allowing NF-κB to enter the nucleus and activate target genes [10]. In the noncanonical NF-κB pathway, tumor necrosis factor receptor (TNFR) superfamily members including BAFF, lymphotoxin-β and CD40 promote proteasomal processing of the p100 (NF-κB2) precursor protein to yield p52, which forms transcriptionally active heterodimers with RelB. The NF-κB inducing kinase (NIK) is a key regulator of this pathway by activating IKKα homodimers which in turn phosphorylate p100 leading to its processing. Tax constitutively activates both the canonical and noncanonical NF-κB pathways, in part by interacting directly with NEMO and IKK [11]–[14]. There is evidence that Tax may require upstream signaling molecules such as the kinase TAK1 to activate canonical NF-κB signaling [15]. Although the proximal signaling components of TNFR and interleukin-1 receptor (IL-1R) are dispensable for Tax to activate NF-κB [16], whether Tax has usurped a distinct NF-κB pathway is unknown.
Tax activation of the canonical and noncanonical NF-κB pathways fosters the aberrant expression of anti-apoptotic and pro-proliferative genes that leads to oncogenesis. Tax mutants defective in NF-κB activation expressed in an infectious HTLV-1 molecular clone are impaired in the immortalization of primary T cells [17]. NF-κB is also required for the survival of HTLV-1 transformed cell lines and patient-derived ATL cells [18]. Therefore, HTLV-1 transformed cell lines and primary ATL leukemic specimens exhibit a strict “addiction” to NF-κB for survival and proliferation, thus establishing the NF-κB pathway as an attractive target for novel ATL therapeutics. However, since Tax expressing cells are vigorously targeted by cytotoxic T cells and other arms of the host immune response, the majority of ATL tumors exhibit downregulated or lost Tax expression by mutations within Tax or deletion or methylation of the 5′ viral long terminal repeat region (LTR) [19]. Thus, Tax likely plays more important roles in the early events of transformation via persistent NF-κB activation, inactivation of p53 and other tumor suppressors and induction of genomic instability and aneuploidy [5]. However, canonical and noncanonical NF-κB signaling remains constitutive in ATL despite the loss of Tax.
The interleukin 17 (IL-17) cytokine family consists of six members (IL-17A-F) that play essential roles in host immunity and inflammatory diseases. IL-17A is the signature IL-17 cytokine and binds to an IL-17RA/IL-17RC receptor complex to orchestrate the host response against bacterial and fungal infections [20]. IL-17A controls the expression of cytokines and chemokines that enhance neutrophil recruitment. Dysregulation of this pathway has been implicated in numerous autoimmune and metabolic diseases and cancer [21]. IL-17E (also known as IL-25) is essential for host defense against parasites by regulating expression of T helper 2 (Th2) cytokines including IL-4, IL-5 and IL-13 that promote eosinophil recruitment [22]. IL-25 has also been linked to allergic airway inflammation and asthma [23]. IL-25 is produced by diverse cell types such as epithelial cells, T cells, eosinophils, mast cells and basophils [24], [25]. IL-25 binds to a heterodimeric receptor composed of IL-17RA and IL-17RB, of which IL-17RB is the specific receptor subunit for IL-25 [26]. IL-17RB is highly expressed in kidney, liver and other peripheral organs as well as memory and effector T lymphocytes [27]. IL-17RB expression can be regulated by IL-4 and TGF-β, however the precise transcriptional regulatory control of IL-17RB is unknown. Upon binding to IL-25, IL-17RB recruits the Act1 (also known as CIKS) adaptor molecule via homotypic SEFIR (similar expression to fibroblast growth factor genes and IL-17R) domain interactions [28], [29]. Act1 activates the ubiquitin ligase TRAF6 and the kinase TAK1 that in turn triggers NF-κB and MAP kinase activation to induce type 2 cytokines IL-4, IL-5 and IL-13 as well as IL-9 [30], [31]. IL-17B also serves as a ligand for IL-17RB, albeit with a lower affinity for IL-17RB compared to IL-25 [32]. In addition to IL-17RB regulation of host defense and allergic airway disease, this pathway can be oncogenic if dysregulated. The IL-17RB locus is a common site of retroviral integration in murine myeloid leukemias, resulting in the upregulation of IL-17RB expression [33]. IL-17RB is also overexpressed in a subset of breast tumors and is associated with poor prognosis [34]. In breast cancer, IL-17RB engagement by IL-17B triggers TRAF6 recruitment to IL-17RB, NF-κB activation and induction of the bcl-2 gene to inhibit apoptosis [34].
Although considerable progress has been made in our understanding of HTLV-1 oncogenesis, the precise mechanisms underlying HTLV-1-induced transformation remain unclear. Previous microarray studies have identified several anti-apoptotic, cell cycle and growth regulatory genes dysregulated by HTLV-1 [35]–[37]. However, due to experimental limitations of these studies and the advent of next-generation sequencing, RNA sequencing (RNA-Seq) has emerged as a powerful tool to evaluate gene expression, differential splicing, noncoding RNAs, RNA editing and gene fusions [38]. In this study, we used RNA-Seq to delineate the transcriptome of primary T lymphocytes immortalized by HTLV-1. This work led to the identification of IL-17RB as an aberrantly overexpressed gene in HTLV-1 transformed cells that was induced by the HTLV-1 Tax protein. Surprisingly, the IL-17RB pathway was required for constitutive NF-κB activation by Tax and in HTLV-1 transformed cell lines. Furthermore, IL-17RB was overexpressed in leukemic cells from acute ATL patients and was essential for NF-κB activation in a subset of Tax-negative ATL cell lines.
To gain insight into the mechanisms of HTLV-1-induced T-cell immortalization, we used a well-established co-culture model [35], [39] whereby primary human CD4+ T cells were purified by immunomagnetic beads from normal donor peripheral blood mononuclear cells (PBMCs) and co-cultured with lethally irradiated HTLV-1 transformed MT-2 cells (to provide a source of HTLV-1). Primary T cells were consistently immortalized in the presence of MT-2 cells between 6–8 weeks. Control T cells cultured in the absence of MT-2 did not proliferate after 4 weeks and were no longer viable at that time. The co-culture assay was performed with T cells from 4 independent blood donors. Of the 4 co-cultures, all produced immortalized T cell clones, however clone #1 ceased proliferation unexpectedly and was excluded from further studies. The immortalized T cell clones (T-MT-2) #2-4 remained dependent on recombinant IL-2 for proliferation and expressed CD3, CD4 and CD25 cell surface markers (Figure 1A).
Total RNA was harvested from T-MT-2 clone #2 (week 12 after co-culture) for RNA-Seq analysis as well as parental primary T cells (week 0), and T cells after 1 week of co-culture. A pure population of viable cells was obtained from the co-culture after removal of dead cells using magnetic labeling and separation. MT-2 RNA was also included as a control for RNA-Seq to confirm that the immortalized T cells expressed a unique genetic signature and were not simply MT-2 contaminants. RNA-Seq and bioinformatics analysis were performed with a total number of reads of 65 million (week 0), 73 million (week 1), 44 million (week 12) and 52 million (MT-2). At 1 week after co-culture, the most abundant induced coding RNAs in the T cells were interferon-stimulated genes (ISGs) such as ISG15, IFI27, OAS1 and MX1 and these results were confirmed by real-time quantitative RT-PCR (qRT-PCR) (Figure 1C and Table S1). Conversely, the HTLV-1-immortalized T cells did not express ISGs, but rather expressed aberrant levels of genes regulating cell growth/cytokines (IL-17RB, IL-5, IL-9, IL-13, CADM1), DNA damage (DDIT4L), cell cycle (CDC14B, CCNA1), metabolism (glycerol kinase 2) and migration/chemokines (CCL1, CXCR7) (Table S2). Also, these immortalized T cells had a distinct genetic signature compared to MT-2 cells (Sequence read archive accession numbers SRS698576 and SRS698477). Notably, many of the aberrantly expressed genes, including IL-17RB, have not previously been linked to transformation by HTLV-1. IL-17RB was one of the most highly induced genes in HTLV-1 immortalized T cells (Figures 1B and S1 and Table S2). IL-17RB mRNA expression was sharply elevated in all 3 independent HTLV-1 immortalized T cell clones as shown by qRT-PCR (Figure 1D). IL-25, the high affinity ligand for IL-17RB, was expressed at variable levels in the three clones (Figure 1E). Aberrant expression of CCL1 (also known as I-309), CXCR7, DDIT4L, IL-9 and IL-13 in HTLV-1-immortalized clones was also confirmed by qRT-PCR (Figure 1D and E). The chemokine CCL1, shown previously to be overexpressed in ATL cells, functions in an anti-apoptotic autocrine loop [40]. Similarly, the chemokine receptor CXCR7 is induced by Tax and regulates the growth and survival of ATL cells [41]. Furthermore, Tax induction of both IL-9 and IL-13 may trigger the autocrine stimulation of HTLV-1 infected cells [42], [43]. Taken together, our RNA-Seq studies have confirmed the dysregulation of known targets of HTLV-1 transformation and have also identified genes, such as IL-17RB, not previously demonstrated to be induced by HTLV-1. Next, the cell surface expression of IL-17RB was examined in HTLV-1 transformed cell lines by flow cytometry. IL-17RB was highly expressed in the HTLV-1 immortalized T-cell clones and most HTLV-1-transformed cell lines, but not in Jurkat T cells (Figure 1F). IL-17RB mRNA was also overexpressed in varying degrees in HTLV-1 transformed and ATL cell lines (Figure 1F). Since IL-17RB forms heterodimers with IL-17RA, the expression of IL-17RA was examined in HTLV-1 transformed and ATL cell lines. IL-17RA and IL-25 mRNAs were also upregulated in a subset of HTLV-1 transformed and ATL cell lines (Figure 1G and H).
A previous study reported a role for TGF-β and IL-4 in the upregulation of IL-17RB expression [31]. Since NF-κB is important for the proliferation and survival of HTLV-1 transformed cells, we hypothesized that NF-κB may regulate IL-17RB induction. Thus, the HTLV-1 transformed T-cell lines C8166 and MT-2 were treated with sc-514, a small molecule inhibitor of IKKβ, and qRT-PCR was performed for IL-17RB and the known NF-κB target gene CD25. Treatment with sc-514 significantly diminished the expression of IL-17RB and CD25 mRNAs in these cells (Figure 2A), thus supporting a role for IKKβ and NF-κB in the expression of IL-17RB in HTLV-1 transformed cells. However, sc-514 treatment had no effect on IL-17RB expression in Jurkat cells (Figure 2A). To provide further evidence for a role of IKK in the regulation of IL-17RB, recombinant lentiviruses expressing either control scrambled short hairpin RNA (shRNA) or two distinct shRNAs specific for IKKα or IKKβ were transduced into C8166 cells. Both IKKα and IKKβ shRNAs strongly suppressed their respective mRNAs as shown by qRT-PCR, and these shRNAs significantly inhibited IL-17RB expression (Figure 2B). Therefore, both IKKα and IKKβ regulate IL-17RB expression in C8166 cells.
The HTLV-1 Tax oncoprotein dysregulates the expression of specific cellular genes as part of its oncogenic mechanism [44]. To determine if Tax was involved in the induction of IL-17RB expression we used a Jurkat cell line inducible for Tax expression (Jurkat Tax Tet-On) by doxycycline (Dox) [45]. Jurkat Tax Tet-On cells were treated with Dox for 1, 2 and 3 days and mRNA was harvested for qRT-PCR for IL-17RB. Indeed, induction of Tax strongly upregulated IL-17RB mRNA (Figure 2C). Conversely, shRNA-mediated knockdown of Tax in C8166 cells diminished the expression of IL-17RB mRNA (Figure 2D). Knockdown of Tax also reduced the expression of IL-17RB protein in C8166 cells (Figure 2E). Thus, both gain-of-function and loss-of-function studies support the hypothesis that Tax is the HTLV-1-encoded protein that promotes the aberrant overexpression of IL-17RB.
Two commonly used Tax mutants M22 (Thr130Leu131->Ala130Ser131) and M47 (Leu319Leu320->Arg319Ser320) can be used to distinguish NF-κB or CREB-specific functions of Tax [46]. Tax M22 is defective for NF-κB and wild-type for CREB activation, whereas Tax M47 is defective for CREB and wild-type for NF-κB activation. Wild-type Tax, Tax M22 and Tax M47 were cloned into a lentiviral vector and recombinant Tax-expressing lentiviruses were used to transduce Jurkat T cells. Wild-type Tax and Tax M47 strongly upregulated IL-17RB mRNA expression as detected by qRT-PCR, however Tax M22 induction of IL-17RB was significantly diminished (Figure 2F). These data further support the notion that Tax requires NF-κB to induce IL-17RB expression. Interestingly Tax induction of IL-17RB was not observed in 293 cells suggesting that this event may be specific for T cells (Figure 2G). Tax activation of NF-κB was also independent of IL-17RB in 293 cells, since knockdown of IL-17RB in 293 cells had no effect on Tax activation of an NF-κB reporter (Figure 2H). Finally, the expression of IL-25 was not regulated by Tax in T cells, therefore Tax induces the expression of IL-17RB but not its high affinity ligand (Figure 2I).
Since IL-17RB signals to NF-κB, we next asked if Tax required IL-17RB to trigger NF-κB signaling in T cells. Jurkat Tax Tet-On cells were transduced with lentiviruses expressing control or IL-17RB shRNA, yielding ∼60–70% knockdown efficiency (Figure 3A). The cells were transiently transfected with NF-κB and HTLV-1 LTR reporters for dual-luciferase assays and also treated with Dox to activate Tax expression. Remarkably, Tax activation of NF-κB, but not the HTLV-1 LTR (which is CREB-mediated), was dependent on IL-17RB (Figure 3A). In agreement with these results, Tax induction of the NF-κB target genes, CD25 and cIAP2, was impaired when IL-17RB expression was suppressed with shRNAs (Figure 3B). Therefore, Tax induces IL-17RB expression to establish a positive feedback loop that is critical for Tax-induced NF-κB activation. Also, the requirement of IL-17RB for Tax-mediated NF-κB activation appears to be T-cell specific.
To determine the role of the IL-17RB pathway in the early events of HTLV-1 transformation of primary human T cells, we conducted an in vitro T-cell immortalization assay with irradiated MT-2 cells and PBMCs from normal donors as described earlier. In this co-culture model, expression of both IL-17RB and IL-25 were significantly increased in primary T cells at early times (1–2 weeks) after co-culture with MT-2 cells (Figure 4A). PBMCs were transduced with lentiviruses expressing control shRNA or shRNAs for IL-17RB or IL-25, co-cultured with irradiated MT-2 cells and puromycin was added to select for cells expressing shRNAs. Both IL-17RB and IL-25 were required for immortalization of primary T cells by HTLV-1 since cells expressing these shRNAs ceased to proliferate after 3 weeks of co-culture (Figure 4B). However, as expected PBMCs expressing control shRNA yielded immortalized CD4+ T cells after 8 weeks (Figure 4C). Taken together, these results suggest that both IL-25 and IL-17RB are required for the early events involved in the immortalization of primary T cells by HTLV-1.
Because Tax required IL-17RB for efficient NF-κB activation and NF-κB is critical for the survival of T cells transformed by HTLV-1, we hypothesized that IL-17RB was essential for NF-κB activation and the viability of established HTLV-1 transformed cell lines. To address this notion, recombinant lentiviruses expressing control or IL-17RB shRNAs were transduced into three distinct Tax-expressing HTLV-1 transformed cell lines (C8166, MT-2 and HUT-102). A total of three independent shRNAs to IL-17RB or scrambled control shRNA were expressed in these cell lines and selected with puromycin. Efficient knockdown of IL-17RB was confirmed by qRT-PCR in MT-2 and C8166 cells (Figure 5B). The CellTiter-Glo Luminescent Cell Viability kit was used to quantify cellular ATP levels to determine cell viability and proliferation. Knockdown of IL-17RB significantly reduced the viability and proliferation of HTLV-1 transformed cell lines (Figure 5A). Next, we examined expression of the NF-κB target genes CD25, cIAP2, IRF4 and IL-9 by qRT-PCR. The expression of each of these genes was significantly attenuated upon IL-17RB knockdown in C8166 and MT-2 cells (Figure 5B). Importantly, Tax expression was unaffected by IL-17RB knockdown in these cell lines (Figure 5B).
An NF-κB DNA binding electrophoretic mobility shift assay (EMSA) was next performed with nuclear extracts from C8166, MT-2 and HUT-102 cells expressing control or IL-17RB shRNA. NF-κB DNA binding was completely abrogated upon IL-17RB suppression in C8166 and MT-2 cells, but not HUT-102 likely due to inefficient lentiviral transduction (Figure 5C). IL-25 was also suppressed by shRNAs in MT-2 cells and shRNA#3 was effective in reducing IL-25 expression (Figure 5D). Knockdown of IL-25 with this shRNA also significantly attenuated NF-κB DNA binding and the expression of CD25 (Figure 5C and D). NF-κB signaling can also be monitored with phospho-specific antibodies for IKK and p65 since these proteins are phosphorylated upon activation. Phosphorylation of IKK and p65 was constitutive in C8166, MT-2 and MT-4 cells but reduced upon knockdown of IL-17RB or IL-25 (Figure 5C and E). IκBα protein was increased upon suppression of IL-17RB (Figure 5C), likely reflecting enhanced stability due to a loss of IKK-induced phosphorylation and proteolysis. IL-17RB knockdown also triggered an apoptotic response in HTLV-1 transformed cells as revealed by PARP and caspase 3 cleavage (Figure 5C). The TNFR cell surface receptors CD40 and OX40 activate NF-κB, are strongly induced by Tax and are overexpressed in HTLV-1 transformed cell lines [47], [48]. However, knockdown of either CD40 or OX40 had no effect on the proliferation or viability of C8166 cells (Figure 5F). Thus, IL-17RB is a receptor that appears to be uniquely required for NF-κB signaling and the survival of HTLV-1 transformed cells.
Our earlier results indicated that HTLV-1 immortalized T-cell clones expressed aberrant levels of IL-9 (Figure 1E). A recent study has demonstrated that the IL-17RB pathway controls IL-9 expression in the context of allergic airway inflammation [31]. Furthermore, Tax has been shown to induce IL-9 expression and IL-9 can regulate the proliferation of primary ATL cells [42]. In light of these findings, we hypothesized that IL-9 may represent a key downstream gene of IL-17RB that governs the proliferation of HTLV-1 transformed T cells. First, to determine if IL-17RB regulated a soluble factor that was important for the proliferation of HTLV-1-transformed T cells, C8166 and MT-2 cells were transduced with lentiviruses expressing control or IL-17RB shRNAs and the media was then replaced with conditioned media from the corresponding cells. As expected, suppression of IL-17RB significantly reduced the proliferation of both C8166 and MT-2 cells (Figure 6A). However, the conditioned media rescued the proliferative defects associated with loss of IL-17RB (Figure 6A), suggesting that a soluble factor(s) is sufficient to restore the growth of these cells.
As described above, IL-9 represented an attractive candidate as a pro-proliferative soluble factor in the conditioned media from HTLV-1 transformed T cells. To determine if IL-9 was necessary to restore the proliferation of IL-17RB knockdown cells, we collected conditioned media from cells transduced with control or IL-9 shRNAs for the culture of C8166 cells expressing control or IL-17RB shRNA. Our results revealed that the conditioned media from cells with suppressed IL-9 expression was unable to restore the proliferation of C8166 cells expressing IL-17RB shRNA (Figure 6B). As expected, conditioned media from cells with control shRNA effectively restored C8166 cell growth (Figure 6B). IL-17RB and IL-9 knockdown were confirmed by qRT-PCR (Figure 6B). Next, to determine if IL-9 was sufficient to rescue the growth defect associated with suppressed IL-17RB expression we provided recombinant IL-9 to the media of HTLV-1 transformed cell lines expressing IL-17RB shRNA. The results indicated that provision of IL-9 was sufficient to restore cell proliferation of both C8166 and MT-2 cells (Figure 6C). Therefore, IL-9 is a key cytokine downstream of IL-17RB that governs the proliferation of HTLV-1-transformed T cells.
IL-17RB can form a heterodimeric receptor complex together with IL-17RA [26], and upon binding to IL-25, the active receptor recruits the Act1 adaptor protein [28]. In addition, the ubiquitin ligase TRAF6 can be directly recruited to IL-17RB via a TRAF6 binding motif and plays a critical role in IL-17RB-mediated NF-κB activation and gene expression [30]. Given the vital role of IL-17RB in NF-κB signaling and survival of HTLV-1 transformed T cells, we sought to determine the requirements of the upstream signaling molecules that constitute this pathway. To this end, knockdown experiments were conducted in HTLV-1 transformed cell lines using shRNAs specific for TRAF6, IL-17RA and Act1. Interestingly, knockdown of TRAF6, but not IL-17RA or Act1, attenuated NF-κB activation as determined by western blotting for phosphorylated forms of IKK, p65 and IκBα (Figures 7A, S2C and S3C). Knockdown of TRAF6 also diminished the expression of NF-κB target genes CD25 and cIAP2 as shown by qRT-PCR in HTLV-1 transformed T-cell lines (Figure 7B). Knockdown of IL-17RA, but not Act1, modestly reduced the expression of CD25 and cIAP2 (Figures S2B and S3B). However, the proliferation of HTLV-1 transformed cell lines was strongly dependent on the expression of both IL-17RA and Act1 (Figures S2A and S3A). These results suggest that IL-17RA and Act1 regulate the proliferation of HTLV-1 transformed cells in an NF-κB independent manner. Since IL-17RA regulates chemokine mRNA stability independently of NF-κB [49], this mode of regulation may explain how IL-17RA and Act1 contribute to the proliferation of HTLV-1 transformed T cells.
Thus far our experiments support a model of a Tax-induced IL-17RB-NF-κB feed-forward autocrine loop that is essential for the in vitro immortalization of primary T cells by HTLV-1 and the proliferation and survival of established HTLV-1 transformed T cell lines. However, the majority of ATL tumors (∼60%) have downregulated or lost Tax expression [19], and these malignant cells have acquired mechanisms to activate NF-κB persistently despite loss of Tax expression [50]. We next asked if IL-17RB played a role in NF-κB activation in ATL cells lacking Tax expression. Tax-negative ATL cell lines ATL-43T, ED40515(-), TL-OM1 and MT-1 were transduced with control or IL-17RB shRNA lentiviruses. Interestingly, proliferation and viability were significantly diminished in ATL-43T and TL-OM1, but not in ED40515(-), MT-1 and control Jurkat cells (Figure 8A). Knockdown of IL-17RB was efficient in all cell lines except ED40515(-), likely due to poor lentiviral transduction (Figure 8B). Thus, IL-17RB appears to be important for some, but not all Tax-negative ATL cell lines since MT-1 cells proliferated normally despite knockdown of IL-17RB (Figure 8A and B). The NF-κB target genes CD25 and cIAP2 were suppressed in TL-OM1 and ATL-43T cells, but not in the other ATL cell lines (Figure 8B). Phosphorylation of IKK and p65 was also inhibited by IL-17RB knockdown in ATL-43T and TL-OM1 cells (Figure 8C). The loss of NF-κB activation also triggered an apoptotic response as shown by PARP and caspase 3 cleavage (Figure 8C).
We next examined the expression of IL-17RB, IL-17RA, IL-25, IL-17B and Tax in eight primary acute ATL leukemic specimens. IL-17RB expression was significantly overexpressed in 3 out of 8 ATL samples compared to normal control PBMCs (Figure 8D). IL-17RA was modestly elevated in all the ATL samples compared to controls (Figure 8D). However, IL-25 expression was not detected in any of the ATL samples (Figure 8D). Tax mRNA was found in a subset of the samples but did not correlate with IL-17RB expression (Figure 8D), suggesting that ATL cells may regulate IL-17RB expression independently of Tax. Surprisingly, IL-17B was overexpressed in the majority of acute ATL samples (Figure 8E), thus raising the possibility that IL-17B may serve as a ligand for IL-17RB in acute ATL samples.
The IL-25-IL-17RB pathway has been linked to allergic airway inflammation and host defense against parasites. Our study has established a novel connection of this pathway to HTLV-1-induced leukemogenesis and also reveal that IL-17RB overexpression can be oncogenic in T cells. Tax promotes the aberrant expression of IL-17RB via NF-κB signaling to establish an IL-17RB-NF-κB feed-forward autocrine loop that drives persistent NF-κB activation in T cells, the natural host cell of HTLV-1. Therefore, Tax has hijacked the IL-17RB-NF-κB signaling axis to sustain high levels of NF-κB and coordinate the induction of a gene program consisting of inflammatory cytokines, chemokines and anti-apoptotic proteins that orchestrates pathogenic T-cell proliferation and survival. Together, these results provide a new framework for how Tax and HTLV-1 persistently activate NF-κB to promote the malignant transformation of T cells.
HTLV-1-induced leukemogenesis is a multi-step process that commences with the IL-2-dependent polyclonal expansion of HTLV-1 infected T cells. The Tax oncoprotein is thought to play critical roles in driving T-cell proliferation and survival in the early events of transformation by HTLV-1. However, at later stages Tax expression is largely dispensable, presumably due to genetic and epigenetic changes that may compensate for the loss of Tax. The HTLV-1-encoded HBZ protein may also exert oncogenic roles in ATL tumors in the absence of Tax expression [51]. Nevertheless, after loss of Tax expression, ATL tumors still exhibit constitutive canonical and noncanonical NF-κB signaling that sustains tumor cell proliferation and survival. However, the mechanisms of Tax-independent NF-κB activation in ATL tumors remain poorly understood. A recent study demonstrated that epigenetic downregulation of the microRNA miR-31 led to overexpression of NIK and activation of noncanonical NF-κB [52]. Our results reveal that IL-17RB drives canonical NF-κB signaling in a subset of ATL cell lines suggesting that the IL-17RB-NF-κB autocrine loop can be maintained in the absence of Tax, most likely by the acquisition of genetic and/or epigenetic changes. Comparative genomic hybridization (CGH) analysis has elucidated specific chromosomal imbalances associated with each of the clinical subtypes of ATL [53]. The highly aggressive acute ATL acquires more frequent chromosomal abnormalities, including characteristic gains at chromosomes 3p, 7q and 14q and losses at chromosomes 6q and 13q [54]. Interestingly, IL-17RB is encoded on chromosome 3p21.1, one of the most frequently amplified regions in acute ATL [54]. We found that IL-17RB is significantly overexpressed in leukemic cells from 3/8 ATL patients (38%), comparable to the 37% of aggressive ATL cases with 3p21 gains [54]. Therefore, IL-17RB overexpression in a subset of acute ATL tumors may potentially regulate the constitutive canonical NF-κB activation in the absence of Tax expression (Figure S4). Nevertheless, additional studies with more ATL patient tumor specimens are warranted to further explore the mechanisms underlying IL-17RB overexpression. It will also be interesting to determine if somatic mutations occur in IL-17RB that render the receptor constitutively active in the absence of ligand. Finally, since IL-17B but not IL-25, was expressed by acute ATL leukemic cells (Figure 8), future studies will need to examine if IL-17B plays a role in IL-17RB signaling, NF-κB activation and proliferation of primary ATL cells. IL-25 expression may potentially be suppressed by active mechanisms in ATL since it exerts pro-apoptotic roles in other tumor types [55].
IL-25 serves as the high affinity ligand for IL-17RB. IL-25 favors Th2 immune responses and orchestrates host defense against parasites by inducing the expression of IL-4, IL-5 and IL-13 [56]. IL-25 signals through a heterodimeric receptor containing IL-17RA/IL-17RB, which in turn recruits Act1 and TRAF6 upon IL-25 stimulation to induce NF-κB and MAPK activation that regulate genes important for Th2 immunity, allergic responses and expulsion of helminths. Our results indicate that HTLV-1 transformed cells are critically dependent on the IL-17RB pathway for proliferation, however only IL-17RB and TRAF6 are essential for NF-κB activation. IL-17RB contains a TRAF6 interaction motif in its intracellular domain that propagates downstream NF-κB activation [30]. Furthermore, a previous study has shown that TAK1, a kinase downstream of TRAF6 in the IL-17RB pathway, is also involved in Tax-mediated NF-κB activation [15]. Therefore, IL-17RB may signal through both TRAF6 and TAK1 to activate IKK in HTLV-1 transformed cells. We have recently identified a consensus TRAF6 interaction motif in the C-terminal region of Tax that mediates TRAF6 interaction and activation [57], thus suggesting that Tax may activate TRAF6 to further enhance the Tax-IL-17RB-NF-κB positive feedback loop in T cells. A previous study claimed that TRAF6 was dispensable for Tax-induced NF-κB activation [58], however they used a cell-free assay system using lysates from murine embryonic fibroblasts. Using intact T cells, we found that TRAF6 indeed plays a role in Tax-mediated NF-κB signaling. Our results also indicate that IL-17RB is dispensable for Tax to activate NF-κB in 293 cells, yet is critical for Tax-mediated NF-κB activation in T cells. Therefore, Tax activation of NF-κB appears to be distinct in T cells compared to other cell types and provides a strong rationale for Tax/NF-κB studies to be conducted in T cells.
Our data has provided new insight into the transcriptional regulation of IL-17RB. Little is known regarding how IL-17RB expression is regulated, although a previous study demonstrated that TGF-β and/or IL-4 can induce IL-17RB expression in mouse T cells [31]. We have provided multiple lines of evidence supporting a role for NF-κB in Tax-induced expression of IL-17RB. First, an IKKβ inhibitor greatly reduced the expression of IL-17RB in HTLV-1 transformed T cell lines (Figure 2A). Second, knockdown of IKKα or IKKβ with shRNAs diminished IL-17RB expression in C8166 cells (Figure 2B). Finally, the Tax M22 mutant, defective for NEMO binding and NF-κB activation, was impaired in the induction of IL-17RB (Figure 2F). Taken together, our data support a two-step model of Tax activation of NF-κB in T cells (Figure S4). First, Tax activation of canonical NF-κB commences through direct NEMO/IKK binding and IKK activation. The precise mechanisms remain poorly understood but may involve IKK oligomerization and inhibition of NEMO-associated phosphatase 2A [59], [60]. Next, Tax and IKK-induced IL-17RB overexpression (and engagement by IL-25) triggers downstream signaling to TRAF6 and further activates IKK to establish a positive feedback loop resulting in strong and sustained NF-κB signaling. It remains unclear whether NF-κB directly regulates the expression of IL-17RB, although we have identified a putative NF-κB site (GGGAATTTCC) ∼3380 base pairs upstream of the human IL-17RB transcriptional start site. Future studies will be necessary to identify important regulatory elements in the IL-17RB promoter.
IL-17RB forms heterodimers with IL-17RA, and although IL-17RA does not directly engage IL-25 it appears to be essential for IL-17RB signaling in untransformed cells [26]. Since IL-17RA and Act1 were largely dispensable for NF-κB activation in HTLV-1 transformed cells (Figures S2 and S3), it is plausible that IL-17RA and Act1 regulate the proliferation of these cells by stabilizing chemokine mRNAs [49]. Because IL-17RB is overexpressed to a much greater degree than IL-17RA in HTLV-1 transformed T cells, it is likely that IL-17RB homodimers constitute the most abundant IL-17R complex that signals to NF-κB in these cells. Further studies are needed to examine the stoichiometry of IL-17R complexes and downstream signaling requirements in HTLV-1 transformed cells.
Although IL-25/IL-17RB signaling has been previously linked to the induction of IL-9 and Th2 cytokines [24], our study has identified additional genes regulated by this pathway that contribute to oncogenesis. We found that knockdown of IL-17RB in HTLV-1 transformed cell lines diminished the expression of cytokines (IL-9), cytokine receptors (CD25), anti-apoptotic genes (cIAP2) and transcription factors (IRF4). Elevated expression of IRF4 in ATL tumors was shown to correlate with resistance to antiviral therapy with zidovudine (AZT) and interferon alpha [61]. Furthermore, cIAP2 was identified as a Tax regulated anti-apoptotic gene that was required for the survival of HTLV-1 transformed T cells [62]. IL-9 was also demonstrated to function as a key proliferative factor for ATL cells [42]. Consistently, we found that IL-9 was both necessary and sufficient to restore the cell proliferation of HTLV-1 transformed T cells with IL-17RB knockdown (Figure 6B and C). These data support the notion that IL-9 is an important downstream target gene of IL-17RB that drives the proliferation of HTLV-1 transformed cells. Additional studies will be necessary to identify the full spectrum of genes regulated by IL-17RB in HTLV-1 transformed T cells that support oncogenic proliferation.
Therapeutic blocking antibodies, such as those targeting HER2 and EGFR, have emerged as an important new treatment option in the clinic for carcinomas of the breast, lung and colon [63], [64]. IL-17RB is overexpressed in a subset of breast tumors and is associated with poor prognosis [34]. Treatment with blocking IL-17RB therapeutic antibodies attenuated the tumorigenicity of breast cancer cells [34]. Given that IL-17RB overexpression can promote oncogenic NF-κB signaling in Tax-negative ATL tumors, this receptor may represent an attractive therapeutic target for ATL. IL-17RB may potentially serve as a biomarker to stratify ATL patients that could benefit from IL-17RB inhibition. Preclinical studies with IL-17RB (or potentially IL-17B) monoclonal blocking antibodies in both in vitro and in vivo ATL models will be required to establish the feasibility of this potential targeted therapy.
Blood from healthy donors was purchased from Biological Specialty Corporation (Colmar, PA). PBMCs were collected from acute ATL patients (n = 8). This study was conducted according to the principles expressed in the Declaration of Helsinki. The study was approved by the Institutional Review Board of Kyoto University (G204). All patients provided written informed consent for the collection of samples and subsequent analysis.
Human embryonic kidney cells (HEK 293T) and Jurkat T cells were purchased from ATCC. The HTLV-1-transformed cell lines MT-2, HUT-102, C8166 and MT-4 were described previously [65], [66]. ED40515(-), MT-1, and TL-OM1 cells are clones of leukemic cells derived from ATL patients, kindly provided by Dr. Michiyuki Maeda (Kyoto University). ATL43T is a Tax-negative ATL cell line that was previously described [67]. Jurkat Tax Tet-On cells were kindly provided by Dr. Warner Greene [45]. 293T cells were cultured in Dulbecco's Modified Eagle's medium (DMEM); Jurkat, MT-2, C8166, MT-4, ATL-43T, HUT-102, ED40515(-), MT-1 and TL-OM1 cells were cultured in RPMI medium. Media was supplemented with fetal bovine serum (FBS; 10%) and penicillin-streptomycin (1×). MISSION shRNAs targeting human IL-17RB, IL-17RA, IL-25, IL-9, Act1, CD40, OX40 and control scrambled shRNA were purchased from Sigma. TRAF6, Tax, IKKα and IKKβ shRNAs were cloned into pYNC352/puro. Target sequences for these shRNAs are listed in Table S3. Tax WT, M22 and M47 were cloned in the pDUET lentiviral vector. Expression vectors encoding κB Luciferase (Luc), pU3R-Luc, pRL-TK (thymidine kinase) have all been described previously [68]. Recombinant human IL-9 was purchased from R&D Systems. The IKKβ inhibitor SC-514 was from EMD Millipore.
The following antibodies were used in this study: anti-hIL-17RB (FAB1207P; R&D Systems), anti-hCD4 (555346; BD Pharmingen), anti-hCD3 (552851; BD Pharmingen), anti-hCD8 (555366; BD Pharmingen), anti-hCD25 (560989; BD Pharmingen), anti-β-actin (AC15; Abcam), anti-IκBα (SC-371; Santa Cruz Biotechnology), anti-phospho-IκBα (14D4; Cell Signaling), anti-p65 (8242S; Cell Signaling), anti-phospho-p65 (3031S; Cell Signaling), anti-IKKβ (2678; Cell Signaling), anti-phospho-IKKα/β (2697S; Cell Signaling), anti-IL-17RB (SC-52925; Santa Cruz Biotechnology), anti-TRAF6 (SC-7221; Santa Cruz Biotechnology), anti-PARP (9542S; Cell Signaling) and anti-caspase-3 (SC-7148; Santa Cruz Biotechnology).
Human PBMCs from healthy donors were prepared from lymphocyte enriched human blood with a Ficoll-Hypaque gradient (Pharmacia Biotech). Samples were tested and found to be negative for hepatitis B virus (HBV), hepatitis C virus (HCV) and human immunodeficiency virus 1 (HIV-1). The cells were stimulated for 36 h with phytohemagglutinin (PHA, 2 µg/ml) and then cultured in RPMI medium supplemented with 20% FBS, 2 mM L-glutamine, penicillin-streptomycin, and 25 units/ml of human recombinant IL-2 (Biological Resources Branch, NCI). Under these conditions, PBMCs continuously grew for up to 4 weeks in the presence of exogenous IL-2. CD4+ T cells were isolated from PBMCs by negative selection using MACS MS Columns (Miltenyi Biotec). The purity of the cells was confirmed by flow cytometry and was>95%.
In vitro transformation of T cells with HTLV-I was performed as previously described [39]. Briefly, PHA-stimulated PBMCs were co-cultured with lethally γ-irradiated (50 Grays (Gy)) HTLV-1 donor cells (MT-2) in IL-2-containing RPMI medium. As expected, the virus-infected T cells became immortalized after about 6 weeks of co-cultivation. These cells proliferated vigorously when exogenous IL-2 was provided, a characteristic of T cells at an early stage of HTLV-1 infection. Under identical culture conditions, the uninfected control T cells or PBMCs typically ceased growth within 4 weeks, and the γ-irradiated MT-2 cells did not proliferate. The HTLV-1-immortalized T cells were maintained in RPMI medium supplemented with IL-2 and used as a bulk population. For shRNA knockdown studies, purified PHA-stimulated PBMCs were first infected with lentiviral particles expressing shRNAs to knockdown the indicated genes and subsequently co-cultured with lethally γ-irradiated (50 Gy) HTLV-1 donor cells (MT-2) in IL-2-containing RPMI medium. Puromycin was added after 3 weeks of co-culture to select for shRNA expressing cells.
Total RNA was prepared from parental primary T cells, HTLV-1-infected cells after 1 week of co-culture, HTLV-1 immortalized T cell clones after 12 weeks of co-culture or MT-2 cells. Dead cells were removed from co-cultured cells after magnetic labeling and separation using the Dead Cell Removal Kit (Miltenyi Biotec). RNA was isolated with RNeasy columns (Qiagen). RNA-Seq and bioinformatics were conducted by the Johns Hopkins Sidney Kimmel Cancer Center next-generation sequencing core. Sequencing analysis was performed by aligning the paired end reads to hg19 using Bioscope. The differential expression analysis was performed using the DEseq R package and the GO enrichment was done with the topGO R package.
Jurkat cells were transfected with TransIT-Jurkat (Mirus) according to the manufacturer's instructions. For lentivirus production, HEK293T cells were transfected with a lentiviral vector and gag/pol-encoding plasmids using GenJet (SignaGen) according to the manufacturer's instructions. Virus was harvested after 48 h by centrifugation at 49,000× g. Cells were transduced with lentivirus by the spinoculation protocol, cultured for 48 h and then selected with puromycin. For luciferase assays, cells were lysed 24 h after transfection using passive lysis buffer (Promega). Luciferase activity was measured with the dual-luciferase assay system according to the manufacturer's instructions (Promega). Firefly luciferase values were normalized based on the Renilla luciferase internal control values. Luciferase values are presented as “fold induction” relative to the shControl (shCTR).
Western blotting was performed essentially as described previously [69]. Whole cell lysates were resolved by SDS-PAGE, transferred to nitrocellulose membranes, blocked in 5% milk or bovine serum albumin (BSA) (for phospho-specific antibodies), incubated with the indicated primary and secondary antibodies, and detected using Western Lightning enhanced chemiluminescence reagent (Perkin Elmer).
Quantitative real-time PCR (qRT-PCR) was performed as described previously [68]. Total RNA was isolated from cells using the RNeasy mini kit (Qiagen). RNA was converted to cDNA using the First Strand cDNA synthesis kit for RT-PCR (avian myeloblastosis virus [AMV]; Roche). Real-time PCR was performed using SYBR Green qPCR (Sigma). Gene expression was normalized to the internal control 18S rRNA. PCR primers are listed in Table S4.
Cell viability and proliferation assay was determined using the CellTiter-Glo Luminescent Cell Viability Assay (Promega). Cells were cultured in 96-well plates and the ATP content was quantified as an indicator of metabolically active cells.
Small-scale nuclear extracts were prepared from cells as described previously [47]. The following sequence was used to generate double-stranded oligonucleotides for electrophoretic mobility shift assays (EMSA): IL-2Rα NF-κB site: 5′-CAACGGCAGGGGAATCTCCCTCTCCTT. Nonradioactive EMSA was performed using LightShift Chemiluminescent EMSA Kit (Thermo Scientific) according to the manufacturer's instructions.
Two-tailed unpaired T test was performed with Prism software. Error bars represent the standard deviation of triplicate samples. The level of significance was defined as: ***P<0.001, **P<0.01, *P<0.05.
|
10.1371/journal.pntd.0004255 | Cissampelos pareira Linn: Natural Source of Potent Antiviral Activity against All Four Dengue Virus Serotypes | Dengue, a mosquito-borne viral disease, poses a significant global public health risk. In tropical countries such as India where periodic dengue outbreaks can be correlated to the high prevalence of the mosquito vector, circulation of all four dengue viruses (DENVs) and the high population density, a drug for dengue is being increasingly recognized as an unmet public health need.
Using the knowledge of traditional Indian medicine, Ayurveda, we developed a systematic bioassay-guided screening approach to explore the indigenous herbal bio-resource to identify plants with pan-DENV inhibitory activity. Our results show that the alcoholic extract of Cissampelos pariera Linn (Cipa extract) was a potent inhibitor of all four DENVs in cell-based assays, assessed in terms of viral NS1 antigen secretion using ELISA, as well as viral replication, based on plaque assays. Virus yield reduction assays showed that Cipa extract could decrease viral titers by an order of magnitude. The extract conferred statistically significant protection against DENV infection using the AG129 mouse model. A preliminary evaluation of the clinical relevance of Cipa extract showed that it had no adverse effects on platelet counts and RBC viability. In addition to inherent antipyretic activity in Wistar rats, it possessed the ability to down-regulate the production of TNF-α, a cytokine implicated in severe dengue disease. Importantly, it showed no evidence of toxicity in Wistar rats, when administered at doses as high as 2g/Kg body weight for up to 1 week.
Our findings above, taken in the context of the human safety of Cipa, based on its use in Indian traditional medicine, warrant further work to explore Cipa as a source for the development of an inexpensive herbal formulation for dengue therapy. This may be of practical relevance to a dengue-endemic resource-poor country such as India.
| India represents ~50% of the global population estimated to be at risk of dengue. Severe dengue, which is potentially fatal, correlates with very high virus load, reduction in platelet counts and haemorrhage. Antiviral therapy to reduce high virus load may be beneficial in attenuating disease severity. We have explored Indian traditional medicine, Ayurveda, to identify plants that could be a source of dengue inhibitory activity. We show that an alcoholic extract prepared from Cissampelos pareira Linn inhibited the replication of dengue viruses in living cells in culture and protected mice against dengue infection. It also showed antipyretic and anti-inflammatory effects. Importantly, this extract did not show any toxic effects in rats and did not affect platelets and red blood cells. This observation is critical as dengue fever is commonly treated with antipyretics. In a dengue-endemic resource-poor country as India, the C. pareira plant may serve as a source for the development of an inexpensive herbal formulation against dengue.
| Dengue disease is a major public health concern around the world. It is spread to humans through the bite of Aedes mosquitoes which serve as carriers of the disease-causing viruses. There are four serotypes of dengue viruses (DENV-1, -2, -3 and -4), belonging to the family Flaviviridae [1], which are prevalent in more than a hundred mosquito-infested countries in the tropical and sub-tropical regions of the globe [2, 3]. According to recent estimates, there are annually ~400 million infections around the world and a fourth of these are associated with symptomatic dengue illness [3]. Symptomatic dengue illness can range from mild dengue fever (DF) to severe and potentially fatal dengue haemorrhagic fever (DHF) and dengue shock syndrome (DSS) [4]. In DF, the appearance of virus in the serum coincides with fever onset, with titers reaching 106 infectious units/ml, at the peak of the febrile phase. In DHF/DSS patients, the viremia can escalate by 1–2 logs [5, 6]. This is accompanied by potentially fatal manifestations of thrombocytopenia, haemorrhage, vascular leakage and shock [1, 2, 7]. Unless hospitalized and given supportive care and fluid replacement, DHF/DSS can be associated with high case fatality rates [4]. Controlling the spread of dengue continues to be an intractable problem due to the inability to eradicate the vector mosquitoes and the lack of safe, potent preventive vaccine [8, 9]. Despite numerous hurdles, persistent efforts over the years have resulted in the development of several live attenuated dengue vaccine candidates [9], one of which has undergone phase III clinical testing [10–12] and is expected to be licensed in the near future. The challenges being faced in dengue vaccine development [13] have emphasized the need for antiviral drugs and spurred new efforts in this direction [14–16].
In regard to drugs, a majority used in treating human illness [17] and several in clinical development [18] can be traced to natural origins. India has a rich herbal repertoire which is used in traditional ethnomedicine. The current work was undertaken with the objective of exploring the possibility of identifying anti-DENV activity that may be associated with indigenous plants in India. To choose plants likely to provide useful leads, the study utilized knowledge from traditional Indian medicine, Ayurveda, a specific branch of Indian ethnomedicine [19]. Ayurveda prescribes medicines based on various herbs for a variety of illnesses. Though dengue as such is not described in Ayurvedic literature, there are illnesses identified by symptoms that can be correlated with some of the clinical manifestations of dengue disease. Plants shortlisted on this basis from the vast indigenous herbal bio-resource available in India were screened for anti-DENV activity. This work describes the screening assays that were set-up to identify anti-DENV activity and presents data identifying two plants that manifested inhibitory potency against all 4 DENVs. One of these, chosen for further study to assess its in vivo efficacy and clinical relevance, was found to be a promising candidate for further development as a pan-DENV inhibitory formulation.
The mosquito cell line C6/36, the monkey kidney cell lines LLCMK2 and Vero, and human hepatoma cell line HepG2 were from American Type Culture Collection, Virginia, U.S.A. C6/36 cells were maintained in Leibovitz L-15 medium supplemented with 0.03% tryptose phosphate broth and 10% heat-inactivated fetal calf serum (ΔFCS) in a CO2-free incubator at 28°C. The monkey kidney cells were maintained in Dulbecco’s Modified Eagle medium (DMEM), supplemented with 10% ΔFCS, in a 5% CO2 humidified incubator, at 37°C. For HepG2 cells, DMEM was replaced by Roswell Park Memorial Institute (RPMI) medium. The remaining conditions were the same. Representatives of the four DENV serotypes (strain and accession numbers are indicated in parentheses) used in this study were: DENV-1 (Nauru Island, U88535); DENV-2 (New Guinea C, AF038403), DENV-3 (H87, M93130) and DENV-4 (Dominica, M14931). These were propagated in C6/36 cells and titrated on LLCMK2 cells using a standard plaque assay as described below.
Plants shortlisted for screening are indicated in Table 1. These were procured through local suppliers and identified and authenticated by the resident herbal taxonomist at Ranbaxy Research Laboratories, Gurgaon, India. Extracts were prepared using part(s) of the plants used in conventional Ayurvedic preparations. In each case, three different extracts, methanolic, hydroalcoholic and aqueous, were prepared using methanol, methanol: water (1:1), and water, respectively. Plant parts were washed, shade-dried and pulverized prior to extraction. To prepare extracts, ~100 g plant material was extracted three times with 500 ml extraction solvent at boiling point for 3 hours, each time. The three extractions were pooled, filtered and concentrated at low pressure and temperature, and dried in a vacuum oven at room temperature (RT) for 16–18 hours. The resultant material was dissolved in dimethylsulfoxide (DMSO), at 20mg/ml and stored at 4°C.
The titer of infectious DENV in virus stocks and in culture supernatants (in the type-3 screening assay, see below) were determined using a standard plaque assay as described previously [20]. Briefly, LLCMK2 monolayers in 6 well plates were infected in duplicate with serial 10-fold dilutions (prepared in DMEM+2%ΔFCS) of the virus-containing samples (250μl/well). Mock-infections were performed in parallel using an equivalent volume of virus diluent alone. Two hours later, the infected monolayers (after aspirating off the virus inoculum) were overlaid with DMEM+6%ΔFCS containing 1% methyl cellulose (2 ml/well), and incubated for 6 days (37°C, 5% CO2). On day 6 post-infection, the overlay was removed and the cells fixed with 4% formaldehyde solution (1 ml/well). Fixed cells were washed and stained with 0.05% (w/v) crystal violet solution in 20% ethanol. Revealed plaques were counted to determine the virus titre, expressed as plaque-forming units (PFUs)/ml.
The initial antiviral screening assay, designated as the type-1 assay, was based on the plaque reduction neutralization test (PRNT) described earlier for determining DENV-neutralizing antibody titres in sera [20]. LLCMK2 cells were seeded in 24-well plates (5x105 cells/well), a day in advance. DENV-1, -2, -3 and -4 (100 PFU each) were separately pre-incubated with serial dilutions of herbal extracts (corresponding to 0–100 μg/ml final concentration) in 300 μl volume, at 4°C overnight. The pre-incubation mixture was diluted with an equal volume of medium (DMEM+ 2% ΔFCS) and used to infect LLCMK2 cells (3 wells for each concentration at 200 μl/well) in the 24-well plate. After 2 hours of adsorption in the incubator (37°C, 5% CO2), infected cells were overlaid with methylcellulose-containing growth medium and processed thereafter as described for the plaque assay above. To assess any potential cytotoxicity, cells were exposed to the herbal extracts (in the same concentration range) in the absence of DENV infection. Additional control experiments, run in parallel, included cells which were either mock-infected (negative control) or infected with DENV in the absence of herbal extract (positive control). The half-maximal inhibitory concentration (IC50 value) for each herbal extract against each DENV serotype, with reference to the positive control which represented 100% infection (or 0% inhibition), was defined as the concentration of herbal extract, in μg/ml, resulting in 50% inhibition of the plaque count.
In the type-2 screening assay, LLCMK2 cells in 24-well plates were infected with DENVs (moi = 0.002) without pre-incubating with the herbal extracts. After 2 hours of adsorption, the virus inoculum was aspirated, the monolayer rinsed, and then fed with complete medium containing the herbal extracts (corresponding to 0–200 μg/ml final concentration). After 24 hours of exposure to the extract, the monolayer was aspirated and overlaid with growth medium containing methyl cellulose and plaques developed 6 days later as above.
The type-3 assay was done using Vero cells, as these cells secrete the viral antigen NS1 and the infectious virions efficiently into the culture medium upon DENV infection. The assay design was similar to the type-2 assay, except that following sequential exposure of cells to DENV and herbal extract, cells were fed with liquid growth medium, instead of the methylcellulose overlay. Aliquots of the culture supernatant were withdrawn at periodic intervals up to 9 days for estimation of NS1 antigen levels (using a commercial ELISA kit) and virus titres (by plaque assay, as described above).
Cytotoxicity was evaluated in two cell lines, LLCMK2 (in which the antiviral activity of the extracts were assayed) and HepG2, a commonly used liver cell surrogate for in vitro cytotoxicity testing. Cells seeded in 96-well plates were exposed to a wide concentration range of Cipa extract (1–200 μg/ml) for 3 days. Cell viability was assessed using a commercial MTT (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide) assay kit (Sigma, cat. # M5655) with reference to control cells that were not exposed to the extract. The half-maximal cytotoxic concentration (CC50 value) for the herbal extract, with reference to the positive control (untreated cells) which represented 100% cell viability (or 0% cytotoxicity), was defined as the concentration of herbal extract, in μg/ml, resulting in 50% cytotoxicity. Selectivity index (SI) of an extract is defined as the ratio of CC50 to IC50 values obtained using the LLCMK2 cell line.
To assess if the duration of pre-incubation of the Cipa extract with DENV influenced antiviral activity in the type-1 assay format, pre-incubation times ranging from 0–24 hours were tested using ~50 PFUs of DENV-3.
To assess the effect of the size of the DENV dose on the anti-DENV efficacy of Cipa extract in the pre-incubation step (at 4°C, overnight), type-1 assays were performed using DENV-3 ranging from 50 to 5000 PFUs. Each dose of DENV-3 was assayed against Cipa extract ranging in concentration from 0–200 μg/ml.
To prepare a stock of challenge virus capable of lethal infection in AG129, we used a previously published strategy [21, 22] with some modifications. Essentially DENV-2 (NGC) was alternately passaged between AG129 (intracranial inoculation of 106 PFU) and C6/36 cells in tissue culture. After 4–5 such cycles of passaging, the virus was tested in AG129 mice to determine the minimum lethal dose (MLD) by i.p. injections. MLD is defined as the dose that can cause clinical symptoms and 90–100% death 3–4 weeks post challenge. The challenge virus stock thus obtained was titrated, aliquotted and stored in liquid N2 until use. To test protective efficacy of Cipa extract, AG129 mice (9–12 weeks old, 20–24 g body weight) were challenged with 106 PFU (per mouse, 0.4 ml, i.p) of the challenge DENV-2 stock described above. Challenged mice were divided into groups (n = 6) and treated orally with vehicle alone (0.25% methyl cellulose) or with two different doses of Cipa extract (at 125 mg and 250 mg/kg body weight). The methanol in the Cipa extracts administered to mice was removed completely by evaporation. The resultant methanol-free Cipa paste was thoroughly re-suspended in 0.25% methyl cellulose water and administered orally to infected mice. The volume of the oral dose was adjusted in accordance with the body weight of each animal (10 ml/Kg/dose) and administered by a trained veterinarian using a specially designed mouse feeder needle fitted with a graduated 1 ml disposable syringe. The treatment was initiated 2 hours post-infection and continued twice daily for 5 consecutive days. Animals were monitored twice daily for a period of 35 days for clinical symptoms and mortality. A control (sham) group that was not virus-challenged, but which received Cipa extract (250 mg/kg), was also tested in parallel. At the end of the experiment, the survival data was used to plot Kaplan Meier survival curves and analysed by the log rank test (Mantel-Cox) test for statistical significance using GraphPad Prism 5 software.
Interaction between paracetamol and Cipa extract was assessed in vitro using type-1 assay format as follows. DENV-3 (~50 PFUs) and Cipa extract (ranging in concentration from 0–50 μg/ml) were pre-incubated overnight at 4°C in a volume of 100 μl, and used to infect LLCMK2 cells in 24-well plates. Parallel infections were set up using pre-incubation mixtures containing paracetamol (1–100 μg/ml), in addition to DENV and Cipa extract. As before, mock-infections and DENV only infections (in the absence of Cipa and paracetamol) were also set up and analysed in parallel.
The in vivo effect of Cipa extract in the presence and absence of paracetamol was assessed using the Wistar rat pyrexia model. Wistar rats (weighing 180–220g) of either sex were used. Basal temperature of the rats was measured using a digital rectal thermometer (Experimetria Ltd., Hungary) and then injected subcutaneously (in the intra-scapular region) with 20% brewer’s yeast (10 ml/kg body weight) and allowed to fast overnight with free access to water. At 18 hours post-injection, rectal temperatures were recorded again to identify animals that registered ≥0.7°C rise in body temperature for inclusion in the study. Groups (n = 7–9) of febrile rats were orally administered paracetamol (200 mg/kg), or Cipa extract (200 mg/kg) or both. Rats in the control group received just the vehicle (0.5% methyl cellulose). This was followed by recording of rectal temperature for 3 hours at 30 minute intervals.
For ex vivo studies, human blood was collected from healthy adult donors after informed consent. Erythrocytes were pelleted down in a centrifuge (1500xg, 5 minutes) from freshly collected heparinized blood, rinsed thoroughly with PBS (pH 7.4), and used to make a 1% cell suspension in PBS. Cipa extract ranging in concentration from 12.5 to 400 mg/L was added to the erythrocyte suspension and incubated at 37°C for 1 hour. After this, the samples were spun down, and the absorbance of the supernatant measured at 576 nm to determine the extent of erythrocyte lysis. Controls wherein erythrocytes were incubated with buffer alone (0% lysis), DMSO alone (Cipa solvent) and 0.1% Triton X-100 (100% lysis) were processed in parallel. Basal platelet count in freshly collected heparinized blood and in blood pre-incubated with DMSO (vehicle) or Cipa extract (2–10 μg/ml) for different durations (1–4 hours) was determined using a Beckman Coulter hematology analyser.
For in vivo studies, four groups (n = 5) of Wistar rats were fasted overnight and administered orally with vehicle (0.25% methyl cellulose) or Cipa extract, at three different dosages (100, 300 and 1000 mg/kg body weight). Blood was collected just before Cipa extract administration (0 hour) and at 1 and 4 hours post-administration. Hematology parameters were measured using ADVIA-120 hematology analyser.
Human peripheral blood mononuclear cells (PBMCs) were obtained as follows. Freshly collected heparinized blood was diluted with an equal amount of RPMI 1640 medium and layered on Ficoll Hypaque 1077 and centrifuged at 2,500 rpm for 25 minutes at RT. The upper layer was discarded and the fluffy layer at the interphase was harvested, rinsed and re-suspended in RPMI 1640 at 5x105 cells/ml. Freshly collected PBMCs were seeded in 96-well plates (105 cells/well) and treated with Cipa extract at different dilutions (in RPMI 1640), followed by 30 minutes incubation at RT on a rotary shaker (200 rpm). Next, wells were treated with 50μl (4μg/ml) lipopolysaccharide (Sigma Cat. # L2654) and allowed to incubate for a further 30 minutes at RT. The volume per well was made up to 200 μl using RPMI+10%FCS and the plates incubated overnight at 37°C in a CO2 incubator. Negative controls (no lipopolysaccharide treatment) were run in parallel. The plates were centrifuged (3000 rpm, 10 minutes) to obtain clarified supernatants for TNF- α (e-Bioscience, cat # BMS223/4TEN,) and IL-1β (BD Bioscience, cat # 557953) determinations using commercial ELISA kits as per the vendors’ instructions.
Groups of 5 adult Wistar rats were orally administered 4 ml 0.25% methyl cellulose (vehicle)/kg or 4 ml vehicle containing 400 mg to 2000mg Cipa extract/kg, once daily for 7 days (in accordance with OECD guidelines- 407). During this period, food intake, body weight, and clinical signs were monitored daily. At the end of the experiment, animals were euthanized, followed by determination of hematological (Hb, WBC count, RBC count, platelet count and hematocrit) and biochemical (SGOT, SGPT, total protein, serum albumin, total cholesterol, urea, creatinine and random sugar) parameters. Necropsy was performed. Organ weights were recorded and histopathology was done.
Human blood from healthy volunteers was collected after written informed consent in strict accordance with approved guidelines of the Human Ethics Committee, Ranbaxy Laboratories, which approved the ‘Clinical protocol for collection of blood samples from healthy adult human subjects for screening of potential new chemical entities to inhibit cytokine release and other in vitro parameters in an assay system’ through Ethical Approval number: 5001_cok_11. All animal experiments were performed in strict accordance with guidelines specified by the Committee for the Purpose of Control and Supervision of Experiments on Animals (CPCSEA) of the Government of India. Animal experimental protocols were approved by the Institutional Animal Ethics Committee (IAEC) of Ranbaxy Research Laboratories, India through Approval number 151/07.
Three bioassays were developed to identify herbal extracts with potential DENV-inhibitory activity, as shown schematically in Fig 1. The type-1 assay was designed to identify herbal extracts that had the ability to block DENV from entering susceptible cells. As a standardized reference is not available to assess the efficacy of DENV inhibitory potency, PRNT data of monoclonal antibodies (mAbs) were used to decide on the cut-off value in the type-1 assay. For example, two pan-DENV mAbs, 4G2 and DB13-19, show half maximal inhibition of DENV-2 plaque counts at 11 and 33 μg/ml, respectively [23]. Based on this observation, an extract which showed an IC50 value ≤ 25μg/ml, a value which falls between these two, was designated to be active as a DENV inhibitor. Considering that this value is for a crude multicomponent extract, it represents a stringent cut-off. In the context of antiviral screening, this assay had the following possible outcomes: the extract could be cytotoxic (compromised monolayer), inactive (no reduction in plaques in comparison to control infection), or active (reduced number of plaques). However, the type-1assay is not likely to necessarily reveal if the active extracts also possessed the ability to inhibit post-entry steps in the DENV life cycle.
The type-2 assay was designed to assess the capacity of the herbal extracts to inhibit DENVs within the cell. Therefore, DENV infection preceded exposure to the herbal extract. As the herbal extract must be internalized before it exerts any inhibitory effect, we decided to use higher amounts of the extract. Accordingly, the cut-off in the type-2 assay was increased to 150μg/ml. Possible outcomes in the type-2 assay are as follows: once again, the extract could be cytotoxic. If not, it may be ineffective due to one of two reasons: the extract failed to enter the cells, or it was ineffective despite successful entry. The final possibility is that the extract would reduce the plaque count, indicating antiviral activity. The type-3 assay was designed to monitor inhibition based on reduction in viral antigen production and virus yields from DENV-infected cells following treatment with herbal extracts. The amounts of viral NS1 antigen and infectious virus released into the culture supernatant were measured over time to gauge the magnitude of inhibition.
In order to identify an effective herbal extract against dengue, it was reasoned that it must target all four DENVs. Further, it would be desirable to be able to inhibit the viruses both before as well as after entry into susceptible cells. Towards this end, all extracts were screened against a single DENV serotype in the type-1 assay, followed by screening of the active extracts against the remaining three serotypes. Next, extracts manifesting pan-DENV inhibitory potential in the type-1 assay were progressed to screening in the type-2 assay. Extracts identified as active against all 4 DENVs in the type-1 and -2 assays were tested in type-3 assay to assess their ability to reduce virus titres by ≥ ten-fold.
Nineteen plants described in Ayurvedic literature as being beneficial in the treatment of illnesses with dengue-like symptoms were shortlisted for bioassay guided screening (Table 1). Initial screening showed that only the methanolic extracts manifested antiviral activity when assayed against DENV-2 or DENV-3. The hydroalcoholic and aqueous extracts of all 19 plants selected for the study did not manifest any antiviral activity when tested against these two DENV serotypes (IC50 >>100μg/ml). As a result, all subsequent studies were carried out using the methanolic extracts.
Methanolic extracts were prepared from each of these and screened against DENV-3 in a type-1 assay. Extracts that manifested IC50 values below the pre-designated cut-off were scored as positive and the remaining as negative for antiviral potency. The results are summarized in Table 1. This screening assay revealed 8 of 19 extracts to possess definite anti-DENV-3 inhibitory activity, while one manifested borderline inhibition. These 9 extracts were then screened against the remaining three DENV serotypes as well in the type-1 assay. These results, shown in Table 2, revealed that 4 of the 9 anti-DENV-3 extracts manifested potent inhibitory activity against the remaining three DENV serotypes as well. The rest were effective at least against two other DENV serotypes. When all these 9 were tested in a type-2 assay against DENV-3 to identify extracts that may inhibit the virus after its entry into cells, two extracts, one obtained from Cissampelos pareira Linn, and the other from Phyllanthus amarus, were effective in curbing DENV-3 (IC50 values <150 μg/ml). Extending the type-2 assay to DENV-1, -2 and -4 revealed that both these extracts possessed the ability to inhibit all four DENVs even after their entry into cells (Table 3). While Phyllanthus amarus has been documented to possess antiviral properties against another flavivirus, Hepatitis C virus [24] as well as many other viruses [25], this is the first report of antiviral activity associated with Cissampelos pareira Linn. Further studies in this work focused on Cissampelos pareira Linn, henceforth designated as Cipa for convenience.
From the perspective of evaluating the availability of starting material, it was considered necessary to ascertain the extent of seasonal and geographical variations in the capacity of the Cipa extract to manifest DENV inhibitory activity. To this end, extracts were prepared from Cipa plants collected during different seasons from multiple geographical locations within India. Analysis of these extracts revealed that the anti-DENV inhibitory activity is evident in plants collected during the flowering season of April to September [26]. In the experiment, the results of which are summarized in Table 4, Cipa extracts were prepared from plants obtained during the flowering season from four different locations, and evaluated using the type-1 assay described above. The data show that Cipa extracts manifested pan-DENV inhibitory activity in both assay formats. Further, the observed mean pan-DENV IC50 values were not significantly different for all four geographic locations tested. Fractionation of the methanolic extracts with different solvents resulted in nominal increase in antiviral efficacy and a significant loss in yield. Data in Table 4 show that among the solvents tested, methanol extraction is the best with yields ranging from 6–13%. In addition, a preliminary LC-MS analysis of extracts from two different locations (S1 Fig), demonstrates that the two Cipa extracts manifest essentially similar profiles. This leads to the conclusion that there is no discernible geographical variation and that the method of extract preparation is standardized and reproducible.
The kinetics of virus inhibition by Cipa extract was analysed in a type-3 assay. Aliquots of the culture supernatant were withdrawn at regular intervals over a period of several days and analysed for the presence of viral NS1 antigen and infectious virus, as shown in Fig 2. In control experiments wherein infected cells were not exposed to Cipa extract, NS1 antigen was detectable after day 2 onwards and rising thereafter during the course of the experiment. In parallel experiments, the exposure of cells to Cipa extract had a dose-dependent inhibitory effect on NS1 antigen secretion. While the inhibition resulting from exposure to a low dose of Cipa extract was manifested after day 4 post-infection, inhibition at higher doses was evident earlier and at relatively higher magnitudes (Fig 2A). In fact, at the highest dose of Cipa extract tested in this experiment (100μg/ml), the inhibition of NS1 antigen was near total for the entire duration of the experiment. The inhibition of viral antigen synthesis evident from this experiment suggests that virus production would also be similarly affected. This notion was substantiated by determination of viral titers in the culture supernatants during the course of the above experiment, as shown in Fig 2B. In the control experiment, viral titers increased steadily reaching a plateau at day 3 post-infection. Cipa extract lowered viral titers in a dose-dependent manner as seen for NS1 secretion. Thus, at the lowest concentration of Cipa extract used, reduction in viral titers became apparent from day 4 onward. Significantly, a small increase in the Cipa dosage resulted in >1 log reduction in viral titer as early as day 3 post-infection. At the highest dose of Cipa extract tested (100μg/ml), the drop in viral titers was ~2 logs. Importantly, the reduction in viral titers was sustained over a period of several days. Interestingly, the magnitude of inhibition appeared to be greater based on NS1 levels compared to viral titers. The data suggest that the Cipa extract may have effects on NS1 antigen synthesis and release that are distinct from its effects on virus replication.
Next, a complementary experiment was performed wherein DENV-3 was pre-incubated with increasing concentrations of Cipa extract for different periods of time before infection (type-1 assay) and overlay. Plaque counts obtained at the end of the experiment revealed a dose- and time-dependent virucidal effect of Cipa extract on DENV-3 as depicted in Fig 3. A converse experiment, again in type-1 format, was carried out to determine the inhibitory efficacy of the extract against DENV-3 stocks whose titers varied over 2 logs. The IC50 values corresponding to DENV-3 dosage of 50, 500 and 5000 PFUs were, respectively, 9.92, 12.5 and 44.45 μg/ml. This leads to the conclusion that the antiviral potency of Cipa extract extends over a wide range of viral loads.
Next, we tested Cipa extract for its efficacy in vivo. For this, we used the AG129 mouse which has emerged in recent years as a promising dengue model [21, 27] for testing potential DENV inhibitors [28]. This mouse is capable of hosting replication of brain-adapted DENV administered intra-peritoneally (i.p.) and succumbs to it at high challenge doses [27]. We developed a mouse brain-passaged DENV-2 (New Guinea C)-derived challenge strain [22]. At a dose of 106 PFU (given i.p.), it was lethal to AG129 mice, killing them 25 days post-challenge. The illness which preceded death was characterized by ruffled fur, lethargy, hunched back and hind limb paralysis (S2 Fig). We found that the median survival time (MST) of challenged mice treated orally with Cipa extract (methanol-free, see ‘Methods‘) twice a day for 5 days post-challenge was increased in a dose-dependent manner. The survival data are present in Fig 4. The MST of challenged mice was 19 days under the experimental conditions. At 125 mg dose, given twice a day for 5 days, survival was 50% and MST was 28 days (p = 0.1). This increased to ~67% when the dosage was doubled. Compared to the placebo-treated (0.25% methyl cellulose) group, the level of protection afforded by 250 mg/Kg dose was statistically significant (p = 0.021). It is to be ascertained in future if the protective efficacy may be further enhanced by extending the drug treatment regimen beyond 5 days.
Since the data so far showed that Cipa extracts have potent pan-DENV inhibitory activity, it was considered worthwhile to explore the feasibility of its therapeutic use. As DF is normally treated with paracetamol, it would be important to ascertain the nature of any interaction between Cipa and this drug. Dengue disease predisposes some patients to hemorrhagic manifestations and tends to be associated with lowered platelet counts. In this context, it also becomes important to assess if Cipa would have any effect on RBCs and platelets. These concerns were addressed in the following experiments. The data from the studies on Cipa extract and paracetamol are depicted in Fig 5. A type-1 assay was carried out in which DENV-3 was pre-incubated with serial dilutions of a Cipa extract. It was observed that DENV-3 infectivity was inhibited progressively as the Cipa extract concentration increased, with an IC50 value of 6.1μg/ml. The addition of up to 100μg/ml paracetamol into the DENV-3/Cipa extract pre-incubation mix did not significantly affect the inhibitory profile of Cipa. The calculated IC50 values in the presence of paracetamol at 1, 10 and 100μg/ml were, respectively, 8.4, 7.4 and 8.5μg/ml (Fig 5A). Paracetamol by itself at all concentrations tested did not have any effect on DENV infectivity (plaque counts obtained with DENV-3 alone and DENV-3 plus paracetamol at 100μg/ml were, 43±3 and 45±4, respectively; n = 3). The next experiment examined the effect of Cipa extract on the antipyretic activity of paracetamol using the Wistar rat pyrexia model. Interestingly, this experiment revealed that Cipa extract possessed intrinsic antipyretic effect (Fig 5B). When rats, in which fever was induced by subcutaneous injection of brewer’s yeast, were treated with Cipa extract, the fever was suppressed at an efficiency that was comparable to that of paracetamol. Interestingly, co-administration of Cipa extract with paracetamol had a synergistic effect, resulting in a more pronounced decrease in body temperature.
In the next set of experiments, whole blood from human volunteers was collected and platelets counts obtained before and after 1–4 hours post-mixing with Cipa extract. The results are shown in Fig 6A. In the control sample, blood mixed with vehicle (saline), platelet counts declined steadily over time. Cipa extract-treated blood samples manifested no statistically significant change in platelet counts with respect to their cognate controls, up to 2 hours (p>0.05). At four hours, the Cipa extract-treated samples displayed significantly higher (p<0.05) platelet counts, compared to corresponding saline-treated control. The apparent platelet-protective effect of Cipa extract is intriguing and merits further investigation. Essentially similar results were observed at 2 and 10 μg/ml Cipa extract concentrations, indicating that Cipa extract did not affect platelets adversely. The effect of Cipa extract on platelets was also evaluated in an in vivo experiment using Wistar rats. In this experiment, platelet counts were determined in blood drawn from rats which had been given Cipa extract orally. The results presented in Fig 6B show that up to 4 hours post-treatment, Cipa extract (up to 1000 mg/kg body weight), did not affect platelet counts significantly (p>0.05 at the highest dose of Cipa extract treatment for 4 hours).
The effect of Cipa extract on erythrocytes was also assessed, both in ex vivo and in vivo assays, as done for platelets above. Incubation of freshly collected human erythrocytes with Cipa extract at concentrations up to 400 μg/L did not cause discernible haemolysis (Fig 7A). The blood samples, withdrawn from the Wistar rats (given Cipa extract, described above), were also analysed for erythrocyte cell counts. This analysis, once again revealed that Cipa extract (at concentrations as high as 1000 mg/kg body weight), did not affect erythrocyte counts in the blood of Wistar rats up to 4 hours post-administration (Fig 7B). The difference in erythrocyte counts between the treated and untreated rats was not statistically significant (p>0.05). We also analysed total leucocyte and differential counts in the blood of Cipa-treated Wistar rats (described in Figs 6B and 7B) and found no significant difference (S3 Fig).
A hallmark of severe dengue disease is the elevation of inflammatory cytokines which are implicated in triggering events that culminate in vascular permeability and haemorrhage [2, 4]. To test if Cipa extract had any effect on the secretion of inflammatory cytokines, PBMCs were isolated from blood and incubated with LPS to induce cytokine release. These cells were then treated with Cipa extract and the release of pro-inflammatory cytokines monitored using commercial ELISA kits. This experiment showed that the secretion of TNF-α and IL-1β was efficiently suppressed by Cipa extract with IC50 values of 6.1±1.3 and 5.7±2.7 μg/ml, respectively. An MTT assay showed that at these concentrations, Cipa extract has no discernible cytotoxicity in both cell lines tested (CC50 = 78.9μg/ml in HepG2; >200μg/ml in LLCMK2). These data suggest that Cipa extract possesses anti-inflammatory activity in addition to the antipyretic activity documented in the experiment above.
An exploratory toxicological study to assess any adverse effect of repeat dosing over a 1 week span with Cipa extract was carried out in Wistar rats. Multiple physiological, haematological, biochemical and clinical parameters were monitored. The results showed that animals treated with up to 2000mg/kg body weight did not manifest any significant changes in any of these parameters compared to vehicle-treated controls (S1–S4 Tables). This essentially corroborates an earlier report that the Cipa extract is essentially non-toxic and well tolerated in the animal model [29].
In the absence of a licensed vaccine or antiviral drug, dengue continues to be a significant global public health problem [3, 4]. This is further accentuated in India by its dense human population which represents ~50% of the global population, estimated by WHO to be at risk of dengue. Recently, indigenous diagnostic tests for the early detection of dengue have become available in India [30]. The utility of early detection would be in being able to provide timely antiviral therapy. Further, in the absence of a vaccine, administering antivirals when an outbreak is detected would offer a way of limiting disease spread. While the level of viremia is generally lower in DF, it increases by an order or two in magnitude in DHF/DSS [5, 6]. This observation has led drug developers to hypothesize that lowering viremia by 1–2 logs may be associated with a favourable prognosis [14].
As many of the modern drugs have been derived from natural precursors [17, 18], it is increasingly being regarded that ethnopharmacology and traditional medicines offer an attractive option for identifying starting material for drug discovery initiatives [19, 31]. This provided the basis for the current work which was undertaken to explore the indigenous herbal bio-resource in India to identify plants that may possess anti-DENV inhibitory activity.
Ayurveda represents a distinct discipline in Indian ethnomedicine and has provided a large number of lead molecules for a variety of indications [19]. To provide a rational basis for plant selection, the ayurvedic literature was examined for instances of illnesses with symptoms similar to dengue. Indigenous plants and herbs prescribed for treating these illnesses were chosen for anti-DENV activity screening. In addition plants documented in the literature to possess antiviral activity were also included. As each of the four DENV serotypes can cause severe dengue disease and all of these co-circulate in the hyper-endemic countries, it was considered desirable for a drug to be effective against all the four serotypes. For this purpose a whole cell-based bioassay-guided screening protocol was developed. In principle, an effective antiviral drug could target extracellular virus and block its entry into susceptible cells or act on the virus that has gained entry into cells by targeting one of the post-entry steps in the viral life cycle. Assays were designed to identify herbal extracts targeting both extracellular and intracellular virus. This resulted in the identification of Cipa extract and the extract from Phyllanthus amarus as possessing pan-DENV inhibitory activity. As the finding of antiviral activity associated with Cipa (SI>45) was a novel one, further work focused on it. Its ability to inhibit DENV in a time and dose-dependent manner in the type-1 assay format suggested that the Cipa extract possesses a virucidal effect. Further, it was effective as an inhibitor over a ten-fold range of DENV titers suggesting its potential to counteract high viremia. Consistent with this, an analysis of the effect of Cipa extract on virus titers demonstrated a >1 log reduction compared to untreated virus controls suggesting its potential utility in altering the course of severe dengue disease to a more favourable outcome. Importantly, Cipa extract manifested statistically significant protective efficacy in the AG129 mouse model [21, 27], which has emerged recently as the only promising small animal model for in vivo DENV inhibitor testing [28].
At this juncture, two options in terms of future course of development could be envisaged. One, the Cipa extract could facilitate the drug discovery process. Two, the Cipa extract could be the starting point for developing a carefully standardized herbal formulation for human use. The former option would entail a systematic fractionation to isolate and characterize the active ingredient which may provide small molecule drug lead(s) for optimization and further development. This is both expensive and time-consuming. Additionally, there is the attendant possibility of loss of activity upon fractionation to isolate pure compounds. Considering that Cipa is in use for treating human ailments in Ayurvedic medicine in India [32, 33] and in traditional medicine in several countries in South America [34], pursuing the latter option could be a lot simpler and affordable, and of practical utility, particularly in the context of the resource-poor dengue-endemic countries. Preliminary in vitro and in vivo experiments addressing the clinical relevance of Cipa extract support the feasibility of pursuing the second option. For example the drug paracetamol used commonly to treat dengue patients, does not affect the antiviral efficacy of Cipa extract. Interestingly, Cipa extract appeared to possess an intrinsic antipyretic activity which could synergize with that of paracetamol in the Wistar rat model. It also possessed the ability to down-regulate the secretion of pro-inflammatory cytokines, particularly TNF- α implicated in the pathogenesis of severe disease [2]. Further, Cipa extract did not have any discernible effect on platelet counts or on erythrocyte viability. Importantly, the extract was not associated with any adverse toxicology. Finally, another factor in favour of the second option is that, Cipa plant is available in several parts of the country and extracts prepared using plants from different geographical locations are fairly similar in terms of their gross overall composition. This suggests that availability is not an issue and extract preparation can be reproducible.
In conclusion, Ayurveda knowledge-based selection of medicinal plants in conjunction with a bioassay-guided screening protocol has resulted in the identification of Cipa extract which manifests potent antiviral activity against all four prevalent DENV serotypes. In addition, this extract also manifested dose-dependent protective efficacy in an in vivo model, and appeared to be compatible with future clinical use. The outcome of these studies should hopefully pave the way to carry out a systematic development of Cipa extract to enable filing of an investigational new drug application with the Drug Controller General of India. Parts of this work have been the subject of patent applications, several of which have been granted [35, 36].
|
10.1371/journal.pcbi.1004779 | Evolution of Cooperation in Social Dilemmas on Complex Networks | Cooperation in social dilemmas is essential for the functioning of systems at multiple levels of complexity, from the simplest biological organisms to the most sophisticated human societies. Cooperation, although widespread, is fundamentally challenging to explain evolutionarily, since natural selection typically favors selfish behavior which is not socially optimal. Here we study the evolution of cooperation in three exemplars of key social dilemmas, representing the prisoner’s dilemma, hawk-dove and coordination classes of games, in structured populations defined by complex networks. Using individual-based simulations of the games on model and empirical networks, we give a detailed comparative study of the effects of the structural properties of a network, such as its average degree, variance in degree distribution, clustering coefficient, and assortativity coefficient, on the promotion of cooperative behavior in all three classes of games.
| Social dilemmas embody the tension between individual self-interest on the one hand and the public good on the other that underlie many of the greatest challenges facing human and animal societies, such as the maintenance of altruism and the responsible use of common resources. Understanding the mechanisms through which cooperative, socially optimal, behavior can be established in social dilemmas is a fundamental problem in evolutionary biology and in many areas of the social sciences. Here we study how cooperative behavior can emerge in three key social dilemmas—known as the donation game, snowdrift game and sculling game—when interactions between individuals form a network. We show that in all three social dilemmas significantly higher levels of cooperative behavior typically emerge in such a situation as compared to what would be the case in the absence of any network structure. In particular, we show that certain structural properties that are common in real-world social networks have a significant effect in increasing the level of cooperative behavior that is present in all three social dilemmas.
| The evolution and maintenance of cooperation in social dilemmas is essential for the formation of systems at all levels of complexity, from replicating molecules to multi-cellular organisms to human societies. A small selection of examples of cooperative behavior in social dilemmas from the bewilderingly large number of possible instances include: assembly of the earliest replicating molecules to form larger replicating entities capable of encoding more information [1, 2]; integration of once free-living prokaryote ancestors of mitochondria and chloroplasts into eukaryotic cells [2]; differential production of intracellular products needed for replication in an RNA phage [3]; vampire bats donating blood meals to roost mates [4]; predator inspection in sticklebacks and guppies [5]; allogrooming in social animals [6]; alarm calls by mammals and birds in response to danger [7]; contribution to providing public goods [8], such as social security programs; restraint in the consumption of common resources [9, 10], including responsible use of fishing grounds, limiting the emission of pollution into the atmosphere, and sharing Internet bandwidth; correct implementation of the TCP protocol so as to avoid congestion in Internet traffic [11]; and sharing files over a peer-to-peer network [12].
In spite of the ubiquity of cooperation in social dilemmas, achieving a satisfactory understanding of the origin and stability of this phenomenon is fundamentally difficult [2, 13–18]. This difficulty resides in the very nature of a social dilemma, which in classical game theory may be defined as a game which possesses at least one socially inefficient Nash equilibrium [16, 19]. At this Nash equilibrium there is no incentive for any individual to change their behavior, and yet because it is socially inefficient there is at least one other outcome in which all individuals would be better of. Here we shall be concerned with the notion of a social dilemma that arises in evolution, which we shall define analogously to be a game which possesses at least one socially inefficient evolutionary attractor. Such an evolutionary attractor could be, for example, an evolutionary stable strategy [20, 21], a stable equilibrium point of the replicator dynamics [22, 23] (where, in fact, the former implies the latter), or an attracting state in stochastic evolutionary dynamics [24–29]. In such a social dilemma, adopting the strategy at the socially inefficient attractor is considered to be defection, while adopting the socially efficient strategy is taken to be cooperation. The nature of the dilemma is now apparent—individuals employing strategies corresponding to the socially inefficient attractor will be trapped there by natural selection, in spite of all individuals being better off if they could adopt socially efficient behavior.
In this paper we study the evolution of cooperation in three symmetric, two-player, two-strategy (i.e., symmetric 2 × 2) games, which are exemplars of social dilemmas in the three fundamental classes of symmetric 2 × 2 games—the prisoner’s dilemma, hawk-dove, and coordination classes.
The first game we consider is the donation game, which is the fundamental exemplar in the prisoner’s dilemma class of games, and provides the basic game theory model for altruism [13, 17, 18]. The second game, the snowdrift game, is an exemplar of a social dilemma in the hawk-dove class of games [15, 18]. While games of hawk-dove type have been extensively studied as models of conflicts and contests [20, 21, 30], the snowdrift game provides an interesting model for certain types of cooperative behavior that differ from pure altruism [15, 17]. The third game that we introduce and study here, which we call the sculling game, is an exemplar of a social dilemma in the coordination class of games. Games in the coordination class have been widely used as models for conventions [24, 25, 31], but they have typically received little attention as models of cooperation, although interesting exceptions to this trend are [32–35]. The sculling game, as we define it, is an exemplar of a game in the coordination class in an analogous sense to that in which the snowdrift game is an exemplar in the hawk-dove class of games. We introduce the sculling game with the intention of using it as a model for certain types of cooperative behavior not described by the donation or snowdrift games. It is worth remarking that in [2] two games were mentioned very briefly in passing, which were referred to as the “rowing game” and the “sculling game”, however, these games are unrelated to the sculling game that we introduce here (in [2] the “rowing game” is a coordination game and the “sculling game” is a prisoner’s dilemma game).
A number of different approaches to understanding the evolution of cooperation in social dilemmas have been studied [36]. These include: kin selection [13], direct reciprocity [14, 37–39], indirect reciprocity [40, 41], evolution in network-structured populations [42–50], and evolution in group structured populations [51, 52].
In this paper we study the evolutionary dynamics of the donation, snowdrift, and sculling games in structured populations modeled by complex networks, where vertices represent individuals in the population and edges between vertices indicate that the corresponding individuals may interact. We use individual-based simulations to investigate the evolutionary dynamics of the three games on a variety of model and empirical networks, and we give a detailed comparative study of the effects of network properties, such as average degree, variance in degree distribution, clustering coefficient, and assortativity coefficient, on the promotion of cooperative behavior in all of these games.
There is an extensive literature concerned with the evolution of cooperation in structured populations, which we will briefly review so as to put the current work in context. Many studies have focused on understanding how cooperation arises in spatially structured populations which can be modeled by regular lattices. A large majority of these works have studied the evolution of cooperation in the prisoner’s dilemma game or variants of it [42, 43, 45, 47–50, 53–60]. A smaller number of studies of lattice structured populations have focused on other games such as the hawk-dove game or the snowdrift game [44, 46, 61, 62]. More recently there has been a shift of attention towards studying the evolution of cooperation in structured populations that are modeled by complex networks [63–66]. Again the vast majority of these investigations have focused on studying various versions of the prisoner’s dilemma game [47–49, 67–73], but with some studies also on other games such as the snowdrift game [48, 74–76] or stag hunt game [34, 35].
Recent work has also focussed on other interesting aspects of the evolution of cooperation in structured populations, including: the use of effective payoffs in the prisoner’s dilemma game [77]; the effect of random and targeted removal of vertices [78]; cooperation in populations with multiple interaction layers [79]; the effect of social influence [80]; the consequences of dynamical linking [81, 82]; and the evolution of cooperation in other games, such as the traveler’s dilemma game and the minimum effort coordination game [19, 83].
No brief overview of the literature on the evolution of cooperation in structured populations can do justice to the enormous body of work on this subject, and for more detailed discussions we recommend that the interested reader consult one or more of the excellent reviews available [50, 84–87].
The main conclusion that emerges from previous studies is that for the prisoner’s dilemma game and its variants structured populations allow cooperation to be maintained under suitable circumstances, which is never possible in a well-mixed population [42, 43, 45, 47–49, 53–55, 57–60, 67–73]. For the hawk-dove game and the snowdrift game the results are less clear-cut. For the hawk-dove game, lattice structured populations typically result in increased levels of cooperation [44], however, for the snowdrift game, cooperation may be inhibited [46] or promoted [48] by different population structures. The small number of studies of the stag hunt game suggest that structured populations can also promote the evolution of cooperation in this game [34, 35].
Despite the large number of studies on the evolution of cooperation in structured populations, surprisingly little is known about how the fundamental properties of a network structured population, such as mean degree, heterogeneity in degree distribution, clustering and assortativity, affect the evolution of cooperation. What little is known in this direction is almost completely confined to the effect of average network degree on the evolution of cooperation in the prisoner’s dilemma game [49, 67], with almost nothing being known about the effect of other structural properties of networks on the evolution of cooperation in the prisoner’s dilemma game or of the effect of almost any structural network properties on the emergence and maintenance of cooperation in games in the hawk-dove or coordination classes. A further limitation of previous studies of cooperation in network structured populations is that there is typically no clear way to compare the levels of cooperation that arise in games representing different classes of social dilemmas. For example, there is no straightforward way to compare the level of cooperation that occurs in the prisoner’s dilemma game [42, 43, 45, 47–49, 53–55, 57–60, 67–73] with those that occur in the snowdrift game [46] or the stag-hunt game [34, 35]. As a result of this lack of comparability, previous work on the evolution of cooperation in structured populations has led to many isolated results, but much less in the way of general understanding that applies across a wide range of social dilemmas.
The purpose of this paper is to provide a systematic and comparative study of the evolution of cooperation for those games which serve as canonical representatives of the three fundamental classes of symmetric 2 × 2 games—the prisoner’s dilemma, hawk-dove, and coordination classes of games—in a wide variety of structured populations. The three canonical games that we study here are the donation game, the snowdrift game and the sculling game, which provide exemplars of the the prisoner’s dilemma, hawk-dove, and coordination classes of games, and with which the evolution of cooperation in all three classes of games in structured populations can be studied in a uniform way.
Aspects of the evolution of cooperation in structured populations have previously been studied for the donation game in [49] and for the snowdrift game in [46, 48]. The sculling game is introduced for the first time here, with the aim of providing a canonical representative game in the coordination class of games that can be used to study cooperative phenomena. We regard the sculling game as playing a similar role in the coordination class of games to that played by the snowdrift game in the hawk-dove class of games, and hope that the introduction of this game will stimulate further study of cooperative behavior in coordination-type games.
The major theme of this paper is to understand how the fundamental structural properties of a network affect the evolution of cooperation in the donation, snowdrift and sculling games. For complex networks the key structural properties include [63–66]: mean degree, heterogeneity of the degree distribution, clustering coefficient, and assortativity coefficient. As noted above, most previous work on cooperation in structured populations has centered on the prisoner’s dilemma game and its variants, but even in this case little is known about the effect of the structural properties of the network on the evolution of cooperation. Still less is known in this direction for the snowdrift and stag hunt games. Determining how the level of cooperation in each of these three games depends on the structural properties of the network, and comparing the levels of cooperation that arise in different games, is a non-trivial problem even in principle, since the level of cooperation that emerges in a particular game on a given network will depend on the parameters that enter into the definition of the game. To overcome this difficulty, we show that for the donation, snowdrift and sculling games it is possible to define a single integrated measure of the total degree of cooperation in the game on a network which is independent of the parameters specifying the game. This measure, which we refer to as the cooperation index, allows the total degree of cooperation in all three games on a network—and, crucially, the dependence of the total degree of cooperation on the structural properties of the network—to be unambiguously computed and compared. These developments represent the major novelty of the present work—we determine in a systematic and uniform fashion, which allows direct comparison, the degree of cooperation that evolves in the donation, snowdrift and sculling games on a wide variety of model and empirical networks, and further study how the level of cooperation in these games depends on the fundamental structural properties of the networks.
In this section we define the three symmetric 2×2 games that we use to model the evolution of cooperation. These games are exemplars of the prisoner’s dilemma, hawk-dove and coordination classes of games, which we will refer to as the donation, snowdrift and sculling games, respectively.
The payoff matrix of a symmetric 2 × 2 game, with strategies C (cooperation) and D (defection) is given by:
π = C D C D [ α β γ δ ] , (1)
where α , β , γ , δ ∈ R. The specific values of the elements of the payoff matrix depend on the game under consideration.
Consider a large, well-mixed population of individuals playing the symmetric 2 × 2 game with payoff matrix given by matrix 1. Let p denote the frequency of individuals playing strategy C, with frequency 1 − p playing strategy D. The evolutionary dynamics of the population is governed by the replicator equation [17, 22, 23]
p˙=p(fC−f¯)=p(1−p)[ p(α−γ)+(1−p)(β−δ) ], (2)
where fC = pπ(C, C) + (1 − p)π(C, D) and fD = pπ(D, C) + (1 − p)π(D, D) denote the fitnesses of the strategies C and D, respectively, and f ¯ = p f C + ( 1 - p ) f D denotes the mean fitness of the population.
In this section, we define a stochastic individual-based model which allows the evolutionary dynamics of symmetric 2 × 2 games to be studied for populations with complex interaction patterns among their members [19, 46, 48, 49]. Consider a population of n individuals, labeled by i = 1, …, n. In order to allow the possibility of complex population interactions, we identify the individuals in the population with the set of vertices in a network Γ. The structure of Γ determines which individuals in the population interact. To be precise, two networks are required to specify the evolutionary dynamics—an interaction network, which specifies whether two individuals in the population can interact by playing the game, and an updating network, which specifies that an individual in the population can update its strategy by comparing its state to the states only of those individuals adjacent to it. Here, for simplicity, we shall assume that the interaction and updating networks are the same, and denote them both by Γ. The set of neighbors of i ∈ Γ (i.e., the set of individuals adjacent to i in Γ) will be denoted by N(i). The degree ki of vertex i is the number of vertices adjacent to i, and the mean degree of the vertices in Γ is defined to be k = 1 n ∑ i = 1 n k i.
The individual-based model is defined as a stochastic process on the network Γ. We shall start with an initial population in which a fraction p0 of the individuals use strategy C and remainder use strategy D. Each iteration of the evolutionary dynamics consists of an asynchronous interaction/update round, which involves sampling the population n times with replacement.
Each interaction/update step is carried out as follows (cf. [46, 48]). First, in the interaction phase we pick at random an individual i ∈ Γ and a random neighbor j ∈ N(i). In addition, we pick a random neighbor k ∈ N(i) of i and a random neighbor l ∈ N(j) of j. We then determine the payoff that player i receives from interacting with player k and the payoff that player j receives from interacting with player l. That is, if p, q, r, and s denote the strategies of i, j, k, and l respectively, then the payoff Pi received by the focal individual i is π(p, r) and the payoff Pj received by the individual j is π(q, s), where π is the payoff matrix for the game under consideration. Second, in the update phase the probability that the focal individual i will inherit j’s strategy, pi ← j, is determined using the Fermi update rule [88–90] as
p i ← j = 1 1 + e - β ( P j - P i ) , (9)
where the parameter β > 0 is the “selection strength” of the update rule. Repeating the interaction/update step n times constitutes one generation of the evolutionary dynamics.
We note that the results of our individual-based simulations (described in the next section) are robust to changes in the update rule. For example, in addition to employing the Fermi update rule given by eq 9, we have also simulated the individual-based model using the replicator update rule [46, 91], in which the probability pi ← j that the focal individual i inherits individual j’s strategy is given by
p i ← j = 0 if P i ≥ P j P j - P i P max - P min otherwise, (10)
where P m a x = max u ∈ Γ P u, and P m i n = min u ∈ Γ P u. We find that the evolutionary dynamics of the symmetric 2 × 2 games that we study here is essentially identical irrespective of which of these update rules we employ. The results presented in the next section are from simulations using the Fermi update rule.
In this section we present the results of individual-based simulations for the donation, snowdrift, and sculling games. For all three games the individual-based model described in the previous section was simulated using the Fermi update rule (see eq 9) on the following model networks [63–66]: random regular networks; networks with exponential degree distribution (constructed using the growing random network model [92]); scale-free networks (constructed using the Barabási-Albert preferential attachment scheme [64]); and scale-free networks with different clustering coefficient C (defined below), generated using the Holme-Kim model [93] and assortativity coefficient r (defined below), generated by applying the rewiring algorithm of [94] to a Barabási-Albert’s scale-free network. The network size was n = 10000 for all these model networks.
A common property of many real-world networks is that they have a non-zero clustering coefficient [63, 66]. The clustering coefficient of a network Γ measures the average probability that two neighbors of a vertex are themselves adjacent. The local clustering coefficient Ci of a vertex i ∈ Γ is defined to be [63, 66]
C i = ( number of pairs of neighbors of i that are adjacent ) ( number of pairs of neighbors of i ) .
The global clustering coefficient C ∈ [0, 1] for the whole network is then defined as the mean of the local clustering coefficients Ci[63, 66]:
C = 1 n ∑ i = 1 n C i .
A network with C = 1 has maximal clustering, while one with C = 0 has no clustering at all.
Another common property of many real-world networks is that they possess some amount of assortativity or disassortativity [95]. Assortative networks have the tendency that high-degree vertices are connected to other high-degree vertices and low-degree vertices are connected to other low-degree ones. In contrast, in disassortative networks, high-degree vertices tend to be connected to low-degree vertices and vice versa. Social networks are often assortative, while biological and technological networks are usually disassortative [95].
The assortativity (or disassortativity) of a network Γ can be quantified by the coefficient of assortativity r ∈ [−1, 1], defined by [95]
r = ∑ i , j = 1 n ( A i j - k i k j / 2 m ) k i k j ∑ i , j = 1 n ( k i δ i j - k i k j / 2 m ) k i k j ,
where A is the adjacency matrix of Γ, ki is the degree of vertex i ∈ Γ, δij is the Kronecker delta, and m is the number of edges in the network. Networks with r > 0 are assortative, while those with r < 0 are disassortative. Networks with r = 0 are neither assortative nor disassortative.
In addition to model networks, we have simulated the donation, snowdrift, and sculling games on the following empirical networks: the network Γhep of coauthorships between scientists who posted preprints on the High-Energy Theory E-Print Archive between January 1, 1995 and December 31, 1999 [96]; the network Γastro of coauthorships between scientists who posted preprints on the Astrophysics E-Print Archive between January 1, 1995 and December 31, 1999 [96]; the largest component, Γfb, of the New Orleans Facebook network of user-to-user links [97]; and a snapshot of the structure of the Internet, Γinternet, at the level of autonomous systems, reconstructed from BGP tables posted by the University of Oregon Route Views Project [66]. The basic properties of these empirical networks are listed in Table 1.
All simulations were carried out for 10000 generations, with the initial fraction p0 of cooperators being 0.5, and with the selection strength β for the Fermi update rule set to 1.0. The results presented here were obtained by averaging the data from 50 independent runs of the model.
It is very useful to have a unitary measure of the total degree of cooperation for the donation, snowdrift, and sculling games on a given network. We will refer to the measure of cooperation that we introduce here as the cooperation index Λ or Λ-index. Let pt(ρ) denote the fraction of cooperators in the population at generation t in a game with cost-to-benefit ratio parameter ρ ∈ (A, B). The long-term mean p ¯ ∞ ( ρ ) of pt(ρ) is defined by
p ¯ ∞ ( ρ ) = lim T → ∞ 1 T ∑ t = 1 T p t ( ρ ) , (11)
and the Λ-index for the three games is defined by
Λ = 1 B - A ∫ A B p ¯ ∞ ( ρ ) d ρ . (12)
The cooperation index Λ ∈ [0, 1], where Λ = 1 represents complete cooperation, while Λ = 0 represents total lack of cooperation. The larger the value of Λ, the more cooperative the system. Thus, the Λ-index serves as a quantitative measure for comparing the totality of long-term cooperation in the population across the different networks. It is informative to compare the results we obtain for the Λ-index on networks with those that can be calculated analytically for the three games from the replicator dynamics in a large well-mixed population, which are as follows: for the donation game, p ¯ ∞ ( ρ ) = 0, A = 0, and B = 1, so Λ = 0; for the snowdrift game, p ¯ ∞ ( ρ ) = 1 - ρ 1 - ρ 2, A = 0, and B = 1, so Λ = 2 − ln 4 ≈ 0.614; and for the sculling game with initial fraction of cooperators p 0 = 1 2, p ¯ ∞ ( ρ ) = 1 if ρ < 1 and 0 otherwise, A = 1 2, and B = 3 2, so Λ = 1 2.
Fig 1(a)–1(c) shows the evolution of cooperation in the donation game on random regular networks. Fig 1(a) shows the variation of the long-term fraction p ¯ ∞ of cooperators with average degree k and cost-to-benefit ratio ρ. In Fig 1(b) the variation of p ¯ ∞ with k is shown when ρ is fixed at 0.1. Fig 1(c) shows the variation of p ¯ ∞ with ρ when k is fixed at 4—the inset in the figure shows the value of the cooperation index Λ. Fig 1(d)–1(f) and 1(g)–(i) show the corresponding results for the donation game on exponential networks and scale-free networks, respectively.
These results indicate that, for all three types of networks, as the average degree k of the network and/or the cost-to-benefit ratio ρ increases p ¯ ∞ undergoes a rapid transition from the totally cooperative state (p ¯ ∞ = 1) to the totally defective state (p ¯ ∞ = 0). The nature of the transition appears to be reasonably similar for all three network types. We observe that the transition from cooperation to defection appears to be governed to a good approximation by the condition that cooperation is favored only if 1 ρ > k - 1 (indicated by a dashed line in Fig 1(a), 1(d) and 1(g)). It follows from this condition that, for given ρ, cooperation only prevails on networks of average degree less than the critical value k c = 1 ρ + 1 (indicated by a dashed line in Fig 1(b), 1(e) and 1(h)), and, for networks of given average degree k, cooperation only prevails for ρ less than the critical value ρ c = 1 k - 1 (indicated by a dashed line in Fig 1(c), 1(f) and 1(i)). It further follows from the latter result that the Λ-index on a network of average degree k is approximately given by Λ = 1 k - 1, which is in reasonably good agreement with the simulation results.
These results imply that, for fixed ρ, as the average degree of a network increases, p ¯ ∞ transitions from the all-cooperator (p ¯ ∞ = 1) state to the all-defector (p ¯ ∞ = 0) state, which is reasonable since a network with a large average degree approximates the complete network, which represents a well-mixed population, for which p ¯ ∞ = 0. We note that the transition from total cooperation to total defection is sharpest for random regular networks and slightly less sharp for exponential and scale-free networks. We also note that the cooperation index Λ is highest for random regular networks and lowest for scale-free network, with exponential networks having an intermediate value of Λ.
Fig 2(a) and 2(b) shows how the cooperation index Λ on scale-free networks with average degrees k = 4 and k = 8 varies with clustering coefficient C (a) and assortativity coefficient r (b). These results suggest that Λ is more or less unaffected by clustering, whereas Λ is higher for significantly assortative networks than for networks that exhibit little to no assortativity. For networks of average degree 4 there is also a slight increase in Λ for disassortative networks compared to those with little or no assortativity.
Fig 3(a)–3(d) show the variation of p ¯ ∞ with ρ on empirical networks: Γhepth (a), Γastro (b), Γfb (c), and Γinternet (d). The inset in the figures shows the value of the cooperation index Λ. The Γhepth and Γinternet networks both have appreciable values of Λ, whereas for both the Γastro and Γfb networks Λ is small. The transition from cooperation to defection on these empirical networks appears also to be governed to a reasonable approximation by the same condition that was found to hold for model networks, namely cooperation is favored only if 1 ρ > k - 1 (indicated by a dashed line in the figures). Since it then follows that Λ is given approximately by Λ = 1 k - 1, the higher values of Λ found for Γhepth and Γinternet and the lower values of Λ found for Γastro and Γfb are consistent with the results that would be expected given the lower average degrees of the former two networks and the higher average degrees of the latter two networks.
Fig 4(a)–4(c) shows the evolution of cooperation in the snowdrift game on random regular networks. Fig 4(a) shows the variation of the long-term fraction p ¯ ∞ of cooperators with average degree k and cost-to-benefit ratio ρ. Fig 4(b) shows the variation of p ¯ ∞ with k when ρ is fixed at 0.75. Fig 4(c) shows the variation of p ¯ ∞ with ρ when k is fixed at 4—the inset in the figure shows the value of the cooperation index Λ, and the dashed line shows the variation of p ¯ ∞ with ρ in a well-mixed population. Fig 4(d)–4(f) and 4(g)–4(i) show the corresponding results on exponential networks and scale-free networks, respectively.
These results show that, for ρ = 0.75, p ¯ ∞ continuously decreases as the average degree of the network increases, approaching a value of 0.4 for k = 20 (indicated by a dashed line in Fig 4(b), 4(e) and 4(h)), which is reasonable because a network with a large average degree approximates the complete network, which represents a well-mixed population, for which p ¯ ∞ = 0 . 4. This behavior seems to be independent of the type of network. For a network with a given average degree k, p ¯ ∞ continuously decreases from higher values to lower values as the cost-to-benefit ratio ρ increases. The cooperation index Λ is not greatly affected by the type of network, although it is slightly higher on random regular networks and progressively lower on exponential and scale-free networks, respectively.
Fig 5(a) and 5(b) show how the cooperation index Λ on scale-free networks with average degrees k = 4 and k = 8 varies with clustering coefficient C (a) and assortativity coefficient r (b). Here again, Λ is more or less unaffected by clustering, whereas Λ is slightly higher for significantly diassortative and assortative networks than for networks that exhibit little to no assortativity. For networks of average degree 4 there is also a slight increase in Λ for disassortative networks compared to those with little or no assortativity.
Fig 6(a)–6(d) shows the variation of p ¯ ∞ with ρ on empirical networks: Γhepth (a), Γastro (b), Γfb (c), and Γinternet (d). The inset in the figures shows the value of the cooperation index Λ, and the dashed line shows the variation of p ¯ ∞ with ρ in a well-mixed population. In all cases the value of Λ is greater than the value Λ = 0.614 for a well-mixed population. The value of Λ for the Γhepth and Γinternet networks is much higher than the value in a well-mixed population, while for the Γastro and Γfb networks, the value is only slightly higher than the value in a well-mixed population. These results are consistent with what would be expected given the differences in the mean degree of the various networks.
Fig 7(a)–7(c) shows the evolution of cooperation in the sculling game on random regular networks. Fig 7(a) shows the variation of the long-term fraction p ¯ ∞ of cooperators with average degree k and cost-to-benefit ratio ρ. Fig 7(b) shows the variation of p ¯ ∞ with k when ρ is fixed at 1.1. Fig 7(c) shows the variation of p ¯ ∞ with ρ when k is fixed at 4—the inset in the figure shows the value of the cooperation index Λ. Fig 7(d)–7(f) and 7(g)–7(i) show the corresponding results on exponential networks and scale-free networks, respectively.
These results show that, for all three types of networks, as the average degree k of the network and/or the cost-to-benefit ratio ρ increases p ¯ ∞ undergoes a rapid transition from the totally cooperative state (p ¯ ∞ = 1) to the totally defective state (p ¯ ∞ = 0). The nature of the transition appears to be quite similar for all three network types. We observe that the transition from cooperation to defection appears to be governed to a good approximation by the condition that, for ρ > 1, cooperation is favored only if 1 ρ - 1 > k - 1 (indicated by a dashed line in Fig 7(a), 7(d) and 7(g)). We note that for ρ < 1, cooperation is both payoff dominant and risk dominant over defection and is thus favored on any network (including the complete network representing a well-mixed population). It follows from this condition that, for given ρ > 1, cooperation only prevails on networks of average degree less than the critical value k c = 1 ρ - 1 + 1 (indicated by a dashed line in Fig 7(b), 7(e) and 7(h)), and, for networks of given average degree k, cooperation only prevails for ρ less than the critical value ρ c = 1 k - 1 + 1 (indicated by a dashed line in Fig 1(c), 1(f) and 1(i)). It further follows from the latter result that the Λ-index on a network of average degree k is approximately given by Λ = 1 k - 1 + 1 2.
It follows from these results that, for fixed ρ > 1, as the average degree of a network increases, p ¯ ∞ transitions from the all-cooperator (p ¯ ∞ = 1) state to the all-defector (p ¯ ∞ = 0) state, which is reasonable since a network with a large average degree approximates the complete network, which represents a well-mixed population, which for ρ > 1 has p ¯ ∞ = 0. We note that, as for the donation game, the transition from total cooperation to total defection is sharpest for random regular networks and slightly less sharp for exponential and scale-free networks. We also note that, as for the donation and snowdrift games, the cooperation index Λ is highest for random regular networks and lowest for scale-free network, with exponential networks having an intermediate value of Λ.
Fig 8(a) and 8(b) show how the cooperation index Λ on scale-free networks with average degrees k = 4 and k = 8 varies with clustering coefficient C (a) and assortativity coefficient r (b). As in the case of the donation and snowdrift games, Λ is more or less unaffected by clustering, whereas Λ is higher for significantly assortative networks than for networks that exhibit little to no assortativity. For networks of average degree 4 there is also a slight increase in Λ for disassortative networks compared to those with little or no assortativity.
Fig 9(a)–9(d) show the variation of p ¯ ∞ with ρ on empirical networks: Γhepth (a), Γastro (b), Γfb (c), and Γinternet (d). The inset in the figures shows the value of the cooperation index Λ. The value of Λ for the Γhep and Γinternet networks is significantly greater than the value in a well-mixed population, while the value of Λ for the Γastro and Γfb networks is only slightly higher than the value in a well-mixed population. The transition from cooperation to defection on these empirical networks appears also to be governed to a reasonable approximation by the same condition that was found to hold for model networks, namely cooperation is only favored, for ρ > 1, if 1 ρ - 1 > k - 1 (indicated by a dashed line in the figures). Since it follows from this relation that Λ is given approximately by Λ = 1 k - 1 + 1 2, the higher values of Λ found for Γhepth and Γinternet and the lower values of Λ found for Γastro and Γfb are consistent with the results that would be expected given the lower average degrees of the former two networks and the higher average degrees of the latter two networks.
In this paper we have studied the evolution of cooperation in social dilemmas in network-structured populations. The social dilemmas that we have focused on—the donation, snowdrift, and sculling games—are exemplars of the fundamental types of social dilemmas represented by symmetric 2 × 2 games, namely, the prisoner’s dilemma, hawk-dove, and coordination classes of games. Here we have investigated the effect of the structure of the network on which the population is modeled on the maintenance of cooperative behavior in each of these three exemplars of social dilemmas. This study yields several conclusions that extend previously known results in a number of significant directions. Interestingly, the results we obtain concerning the influence of network structure on the evolution of cooperation in each of the donation, snowdrift, and sculling games are all in good overall agreement and allow a number of general conclusions to be drawn. The most basic conclusion is that for all these social dilemmas the propensity for cooperation is increased on complex networks compared to a well-mixed population. This conclusion is generally well accepted for the donation game and other forms of the prisoner’s dilemma in network-structured populations [42, 43, 45, 47–49, 56]. For the case of the snowdrift game, this conclusion is consistent with the results found in [48, 98], and the inhibiting effect of spatial structure found in [46] appears to be limited to lattice networks. The evolution of cooperation in games of coordination class on networks has received relatively little attention, with the interesting exceptions of [32, 34, 35], and this paucity has motivated our introduction and study of the sculling game. Again for this game we find that network structure promotes the evolution of cooperation compared to a well-mixed population.
For the donation, snowdrift, and sculling games on networks the most basic network characteristic that affects the evolution of cooperation is the average degree of the network. For all these social dilemmas the level of cooperation decreases monotonically as the average degree of the network increases. This behavior is of course completely reasonable since as the average degree increases the network approximates more and more closely to the complete network, which represents a well-mixed population. For the donation and sculling games there are reasonably sharp transitions from cooperation to defection determined by the average degree k and the cost-to-benefit ratio ρ. In both cases the transition from cooperation to defection is apparently governed to a good approximation by simple relations—for the donation game cooperation only prevails if 1 ρ > k - 1 (with defection dominating otherwise) and for the sculling game cooperation only prevails, for ρ > 1, if 1 ρ - 1 > k - 1 (with defection dominating otherwise). These relations hold to a good approximation on a variety of model networks and also to a reasonable approximation on a range of empirical networks. We observe that among model networks the relations are most exactly obeyed for both games on random regular networks and least exactly obeyed on scale-free networks, which suggests that the accuracy with which these relations govern the evolution of cooperation in both games decreases as the variance in the degree distribution of the network increases. It is also the case that the overall level of cooperation in the donation and sculling games, as measured by the Λ-index, decreases as the variance of the degree distribution of the network increases. The relation we find governing the evolution of cooperation in the donation game (i.e., 1 ρ > k - 1) is clearly analogous to those found in [49, 98] for different update procedures. For example, it was found in [49, 98] that cooperation is favored with death-birth updating only if 1 ρ > k and with imitation updating only if 1 ρ > k + 2. Thus, we note that the Fermi update procedure we have used favors the evolution of cooperation at lower values of the average degree than the death-birth or imitation updating used in [49, 98]. The evolution of cooperation in the snowdrift game on complex networks does not display the sharp transition from cooperation to defection found in the donation and sculling games, but the same general pattern of decreasing cooperation with increasing k and/or ρ still holds, as does the decrease in cooperation as the variance in the degree distribution increases.
The general form of the dependence of the level of cooperation in all three games on the mean degree of the network can be understood by recognizing that the key determinant of the level of cooperation that is maintained in a game on a network is the extent to which the network structure results in assortative interactions occurring between the strategies used by the individuals in the population—that is, the extent to which the network structure results in individuals interacting more often with their own strategy type than would occur in a randomly interacting population. It is clear that the degree to which assortative interactions will be induced in a network structured population will increase as the average number of neighbors in the population decreases. Thus, the decrease in the levels of cooperation in all three games with increasing mean network degree is a consequence of the decreasing degree of strategic assortativity that arises as the average degree increases. This point of view also suggests a possible explanation for our finding that for all three games the level of cooperation that evolves decreases as the heterogeneity of the degree distribution of the network increases. Networks with highly heterogeneous degree distributions will have some vertices with very large degrees, and by definition these high degree vertices will be the neighbors of many other vertices in the network. However, high degree vertices will have low levels of assortative interactions which will result in them generally being less able to support cooperative strategies, and this in turn implies that many other vertices in the network will have such uncooperative individuals as their neighbors, which will inhibit the formation of clusters of cooperators and thereby reduce the overall level of cooperation.
We have also investigated the effect of the network properties of clustering and assortativity on the maintenance of cooperation in the donation, snowdrift and sculling games. The results results are again largely consistent for all three games and allow some general conclusions to be drawn. One very clear finding is that for all three social dilemmas the clustering coefficient of the network has no appreciable effect on the level of cooperation that is maintained. This is at first sight surprising as it might seem a priori that the increased local density of short closed paths that exist in networks with higher clustering coefficient could facilitate the emergence of local clusters of cooperators and thus increase the level of cooperation in such networks [56, 99]. A possible explanation for the result that an increased clustering coefficient does not lead to an increase in cooperation may be that while an increased clustering coefficient results in a high local density of short closed paths it necessarily also results in there being fewer edges between these dense local clusters [100], and this sparsity of connecting edges may inhibit the formation of large clusters of cooperators which are necessary to produce high levels of assortative interactions, thereby counteracting at a global level what may be a positive local effect of network clustering. The further study of the mechanisms through which clustering affects the evolution of cooperation in these three games seems to be an interesting topic for future research.
Another rather clear conclusion that we can draw from this study is that higher network assortativity results in higher levels of cooperation in all three games. A possible explanation of this phenomenon is suggested by the relation between the degree of vertices and the extent to which they can support assortative strategic interactions that was mentioned above. In networks that are degree assortative low degree vertices are preferentially connected to other low degree vertices. This results in a connected cluster of low degree vertices, which by this very fact, is able to maintain high levels of assortative strategic interactions, and therefore also maintain high levels of cooperation. It is interesting to note that according to the point of view that we are advocating, the evolution of cooperation in a network structured population is driven by the low degree vertices, and therefore mechanisms that serve to segregate the low degree vertices may be anticipated to have a positive effect on the level of cooperation that can be maintained. The positive effect of network assortativity on the level of cooperation that evolves in all three social dilemmas in network structured populations is intriguing in view of the fact that social networks are almost always significantly assortative. It appears, therefore, that real-world social networks possess an important structural property which promotes the level of cooperation that evolves in social dilemmas on these networks. While it may be that the underlying causes of assortativity in empirical social networks are unrelated to the effect that it has on the evolution of cooperation, it is nevertheless interesting to speculate that an important causative factor may be between group selection acting on social dilemmas in network structured populations, which favors those populations with higher levels of cooperation, and thus also favors those social networks with higher assortativity. The further investigation of the relation between network assortativity and the evolution of cooperation in social dilemmas in network structured populations is an interesting and potentially important topic for future study.
Obtaining analytical insights into the results we have established here using individual-based simulations seems to be a natural topic for further investigation. While it is beyond the scope of the present paper to discuss analytical methods in any detail it nevertheless seems worthwhile to comment briefly on possible analytical developments. First, it should be noted that no very general theoretical results are to be expected for frequency dependent systems, such as the social dilemmas we have studied here, on arbitrary networks [101]. However, it seems likely that analytical methods can be successfully applied to the problems studied here in special cases. In particular, it seems reasonable to anticipate that the conditions we have found that cooperation is favored in the donation game only if 1 ρ - 1 > k - 1 and cooperation is favored in the sculling game, for ρ > 1, only if 1 ρ - 1 > k - 1, can be established analytically in the weak selection limit β → 0 on Cayley trees using pair approximation techniques as in [49, 98]. This would require extending the pair approximation methods used in [49, 98] to apply to the form of the update rule used here. Obtaining analytical results for the dependence of the cooperation index Λ on the clustering coefficient or assortativity coefficient appears to be considerably more challenging theoretically. This is a consequence of the fact that networks with clustering or assortativity will contain many short closed paths which violate the key property required for the pair approximation to be accurate [66]. Thus, it seems likely that new analytical techniques will have to be developed in order to understand the dependance of the cooperation index on clustering and assortativity. The development of such methods may present an exciting challenge for future research.
Finally, we remark that a potentially important avenue for further research may be to understand at a deeper level the intriguing and unexpected parallel that has emerged between the donation game and the sculling game on networks, and to explain why the conditions for cooperation to evolve appear to be so similar for two games which are conceptually quite distinct.
|
10.1371/journal.ppat.1000023 | Cationic Amino Acid Transporter-2 Regulates Immunity by Modulating Arginase Activity | Cationic amino acid transporters (CAT) are important regulators of NOS2 and ARG1 activity because they regulate L-arginine availability. However, their role in the development of Th1/Th2 effector functions following infection has not been investigated. Here we dissect the function of CAT2 by studying two infectious disease models characterized by the development of polarized Th1 or Th2-type responses. We show that CAT2−/− mice are significantly more susceptible to the Th1-inducing pathogen Toxoplasma gondii. Although T. gondii infected CAT2−/− mice developed stronger IFN-γ responses, nitric oxide (NO) production was significantly impaired, which contributed to their enhanced susceptibility. In contrast, CAT2−/− mice infected with the Th2-inducing pathogen Schistosoma mansoni displayed no change in susceptibility to infection, although they succumbed to schistosomiasis at an accelerated rate. Granuloma formation and fibrosis, pathological features regulated by Th2 cytokines, were also exacerbated even though their Th2 response was reduced. Finally, while IL-13 blockade was highly efficacious in wild-type mice, the development of fibrosis in CAT2−/− mice was largely IL-13-independent. Instead, the exacerbated pathology was associated with increased arginase activity in fibroblasts and alternatively activated macrophages, both in vitro and in vivo. Thus, by controlling NOS2 and arginase activity, CAT2 functions as a potent regulator of immunity.
| Recent studies conducted with amino transporter Slc7a2-deficient mice (CAT2) demonstrated that NOS2 activity in macrophages is regulated by CAT2. NOS2, which synthesizes nitric oxide, regulates numerous important activities, including resistance to infectious organisms, tumor development, and autoimmune diseases. It also competes with the enzyme Arginase-1 (Arg1) for the common substrate L-arginine. However, the role CAT2 in the regulation of Arg1 activity has not been previously examined. Therefore, we infected CAT2-deficient mice with the helminth parasite Schistosoma mansoni or with the protozoan pathogen Toxoplasma gondii, two organisms that trigger highly divergent host immune responses. Strikingly, following infection with S. mansoni, CAT2−/− mice developed parasite egg–induced lesions in the liver that were 3 to 4 times larger than wild type and hepatic fibrosis (a feature of severe schistosomiasis) was exacerbated, indicating a general worsening of disease in the absence of CAT2. The CAT2−/− mice were also more susceptible to T. gondii infection, demonstrating that CAT2 is critical for the development of protective cell-mediated immunity. Thus, these studies identify CAT2 as a powerful regulator of host immune responses, which may have major implications for a variety of infectious, inflammatory, and autoimmune diseases.
| Tissue macrophages comprise a heterogeneous population of cells, recently separated into three major categories based on their unique functional capabilities. The TH2 cytokines IL-4 and IL-13 trigger a characteristic ‘alternative’ state of activation in macrophages that is distinct from the ‘classical’ TH1-type activation by IFN-γ or deactivation phenotype associated with IL-10 and TGF-β [1]. In contrast to classically activated macrophages (CAMø), which regulate cellular immunity to intracellular pathogens, alternatively-activated macrophages (AAMø) are associated with chronic helminth infections and allergic disease [2],[3],[4]. AAMø's participate in humoral immune responses, facilitate clearance and presentation of antigens, and regulate the important process of tissue repair [1],[5].
In the murine model of schistosomiasis, mice chronically infected with Schistosoma mansoni develop severe liver pathology characterized by the formation of eosinophil-rich granulomas and fibrosis, which leads to portal hypertension, bleeding from collateral vessels, and ultimately death [6]. As with many helminth infections the immune response to S. mansoni is TH2-biased [7]. Consequently, AAMø's are the major macrophage subpopulation observed in schistosomiasis [8],[9], with recent studies suggesting their development is critical to the long-term survival of the infected host [10]. Although the exact role of AAMø's in inflammation and fibrosis remains unclear, numerous studies including our own have suggested they are important regulators of wound healing. This hypothesis is based on the observation that AAMø's express a number of genes known to be involved in cell proliferation and collagen synthesis, most prominent being the enzyme arginase-1 (Arg-1) [2],[9],[11],[12],[13],[14].
In contrast to iNOS, a widely investigated enzyme critically involved in many aspects of host immunity [15], much less is known about the role of arginase in infectious disease models [16]. Although it is known that arginases can antagonize NO synthesis by competing for L-arginine [17],[18],[19], the inducible Arg-1 isoform is believed to regulate other important functions as well. One of the major products of arginase is L-ornithine, a precursor in the production of polyamines and proline, which control cell proliferation and collagen production, respectively [16],[19]. It is thought that extracellular L-ornithine and L-proline, secreted from arginase expressing cells (AAMø's), are transported into fibroblasts, where they subsequently become incorporated into collagen [20]. Therefore, Arg-1 expressing cells have been hypothesized to be critical regulators of fibrosis. Thus, a better understanding of the mechanisms regulating Arg-1 activity could reveal novel strategies to control fibroproliferative diseases.
Since extracellular L-arginine is required for sustained NO and L-ornithine production [21], mechanisms controlling L-arginine transport may critically regulate iNOS and Arg-1 activity. Among the transport systems that facilitate L-arginine uptake, system y+ is considered to be the major L-arginine transporter in most cells and tissues [22]. Encoded by the solute carrier 7a1-3 (Slc7a1-3) family of genes, y+ is a Na+-independent high affinity amino acid transport system. CAT2 is the most dynamically regulated of the three transporters, with CAT1 operating as the product of a constitutively expressed “housekeeping” gene and CAT3 expressed primarily in the brain [23],[24]. Several pro-inflammatory mediators including LPS can regulate the expression of CAT2; thus, it likely functions as the key L-arginine transporter during inflammatory responses. Recent studies with CAT2-deficient mice showed sustained NO production in macrophages is dependent on CAT2 [25]. Thus, it appears to be the essential L-arginine transporter in macrophages. However, while CAT2 has been studied in the context of iNOS activity [25], no studies have addressed its role in the regulation of Arg-1 activity following infection.
To elucidate the function of the Slc7a2 gene in vivo, we infected CAT2−/− mice with either Schistosoma mansoni or Toxoplasma gondii; pathogens that induce highly polarized Th2 and Th1 responses, respectively [7]. Strikingly, following infection with S. mansoni, CAT2−/− mice developed granulomas that were 3- to 4-times larger than WT and hepatic fibrosis (a feature of severe disease) was significantly exacerbated in chronically infected mice [26],[27],[28],[29], indicating a general worsening of Th2-associated pathologies in the absence of CAT2. The CAT2−/− mice were also more susceptible to T. gondii infection, demonstrating that CAT2 is critical for the development of protective Th1-dependent immunity. Thus, these studies identify CAT2 as a powerful regulator of TH1 and TH2 effector responses, which may have major implications for a variety of infectious diseases.
To determine whether CAT2 plays a regulatory role during an acute Th2 response, we exploited the S. mansoni pulmonary granuloma model [30]. In this model, schistosome eggs are delivered to the lungs of mice via tail vein injection. The eggs are deposited in the pulmonary vasculature where they induce an eosinophil rich, CD4+ Th2 cell-dependent granulomatous response [7]. Wild-type (WT) and CAT2−/− mice were sensitized and challenged with 5000 live S. mansoni eggs and on day 4 and 7 post-challenge, animals were sacrificed and the effects of CAT2 deficiency were examined microscopically in the lung. Although there were no significant differences in granuloma size or composition on day 4, the CAT2−/− mice displayed an average 37% increase in granuloma size on day 7 (Fig. 1A), the peak of the granulomatous response [31]. The increase in peak granuloma size was also associated with significant picrosirius red staining of histological sections (Fig. 1B), providing evidence of increased fibrosis in the CAT2−/− lung.
To explore the role of CAT2 during a chronic Th2-driven inflammatory response, WT and CAT2−/− mice were infected with S. mansoni cercariae and the granulomatous response was examined in the liver 8, 12, and 24 weeks post-infection. As expected, peak granuloma size was observed at the acute time point (wk 8), with subsequent down-modulation in granuloma formation in chronically infected (wk 12–24) animals (Fig. 2A). When the responses in WT and CAT2−/− mice were compared, however, it was clear that the CAT2−/− mice developed granulomas 2- to 3-times larger than WT at all time points (Fig. 2A). There was also a small but significant increase in tissue eosinophils on wk 12 (Fig. 2B) and a consistent increase in mast cells in the CAT2−/− granulomas (Fig. 2C). The representative photomicrographs shown in panels 2D (WT) and 2E (CAT2−/−) illustrate the exacerbated inflammatory response in the CAT2−/− liver. When infected with a high dose of parasites, the CAT2−/− mice also succumbed significantly faster than WT animals (Fig. 2F). However, at low doses, the survival of CAT2−/− was not significantly different from WT through at least 24 wk of infection (not shown).
Liver fibrosis is the primary cause of chronic morbidity in S. mansoni infections [32]. To determine whether CAT2 regulates tissue fibrogenesis, liver tissue was taken at various time points post-infection and collagen content was measured as hydroxyproline [33]. Although both groups developed significant fibrosis, hydroxyproline levels were markedly increased in the CAT2−/− livers, particularly at the chronic time points (Fig. 3A, 3B). Collagen deposition was also examined histologically with Masson's trichrome (Figs. 3C and 3D) and picrosirius red stains (not shown), and thick bands of collagen were seen throughout the livers of the infected CAT2−/− mice. In contrast, collagen deposition was primarily in areas surrounding the granulomas in WT animals.
Serum AST (SGOT) and ALT (SGPT) levels were similarly increased in both groups following infection (Fig. 4A, 4B), indicating there was no evidence of significant egg-induced hepatotoxicity in the CAT2−/− animals. In fact, AST/ALT levels were slightly reduced in the CAT2−/− mice at the 8 wk time point. However, the CAT2−/− mice displayed significant hepatomegaly, particularly at the acute and early chronic time points (Fig. 4C). There was also marked splenomegaly in the absence of CAT2 (data not shown). Thus, in contrast to the enhanced liver toxicity observed in IL-4−/−, IL-4Rα−/−, and LysMCreIL-4R−/− mice [10],[29],[34], CAT2−/− mice developed significant liver fibrosis, portal hypertension, and collateral vessels, which are features of severe hepatosplenic disease.
Importantly, the increased pathological responses in the CAT2−/− mice were not attributed to differences in parasite burden since similar numbers of eggs and paired adult parasites were found in the tissues of both groups at all time points (Table S1).
Granuloma formation and fibrosis are tightly controlled by the egg-induced Th2 response [5],[7]. Therefore, to determine whether local or systemic changes in Th2 cytokine production were responsible for the severe pathological reactions in CAT2−/− mice, granuloma-associated lymphocytes were isolated from the livers of individual mice (wk 8) and IL-5, IL-13, and IFN-γ production was assayed by intracellular cytokine staining (ICS). Surprisingly, despite displaying a significant increase in pathology, the frequency of IL-5 and IL-13-producing CD4+ T cells was markedly reduced in the livers of infected CAT2−/− mice (Fig. 5A). IL-13 production was also reduced in the non-CD4+ T cell population (Fig. 5B). The reduction in type-2 cytokines did not result from an increased type-1 response because the frequency of IFN-γ producing cells (CD4+ and CD4− cells) was also reduced in the CAT2−/− livers, but not to the same magnitude as the type-2 cytokine producing cells.
We also isolated RNA from the liver and examined IFN-γ, IL-5, IL-13, IL-4, and IL-10 mRNA responses by real-time PCR at 8, 12, and 24 wk (Fig. 5C). As expected [35], there was a marked increase in IL-4, IL-5, IL-10, IL-13, and IFN-γ mRNA in the livers of infected WT mice. However, consistent with the ICS results, IL-13 mRNA levels were significantly reduced in the CAT2−/− mice at all time points post-infection. Similar results were seen with IL-4, although IL-5 mRNA was only slightly reduced in the knockout mice. Also consistent with the ICS studies, IFN-γ mRNA expression was reduced in the CAT2−/− liver, but only significantly at the 8 wk time point. In contrast to the other cytokines, IL-10 mRNA levels increased to a similar extent in both groups at all time points post-infection. Together, these results indicate that CAT2 expression ensures maximal development of Th2 cytokine responses in vivo.
To explore mechanisms by which CAT2 regulates Th2 response development in vivo, we investigated whether the proliferation of cytokine-producing cells was affected by CAT2 deficiency. Purified lymphocytes isolated from the granulomatous livers (Fig. 6A) and mesenteric lymph nodes (Fig. 6B) were CFSE-labeled and stimulated polyclonally with ConA for 72 hr. Following stimulation, cells were assayed by intracellular cytokine staining for IFN-γ and IL-13, as markers of Th1 and Th2 effector cells, respectively. In the liver (Fig 6A), there was significant proliferation without additional ConA stimulation, indicating the presence of a large population of antigen-activated T cells in the granulomatous tissues of both WT and CAT2−/− mice. 34.7% of the lymphocytes in the unstimulated WT group were also producing IL-13, which increased to 45.1% after Con A stimulation. The majority of the cytokine producing cells were also proliferating (20.6% before and 33.1% after Con A stimulation), indicating the presence of a large population of effector Th2 cells in the infected WT liver. In contrast, only 21.1% of the CAT2−/− lymphocytes were producing IL-13 and no increase was observed after Con A stimulation. The CAT2−/− IL-13-producing cells were also proliferating at a much slower pace (8.5% before and 10.3% after ConA stimulation). Similar results were seen for IFN-γ (right panels), although in general there were more IL-13 than IFN-γ producers in the liver. The number of proliferating IFN-γ producing cells in WT liver was 24%, which decreased to less than 5% in CAT2−/− mice (Con A stimulated), demonstrating that both the proliferative and cytokine producing capabilities of granuloma-associated lymphocytes were diminished in the absence of CAT2. As expected, the frequency of cytokine producing cells was much lower in the MLN (Fig. 6B). Moreover, although the frequency of cytokine-producing cells increased following Con A treatment, there were no significant differences between the two groups, suggesting that the impaired cytokine and proliferative responses of CAT2−/− mice were restricted to the granulomatous tissues.
In addition to Th1/Th2 cytokines, we also examined whether FoxP3, IL-17, and TGF-β1 expression were altered in the infected CAT2−/− mice. In contrast to the marked effect observed on Th2 cytokine expression, however, granuloma-associated CD4+ T cells from CAT2−/− and WT mice displayed similar IL-17, FoxP3, and TGF-β1 responses. In fact, the Th17 response was weak when compared with the Th2 cytokine response. For example, at 9 wk post-infection, the percentage of CD4+ T cells that were IL-13 positive was 17.1% and 11.8% in WT and CAT2−/− mice, respectively, while only 0.19% and 0.2% were IL-17 positive. Although we observed significant FoxP3 expression in the liver, the responses in WT and CAT2−/− were again nearly identical, with 6.87% of WT and 6.19% of CAT2−/− CD4+ T cells expressing FoxP3. There were also no significant difference in TGF-β1 mRNA expression in the livers of infected WT and CAT2−/− mice (not shown).
Numerous studies have demonstrated that granuloma formation and hepatic fibrosis are dependent on Th2 cytokines [26],[29],[36]; therefore, it was surprising to find a markedly reduced Th2 cytokine response in the granulomatous tissues of the CAT2−/− mice, since immunopathology increased significantly in these animals. Because Arg-1 and iNOS activities are regulated by the availability of L-arginine [17] and alternatively-activated macrophages play an important role in the pathogenesis of schistosomiasis [9],[10], we examined whether CAT2 deficiency was regulating the function of alternatively-(AA) or classically- (CL)-activated macrophages. CAT2 mRNA levels were increased 20- to 40-fold in both classically and alternatively activated macrophages, suggesting that CAT2 activity is not restricted to a Th1- or Th2-polarized response (Fig. 7A). CAT2 mRNA levels were also increased over 10-fold in the granulomatous tissues of infected mice (data not shown). Interestingly, however, when nitric oxide and urea levels (a quantitative measure of arginase activity) were measured, the NO producing ability of macrophages was decreased in the absence of CAT2, regardless of the activation stimuli used (Fig. 7B). In marked contrast, urea production was significantly increased in alternatively-activated CAT2−/− macrophages (Fig. 7C). IL-4, IL-13, IL-21, and GM-CSF have all been shown to increase arginase activity in macrophages [13],[37],[38]. Interestingly, the CAT2−/− macrophages displayed enhanced arginase activity with nearly every stimulus examined (Fig. 7D). Finally, there were also significantly more macrophages in the CAT2−/− granulomas (Fig 7E), suggesting that the increase in granuloma size was due in part to the increased recruitment of macrophages to the liver.
Next we investigated fibroblast activity. For these studies, primary fibroblasts were generated from lung tissue and in initial studies, the production of NO and urea was compared in WT and CAT2−/− fibroblasts following classical or alternative activation. In contrast to classically activated macrophages, where NO expression was only partly CAT2 dependent (Fig. 7B), production of NO by CAT2−/− fibroblasts was almost entirely dependent on CAT2 activity (Fig. 8A). Nevertheless, the amount of NO produced by CL-activated fibroblasts was nearly ten-fold lower than macrophages plated at the same density (Figs 7B and 8A). Unlike macrophages, in which IFN-γ/LPS was strictly required for iNOS activity and IL-4/IL-13 for arginase activity (Fig. 7B and 7C), we observed significant spontaneous arginase activity in primary fibroblasts. Indeed, unstimulated WT fibroblasts (Fig. 8B) produced nearly the same amount of urea as alternatively-activated WT macrophages (Fig. 7C). There was also no evidence of enhanced arginase activity in fibroblasts following stimulation with IL-4, IL-13, IL-21, or GM-CSF (Fig. 8B). Most striking however, was the 4- to 5-fold increase in arginase activity in the CAT2−/− fibroblasts. The CAT2−/− fibroblasts also proliferated more rapidly, both spontaneously and in response to FGF stimulation (Fig. 8C). In addition, production of IL-6, a key cytokine in fibroblast proliferation and activation was also increased in the CAT2−/− fibroblasts, both at baseline and in response to IL-4/IL-13 stimulation (Fig. 8D). Consistent with these in vitro observations, we detected significantly more fibroblasts in CAT2−/− liver granulomas at both 8 and 12 wk post-infection (Fig. 8E). Finally, to provide evidence that alternative activation was increased in vivo, we injected WT and CAT2−/− mice intravenously with 5000 viable S. mansoni eggs and examined the expression of Arg1 and Retlna (RELM-α/Fizz1) mRNA in the lung at 4 and 7 days post-injection. As shown in Figure 8F, both genes associated with alternative activation were significantly upregulated in the CAT2−/− lung. Finally, we also stained liver sections from infected mice with antibodies to Arg1, alpha smooth muscle actin (α-SMA), and F4/80, to characterize the pattern of Arg1 expression in vivo. Consistent with the enhanced fibroblast activity observed in vitro, the CAT2−/− granulomas showed much greater staining for α-SMA, a marker of activated myofibroblasts. They also displayed much stronger staining for Arg1 and the overlay (purple staining) suggested that the majority of Arg1 was associated with myofibroblasts, with lesser staining observed in macrophages (Fig. 9).
The IL-13 receptor alpha 2 functions as a decoy receptor for IL-13 [39], and studies conducted with IL-13Rα2−/− mice demonstrated that the decoy receptor inhibits the development of hepatic fibrosis in schistosomiasis [40],[41]. Because fibroblasts are believed to be the key producers of sIL-13Rα2 and fibroblast function was altered in the absence of CAT2 (Fig. 8), we measured the circulating levels of IL-13Rα2 in infected CAT2−/− mice, since changes in IL-13Rα2 expression might be contributing to their exacerbated IL-13-associated pathologies. Surprisingly however, we found either similar, or at some time points, increased levels of sIL-13Rα2 in the infected CAT2−/− mice (Fig. 8G). Thus, despite displaying decreased IL-13 responses (Figs. 5–6) and increased IL-13 decoy receptor levels (Fig. 8G), the CAT2−/− mice developed an exacerbated fibrotic response.
Next, to determine whether the severe liver pathology in the infected CAT2−/− mice was in fact dependent on IL-13, we infected WT and CAT2−/− mice with S. mansoni and inhibited IL-13 with a neutralizing mAb. As expected [26], IL-13 blockade significantly decreased fibrosis in WT mice (Fig. 10A) without affecting the overall magnitude of the granulomatous response (Fig. 10B). Surprisingly however, IL-13 blockade was completely ineffective in CAT2−/− mice (Fig. 10A), suggesting that their fibrotic response was independent of IL-13 activity. Importantly, similar numbers of eggs and paired adult parasites were found in the tissues of all groups (Table S2).
In a final series of experiments, we investigated whether CAT2 is required for the development of Th1-dependent immunity, since NO production was impaired in CAT2−/− macrophages (Fig. 7B), as well as in fibroblasts (Fig. 8A). In these studies, the Toxoplasma gondii model was used, since resistance is known to be mediated by an IFN-γ and NO-dependent mechanism [42]. Initially, we examined whether susceptibility was altered in the CAT2−/− mice by monitoring host survival following infection with T. gondii. IFN-γ−/− and NOS2−/− mice were included as controls. As expected, WT mice were much more resistant than either IFN-γ−/− and NOS2−/− mice, with approximately 50% of the WT animals surviving through day 50 (Fig. 11A). In contrast, 100% of the IFN-γ−/− and NOS2−/− mice succumbed between days 7–10 post-infection, while CAT2−/− animals displayed an intermediate phenotype, with 100% mortality observed by day 42. We also infected WT and CAT2−/− mice and isolated peritoneal exudate cells (PECs) on day 7 to quantify the number of infected cells and to examine IFN-γ and NO responses ex vivo. Consistent with their enhanced susceptibility, the percentage of infected cells increased in the CAT2−/− mice (Fig. 11B). This was also associated with a significant increase in IFN-γ production, both at baseline and following stimulation with soluble T. gondii antigen (STAG). Nevertheless, despite displaying much stronger IFN-γ responses, production of NO was markedly decreased in the CAT2−/− PECs, which likely explains their enhanced susceptibility.
Previous studies have suggested that CAT2 controls NOS2 and arginase activity by regulating arginine transport into cells [43], however the relative importance of CAT2 in the development of Th1 and Th2 effector functions was not investigated. Here, we examined the role of CAT2 encoded by the Slc7a2 gene in vivo by studying two well-established infectious disease models, characterized by the development of either protective Th1- or pathogenic Th2-type immune responses [5],[42]. We found that CAT2-deficient mice were significantly more susceptible to the Th1-inducing pathogen T. gondii. The increased susceptibility was attributed to the attenuated NO response, which led to uncontrolled parasite replication. When CAT2−/− mice were challenged with the Th2-inducing pathogen S. mansoni, the animals developed significantly worse Th2-associated pathology, despite displaying weaker Th2 responses. Importantly, the pathological changes in the CAT2−/− mice were associated with increased arginase activity in fibroblasts and alternatively activated macrophages. These results reveal an essential role for CAT2 in the development of Th1 immunity. However, they also suggest that CAT2 functions as a potent negative regulator of Th2-associated pathology, most likely by limiting arginase activity in important effector cells like fibroblasts and macrophages.
NO production by iNOS contributes to normal cellular processes, resistance to intracellular pathogens, and pathophysiological conditions [23],[44]. MacLeod and colleagues found that CAT2 is induced coordinately with iNOS in numerous cell types and studies conducted with CAT2-deficient cells demonstrated that arginine uptake via CAT2 is required for sustained NO production in macrophages [25] and to a lesser extent in astrocytes [45]. However, NO synthesis in fibroblasts was only partially dependent on CAT2, suggesting that other compensating transporters can provide arginine for iNOS-mediated NO synthesis [46]. Thus, the dependence on CAT2-mediated L-arginine transport for NO production appears to vary in different cell types. Moreover, the relative importance of CAT2 in the development of NO-dependent immunity in vivo was previously unknown.
To evaluate the function of CAT2 in vivo, we infected CAT2−/− mice with the intracellular pathogen T. gondii. Resistance to T. gondii is mediated by an IFN-γ and NOS2-dependent mechanism [42]. Therefore, we compared CAT2−/− mice with IFN-γ- and NOS2-deficient animals, since they are known to rapidly succumb to T. gondii infection. Interestingly, despite developing a significantly stronger IFN-γ response (due to the higher parasite burdens), the CAT2−/− mice were much more susceptible to T. gondii, with all animals succumbing within 6 weeks. The increased susceptibility was associated with a markedly attenuated NO response, suggesting that CAT2 is critically important to the development of Th1-associated immunity. However, the fact that NOS2−/− mice died earlier than the CAT2−/− animals suggests that NO synthesis in vivo is only partly dependent on CAT2. This was consistent with the reduced but not completely ablated NO responses of CAT2−/− peritoneal exudate cells.
Since NOS2 and Arg-1 both require L-arginine as a substrate [1], we hypothesized that CAT2 might also regulate important Th2 effector functions. Recent in vitro studies with bone marrow-derived macrophages demonstrated that CAT2 is induced by both Th1- and Th2-type stimuli [43],[47], which was consistent with our observations. Moreover, studies conducted with macrophages showed that L-arginine transport is significantly impaired in the absence of CAT2, regardless of the stimuli used to activate the cells [47]. Thus, it was suggested that CAT2 regulates both the classical and alternative activation of macrophages [43]. Because Th2-driven alternative macrophage activation plays a critical role in the pathogenesis of schistosomiasis [9],[10], we investigated the function of the CAT2 gene in the murine model of schistosomiasis. Strikingly, although CAT2 deficiency did not affect the establishment of S. mansoni infection, Th2-associated pathology in the liver was exacerbated and the animals died at a significantly accelerated rate when compared with WT mice. Indeed, granuloma size increased more than 3-fold and development of hepatic fibrosis was exacerbated. Similar results were also obtained with the S. mansoni pulmonary granuloma model.
Because granuloma formation and fibrosis are driven by the Th2 cytokine response [7],[10],[26],[27],[29],[35],[36],[48], we initially examined whether CAT2−/− mice were developing stronger Th2 responses. Unexpectedly, despite displaying a significant increase in Th2-associated pathology [6], the frequency of cytokine-producing CD4+ Th2 cells was markedly reduced in the livers of the infected CAT2−/− mice (Fig. 5A). The granuloma associated CD4+ Th2 lymphocytes also proliferated less when restimulated in vitro. Together, these results indicate that CAT2 is required for the maximal development of Th2 responses. Thus, the severe pathological changes in the CAT2−/− mice were paradoxically associated with reduced rather than enhanced Th2 cytokine production.
Previous studies with IL-4Rα−/− and some IL-4-deficient mice demonstrated that development of the Th2 response is critical for survival in schistosomiasis, especially during the early stages of infection [10],[27],[49],[50]. In addition, recent studies with macrophage/neutrophil-specific IL-4Rα-deficient mice suggested that the development of alternatively activated macrophages, in particular, is critically important for host survival. [10]. Despite developing significantly weaker Th2 responses, however, the CAT2−/− mice showed no signs of increased susceptibility to S. mansoni when infected with a low dose of parasites, with all of the knockout animals successfully establishing chronic infections. Moreover, in contrast to infected IL-4Rα and LysM(Cre)IL-4Rα(−/flox) animals [10], the CAT2−/− mice did not default to a Th1-type immune response. They also displayed no evidence of significant hepatoxicity as determined by their serum AST/ALT responses. In fact, liver enzymes were slightly reduced in the CAT2−/− mice when compared with infected WT animals. These data, when combined with the histological findings discussed above, suggest that alternative macrophage activation is not significantly impaired in the infected CAT2−/− mice. In fact, evidence was obtained both in vitro and in vivo that alternative activation increased in the absence of CAT2.
To investigate this hypothesis further, we stimulated WT and CAT2−/− bone marrow-derived macrophages with cytokines that are known to promote alternative macrophage activation including, IL-4, IL-13, IL-21, and GM-CSF [13],[37],[38] and examined the induction of arginase activity, a key feature of AAMøs [1],[51],[52]. As expected, the Th2-associated cytokines triggered significant arginase activity in WT macrophages. However, the macrophages generated from CAT2−/− mice consistently displayed a markedly exaggerated response. These data demonstrate that CAT2 functions as a negative regulator of arginase activity in macrophages, which may in part explain their exacerbated fibrotic response. Thus, although it was recently suggested that CAT2 could regulate both the classical and alternative activation of macrophages [43], our combined in vitro and in vivo data indicate that the primary role of CAT2 is to optimize NO production in classically-activated macrophages, while limiting arginase activity in alternatively-activated cells. The maintenance of arginine transport by the constitutive arginine transporter CAT1 could explain the preservation of arginase activity in the CAT2−/− mice.
Although alternatively-activated macrophages are believed to be important regulators of wound healing and fibrosis [1],[5], fibroblasts are the primary collagen secreting cells. While it was previously shown that CAT2 has only minimal effect on NO production in classically-activated fibroblasts [46], the effects of CAT2-deficiency on arg1 activity was not examined. To determine whether CAT2 regulates fibroblast activation, we generated primary lung fibroblasts from WT and CAT2−/− mice and stimulated the cells with various Th1 or Th2-type stimuli. Surprisingly, although classically-activated fibroblasts were in general less potent producers of NO than macrophages, we observed a significant (>75%) reduction in NO production in CAT2−/− fibroblasts. In combination with earlier studies focused on embryonic fibroblasts that reported a minimal (<20%) effect on NO production [46] , our data suggest the dependence on CAT2 for NO synthesis varies in different fibroblast subpopulations.
We also examined the effects of CAT2 deficiency on arginase activity. In contrast to macrophages, however, where arginase activity was strictly dependent on Th2 cytokine stimulation, fibroblasts displayed no significant cytokine inducible arginase response, even when stimulated with an optimal combination of Th2-type cytokines [37],[38]. Nevertheless, when arginase activity in WT and CAT2−/− fibroblasts was compared, the CAT2−/− fibroblasts exhibited much greater arginase activity at baseline. The fibroblasts from CAT2−/− mice also proliferated faster and produced significantly more of the autocrine growth factor IL-6 [53], both before and after stimulation with IL-4 and IL-13 [54],[55],[56]. Thus, the enhanced arginase activity in CAT2−/− fibroblasts and AAMø's likely contributed to their exacerbated inflammatory and fibrotic responses following infection with S. mansoni. The increased arginase response may also explain the suppressed CD4+ Th2 cell responses in the granulomatous tissues. Because T cells, macrophages, and fibroblasts all compete for arginine, the increased arginase activity in CAT2−/− macrophages and fibroblasts may have reduced arginine levels in the granulomatous tissues, resulting in the local suppression of CD4+ T cell responses, as has been postulated recently in related studies [17],[57],[58]. In contrast to classically activated macrophages, AAMø's are also known to be inefficient stimulators of T cell proliferation [59], with F4/80+ alternatively activated macrophages functioning as potent inhibitors of antigen-specific CD4+ T cell proliferative responses in vivo [58].
Fibroblasts are also important producers of the soluble IL-13Rα2 [40],[60], which can function as a decoy receptor for IL-13 and was recently shown to inhibit the development of fibrosis in schistosomiasis [39],[40],[41]. Since CAT2 fibroblasts displayed an unusual activated phenotype, we examined whether production of the sIL-13Rα2 was also altered in the CAT2−/− mice. Decreased production of the sIL-13Rα2 could provide a simple and straightforward explanation for their exacerbated IL-13-driven pathological responses. Surprisingly, the exact opposite was observed. Indeed, serum levels of sIL-13Rα2 were either the same or slightly increased in the infected CAT2−/− mice. Thus, we questioned whether the development of fibrosis in CAT2−/− mice was in fact dependent on IL-13. To formally address this question, we performed a series of studies with neutralizing antibodies to IL-13 [61]. Strikingly, although IL-13 blockade had a highly significant anti-fibrotic effect in WT mice, CAT2−/− mice were unresponsive. These observations, when combined with the reduced IL-13 and enhanced IL-13Rα2 responses, suggest that the development of fibrosis in CAT2−/− mice is to a large extent IL-13-independent.
When viewed together, the data point to fibroblasts and AAMø's as the key mediators of the exacerbated pathological response, since arginase activity was increased in the CAT2−/− cells. In the case of fibroblasts, the enhanced arginase response also appeared to be independent of Th2 cytokine stimulation. Arg1 and α-SMA expression also colocalized in the granulomatous livers and both proteins were expressed at much higher levels in the infected CAT2−/− mice, confirming that there were more activated myofibroblasts. However, there was no evidence of spontaneous liver fibrosis in uninfected CAT2−/− mice, suggesting that a chronic inflammatory stimulus or some type of tissue damage was needed to initiate the fibroproliferative response. Nevertheless, a recent study found that CAT2-deficient mice are susceptible to the development of spontaneous inflammation in the lung [62]. The same group also showed that CAT2 expression is linked with the development of asthma [63]. Thus, mucosal tissues, which are repeatedly exposed to irritants, may be particularly sensitive to changes in CAT2 activity. As such, CAT2 may be involved in the regulation of a wide variety of diseases that are normally associated with chronic Th2 responses.
In conclusion, these studies demonstrate for the first time that CAT2 is critically important for the development of IFN-γ/NO-dependent immunity to the intracellular protozoan pathogen T. gondii. In addition, by inhibiting arginase activity in fibroblasts and alternatively-activated macrophages, CAT2 functions as a powerful negative regulator of type-2 cytokine-driven pathology. Thus, these findings may have major implications for a wide variety of infectious and inflammatory diseases.
Female C57BL/6 were obtained from Taconic Farms (Germantown, NY) [64]. Breeding pairs of C57BL/6 CAT2−/− mice were obtained from the UCSD Cancer Center (La Jolla, CA) [25]. Mice were housed under specific pathogen-free conditions at the National Institutes of Health in an American Association for the Accreditation of Laboratory Animal Care approved facility. The NIAID animal care and use committee approved all experimental procedures. S. mansoni eggs were extracted from the livers of infected mice (Biomedical Research Institute, Rockville, MD) as previously described [35]. For the induction of secondary granulomas, mice were sensitized intraperitoneally (i.p.) with 5000 live eggs, and then challenged with 5,000 live eggs i.v [31]. In the infection experiments, mice were infected percutaneously via the tail with 30–35 cercariae of a Puerto-Rican Strain of S. mansoni (NMRI) obtained from infected Biomphalaria glabrata snails (Biomedical Research Institute). Soluble egg Antigen (SEA) was obtained from purified and homogenized S. mansoni eggs [33]. All animals underwent perfusion at the time of sacrifice so that worm and tissue egg burdens could be determined [33]. 20 cysts of the avirulent ME49 strain were inoculated i.p into C57BL/6, CAT2−/− , NOS2−/− (Taconic), and IFN-γ−/− (Taconic) mice for morbundity studies. In some studies, mice were sacrificed day 7 post-inoculation and PECs were harvested and set up in culture for 24 and 48 hours in media and with soluble tachyzoite Ag (Stag), which was prepared as described [65].
The sizes of pulmonary and hepatic granulomas were determined on histological sections that were stained with Wright's Giemsa stain (Histopath of America, Clinton, MD). Approximately 30 granulomas per mouse were included in all analyses. A skilled pathologist evaluated the percentages of eosinophils, mast cells, and other types of cells in the same sections. The number of schistosome eggs in the liver and the gut and the collagen content of the liver, as measured by hydroxyproline levels, were determined as previously described [33]. Specifically, hepatic collagen was measure as hydroxyproline by the technique of Bergman and Loxley [66]. The increase in hepatic hydroxyproline was positively related to egg numbers in all experiments and hepatic collagen is reported as the increase above normal liver collagen in µmoles per 10,000 eggs; (infected liver collagen – normal liver collagen)/liver eggs × 10−4 or µmoles per worm pair. At late chronic time points, fibrosis is reported as total liver collagen per liver. The same individual scored all histological features and had no knowledge of the experimental design.
Mesenteric lymph nodes (MLN) and about 200 mg of granulomatous liver tissue was disrupted into single cell suspension by grinding through a 100 µm nylon mesh. The WBCs from liver cells were separated on a 34% percoll gradient (350 g for 20 min) (Fluka). MLN and Liver WBCs were treated with 2 ml of ACK lysis buffer (Quality Biological) for 2 min. Purified leukocytes were stained with 5 mM CFSE (Molecular Probes) for 5 min at RT. Excess CFSE was quenched by washing the cells in RPMI supplemented with 10% FBS. 3×106 cells were cultured in 24 well plates and were either left unstimulated or stimulated with 1 µg/ml of Con A for 72 hours. (note, the WBCs separated from the liver contain live S. mansoni eggs and therefore all liver leukocyte cultures are exposed to soluble egg antigens as well). ICC: Liver leukocytes either freshly isolated (ex vivo) or restimulated for 72 hrs were stimulated with PMA (10 ng/ml), Ionomycin (1 µg/ml) and BFA (10 µg/ml) (Sigma) for 3 hrs. Cells were surface stained for CD4 PE-Cy5 (BD Biosciences), fixed in 2% formaldehyde for 20 min at RT, permeabolized with 0.1% saponin buffer (Sigma) and stained for IFN-γ (APC or FITC), IL-5 APC (BD Biosciences) and IL-13 PE (Centocor) and acquired with FACS Calibur®. Data were analyzed in Flowjo® V8.
Lung and liver tissue samples were placed individually in 500 µl of RNAlater and frozen at −20°C (Ambion). Samples were removed from RNAlater and placed in 500 µl TRIzol reagent (Invitrogen) to purify RNA. Total RNA was further purified using RNeasy Mini Kit from Qiagen (Qiagen Sciences). RNA (1 µg) was reverse-transcribed using Superscript II (Invitrogen, Carlsbad, CA) and quantification of transcripts was performed using Applied Biosystems (Foster City, CA) pre-designed gene expression assays for IFN-γ, IL-5, IL-13, IL-4, and IL-10. Each Taqman assay was run in duplicate. For each sample 5 ul of a 1∶30 dilution of cDNA reaction cocktail in a 20 ul final volume TaqMan reaction was used for each assay. Reaction preparation and thermal cycling were carried out following the manufacturer's protocol with a modification of increasing qPCR cycles to 50. Assay samples were normalized to HPRT expression and compared to uninfected controls according to comparative CT method (Applied Biosystems).
Bone marrow was recovered from female C57BL/6 and CAT2−/− mice and cultured in Petri dishes (100 × 15 mm) containing supplemented DMEM media (20% L929 conditioned medium) for a period of 6 days. After six days cells were harvested and seeded at a concentration of 5 × 105 cells/well in 24 well plates containing supplemented DMEM media (10% FBS, 2 mM L-glutamine, 100 U/mL penicillin, and 100 ug/mL streptomycin). Cells were stimulated for 16 hr with combinations of IL-4, IL-13, and GMCSF (20 ng/ml), IFNγ (100 U/mL), or LPS (100 ug/mL)( Peprotech). In some cases, the cells were pretreated with IL-21 (R&D) for a period of 6 hr. Supernatants were collected for NO analysis and cells were lysed for arginase activity and RNA isolations. Real-time RT-PCR was performed on an ABI PRISM 7900HT Sequence Detection System (Applied Biosystems). Relative quantities of mRNA was determined using SYBR Green PCR Master Mix (Applied Biosystems) and by the comparative threshold cycle method. In this method, mRNA levels for each sample were normalized to hypoxanthine guanine phosphoribosyl transferase (HPRT) mRNA levels and then expressed as a relative increase or decrease compared with levels in media only controls. Primers were designed using Primer Express software (version 2.0; Applied Biosystems). Primers for CAT2-F CTC CTG GGT GCT CTG AAC CA and CAT2-R CTT CTC CCC TCC CGT TGA AC.
Whole lungs were harvested in cRPMI supplemented with 10% FBS (Hyclone), 2 mM- L-Glutamine, 100 µg/ml penicillin–streptomycin (Gibco), 50 uM ß-mercaptoethanol (Sigma), minced into small pieces, and exposed to Collagenase D (1 mg/ml) (Roche) and 4 U/ml DNase I (Sigma) for 40 mins at 37°C with shaking. Tissues were disrupted by straining through a 100 micron nylon mesh (BD Falcon). The single cell suspensions were plated in Iscove's Modified Dulbecco's Medium with 2 mM L-glutamine, 5% FCS, 25 mM HEPES, 100 µg/mL Streptomycin, 100 U/mL Penicillin, 50 uM 2-ME. (3 lungs plated on 3–100 × 15 mm Petri dishes). 50% of media was changed on day 7 and cells were recovered on day 14 by adding 4 mL of HyQtase (Hyclone) reagent for 20 min and rigorously pipetting repeatedly to remove cells. Cells were then cultured at 5 × 105 cells/well in 24 well plates. After activation with cytokines, supernatants were collected for NO analysis and IL-6 determination. Other cells were lysed to determine arginase activity or cultured for fibroblast proliferation.
IL-13Ra2 levels were determined by ELISA as previously described [41]. The concentration of IL-13Rα2 in the sample was determined from a serial-fold diluted standard of rmIL-13Rα2 Fc/chimera (R & D Systems). The sensitivity of the assay was approx. 98 pg/ml. IL-6 levels were measured using murine IL-6 DuoSet ELISA Development System (R&D Systems) according to the manufacturer's protocol. IFN-γ was assayed by sandwich ELISA, as previously described [41], and quantitated by comparison with standard curves generated with rIFN-γ (provided by Genentech, San Francisco, CA, and Genetics Institute, Cambridge, MA, respectively).
The concentration of nitrite in supernatants of primary lung fibroblasts and bone-marrow derived macrophages stimulated in vitro was determined spectrophotometrically by using the Griess reagent. Supernatants were collected after 16 hours, mixed 1/1 with Griess reagent, and absorbance measured at 543 nm using a SpectraMax 190 (Molecular Devices, Sunnyvale, CA). The nitrite concentration was determined using sodium nitrite as standard. In the arginase assays, cells were plated at 5 × 105 per well in 96 well tissue culture plates and stimulated with combinations of IL-4, IL-13, and IL-21 (20 ng/mL). IL-21 was added 6 hours prior to IL-4 or IL-13 stimulation. Following stimulation, cells were washed with PBS and lysed with 0.1% TritonX-100 containing protease inhibitor (Roche). Lysates were transferred into a 96 well PCR plate and incubated with 10 mM MnCl2 and 50 mM Tris HCl (pH 7.5) to activate enzyme for 10 min at 55°C. After enzyme activation, 25 µl of lysate was removed and added to 25 µl 0.5 M arginine (pH 9.7) in a new PCR plate and incubated for 1–2 hours at 37 C. 5 µl of each sample was added in duplicate to a 96 well ELISA plate along with 5 µl of each standard, diluted in same assay conditions, starting at 100 mg/dL. Urea determination reagent from BioAssay Systems Quantichrome Urea Assay Kit was used according to the manufacturer's protocol.
C57BL/6 (10/group) mice were infected percutaneously via the tail with 30–35 S. mansoni cercariae. Beginning on wk 5 post-infection, mice were treated with either mouse anti-IL-13 mAb [61] or GL113 control antibody (Harlan Bioproducts). Each mouse received one 0.5 mg dose/wk via i.p. injection on wk 5, 6, 7 and 8. Mice were sacrificed on wk 9 post-infection.
Acute tachyzoite growth was assessed using cytocentrifuge smears of peritoneal cells as previously described [65]. Differential analyses, including assessment of intracellular T. gondii infection, were performed on 700 or more cells per animal. Single-cell suspensions were prepared from peritoneal exudates and washed in CRPMI. Peritoneal cells were cultured at 2.5 × 106 cells/mL in 200 µl of RPMI 1640 supplemented with 10% FBS, penicillin (100 U/ml), streptomycin (100 mg/ml), L-glutamine (2 mM), HEPES (10 mM), and 2-ME (Sigma) in the presence or absence of soluble tachyzoite Ag (5 µg/ml). Supernatants were harvested 24 and 48 hr later for determination of levels of IFN-γ and NO.
WT and CAT2−/− primary lung fibroblasts (1 × 105/well) were plated in Iscove's Modified Dulbecco's Media in 96 well flat bottom plates (BD Falcon) for 16 hr at 37°C. 1 µCi/well of [3H]thymidine (Amersham) was added for 24–72 hrs. Cells were frozen at −20 C and were later collected onto a glass fiber filter pads (LKB Wallac, Turku, Finland) using a 96-well harvester (Tomtec, Orange, CT). Scintillation cocktail was (XSC/9200; LKB Wallac) added, and radioactivity determined on a liquid scintillation counter (Betaplate model 1205, LKB Wallac). Some cells were stimulated with FGFβ at 50 ng/ml (Invitrogen). Each sample was set up in triplicate.
8 um liver sections were taken from WT and CAT2−/− 10 wk S. mansoni infected livers. Sections were fixed in cold acetone for 10 minutes and stored at −20°C. Slides were washed 3× in DPBS and incubated with F4/80 conjugated with Alexa 488 (Caltag Clone CI:A3-1), Alpha Smooth Muscle Actin (Sigma-Aldrich Clone 1A4) conjugated with Texas Red, and Arginase1 (Santa Cruz Biotechnology Clone V-20) conjugated with Alexa 647. Images were collected on a Leica SP5 confocal microscope (Leica Microsystems, Exton, PA USA) using a 20× oil immersion objective NA 0.70. Fluorochromes were excited using an Argon laser for Alexa 488, an Orange Helium-Neon laser for Alexa 594 and a Red Helium-Neon laser for Alexa 647. To avoid possible crosstalk the wavelengths were collected separately and later merged. Images were processed using Leica LAS-AF software (version 1.7.0 build 1111).
Hepatic fibrosis (adjusted for egg number) decreases with increasing intensity of infection (worm pairs). Therefore, these variables were compared by analysis of covariance, using the logarithm of total liver eggs as the covariate and the logarithm of hydroxyproline content per egg. Variables that did not change with infection intensity were compared by one-way ANOVA or Student's t test [33]. Changes in cytokine mRNA expression and granuloma size were evaluated using ANOVA. Differences were considered significant when p<0.05. |
10.1371/journal.pntd.0004808 | Antibody-Mediated Neutralization of the Exotoxin Mycolactone, the Main Virulence Factor Produced by Mycobacterium ulcerans | Mycolactone, the macrolide exotoxin produced by Mycobacterium ulcerans, causes extensive tissue destruction by inducing apoptosis of host cells. In this study, we aimed at the production of antibodies that could neutralize the cytotoxic activities of mycolactone.
Using the B cell hybridoma technology, we generated a series of monoclonal antibodies with specificity for mycolactone from spleen cells of mice immunized with the protein conjugate of a truncated synthetic mycolactone derivative. L929 fibroblasts were used as a model system to investigate whether these antibodies can inhibit the biological effects of mycolactone. By measuring the metabolic activity of the fibroblasts, we found that anti-mycolactone mAbs can completely neutralize the cytotoxic activity of mycolactone.
The toxin neutralizing capacity of anti-mycolactone mAbs supports the concept of evaluating the macrolide toxin as vaccine target.
| Mycolactones A/B (also simply referred to as mycolactone) are macrolide exotoxins produced by Mycobacterium ulcerans, the causative agent of the neglected tropical skin disease Buruli ulcer (BU). The potent cytotoxic and anti-inflammatory activities of mycolactones lead to severe destruction of the subcutaneous fat tissue with minimal inflammation in the core of the lesion. Due to the lipid-like nature of mycolactones, the production of antibodies against these molecules has so far been unsuccessful. By using the classical approach of fusing immortal cells with spleen cells of mice immunized with a protein conjugate of a truncated non-toxic synthetic mycolactone, we were able to generate hybridoma cells producing mycolactone-specific monoclonal antibodies. Mammalian cell-based cytotoxicity assays demonstrated that these antibodies have neutralizing activity and can fully block the cytotoxic activity of mycolactone, resulting in survival of the target cells. These findings support the concept to target mycolactone by a vaccine. Furthermore, the anti-mycolactone antibodies may represent useful tools for BU diagnostics development.
| Mycobacterium ulcerans, the causative agent of the neglected tropical skin disease Buruli ulcer (BU), produces the exotoxin mycolactone, which is responsible for the formation of chronic necrotizing skin lesions [1, 2]. Early diagnosis followed by rapid initiation of the currently recommended treatment with rifampicin and streptomycin [3] is crucial to avoid massive tissue destruction and long-term disabilities.
Mycolactones consist of a 12-membered macrolide core structure, a short C-linked upper side chain (comprising C12–C20) and a longer C5-O-linked lower polyunsaturated acyl side chain. Mycolactones produced by different M. ulcerans lineages differ in the structure of the lower side chain, but are otherwise conserved [4]. For the lower side chain variations in length, the number of double bonds and the number and localization of hydroxyl groups have been described. While M. ulcerans strains may produce mixtures of several mycolactone species, the composition of these pools seems to be highly conserved for a particular M. ulcerans lineage [5]. Strains belonging to the genomically monomorphic classical African lineage produce the most toxic variant, mycolactone A/B [6]. This lineage is responsible for > 95% of the BU cases reported worldwide [7].
Mycolactone-deficient M. ulcerans mutants are less virulent and intradermal injection of the toxin in animal models is sufficient to induce the formation of BU-like lesions [4, 8]. Since mycolactones play such a central role in the pathogenesis of BU [1, 2, 9], they may represent a suitable target for both vaccine development and passive immunotherapy. However, due to their lipid-like nature, as well as their cytotoxic and immunosuppressive [10–12] activities, attempts to raise antibody responses against mycolactones have failed so far. While extracts from M. ulcerans cultures have been the only source of mycolactones for a long time, highly defined synthetic mycolactones have become available more recently [6, 13, 14], greatly facilitating experimental work with these compounds. Here, we generated mycolactone specific immune-sera and monoclonal antibodies (mAbs) by immunizing mice with a protein conjugate of a non-toxic synthetic truncated mycolactone derivative.
Animal experiments performed were approved by the animal welfare committee of the Canton of Basel (authorization number 2375) and were conducted in compliance with the Swiss Animal Welfare Act (TSchG), Animal Welfare Ordinance (TSchV), and the Animal Experimentation Ordinance (TVV).
Natural mycolactones A/B, C and F as well as variants PG-157, PG-165, PG-182 and core PG-119 were produced as described previously [6, 14]. The synthesis of biotin-conjugate PG-204 and the immunogen PG-203 will be described elsewhere. All compounds were HPLC-purified. For biological testing 0.5 mg/ml mycolactone stock solutions were made by adding cell culture grade DMSO (Sigma).
Extracts of mycolactone were prepared using mycobacterial pellets from cultures of strains isolated from lesions of Cameroonian BU patients (S1013, S1019, S1047). Briefly, 1 ml of a chloroform-methanol solution (2:1, v/v) was added to the individual pellets. Bacteria were resuspended by vortexing and lipids were extracted by incubating the samples shaking at RT for 2 h. Then, 200 μL ddH2O were added to induce a phase separation. After vigorous vortexing, the samples were centrifuged for 10 min and the lower organic phase was transferred to a fresh tube. Samples were placed in a Speedvac (Thermo Scientific) for complete drying. 50 μL acetone were added and after an additional spin for 10 min at maximal speed, the acetone-soluble lipid fraction containing mycolactones was separated from the precipitate, collected in fresh tubes and again dried using the Speedvac device. For biological testing, the acetone-soluble lipid fractions were resuspended in cell culture grade DMSO (Sigma).
2 mg of mycolactone PG-203 were coupled to 2 mg of BSA using the Imject EDC BSA spin kit (Pierce). NMRI mice (Harlan Laboratories) were immunized thrice in 3 week intervals by subcutaneous injection of 40 μg of the coupled product emulsified in Sigma adjuvant. Serum antibody titers against the biotinylated mycolactone derivative PG-204 were tested by ELISA. Based on the ELISA results, one NMRI mouse was selected to receive a final intravenous injection of 40 μg of the PG-203-BSA conjugate without adjuvant. Three days after this last booster dose, hybridoma cell lines were generated as described below.
Mice were euthanized and the spleen was aseptically removed. Spleen tissue was mashed and the cells were fused with PAI myeloma cells in the presence of PEG 1500 (Roche) [15]. Mother cell line culture supernatants were tested by ELISA for the presence of anti-mycolactone antibodies and positive lines were cloned by limiting dilution. mAbs were purified from hybridoma culture supernatants by affinity chromatography using a HiTrap Protein A HP column (GE Healthcare). Isotypes were determined by ELISA with isotype specific reagents (SouthernBiotech).
NeutrAvidin Coated High Capacity plates (Thermo Scientific) were incubated with 2 μg/ml biotinylated mycolactone PG-204 in PBS for 2 h at 37°C in the dark. The PG-204 solution was then replaced by SuperBlock T20 (TBS) blocking buffer (Thermo Scientific). Plates were incubated for 1 h at 37°C in the dark, then washed twice with PBS 0.05% Tween. Hybridoma supernatants were applied to the plates for 2 h at 37°C in the dark. Plates were washed four times with PBS 0.05% Tween. Bound anti-mycolactone antibodies were detected using an Alkaline Phosphatase (AP) conjugated goat anti-mouse IgG antibody (Sigma) diluted 1:20,000 in blocking buffer. The plate was incubated for 1 h at 37°C in the dark, and then washed four times with PBS 0.05% Tween. Development was done using the Alkaline Phosphatase Yellow (pNPP) Liquid Substrate System (Sigma). The absorbance was measured at 405 nm with a microplate reader (Tecan Sunrise) and the values were plotted against the concentration.
For competition experiments Maxisorp plates were coated with mAbs (10 μg/ml) o/n and then blocked with SuperBlock T20 (TBS) blocking buffer (Thermo Scientific) for 1 h at 37°C. After washing twice with PBS 0.05% Tween, serial dilutions of synthetic mycolactone compounds were added and incubated for 2 h at 37°C in the dark. DMSO served as negative control. Subsequently biotinylated PG-204 was added (100 pg/ml) without washing steps and incubated for an additional 30 min. Plates were washed four times with PBS 0.05% Tween. Bound PG-204 was detected with alkaline phosphatase coupled streptavidin (SouthernBiotech) after 45 min incubation at 37°C. The development was performed as described above.
Murine L929 fibroblasts were cultivated at 37°C and 5% CO2 in RPMI medium (Gibco) supplemented with 10% FCS (Sigma), 2 mM glutamine (Gibco) and 0.05 mM β-mercaptoethanol (Gibco). For the assay, L929 cells (24,000 per well) were seeded in 24-well plates (Falcon) and incubated o/n at 37°C. Medium was aliquoted and mixed with 15 ng/ml synthetic mycolactone A/B, mycolactone core PG-119 or varying amounts of mycolactone extracts. Dependent on the assay format, fixed or increasing concentrations of anti-mycolactone antibody or isotype-matching control antibody (JD4.1) were added and incubated for 10 min. The medium in the 24-well plates was aspirated and replaced by 500 μl medium containing the mycolactone-antibody mixes. After an incubation period of 48 h, Resazurin solution (Sigma) was added to the wells (10% v/v), and the cells were further incubated for 2 hours at 37°C and 5% CO2. In order to quantitatively assess metabolic activities, the fluorescence intensities were measured using a SpectraMax Gemini XS (Molecular Devices), and the obtained values were normalized for the DMSO control. The experiments were set up in duplicates and performed at least three times. Fluorescence intensities were plotted against the log concentration of neutralizing antibody.
SPR experiments were performed on Biacore 3000 or T200 instruments (GE Healthcare, Uppsala, Sweden) at 25°C. HEPES buffer (10 mM HEPES, 150 mM NaCl, 0.05% (w/v) Tween 20, pH 7.4) was used as running buffer and flow rates of 5 or 50 μl min-1 were used for the immobilization and binding assays, respectively. Anti-mycolactone mAbs were captured on the chip surface by primary antibodies. First, the primary antibody was immobilized covalently on the carboxymethyldextran surface of a CM5 sensor (GE Healthcare, Uppsala, Sweden). The carboxymethyl groups of the dextran sensor surface were activated for 600 s with a mixture of 0.2 M (3-dimethylaminopropyl)carbodiimide (EDC) and 0.05 M N-hydroxysuccinimide (NHS). Next, the primary antibody, a rat anti-mouse IgG1 mAb (SouthernBiotech, AL, USA) diluted in 10 mM borate buffer (pH 8.2) to a concentration of 20 μg/ml was contacted for 120 s with the activated sensor surface to achieve antibody densities of 8000–12000 RUs. Subsequently, remaining active ester molecules on the sensor were deactivated by applying 1M ethanolamine (pH 8.0) for 420 s. Anti-mycolactone mAbs were diluted in running buffer to a concentration of 200 nM and captured by the primary antibodies to achieve antibody densities in the range of 2500–4000 RUs.
Mycolactone derivative PG-204 was titrated over the surfaces with the immobilized anti-mycolactone mAbs. Association and dissociation phases were monitored for 240 and 3600 s, respectively. Bound mycolactone was washed out from the surface in regeneration steps (injection of 10 mM glycine, pH 2.2).
For the generation of mycolactone specific antibody responses we have immunized mice with PG-203, a truncated and non-cytotoxic mycolactone derivative coupled to BSA via a diethylene glycol-based linker replacing the C5-O-linked polyunsaturated acyl side chain (Fig 1A).
Mice had developed strong anti-mycolactone IgG responses after two immunizations with the adjuvanted protein carrier conjugate and a third immunization boosted the immune response further (Fig 2).
Spleen cells of mice immunized with the PG-203-BSA conjugate were used for the generation of hybridomas producing anti-mycolactone mAbs by conventional B cell hybridoma technology. Hybridoma cell culture supernatants were screened by ELISA using NeutrAvidin plates coated with PG-204 (Fig 1C), a biotinylated derivative of PG-203. Screening led to the identification of 12 hybridoma clones producing the anti-mycolactone mAbs JD5.1 to JD5.12. All twelve mAbs were of the IgG1 isotype and showed binding to PG-204, but not to PG-183 (Fig 3), a mycolactone derivative biotinylated at the C-linked short upper side chain (Fig 1D).
The fine specificity of mAbs was further characterized by competition ELISAs. Plates were coated with anti-mycolactone mAbs and the biotinylated mycolactone derivative PG-204 was added as reporter molecule. By titrating in increasing amounts of synthetic natural variants of mycolactone (mycolactone A/B, C and F), a truncated synthetic derivative (PG-119), or derivatives with modifications in the C-linked upper side chain (PG-157, PG-165 and PG-182) we were able to assess the fine specificity of the binding. All mAbs showed very similar fine specificity patterns (depicted for the prototype mAb JD5.1 in Fig 4). As predicted from the structure of the immunizing truncated mycolactone derivative PG-203, the presence/absence or detailed structure of the C5-O-linked polyunsaturated lower acyl side chain had no significant effect on binding to the mAbs. PG-119, a truncated mycolactone lacking the lower fatty acid acyl side chain was thus as efficient as the complete mycolactones A/B, C and F in competing with the reporter molecule PG-204 (Fig 4).
In contrast, modifications in the C-linked upper side chain had major effects on the efficiency as competitor. The more complex the modifications were, the less competition was observed (Fig 4), suggesting that the upper side chain along with the mycolactone core constitutes part of the epitope of the mAbs.
For all mAbs, addition of a terminal hydroxyl group in PG-165 reduced binding and a major extension of the upper side chain in PG-182 abrogated competition completely. Only for PG-157, a derivative with a slight extension of the upper side chain, a difference in the competition pattern was observed, with only mAbs 5.9 and 5.11 showing slight inhibition (the difference is depicted for the prototype mAbs JD5.1 and JD5.11 in Fig 5).
In Surface Plasmon Resonance (SPR) analyses the anti-mycolactone mAbs immobilized on the surface of the sensor chip showed very tight binding to the biotinylated mycolactone derivative PG-204. The association and dissociation rate constants could not be precisely determined as the dissociation rate constants were outside the limitation of the measuring technology (Fig 6). For some of the mAbs (JD5.1, JD5.2, JD5.5, JD5.8 and JD5.10) dissociation rates were extremely slow with the monitored dissociation rate constant outside the limitation of the instrument resolution (koff < 10−6 s-1). Also the other mAbs showed slow dissociation rates with an estimated koff approximating the resolution limit (koff between 10−5 s-1 and 10−6 s-1). Measurements with mycolactone A/B failed, most likely due to aggregation of the more hydrophobic complete toxin molecule.
To investigate whether mycolactone specific antibodies can neutralize the cytotoxic activity of the macrolide toxin, L929 fibroblasts were incubated with synthetic mycolactone A/B at the cytotoxic concentration of 15 ng/ml (20 nM) in the presence of serially diluted anti-mycolactone mAbs. After 48 h of incubation, the metabolic activity of the L929 cells was assessed by performing Resazurin-based assays. While all anti-mycolactone mAbs showed toxin-neutralizing activity, the molar ratio of antibody versus mycolactone required for neutralization varied (Fig 7). While several mAbs neutralized the toxin completely already at a molar ratio of 2.5, only partial inhibition was observed at a ratio of 12.5 for the least active mAb JD5.2 (Fig 7). The morphology of L929 cells treated with mycolactone and sufficient amounts of toxin-neutralizing antibody was indistinguishable from that of DMSO-treated control cells. In contrast, cells treated with no or insufficient amounts of antibody showed characteristic signs of apoptosis. The anti-mycolactone mAbs also neutralized the mycolactones produced by cultivated M. ulcerans bacteria (S1 Fig).
Competition with the truncated non-toxic mycolactone derivative PG-119 lacking the lower acyl side chain (Fig 1) reconfirmed specificity of the neutralizing activity. At a concentration of 40 ng/ml (86 nM), PG-119 completely abrogated the toxin neutralizing activity of mAb JD5.1 (Fig 8).
Here we describe for the first time the production of antibodies against mycolactone, the main virulence factor of M. ulcerans. So far, no antibodies against mycolactone were detected in mice or humans infected with M. ulcerans [16] and attempts by several groups to generate mycolactone-specific antisera or mAbs by immunization with protein carrier conjugates of chemically modified mycolactone extracted from M. ulcerans cultures have failed before. This may be related to residual cytotoxicity of the conjugates and killing of B cells that incorporated them after binding to their mycolactone specific surface immunoglobulins. With the availability of synthetic mycolactone derivatives it has become possible to both generate highly defined protein conjugates for immunization and to develop reliable assays for the detection of mycolactone specific antibodies. For immunization we have used here a carrier conjugate of a mycolactone derivative in which the C5-O-linked polyunsaturated acyl side chain was replaced by a diethylene glycol-based linker. Since the structure of the lower side chain is crucially important for the cytotoxic activity of mycolactone [6], this compound (PG-203) was expected to be non-cytotoxic, even if it were released from the carrier protein after massive uptake by mycolactone specific B cells.
Coupling of the synthetic mycolactone derivative to the carrier protein BSA ensured T cell help for the mycolactone specific B cells, leading to clonal expansion, affinity maturation and isotype switching. As a result high anti-mycolactone IgG titers were elicited in mice already by two immunizations with the adjuvanted PG-203 carrier conjugate. All mAbs generated from the immunized mice were of the IgG1 subclass and exhibited high affinity and specificity for mycolactone. Competition assays indicated that the epitope recognized includes elements of the upper short side chain that is C-linked to the core structure. As expected [6], structural variation or absence of the lower C5-O-linked lower polyunsaturated acyl side chain had no effect on antibody binding. All mAbs exhibited high affinity binding with very slow dissociation rates which did not permit exact determination of the affinity constant by SPR analyses.
Currently, there is no highly effective vaccine against the major mycobacterial diseases tuberculosis, leprosy and BU available. BCG, originally developed against tuberculosis, may offer only partial and short-lasting [17] or no [18] protection against BU. Attempts to develop a subunit vaccine [16, 19, 20] or a live vaccine based on mycolactone-deficient M. ulcerans [21] had limited success. All twelve mycolactone-specific mAbs generated here showed, albeit to a varying degree, the capacity to neutralize mycolactone and to rescue mammalian cells from apoptosis in an in vitro assay. This supports the concept to target mycolactone in BU vaccine design. Both prophylaxis and therapy with toxin-neutralizing antibodies and active immunization with toxoids are highly successful strategies for protection against pathogens such as diphtheria or tetanus bacteria that produce a toxin as key virulence factor.
M. ulcerans has evolved from a common ancestor with M. marinum by acquisition of a plasmid designated pMUM, which encodes the polyketide synthases required for mycolactone biosynthesis [22–24]. None of the other pathogenic mycobacteria produce a macrolide toxin, making mycolactone an excellent target for the development of a species specific diagnostic test. Such an assay would also have potential for the monitoring of treatment success by measuring mycolactone levels in fine needle aspirates from closed BU lesions or swab samples from ulcerative lesions. The availability of the mAbs described here is enabling the development of a competition assay for the quantification of mycolactone. If non-competing mAbs specific for the lower part of the core and the lower side chain can be generated, development of an antigen capture assay may become possible.
Taken together, the first successful generation of mycolactone specific antibodies described in this report will stimulate development of new tools for research and control of BU.
|
10.1371/journal.pcbi.1007319 | From Escherichia coli mutant 13C labeling data to a core kinetic model: A kinetic model parameterization pipeline | Kinetic models of metabolic networks offer the promise of quantitative phenotype prediction. The mechanistic characterization of enzyme catalyzed reactions allows for tracing the effect of perturbations in metabolite concentrations and reaction fluxes in response to genetic and environmental perturbation that are beyond the scope of stoichiometric models. In this study, we develop a two-step computational pipeline for the rapid parameterization of kinetic models of metabolic networks using a curated metabolic model and available 13C-labeling distributions under multiple genetic and environmental perturbations. The first step involves the elucidation of all intracellular fluxes in a core model of E. coli containing 74 reactions and 61 metabolites using 13C-Metabolic Flux Analysis (13C-MFA). Here, fluxes corresponding to the mid-exponential growth phase are elucidated for seven single gene deletion mutants from upper glycolysis, pentose phosphate pathway and the Entner-Doudoroff pathway. The computed flux ranges are then used to parameterize the same (i.e., k-ecoli74) core kinetic model for E. coli with 55 substrate-level regulations using the newly developed K-FIT parameterization algorithm. The K-FIT algorithm employs a combination of equation decomposition and iterative solution techniques to evaluate steady-state fluxes in response to genetic perturbations. k-ecoli74 predicted 86% of flux values for strains used during fitting within a single standard deviation of 13C-MFA estimated values. By performing both tasks using the same network, errors associated with lack of congruity between the two networks are avoided, allowing for seamless integration of data with model building. Product yield predictions and comparison with previously developed kinetic models indicate shifts in flux ranges and the presence or absence of mutant strains delivering flux towards pathways of interest from training data significantly impact predictive capabilities. Using this workflow, the impact of completeness of fluxomic datasets and the importance of specific genetic perturbations on uncertainties in kinetic parameter estimation are evaluated.
| Microbial production hosts are used for production of a wide range of commodity chemicals. Improving the conversion efficiency of microbial strains is critical to the economic viability and the continued push towards the use of environmentally neutral bioprocesses as a means for producing the chemicals society depends on. Metabolic models have played a key role in helping us to predict metabolic behavior in response to environmental and genetic perturbation that can maximize efficiency. Recently, kinetic models of metabolism have re-emerged as a means for characterizing metabolism, offering improvements over their stoichiometric counterparts in both the type of information that can be gleaned from them, and in prediction accuracy. Despite recent developments, a lack of raw experimental data needed for flux elucidation and, subsequently, kinetic parameterization, and high computation cost have prevented the development of a uniform workflow for construction of the most informative kinetic models. Here, we have incorporated raw 13C-isotopic labeling data and a computationally inexpensive parameterization algorithm into a kinetic parameterization pipeline to ensure that the resulting kinetic model (k-ecoli74) conforms to experimental data. We show how the use of an identical metabolic network for flux elucidation and kinetic parameterization influences predictive capabilities.
| The standardization and automation of genome characterization and editing techniques has been accompanied by a rapid increase in the number of prokaryotic and eukaryotic microbial organisms available for engineering for overproduction of target commodity metabolites. With annotated genomes and CRISPR-Cas toolboxes consolidated in organism-specific biofoundries [1–5], the demand for biologically robust genetic intervention strategies for target metabolite overproduction has also increased. This has created a need for a standardized computational workflow capable of reliably predicting phenotype based on genetic intervention strategies. Traditional stoichiometric models of metabolic networks and integer programming design algorithms such as OptKnock [6] have provided insight into metabolic state as a function of genetic perturbation. These tools provide information on how an organism may behave under a specific genetic condition. However, the types of information that can be gleaned from them is limited to what can be deduced from reaction flux distributions, and flux ranges predicted via stoichiometry-based models are generally broad and subject to variability based on a user-defined cellular objective (i.e. maximum butyric acid production [7], maximum biomass [8], MOMA [9]). In recent years kinetic models of metabolism have (re)-emerged as a promising modeling paradigm offering a number of advantages over their stoichiometric counterparts albeit with a much higher effort associated with their construction. Kinetic models incorporate the mechanistic details of enzyme catalyzed reactions in metabolic networks to characterize a metabolite concentration/reaction flux pair as a function of physiological state. Kinetic models developed to date have primarily focused on characterizing either metabolic pathway behavior [10–13] or core metabolic function [14–17], as the computational burden and data needs associated with parameter estimation has been a limiting factor in both the rate of kinetic model development and scale-up of metabolic network. A number of kinetic formalisms and parameterization methods have been used to characterize and predict the dynamic behavior of metabolic systems. Mass action [18–20], S-system [21, 22], and log-transformed kinetic [10, 23–25] models have used canonical kinetic rate expressions to describe enzyme-catalyzed reactions. A number of models have also used mechanistic or approximate mechanistic expressions to characterize behavior of metabolic pathways [11–13, 26] and central carbon metabolism [14, 16, 17]. Both gradient-based [27–29] and stochastic [14, 30, 31] optimization methods have been developed for in silico identification of optimal sets of kinetic parameters. However, probabilistic [12] and meta-heuristic [30] parameterization methods have been at the forefront of recent kinetic model development [12, 14–16, 32] to bypass the computational challenges arising from the nonconvexity of the constraints and interdependence of kinetic parameters. Off the shelf solvers are ill-equipped to address the kinetic parametrization problem for these reasons, as the non-linearity of algebraic equality constraints required to ensure conservation of mass makes finding even an initial feasible point challenging. Furthermore, evaluation of mutant strain metabolite concentration, enzyme level, and reaction flux requires integration of a system of ordinary differential equations that tends to be stiff and prone to failure. The ensemble modeling paradigm [30] was introduced to address these challenges, and incorporated mechanistic rate expressions. However, application of this method to large metabolic networks requires very significant computational resources for parametrization rendering follow-up analysis of parameter robustness and sensitivity analysis prohibitive. This is due to the costly integration step needed each time a new steady-state is evaluated and the many thousands of recombination operations needed for convergence due to the non-inclusion of gradient information to guide the search. Greene et al. [33] have demonstrated how conservation and stability analysis on kinetic models in an ensemble can significantly improve parameterization time in the ensemble modeling paradigm by reducing both the number of model evaluations required to parameterize a kinetic model and the time required for a single model evaluation. Their complete methodology, however, has been limited in application to toy networks and a core kinetic model. Lee et al. [34] have used first order partial derivatives with the ensemble modeling paradigm to characterize the robustness of synthetic metabolic pathways by perturbing Michaelis-Menten (Km) and maximum rate of reaction (Vmax) parameters across all models in an ensemble and determining the probability of failure. Their analysis, however, was limited to systems with less than 20 reactions and did not require experimental training data.
Kinetic model development has been further hindered by a lack of experimental datasets to use in parametrization. Fluxomic data (in the case of ensemble modeling paradigm) across a range of single or multiple gene knockout conditions, and coverage across the entire metabolic network considered in a kinetic model, is required to generate a set of kinetic parameters capable of predicting metabolic state for any given condition. Mechanistic microbial kinetic models have been developed for core [14, 17] and genome-wide [32] metabolism of E. coli as well as core metabolism for C. thermocellum [16], while canonical models have been shown to be scalable to genome scale size [18].
A prominent tool for characterizing reaction flux distribution in living cells is 13C metabolic flux analysis (13C-MFA) [35–41]. The workflow for 13C-MFA is carried out in experimental and computational stages. The experimental stage is performed by first introducing an isotopically labeled substrate to a growing cell culture. Then the labeling distribution of mass isotopmers of labeled metabolites produced by the cell is measured using gas chromatography-mass spectrometry [42], liquid chromatography-mass spectrometry [43], or nuclear magnetic resonance spectroscopy [44]. Proteinogenic amino acid fragments and metabolites from central carbon metabolism are prominently featured in isotopic labeling datasets [42, 43, 45]. The 13C-MFA computational workflow consists of a least-squares fitting problem, whereby a metabolic flux distribution is estimated by minimizing the variance weighted sum of squared residuals (SSR) between the experimentally measured isotopic labeling distribution and a predicted isotopic labeling distribution inferred via the estimated flux distribution [46]. Application of 13C-MFA has yielded quantitative core metabolic characterization of a plethora of prokaryotic and eukaryotic organisms and cell types [37, 47–52]. It has also shed light on flux redirection under genetically and environmentally perturbed conditions [53–55] and revealed previously unknown pathway activity usage [56, 57]. Elucidation of atom mappings for peripheral carbon pathways and more elegant methods for mapping carbon flow (i.e. the EMU framework [58]) has allowed for scale-up of 13C-MFA to the genome-scale in three organism: E. coli [40], Synechocystis PCC 6803 [59], and Synechococcus PCC 7923 [60].
In order to accelerate the emergence of kinetic metabolic models as a viable tool for use in microbial strain design, we have developed a pipeline for rapid kinetic parameterization. By coupling 13C-MFA and kinetic parameterization computational methods using the same metabolic network, we acknowledge the intrinsic dependence of kinetic modeling on the metabolic network and 13C-glucose labeling datasets used to elucidate the flux distributions required for kinetic parametrization. We also provide a customizable framework for generating kinetic models that are consistent with reported flux ranges and applicable to any microbial metabolic network for which a set of isotopic labeling data across multiple genetic or environmental conditions can be procured. Our workflow for rapidly generating kinetic models of metabolic networks was carried out in two phases: flux elucidation was carried out via 13C-MFA and kinetic model parameterization using the gradient-based K-FIT algorithm developed by Gopalakrishnan et al. [61].
K-FIT differs from previously developed elementary decomposition approaches to kinetic parameterization by optimizing the model on the space of wild-type enzyme fractions and reverse elementary fluxes. Net fluxes and concentrations for the mutant networks are then recovered based on an iterative decomposition approach. The inclusion of gradient information in K- FIT also allows for the direct assessment kinetic parameter sensitivities. Taking advantage of 13C labeling data available for E. coli generated using glucose feedstock labeled at the first two carbon positions ([1,2-13C]glucose, known to yield precise flux estimations in E. coli core metabolism by 13C-MFA when compared to other single-tracer experiments [62]), we have applied our seamless workflow to the development of a kinetic model of E. coli core metabolism. Our model can predict metabolite pool size and metabolic flux distribution, satisfies flux distributions for wild-type and seven single gene deletion mutants from upper glycolysis, PP pathway, and Entner-Doudoroff (ED) pathway under mid-exponential growth conditions, and recapitulates carbon uptake kinetics.
We elucidated flux distributions and 95% confidence ranges for wild-type and seven single gene deletion mutant strains of E. coli (Δfbp, Δedd, Δeda, Δpgi, Δrpe, Δzwf, and Δgnd). For strains Δfbp, Δedd, and Δeda, the flux distributions were similar to the wild-type strain, with statistically insignificant variations from the wild-type strain in glucose uptake rate. Strains Δpgi, Δrpe, Δzwf (glucose-6-phosphate dehydrogenase (G6PDH2r) knock-out), and Δgnd each exhibited flux redirections compared to the wild-type strain. Given the obtained flux datasets, we then parameterized a core kinetic model using a metabolic network identical to that used for flux elucidation with 74 reactions, 61 metabolites, and 55 substrate-level inhibitions. Although activation is a prevalent regulatory mechanism in metabolism [63], it was not included in the model due to the absence of complete cofactor balance and known inaccuracy of energy metabolism representation in core metabolism flux distributions. The use of identical metabolic networks for both flux elucidation and kinetic parameterization safeguards against information loss in the form of feasible flux distributions due to flux projection from a core model to a larger model stemming from incomplete atom mapping and stoichiometric information.
Kinetic parameterization time was reduced by approximately 80% over the ensemble modeling (EM) method employed by Khodayari et al. [14] from more than a week to 36 hours (real time, due to evaluation of locally unstable parameter sets) for a core kinetic model. The average parameterization time per random initialization during k-ecoli74 parameterization was approximately four hours. The model constructed in this study (k-ecoli74) predicted 86% of reaction fluxes within a single standard deviation (SD), 95% within two SDs, and 99% within three SDs of 13C-MFA estimated flux values for mutant strains used in fitting.
k-ecoli74 was validated, and its predictive capabilities tested by comparing product yields for seven metabolites produced by nine engineered strains with experimental yield values and those reported for the previously developed k-ecoli457 kinetic model [32]. k-ecoli74 predicted product yields well overall, significantly outperforming the predictive capabilities of k-ecoli457 for malate and acetate production by engineered strains. This was due to similarity in experimental conditions between the strain engineering studies and those used for 13C-labeling data generation in this study (i.e. glucose-rich batch culture, mid-exponential growth phase). Metabolites not included in k-ecoli74 were systematically overpredicted due to the use of central carbon metabolism drains as proxies for pathways not included in the model. For example, 2,3-butanediol was over-predicted due to the use of pyruvate dehydrogenase (PDH) as a proxy for the heterologous 2,3-butanediol synthesis pathway (not included in k-ecoli74).
When flux data generated using a simplified metabolic network was used to parameterize a kinetic model with a metabolic network identical to k-ecoli74, discrepancies in flux predictions were observed in amino acid metabolism. In particular, there were significant differences in both the magnitude and directionality of reactions in serine, glycine, threonine, and glutamate metabolism. A significant decrease in predictive capability was observed when the kinetic model parameterized using simplified flux dataset was used to predict metabolite yields for the nine validation strains. The model generated with the reduced flux set was only able to predict feasible flux distributions for five of nine validation strains tested, and four out of those five strains yielded predictions similar to or worse than k-ecoli74 predictions.
The method developed in this study provides a framework for constructing kinetic models of metabolic networks from experimental data that ensures all pathways with resolvable flux ranges are accounted for in parameter estimation, and carbon and energy balance are characterized as accurately as possible. In addition, the relative computational tractability of the kinetic parameterization method used in this approach allows for the a posteriori analyses on kinetic parameter identifiability and sensitivity. Application of kinetic parameterization pipeline developed in this study to any organism or metabolic network requires a set of 13C labeling data, an identical metabolic network for flux elucidation and kinetic parameterization, a 13C-MFA software package for flux elucidation, and the K-FIT algorithm for kinetic parametrization.
The developed workflow for kinetic model construction relies on identical metabolic networks for flux elucidation and kinetic parameterization. This circumvents the information loss in the form of feasible solutions associated with the projection of the core model flux distributions onto the genome scale metabolic model and allows for the seamless integration of biomass yield information on precursor pathway drains. A pictorial representation of the kinetic parameterization pipeline is presented in Fig 1. The steps for constructing a kinetic model using the pipeline are as follows: first the stoichiometric model (Fig 1, step 1B) and corresponding atom mapping model are assembled (Fig 1, step 1A). Then they are used for flux elucidation of wild-type and genetic mutant strains of the organism of interest via 13C-MFA from 13C-isotopic labeling data (Fig 1, step 2). Finally, using the constructed stoichiometric model, the elucidated flux ranges, and any substrate level regulatory events identified in literature or inferred via computational methods (e.g. SIMMER [41] or model-based identification [64]), the kinetic model is parameterized using the K-FIT kinetic parameterization algorithm (Fig 1, step 3).
The core metabolic network used for 13C-MFA in this study (Fig 2) contains 74 reactions and 61 metabolites. The metabolic network and atom mapping model developed by Leighty and Antoniewicz [42] was used as a basis with the addition of L-serine deaminase (SERD-L). Pyruvate kinase (PYK) was also allowed to carry reverse flux to account for the significant flux converting pyruvate to phosphoenolpyruvate by the terminal phosphotransferase in the PTS system observed in vivo [65] and phosphenolpyruvate synthase activity. Atom transitions for SERD-L were taken from the imEco726 genome-scale atom mapping model [40]. The network included glycolytic, pentose phosphate (PP) pathway, and tricarboxylic acid (TCA) cycle pathways, as well as anaplerotic and cataplerotic reactions, lumped amino acid synthesis pathways, glycine cleavage, energy metabolism, acetate metabolism, and a biomass sink reaction. The metabolic network used for kinetic parameterization included identical reactions to those used for 13C-MFA. However, the biomass sink reaction was decomposed into individual metabolite sinks for each biomass precursor. A total of 55 substrate level regulations on central carbon metabolism reactions curated from the BRENDA [66] and EcoCyc [67] databases were included in the kinetic model, and are depicted in Fig 3. Substrate level regulations included competitive, uncompetitive, and noncompetitive inhibition. The reactions, metabolites, allosteric regulations, and atom mapping model used in this study are provided in S4 File.
13C isotopic labeling datasets for wild-type and seven single gene deletion mutant strains with glucose feedstock labeled at the first two carbon positions (100% [1,2-13C] glucose) generated by Long and Antoniewicz [68] was used as input data for the kinetic parameterization pipeline. Mass isotopomer distributions for 22 metabolite fragments derived from 10 amino acids (alanine, glycine, valine, leucine, serine, threonine, phenylalanine, aspartate, glutamate, tyrosine) and two sugar phosphates (ribose 5-phosphate, glucose 6-phosphate) were included in each labeling dataset. The seven mutant strains with available 13C isotopic labeling data included pgi, fbp, zwf (glucose-6-phosphate dehydrogenase (G6PDH2r) knock-out), gnd, rpe, edd, and eda knockout strains. Fig 2 shows the location of reactions in upper glycolysis, PP pathway, and ED pathway inactivated by genetic knockouts in strains use for parameterization.
Metabolite yield data from a series of genetically engineered overproducing strains was procured from literature [69–78] and used for model validation and testing predictive capabilities under conditions not included in the training data. Model validation strains included both up and downregulation of central carbon metabolism reactions as well as genetic knockouts. Genetic perturbation strategies, metabolites whose yields were tested, and experimental yield values are listed in S9 File, and a visual representation of perturbation strategies are provided in Fig D in S3 File. Strains designed to overproduce malate, acetate, L-valine, naringenin, lactic acid, 2,3-butanediol, and glucaric acid were included in the validation set. The malate overproduction strain was characterized by a downregulation of phosphotransacetylase (PTAr) and upregulation of phosphoenolpyruvate carboxylase (PPC). The acetate overproduction strain was characterized by a downregulation of ribose-5-phosphate isomerase (RPI). Two naringenin overproducing strains were considered, one characterized by succinyl-CoA synthetase (SUCOAS) knockout and fumarase (FUM) downregulation, and the other malate dehydrogenase (MDH) knockout and SUCOAS downregulation. Two lactic acid overproduction strains were also included, one characterized by acetate kinase (ACKr) downregulation and the other by ACKr knockout. One 2,3-butanediol overproduction strain was characterized by PYK overexpression, and one glucaric acid overproduction strain was characterized by NAD transhydrogenase (NADTRHD) overexpression.
Flux elucidation and 95% confidence interval estimation was performed for wild-type and each of the seven mutant strains using [1,2-13C]glucose isotope tracer data. The atom mapping model assembled from the atom transitions gleaned from Leighty and Antoniewicz [42] and the imEco726 model [40] was used to construct the elementary metabolite unit (EMU) network. An EMU is a subset of carbon atoms of any metabolite included in the stoichiometric model, and the EMU network characterizes how these subsets of carbon travel through the reactions in the network. The EMU network allows for characterization of the mass isotopmoer distribution of each metabolite in the metabolic network based on the isotope labeling scheme of the substrate upon estimation of a steady-state flux distribution [58]. Strain-specific biomass composition and acetate yields determined by Long et al. [79] were used for flux fitting. EMU decomposition, flux elucidation, and confidence interval estimation were performed according to the procedure outlined by Gopalakrishnan and Maranas [40], and glucose uptake was normalized to 100 flux units as a basis for each fitting. A summary of the 13C-MFA computational procedure is provided in S1 File. In order to ensure the best flux distribution was selected for use in the kinetic parameterization procedure, 100 randomly initialized multi-starts were performed for each strain. The minimized 13C-MFA objective was the variance weighted SSR between experimentally measured mass isotopomer distributions and mass isotopomer distributions inferred using the EMU network and steady-state flux distribution. A solution was accepted only if the algorithm converged to that solution at least 50% of the time. This does not guarantee convergence to the true global minimum but provides a practical safeguard against accepting local minima as solutions.
Kinetic parameterization was performed using the flux distributions estimated via 13C-MFA for wild-type and all mutant strains as training data. The gradient-based K-FIT algorithm [61] was used to parameterize the kinetic model. The wild-type flux distribution was used to estimate a set of elementary kinetic parameters (i.e. a set of kinetic parameters satisfying the wild-type flux distribution was generated). This was done to ensure the set of elementary parameters corresponds to a feasible steady-state solution in the wild-type strain. The elementary kinetic parameters were then used to estimate mutant flux distributions, and calculate the variance weighted sum of squared residual error (SSR) between all 13C-MFA mutant flux distributions and mutant flux distributions predicted using the estimated elementary parameters. Kinetic parameters were updated using gradient-based optimization, and the process was repeated until a local minimum was reached. To ensure reactions carrying little flux with narrow flux ranges were not over-weighted in the objective function, standard deviation used for weighting of residual errors was defined as the maximum value of either 1.0, five percent of the corresponding flux value, or the standard deviation value calculated according to the 13C-MFA 95% confidence interval. A summary of the K-FIT optimization algorithm is included in S1 File. Due to the nonconvexity of the resultant optimization model, 500 randomly-initialized multi-starts were performed. A threshold for change in concentration of any metabolite with respect to time was set at 10−6 flux unit to ensure strict adherence to the pseudo-steady state assumption [80]. All fluxes used as training data were scaled by the ratio of absolute mutant glucose uptake rate to absolute wild-type glucose uptake rate.
Model acceptance criteria was based on the SSR value, flux distribution reproducibility, and ability to predict flux distributions for genetic conditions not used for parameterization. In order for a model to be selected as a best model, two criteria had to be satisfied: first, the model had to yield the lowest SSR with at least one other model yielding a local minimum SSR value within 10% of the best model’s value (to ensure a reproducible solution). Because elementary kinetic parameters are the product of two distinct optimization variables (reverse elementary reaction flux and enzyme complex fractional abundance), local minima with similar SSR values but different elementary parameters may exist. Selecting a model with SSR similar to other local minima (i.e. within 10% of optimal SSR value) allows for the assessment of the sensitivity of Km and Vmax parameters using models yielding similar flux distributions through comparison of parameters generated by different models with similar fitness to data. The second condition for model acceptance was the capability to estimate steady-state flux distributions under genetic conditions not used for parameterization [81].
A total of 896 elementary kinetic parameters and 78 inhibitor constants were estimated corresponding to the 74 reactions, 34 biomass precursor sink reactions, and 55 substrate-level regulations in the metabolic network. Fluxes corresponding to central carbon metabolism, amino acid synthesis and degradation, and biomass formation were fitted. Fructose bisphosphatase (FBP) and phosphofructokinase (PFK) fluxes were excluded from fitting due to unresolvability (i.e., very wide ranges) stemming from simplifications made to energy metabolism in the core model. Energy metabolism and nutrient uptake reactions were also disregarded in the fitting due to simplifications and unavoidable inaccuracy of energy metabolism fluxes due to the nature of core metabolism 13C-MFA [40]. A total of 94 fluxes were fitted per mutant strain. The best-fitting model across seven mutant strains that also exhibited model stability across all strains for which metabolite yields were predicted had a SSR of 338 and an average weighted squared residual per mutant reaction flux of 0.52 (SSR was calculated for the seven single gene deletion strains used for training. The nine non-inclusion strains used to validate the model and evaluate predictive capability were not included in the calculation of SSR or evaluation of model fitness). Two additional models at neighboring local minima yielded SSR values within 10% of the optimal SSR value. The average percent error for reactions whose SD was within 20% of the experimental flux value (210 of 665 reactions) was 5.7%. Of those reactions whose SD was greater than 20% of the experimental flux value (455 of 665 reactions), 91% of predicted values differed from the corresponding experimental value by less than 1 mmol/100 mmol wild-type glucose uptake, and the average deviation was 0.36 mmol/100 mmol wild-type glucose uptake. The contribution of each strain to overall lack of fitness is shown in Fig 4. Δeda was the worst fitting strain, and contributed 36% of the overall SSR, while Δedd contributed to 23% of the overall SSR. Δpgi contributed to 19% of the total SSR, and no other strain contributed more than 8% of the total SSR. The best fitting strain was Δzwf, and contributed only 2% of overall SSR. Fig 5 shows a comparison of model-predicted flux values and 13C-MFA-estimated flux values. The plotted data yielded a Pearson correlation coefficient of 0.997, indicating a strong positive correlation between model predictions and 13C-MFA values. No single flux in the metabolic network deviated from the 13C-MFA value by more than a single SD for more than three fitted strains. Five predicted fluxes deviated from 13C-MFA values by more than a single SD across three strains. Lower glycolytic reaction PDH deviated by more than a single SD in Δpgi, Δedd, and Δeda. Fig E in S3 File depicts the relative contribution of each reaction in each strain to the total SSR. Acetate exchange, deviated by more than a single SD in Δpgi, Δeda, and Δgnd, and TCA cycle reactions citrate synthase (CS) and aconitase (ACONT) deviated by more than a single SD in Δedd, Δeda, and Δfbp, while isocitrate dehydrogenase (ICDHyr) and alpha ketoglutarate dehydrogenase (AKGDH), deviated by more than a single SD in Δpgi, Δedd, and Δeda. Fig 6 shows the number of fluxes falling within one, two, three, or four SDs of 13C-MFA values across strains. The results indicate 86% of all fluxes fitted fell within a single SD, 96% fell within two SD, and 99% within three SDs of their corresponding 13C-MFA values.
Glycolytic reaction fluxes were underpredicted in Δpgi due to the misdirection of flux towards the biomass sink reaction for ribose 5-phosphate and through glycine cleavage (GLYCL) instead of through L-serine deaminase (SERD-L). k-ecoli74 predicted Δpgi succinyl-CoA synthetase (SUCOAS) flux in the opposite direction of the 13C-MFA flux. Because SUCOAS flux was in the positive direction in all other strains used for parameterization, the fitted kinetic parameters were incapable of delivering reverse flux through SUCOAS. Serine hydroxymethyltransferase (SHMT) and GLYCL flux were also overpredicted by more than one SD in Δpgi. k-ecoli74 underpredicted SERD-L flux, while serine synthesis flux was predicted higher than 13C-MFA flux to satisfy biomass precursor demand. GLYCL, therefore, served as a sink for carbon that should have been delivered back to glycolysis. Most Δpgi PP and ED pathway fluxes deviated from 13C-MFA values by less than a single SD. k-ecoli74 predicted a 98% reduction in atp concentration and a 76% reduction in nadh concentration (competitive inhibitors of G6PDH2r) in Δpgi relative to the wild-type strain. This caused an increase in enzyme available to catalyze the G6PDH2r reaction, and k-ecoli74 was, therefore, able to successfully re-direct the entirety of carbon flux through the PP pathway. Acetate exchange was a significant carbon sink in all strains except for Δpgi. The optimal set of kinetic parameters, therefore, were suited for delivering significant flux towards acetate secretion, and that flux was overpredicted in the Δpgi strain.
Due to ED pathway irreversibility, the effect of Δeda and Δedd knockouts on the predicted flux distributions were almost identical, despite having different 13C-MFA flux distributions when scaled by absolute glucose uptake rate. In Δedd, reactions whose predicted fluxes were greater than a single SD from the corresponding 13C-MFA value (and contributed most to SSR) were primarily found in glycolysis and the TCA cycle. Fluxes were overpredicted in both pathways. Δedd glucose uptake rate when scaled to 100 mmol of wild-type glucose uptake was 8.7 mmol/100 mmol wild-type glucose uptake lower than wild-type, thus fluxes were overpredicted. In the Δeda strain, TCA cycle reactions, acetate formation and excretion, and PDH contributed most to SSR. Scaled Δeda glucose uptake rate was 2.4 mmol/100 mmol wild-type glucose uptake greater than in the wild-type strain. Despite their differences in glucose uptake rates from the wild-type strain and each other, glucose uptake rate did not contribute significantly to SSR in either strain compared to the aforementioned reactions. This is because standard deviations for glucose uptake rate in both strains was large, ensuring glucose uptake was not a primary source of error.
In Δfbp, only TCA cycle predicted fluxes deviated from 13C-MFA values by more than a single SD. Specifically, upper TCA cycle fluxes were overpredicted by the model. Because CS flux was increased, ACONT flux was increased as well. Isocitrate lyase (ICL) showed increased activity compared to the experimental data, redirecting carbon flowing through the TCA cycle. Glutamate dehydrogenase (GLUDy) also had a higher predicted flux than the 13C-MFA value in order to drain the excess carbon. While FBP and PFK reactions were not used to fit the kinetic model because of their unresolvability across all other strains during 13C-MFA, leaving them out of the fitting did not have an impact on the resulting fitness of the Δfbp strain. FBP flux was fixed to zero in Δfbp, and the flux through PFK was constrained by the other reactions in the network producing and consuming f6p (glucose-6-phosphate isomerase (PGI), transaldolase (TALA), fructose bisphosphate aldolase (FBA)) to ensure conservation of mass. Because flux through these reactions fit the data well in Δfbp, so did PFK flux. This was confirmed upon comparison of PFK flux with 13C-MFA value (PFK residual error in the Δfbp strain was 0.2).
In Δrpe, Δzwf, and Δgnd, no more than one predicted flux deviated from 13C-MFA values by more than a single SD. In Δrpe, arginine synthesis flux was underpredicted, and was the only non-biomass reactions with predicted fluxes outside of a single SD from the 13C-MFA value. No predicted Δzwf fluxes deviated from the 13C-MFA flux values by more than a single SD. In Δgnd, only acetate exchange deviated from the 13C-MFA flux values by more than a single SD.
Changes in metabolite concentration due to genetic perturbation were also assessed. The scaled metabolite concentrations for all metabolites with corresponding wild-type experimental data across each mutant strain used for parameterization are compared to wild-type metabolite concentrations in Fig 7. Significant changes in concentration were observed across all strains except Δedd. The only notable pool size changes in Δedd was a general decrease in amino acid concentrations. In Δpgi, a significant decrease in metabolite pool sizes for most metabolites across the metabolic networks was observed. The only metabolites with increased pool sizes were aspartate (>1000%) and nadph (+76%). Δrpe also exhibited significant decreases in pool size for most metabolites in the network. The only increase in concentration relative to the wild-type strain was for aspartate (>1000%). In Δeda, 2-dehydro-3-deoxy-D-gluconate 6-phosphate (kdpg) concentration increased significantly (>1000%), and there was a general decrease in amino acid concentrations. In Δfbp, glycolytic intermediates glucose-6-phosphate (g6p) (>1000%) and glyceraldehyde-3-phosphate (g3p) (+350%) and PP pathway intermediates 6-phospho-D-gluconate (6pg) (>1000%), D-ribulose-5-phosphate (ru5p) (+486%), and D-xylulose-5-phosphate (x5p) (+506%) increased. A redistribution of amino acid pool sizes was also observed: aspartate (-72%), histidine (-72%), isoleucine (-99%), leucine (-73%), lysine (-73%), threonine (-94%), tryptophan (-71%), and valine (-83%) concentrations all decreased significantly, while arginine (>1000%), glutamine (+174%), methionine (+117%), phenylalanine (+190%), and tyrosine (+249%) concentration increased. In Δzwf, a general increase in the pool size of aspartate and aspartate-derived amino acids and a decrease in pyruvate and pyruvate derived amino acid pool sizes was observed. Histidine (>1000%) concentration also increased, as well as the pool sizes for glycolytic intermediates dihydroxyacetone phosphate (dhap) (688%), fructose 1,6-bisphosphate (fdp) (875%), g6p (457%), and g3p (212%). In Δgnd, both g6p (481%) and 6pg (+933%) concentrations increased significantly, causing ED pathway activity to increase compared to wild-type and all other mutant strains. PP pathway intermediates erythrose-4-phosphate (e4p) (-86%) and sedoheptulose-7-phsophate (s7p) (-90%) pools sizes decreased, and a general decrease in amino acid pool sizes was observed.
While the majority of metabolite concentration values estimated by k-ecoli74 fell within ranges consistent with experimental data [85], the loose constraints placed on optimization variables (enzyme complex fractional abundance (1*e−3 < [e] < 1) and reverse reaction flux (0 < vr < 10000)) allowed elementary kinetic parameters to assume a broad range of values (0 < k < 1000000). As a result, mutant strain metabolite fold changes also had the potential to assume large values upon steady-state evaluation that were not always physiologically relevant. This occured when the magnitudes of the estimated forward and reverse elementary kinetic parameters for any elementary step in the network were different by several orders of magnitude. The fold changes predicted for aspartate in the Δpgi, Δrpe, and Δzwf strain reflected this, as the elementary parameters for production and consumption of aspartate by ASPTA differ by several orders of magnitude. Small ASPTA free enzyme fractional abundance in those strains relative to other enzyme complexes ensured aspartate concentration had to be large to recapitulate flux. Another reason for large metabolite concentrations was to help to characterize flux redirections in mutant strains. In Δgnd, 6pg and kdpg concentrations were increased to values >1000% of the wild-type concentration to help characterize the increased ED pathway flux that was dependent upon their concentrations, and only observed in the Δgnd strain. In Δpgi, extracellular glucose concentration was increased because glycolysis was blocked and growth rate was low, causing the model to predict an accumulation of glucose. In other instances, large metabolite concentrations were an artifact of the assumptions made in the modeling framework. kdpg was only produced and consumed in the ED pathway, and was not an inhibitor of any reactions. As a result, in the Δedd and Δeda knockout strains, because only EDD or EDA enzyme level was forced to zero, kdpg was able to assume large values as long as the net flux through the reaction that was not knocked out in the pathway was zero. These predictions are a limitation of the elementary decomposition approach for kinetic parameterization in the absence of metabolomics training data, as the mathematical framework of K-FIT allows for flexibility in metabolite concentration predictions when only fluxes are fitted.
To test how using the kinetic parameterization pipeline changes training data recapitulation and impacts kinetic model predictive capabilities, we parameterized a kinetic model with metabolic network coverage identical to k-ecoli74, but using flux data for wild-type and the seven mutant strains of E. coli generated with a reduced metabolic network. Our first attempt at parameterizing a kinetic model using a reduced network was with a set of fluxes generated using the minimal metabolic network used by Ishii et al. [36]. K-FIT, however, was unable to converge to a solution because simplifications to the Ishii et al. network and biomass equation when compared to k-ecoli74 were too drastic. Flux redirections upon flux projection associated with differences in the biomass formulation caused the wild-type and mutant strain flux distributions to fall outside of the nullspace of the k-ecoli74 stoichiometric matrix, and the gradient-based search tended to go towards unstable parameter sets.
The reduced network used for flux elucidation, therefore, was a network simplified from the k-ecoli74 metabolic network (see S4 File). It did not contain amino acid degradation pathways, and amino acid synthesis pathways were simplified to reflect only the transfer of carbon from central carbon metabolites to the amino acids being synthesized. All metabolism that did not contribute to carbon mapping (nitrogen, oxygen, and sulfur metabolism, energy metabolism and cofactors) was also removed from the metabolic network. Flux elucidation with these network simplifications using the same labeling data resulted in 95% confidence intervals that did not contain a steady-state flux distribution for the k-ecoli74 network upon flux projection. Flux projection was, therefore, performed by minimizing the violation of 95% confidence intervals of the reduced network by the wild-type flux distribution at metabolic steady-state. Differences between the projected flux distribution and the reduced network 95% confidence intervals were confined to amino acid synthesis and degradation pathways. The minimum total violation of reduced network 95% confidence intervals in the wild-type strain was 53 mmol/100 mmol glucose uptake, and the most notable rearrangement was in GLUDy flux (as GLUDy carried only the flux towards glutamate required for biomass formation). Serine, and glycine metabolism also underwent shifts in fluxes (due to the absence of ammonium from those pathways in the reduced network). SSR for each flux elucidation and the kinetic parameterization are provided in Table 1. The kinetic model parameterized using the reduced flux dataset that yielded the best SSR out of 30 multi-starts, had an SSR of 559, and an average SSR per data point of 2.5. An increase in SSR across conditions was observed, and is expected due to simplifications in the metabolic network used for flux elucidation. While 30 parameterizations converged to local minima, over 500 parameterizations were initialized. K-FIT encountered many unstable models during parameterization, but a limited number of initializations were able to converge, which was an improvement over parameterization using flux data generated with the Ishii et al. [36] network.
Predictions by the model parameterized with the reduced network flux dataset were inconsistent with k-ecoli74 predictions for several reactions peripheral to central carbon metabolism. Table 2 highlights the significant inconsistencies between fluxes estimated using the kinetic model parameterized using the reduced network flux dataset and k-ecoli74. We included only central carbon metabolism reactions with 13C-MFA-elucidated fluxes in reduced network mutant training datasets. When compared to the flux ranges predicted by k-ecoli74, the flux ranges predicted by the core model parameterized with reduced flux dataset were consistent across glycolysis, PP pathway, and TCA cycle. However, due to deviation from k-ecoli74 training data observed in the wild-type strain upon flux projection, both the directionality and magnitude of amino acid synthesis and degradation reactions was impacted across all strains used for parameterization. In the projected wild-type flux distribution, reverse flux through SERD-L was required to generate a flux distribution at metabolic steady state with minimum violation of the bounds defined using the reduced network 95% confidence intervals. As a result, SERD-L flux was predicted to be reversed across all strains used for parameterization. In Δzwf, Δgnd, and Δrpe, the magnitude of SERD-L flux was so large that the lumped serine synthesis reaction was also predicted to carry reverse flux. GLYCL and threonine degradation were also required to carry negative flux in order to minimize deviation from the reduced network 95% confidence flux bounds in the wild-type strain. Because glutamate did not re-enter central carbon metabolism as 2-oxoglutarate (akg) during flux elucidation, glutamate demand across all strains was significantly reduced when compared to k-ecoli74 training data. The behavior of the parameterized model reflected these discrepancies. The most notable underpredictions occurred in Δpgi, Δrpe, Δzwf. In each of these strains, GLUDy flux was underpredicted by more than 50%. The observed discrepancies between the model generated using a projected wild-type flux distribution and the flux distribution used for k-ecoli74 indicates that the information lost due to the absence of atom mappings and stoichiometric information for omitted and simplified reactions is significant. When flux information for these reactions is generated via constraint-based flux projection in a manner consistent with that employed by Khodayari et al. [14, 32] flux predictions in amino acid metabolism are greatly affected. It is, therefore, even for a core metabolism kinetic model, critical that the metabolic network used for flux elucidation include cofactor metabolism and carbon metabolism that are consistent with the kinetic model metabolic network to ensure construction of the most accurate and informative kinetic model possible.
Target product overproduction was evaluated using k-ecoli74 for seven metabolites using nine engineered strains not used during parameterization. All genetic intervention strategies, target metabolites for overproduction, target reactions used to estimate product yield, predicted product yields, and comparisons with experimental values and k-ecoli457 predictions are reported in Table 3. The reduced model was also used to predict metabolite yields. Reduced model metabolite yields are reported alongside k-ecoli74 metabolite yield predictions in Table 4. Fig D in S3 File depicts the perturbation strategies. Over and under expression were modeled as a fold-change in enzyme level applied to all enzyme complexes of the effected reaction. Because a model for protein synthesis is currently beyond the scope of K-FIT, an x-fold change in gene expression was modeled as an x-fold change in enzyme level. Only enzymes corresponding to perturbed genes had their fold-changes adjusted in the mutant strains. Out of the seven products tested, three were included in k-ecoli74, and four were not. Products present in k-ecoli74 included L-valine, acetate, and malate. L-valine yield was estimated using a lumped L-valine synthesis reaction flux. Acetate exchange flux was used directly to evaluate acetate yield, while MDH flux was used to evaluate malate yield. Products not included in the metabolic network included naringenin, lactic acid, 2,3-butanediol, and glucaric acid. Yields for these metabolites were estimated using flux through reactions producing or consuming precursor metabolites as proxies for heterologous pathways or, in the case of lactic acid, the lactic acid secretion pathway.
Kinetic model predictions for malate and acetate yields were consistent with experimental observations. In both cases, the kinetic model outperformed k-ecoli457 in estimating product yield. In the case of malate, although an approximate three-fold increase in the predicted value was observed compared to the experimental value, TCA cycle flux is required for strain viability, and therefore only a fraction of that flux would be directed towards malate secretion. The reported value is, therefore, a maximum theoretical yield. These results indicate that k-ecoli74 is better suited for predicting central carbon metabolite yields under glucose-rich batch conditions than k-ecoli457.
As expected, L-valine yield was underpredicted by k-ecoli74 by an order of magnitude (similar to k-ecoli457). This underprediction was due to the absence of an excretion pathway known to exist in E. coli [67], and the absence of flux distributions delivering significant flux towards the L-valine synthesis reaction from the training data. The only reaction consuming L-valine was the L-valine biomass sink reaction. The incomplete pathway coverage offered no drain for flux directed towards L-valine synthesis, and the model was thus incapable of carrying significant flux in that pathway.
k-ecoli74 overpredicted yield for metabolites not included in the network for all six strains tested. This was due to the use of central carbon drains as proxies for pathways not included in the model. This overprediction was expected, since only a fraction of central carbon flux can be directed towards branched pathways if the strain is viable due to the need for carbon flux towards biomass precursors synthesis and energy generation reactions. In three of those six strains, the engineered strain product yield was higher than the wild-type product yield, indicating a favorable re-direction of flux resulting from the genetic perturbations. k-ecoli457, however, outperformed k-ecoli74 when predicting yields in those strains. This was due to the inclusion of pathways peripheral to central carbon metabolism in k-ecoli457, and anaerobic conditions included in the training data which delivered significant flux towards lactic acid secretion, allowing for the generation of a kinetic parameter set better suited to predict lactic acid yield.
The reduced model was only capable of predicting metabolite yields for five of the nine model validation strains tested. In both naringenin overproduction strains and in the glucaric acid overproduction strain, the reduced model predicted no carbon uptake or growth. In the acetate overproducing strain, the model became unstable upon downregulation of RPI expression, and a steady-state flux distribution could not be reached in the perturbed state. The 2,3-butanediol yield prediction was similar to k-ecoli74 prediction, as was L-valine yield, and lactic acid yield under ACKr downregulation. k-ecoli74 outperformed the reduced model significantly, however, when predicting lactic acid yield under the ACKr knockout condition. This indicates that k-ecoli74 is more sensitive to small changes in enzyme level than the model parameterized with the reduced network. The reduced model outperformed k-ecoli74 significantly when predicting malate yield. Reduced model malate yield was within 9% of the experimental value, while k-ecoli74 overpredicted malate yield by 180%. Overall, the reduced model had difficulty predicting metabolite yields when TCA cycle, PP pathway, or oxidative phosphorylation enzyme levels were perturbed. Thus, the feasible prediction space of the core model parameterized by the reduced flux dataset is substantially reduced compared to k-ecoli74, as these pathways represent significant components of core metabolism and potential targets for perturbation strategies that kinetic model of core metabolism are designed to predict.
Metabolite yield variability was also assessed by comparing the upper and lower bounds of metabolite yields predicted by the three top performing model with experimental ranges. Fig 8 compares the ranges of predicted metabolite yields with experimental ranges. Because the k-ecoli457 model did not provide any information on parameter uncertainty or alternate models with similar fitness, k-ecoli457 yield predictions are represented with a single point. Only naringenin yield produced by the MDH knockout, SUCOAS downregulation strain expanded considerably when the three best models were considered. All other ranges spanned less than 25% of the mean yield value. Overall, narrow ranges of metabolite yields across models confirms the accuracy of k-ecoli74.
To test k-ecoli74’s ability to predict flux in strains not used for training and determine the impact of parameterization with and without unique flux phenotype in the training data, leave-one-out cross validation was performed. For each cross-validation, a unique kinetic model was parameterized while removing one mutant strain from the training data at a time. Parameterizations were either initialized until a model was generated with SSR less than the best identified model with all data included, or initialized until the recovery rate for a single model fell below the recovery rate obtained during the primary parameterization. In the latter case, the best model identified was used for cross validation analysis. Optimality was tested after each parameterization to ensure a local minimum was reached. Number of parameterizations, recovery rates and minimum SSR for each cross validation are listed in Table 5.
Δpgi was the only strain for which a cross-validation model with SSR less than the best model parameterized with all strains could not be identified. Fig 9A shows the residual error of each strain when the model parameterized without it was used to predict its flux distribution compared with its fitted flux in the full parameterization model. Δrpe, Δpgi, and Δgnd performed the worst during cross-validation, as there were phenotypic attributes of these strains that were unique when compared with the other training datasets. In each of these cases, the cross-validation model was incapable of predicting glucose uptake accurately, and the associated error propagated throughout the entire metabolic network. Fig 9B compares the glucose uptake rate of the cross-validation predictions.
In Δpgi, the cross-validation model predicted negligible glucose uptake because the PP pathway kinetic parameters were not well suited for carrying the entirety of flux from g6p. The Δrpe cross-validation model wasn’t well-suited for predicting the reduced glucose uptake observed experimentally in the absence of similar training data. Rather than predicting reduced glucose uptake, the Δrpe cross-validation predicted increased glycolytic flux compared to experimental fluxes and increased ED pathway flux to compensate for the partial blockage of the nonoxidative PP pathway. Because the training data in the Δgnd cross validation lacked a strain delivering significant flux through the ED pathway, the PP pathway and glycolytic kinetic parameters were unable to accommodate the full amount of glucose uptake observed, and the ED pathway was incapable of carrying significant flux. Δeda and Δedd cross validations models each performed well due to their phenotypic similarities, which ensured that a similar phenotype was included in the training data when one was left out. The resulting models were able to predict the flux distributions of interest well, even though they were the worst fitting strains when all sets were included in the training data. Δfbp and Δzwf each performed marginally well when compared with the other strains. In Δfbp, glycolytic reactions and GLUDy had the highest associated residual errors. The Δfbp cross-validation model predicted an increase in glucose uptake flux relative to the wild-type strain, which was propagated through glycolysis. The excess carbon fed into the network was sinked towards glutamate and glutamate-derived amino acid synthesis, rather than the metabolic byproduct, acetate, resulting in the high residual error for GLUDy. In Δzwf, a decrease in glucose uptake rate was predicted by the cross-validation model relative to the wild-type strain. This decrease in carbon availability was propagated through glycolysis, where all reactions had high residual error. Similar to Δfbp, GLUDy was the central carbon sink reaction with the highest residual error to compensate for the low glucose uptake flux. The results indicate that the model isn’t able to predict metabolic fluxes well when the strains being predicted have phenotypic characteristics that are drastically different from the training datasets.
Overall, the cross-validation results, as well as the full parameterization results, parameterization with reduced flux dataset results, and yield prediction results highlight the large influence that mutant flux dataset selection has on the value of the inferred parameters. Ideally, one would want to assemble a set of mutant flux datasets that uniformly perturb the flux in all major pathways. Unfortunately, this is difficult to a priori achieve due to the scarcity of experimental data. Metabolic flux magnitude and directionality dominant in the datasets is often reflected in resultant kinetic parameters, causing inaccurate prediction of unique flux redirections in the context of those used for training data. This implies that parameterization results must be carefully interpreted and flux datasets revealing unique flux redirections may have to be more heavily weighted during the parameterization process.
To assess the necessity of regulatory mechanisms included in k-ecoli74 and identify nonessential regulations, we compared the locally approximated standard deviation of the inhibition parameters to their parameter values. Regulations were identified as nonessential if the parameter range characterized by the standard deviation had a lower bound of zero. Fig 10 illustrates all regulations determined to be non-essential for model fitness to training data. It was determined that 26 of 55 regulations were essential to model fitness. The results indicate that regulation on one or two key enzymes in a pathway were sufficient to control flux through the entire pathway. The results also indicate that the gradient-based parametrization tended to drive non-essential inhibition constants towards small values (less than one), while essential inhibitions were driven towards large values, as there was a five order of magnitude difference between the average essential inhibition constant value and the average non-essential inhibition constant value. Regulations on the first two glycolytic steps (PGI and PFK) were determined to be essential, while all other glycolytic regulations (on FBP, FBA, GAPD/PGK, and PYK) were dispensable. All regulatory mechanisms in the PP pathway were identified as dispensable except for competitive inhibition of TALA by so4. Regulation of EDA by 6pg and 3pg were sufficient to control ED pathway flux, while EDD regulation by o2 was dispensable. The TCA cycle contained regulations on ICDHyr, SUCCOAS, and FUM. Out of these, SUCOAS regulations were dispensable, with flux controlled by regulation of ICDHyr and FUM. All glyoxylate shunt regulations, and at least one regulation on all anaplerotic reaction with included regulatory mechanisms were identified as essential to model fitness. The results indicate that the inclusion of substrate level regulations in the model is critical for characterizing metabolite pool sizes and enzyme complex fractional abundances, and consequently flux distribution. However, the number of regulatory mechanisms actively controlling flux through the network is limited to a few per pathway controlling key reactions and those peripheral reactions that serve a condition-dependent purpose, such as ME2, PPC, or ICL.
The kinetic parameterization pipeline developed in this study and applied to the development of k-ecoli74 is unique compared to other frameworks used for kinetic model construction in that a single metabolic network was used for flux elucidation and kinetic parameterization. Recent kinetic models of E. coli metabolism have either relied on experimentally determined kinetic parameters gleaned from a database and a combination of metabolomics and fluxomics data gleaned from a number of sources [87], or relied on fluxomics data generated using a metabolic network inconsistent with that used for kinetic parameterization [14, 32]. A comparison of k-ecoli74 kinetic parameters with those of other models and experimental parameters indicates that the parameterization method can significantly affect the resultant model.
When Km parameters were derived from the regressed elementary rate constants and compared to those generated using a previously developed core E. coli kinetic model [14], significant differences were observed (see S2 File). These can be attributed to the differences between growth conditions used in training data, differences in substrate binding and product release orders defined during elementary decomposition between the two models, and the absence of complete cofactor balances in the k-ecoli74 model. Discrepancies were also observed between Km parameters calculated from both models and experimental ranges, indicating that the optimal set of kinetic parameters for predicting in-vivo flux differs from the optimal kinetic parameters determined from in-vitro experiments. While no definitive reason can be identified based on the results of this study, possible reasons include differences in physiological states in vivo and in vitro (such as macromolecular crowding effects in vivo) [88–93], and differences in physiological state between strains (such as variable enzyme level [36]). The developed workflow, therefore, offers an alternative over kinetic modeling efforts that extract kinetic rate constants from databases because the rate constants are derived from data that is consistent with the conditions for which the model is designed to predict.
The use of the K-FIT algorithm has improved parameterization time by almost an order of magnitude over EM-based parameterization of a core E. coli model performed by Khodayari et al. [14]. Whereas parameterization of a core kinetic model using the ensemble modeling method generally takes more than a week, kinetic parameterization of k-ecoli74 took less than two days. The parameterization time for k-ecoli74 was greater than that reported by Gopalakrishnan et al. [61] for a similar core model parameterized with a set of toy data. The increase in computational expense was the result of inclusion of experimental training data with competing objectives, as solution reproducibility and parameterization time both improved when leave-one-out cross-validation was performed. This highlights the challenges that can arise when applying modeling frameworks steeped in assumptions to data taken directly from physical systems, and the influence that experimental uncertainty can have on parameterization. Model parameterization time was greater than the reported parameterization time for a core E. coli model with metabolic network conservation analysis and model stability check performed by Greene et al. [33]. Greene et al. were able to reduce ensemble modeling parameterization time by reducing the solution space based on a pre-evaluation stability analysis for all models in their ensembles, while the K-FIT algorithm reduces parameterization time using a gradient-based search and customized algebraic solvers.
Compared to the previously developed ensemble modeling-parameterized kinetic model of E. coli core metabolism constructed using an elementary decomposition approach, the model developed here shows improvement in fitness to 13C-MFA-derived flux distributions. In this study, 86% of fitted fluxes fell within a single standard deviation of their corresponding 13C-MFA value. The ensemble modeling-parameterized core kinetic model of E. coli, parameterized using the same number of mutant strains, estimated 78% of fitted flux values within a single standard deviation of 13C-MFA values [14].
A comparison of the yields for acetate and malate predicted by k-ecoli74 and those predicted by k-ecoli457 highlight the impact that the biological conditions of the cell culture used to generate the labeling data required for for 13C-MFA (and consequently, shifts in flux ranges) can have on the resulting kinetic model and prediction accuracy. k-ecoli457 was parameterized using flux distributions for wild-type and mutant strains grown primarily under chemostat conditions with growth rate fixed at a uniform, arbitrarily low value (0.2h-1) [36, 94]. The flux distributions exhibited zero acetate excretion across strains, and increased TCA cycle flux (wild-type TCA cycle flux five times higher than in the wild-type flux distribution generated in this study) [36]. Experimental growth conditions under which the acetate [71] and malate [69] yields were measured in the engineered strains were similar to experimental conditions used to generate 13C labeling data for this study (i.e. glucose-rich batch culture, mid-exponential growth phase). As a result, k-ecoli457 overpredicted malate yield by more than 540% and under predicted acetate yield by 75%, while k-ecoli74 overpredicted malate yield by only 180% and predicted acetate yield within 18% of the experimental value. While k-ecoli74 performed better than k-ecoli457 in predicting product yields for metabolites in central carbon metabolism, it was limited in its ability to predict yields for metabolites outside of the k-ecoli74 network. k-ecoli457 outperformed k-ecoli74 when predicting yields for 2,3-butanediol and glucaric acid due to the inclusion of pathways peripheral to central carbon metabolism in the model. k-ecoli457 also outperformed k-ecoli74 when predicting lactic acid yields. This was because the k-ecoli457 training data included datasets generated under anaerobic conditions which exhibited significant flux towards lactic acid secretion. The limited coverage of k-ecoli74 also limited the number of metabolite yields that could be effectively predicted. Whereas k-ecoli457 was able to predict metabolite yield for 24 metabolites across 320 genetic conditions, k-ecoli74 was only able to predict yields for seven metabolites across nine of those conditions. This was due to the non-inclusion of reactions that were perturbed in mutant strains and the non-inclusion of precursor metabolites to the pathway producing the metabolite of interest.
Our results from parameterization with a reduced flux dataset indicate that there is potential for significant discrepancies between kinetics models parameterized using fluxes inferred using the same network compared to fluxes inferred from a simplified model. The differences in the directionality and magnitude of amino acid synthesis and degradation reactions between k-ecoli74 and the kinetic model parameterized with reduced flux dataset demonstrate this. The reduced flux dataset parameterization also show a clear loss in predictive capabilities when the full stoichiometric model is not considered during flux elucidation, as the model parameterized with the reduced flux dataset failed to predict a feasible flux distribution for four of nine validation strains tested. Thus, the inclusion of the full metabolic network in the flux elucidation step of the kinetic parameterization pipeline was essential to successful k-ecoli74 parameterization.
The flux data used for kinetic parameterization also had an impact on the recovery rate of the best solution using the gradient-based methodology. Only 0.6% of solutions recovered agreed with our best model, despite all solutions satisfying local optimality criteria (i.e. zero gradient, non-negative Hessian). Recovery rate increased to 4% and 7%, respectively, when cross-validation was performed for Δrpe and Δgnd, and in the Δrpe cross-validation, 12.5% of solutions yielded SSR values within 10% of the best model when all data was used in fitting. This indicates that the existence of unique phenotypic behavior in mutant strains can lead to the existence of a large number of local minima and decreased recovery rate. It is important to note that our criteria for reproducibility was stringent. The maximum average square residual deviation from the best model was 0.05 per reaction flux. In both cross-validation parameterizations and the primary k-ecoli74 parameterization, many models were generated that produced similar flux distributions to the best model, but did not satisfy the specified reproducibility criteria.
While the developed kinetic parameterization pipeline addresses some issues that have slowed the development and application of kinetic models of metabolism in strain design and lays a framework for kinetic model scale-up in E. coli, a number of issues still exist. To construct a genome-scale kinetic model using 13C-labeling data for parameterization, a comprehensive set of genetic knockout strains across not only upper glycolysis and the PP pathway, but also lower glycolysis and the TCA cycle is required to generate the informative parameter sets for peripheral pathways branching from lower glycolytic and TCA cycle metabolites (such as pyruvate, acetyl-CoA, succinate, oxaloacetate, and 2-oxoglutarate). Another limitation requiring attention is that kinetic models constructed using elementary decomposition methods [30] are able to exclude enzyme concentration from kinetic expressions by assuming that enzyme concentrations do not change from wild type (except for enzymes coded by deleted genes). This assumption simplifies calculations in the absence of comprehensive proteomics data across multiple genetic and/or environmental conditions. Incorporation of kinetic descriptions of transcriptional and translational events into kinetic parameterization procedures would allow for the decoupling of enzyme concentration and elementary kinetic parameters. As an alternative to direct description of protein synthesis events in the cell, protein cost studies have shown that cellular enzyme concentration can be reasonably predicted using kinetic rate expressions for metabolic reactions [95, 96]. It may, therefore, be possible to use kinetic rate expressions for enzyme catalyzed reactions directly to characterize changes in enzyme level across conditions rather than kinetic expressions for transcription and translation. Proteomics data also suggest that changes in enzyme concentration in identical strains across different growth conditions is not simply proportional to the change in growth rate [36, 97], confirming that kinetic parameters regressed using the current framework are growth condition-specific. While the experimental data used in this study was taken from mutant strains with identical growth conditions (i.e. glucose-rich batch culture, mid-exponential growth phase), a systematic method for updating enzyme level according to growth rate or growth condition would allow for a broader range of applications. This would be a step towards the development of mechanistic kinetic model of metabolism with universal application.
|
10.1371/journal.pntd.0003637 | Validation of Serological Tests for the Detection of Antibodies Against Treponema pallidum in Nonhuman Primates | There is evidence to suggest that the yaws bacterium (Treponema pallidum ssp. pertenue) may exist in non-human primate populations residing in regions where yaws is endemic in humans. Especially in light of the fact that the World Health Organizaiton (WHO) recently launched its second yaws eradication campaign, there is a considerable need for reliable tools to identify treponemal infection in our closest relatives, African monkeys and great apes. It was hypothesized that commercially available serological tests detect simian anti-T. pallidum antibody in serum samples of baboons, with comparable sensitivity and specificity to their results on human sera. Test performances of five different treponemal tests (TTs) and two non-treponemal tests (NTTs) were evaluated using serum samples of 57 naturally T. pallidum-infected olive baboons (Papio anubis) from Lake Manyara National Park in Tanzania. The T. pallidum particle agglutination assay (TP-PA) was used as a gold standard for comparison. In addition, the overall infection status of the animals was used to further validate test performances. For most accurate results, only samples that originated from baboons of known infection status, as verified in a previous study by clinical inspection, PCR and immunohistochemistry, were included. All tests, TTs and NTTs, used in this study were able to reliably detect antibodies against T. pallidum in serum samples of infected baboons. The sensitivity of TTs ranged from 97.7-100%, while specificity was between 88.0-100.0%. The two NTTs detected anti-lipoidal antibodies in serum samples of infected baboons with a sensitivity of 83.3% whereas specificity was 100%. For screening purposes, the TT Espline TP provided the highest sensitivity and specificity and at the same time provided the most suitable format for use in the field. The enzyme immune assay Mastblot TP (IgG), however, could be considered as a confirmatory test.
| The success of any disease eradication campaign depends on considering possible non-human reservoirs of the disease. Although the first report of T. pallidum infection in baboons was published in the 1970’s and the zoonotic potential was demonstrated by inoculation of a West African simian strain into humans, nonhuman primates have not yet been considered as a possible reservoir for re-emerging yaws in Africa. Simian strains are genetically most closely related to the strains that cause yaws in humans. The identification of baboons as a reservoir for human infection in Africa would be revolutionary and aid important aspects to yaws eradication programs. Reliable serological tests and a useful standardized test algorithm for the screening of wild baboon populations are essential for studying potential transmission events between monkeys and humans.
| Treponema pallidum is the bacterium that causes venereal syphilis (ssp. pallidum) and the non-venereal diseases yaws (ssp. pertenue) and endemic syphilis (ssp. endemicum) in humans [1]. The spirochete is able to cause a life-long chronic infection in untreated individuals [2] and elicits a strong adaptive immune response against a wide array of antigens [3–4] with strong serum IgM and IgG response [5–7] towards a number of lipoproteins (e.g., Tp15, 17, and 47), endoflagellar proteins (e.g., FlaA, FlaB1, 2, and 3), and the Tpr family proteins (e.g., TprA-TprL) [6]. Furthermore, infection-related cellular damage is known to induce the production of non-treponemal antibodies mainly directed against cardiolipids [8,9].
Recently, we have reported that T. pallidum can infect large numbers of African monkeys and great apes [10]. To date, all simian isolates seem to be closely related to T. pallidum ssp. pertenue, the pathogen causing human yaws [11,12] and at least the Fribourg-Blanc simian strain, which was isolated from a baboon in Guinea [13], has the potential to cause sustainable infection in humans [14]. Thus, there is evidence to suggest that yaws exists in non-human primate populations residing in regions where humans are also infected [15]. The clinical manifestations in non-human primates (NHPs) however, show regional differences. While West African simian strains of T. pallidum mostly cause no clinical signs [16], gorillas in the Republic of the Congo show yaws-like lesions [17] and baboons in East Africa are known to develop severe genital ulceration [11,18]. However, independent of the clinical manifestations simian strains induce a pronounced serological response in the respective host [10], which may be used to screen and identify host populations for their potential as a natural reservoir.
In the context of the possible zoonotic potential of simian strains [14], the identification and knowledge of a nonhuman reservoir for T. pallidum is crucial to disease elimination or eradication efforts and could help to identify hot spots for potential simian-to-human disease transmission. There is therefore considerable need to validate treponemal tests (TTs) and non-treponemal (NTTs) for their use in NHPs. Due to the close relationship of simian and human treponemes [12], we hypothesized that A) commercially available serological tests are able to detect simian anti-T. pallidum IgM and IgG in serum samples of baboons, a NHP species with high infection rates and B) that the serological tests will be equally reliable in terms of sensitivity and specificity in baboon sera compared to the human sera.
Baboon serum samples were taken in accordance with the Tanzania Wildlife Research Institute’s Guidelines for Conducting Wildlife Research (2001) and with permission of Tanzania National Parks (TNP/HQ/E.20/08B) as well as Commission for Science and Technology in Tanzania (2007-56-NA-2006-176). The committee of Tanzania National Parks and Tanzania Wildlife Research Institute approved sample collection. Baboon serum samples from the German Primate Center were granted from the institute’s bio bank and originated from healthy animals that were sampled during post-mortem examination. The Animal Welfare and Ethics Committee of the German Primate Center approved the use of samples for this study.
In a previous study, we were able to detect T. pallidum infection in wild olive baboons (Papio anubis) at Lake Manyara National Park in Tanzania [18]. Although the isolated strain is most closely related to T. pallidum ssp. pertenue [11], the pathogen causes severe genital ulceration. Diagnosis was based on gross pathology, histology, and molecular biological tests. The latter included quantitative [19] and qualitative PCR [20], targeting the polA gene of T. pallidum. DNA was extracted from skin tissue samples [18]. Data and corresponding serum samples that were constantly stored at -80°C of 57 untreated baboons from this study were available for analysis in 2013. An additional set of 11 serum samples of healthy captive olive baboons from the German Primate Center were included as negative control. The extent of genital ulceration was used to classify and group animals as clinically healthy, initially-, moderately-, or severely-infected (Fig. 1). It is not known whether simian infection develops in stages similar to human infection.
Generally, the definition of a test result was based on the individual’s overall infection status (Table 2). Details of infection status including genital ulceration status and each test’s interpretation can be found in the Supporting Information (S1 Table).
Statistical analyses were performed using Prism 6.0 (GraphPad Software). Results of the TTs were first compared to the result of the Serodia TP-PA and second to the consensus of infection status, as it is defined in Table 2, and which takes into account the appearance of clinical symptoms (genital health status), IHC and skin tissue PCR results of the same animals as published elsewhere [18]. While the TP-PA was used as the gold standard for TTs, we compared results to the baboon’s infection status for further verification of test results and accuracy. With regard to NTTs it was assumed that a significant proportion of tests might become nonreactive in chronically infected baboons, as it is described for untreated human syphilis infection [25–27]. Test performances of the NTTs were therefore evaluated exclusively on the basis of the consensus of infection.
A non-parametric test for nominal scale data, the two-tailed Fisher’s Exact Test, was used to compare the proportions among the serological tests, skin tissue PCR results and clinical signs of infection as well as for the analysis of sensitivity and specificity of the serological tests.
Tables 3 and 4 summarize sample size, proportions, and performance characteristics of the serological tests that were used in the 57 baboon serum samples from Lake Manyara National Park and an additional set of 11 serum samples from olive baboons of the German Primate Center in Germany. The latter were included for the purpose of additional negative control.
When comparing TT performances with the TP-PA, we observed test sensitivity in the range of 91.3–100%, and specificity ranging from 94.7–96.0% (Table 3). When test results were compared to the consensus of all test results including PCR, however, the observed sensitivity of the TTs was in the range of 97.7–100%; whereas the specificity reduced slightly to the range of 88.0–100% (Table 4). This reduction was almost exclusively caused by the test performance of Syphilitop Optima (Table 4). NTT performances were not compared to TP-PA results since positive TT results in untreated, chronically infected patients do not necessarily predict reactivity of the corresponding NTT [25–27]. However, both NTTs used in this study reliably detected anti-lipoidal antibodies in serum (Table 4) or plasma samples (S2 Table) of baboons. When infection status was considered in the context of all test results including PCR analysis, NTT sensitivity in serum samples was lower (83.3%) than the average of the TTs (99.54%). The specificity of VDRL and RPR in serum samples was 100% and therefore higher than the performance of the standard TP-PA (92.0%) and Syphilitop-Optima rapid test (88.0%). The performance data of NTTs for plasma samples are listed in S2 Table.
Anti-T. pallidum antibodies were found in 97.3% of baboons with genital ulceration and in 6 of 20 animals that were clinically healthy (30.0%, Table 5, Fig. 1). For comparison, a remarkable proportion of genital-ulcerated baboons (13.5%; n = 5/37, Table 5) had a negative PCR outcome of their respective skin tissue biopsy. Vice versa, we found 30% (n = 6/20) genital non-genital-ulcerated baboons with positive T. pallidum PCR of their corresponding skin biopsy. Yet, genital-ulcerated baboons were significantly more often reactive for T. pallidum in serology (TTs, p < 0.0001) and skin tissue PCR (p < 0.0001) compared to clinically healthy and thus non-ulcerated baboons. Both treponemal rapid tests, Espline TP and Syphilitop Optima, were more sensitive than the Serodia TP-PA (Two-tailed Fisher’s exact test, p < 0.0001, Table 4), although specificity in the Syphilitop Optima was much lower than Serodia TP-PA. The same applied, when Mastafluor FTA-ABS IgG and Mastablot TP IgG were compared to the Serodia TP-PA particle agglutination assay. In both tests the proportion of positive results matched the results of the Serodia TP-PA (Two-tailed Fisher’s exact test, p < 0.0001, Table 4).
No correlation was found when anti-T. pallidum IgG positive serum samples were tested in the immunoblot assay Mastablot TP for the presence of IgM antibodies against T. pallidum. Only a limited number of animals, 6 out of 59 (10.2%), tested positive for both anti-T. pallidum IgG and IgM antibodies. We did not find any samples that were positive for anti-T. pallidum IgM only.
Even after log10-transformation of antibody titers, the clinically healthy and the initial stage genital-ulcerated group (Fig. 1; it is not known whether NHPs develop three stages similar to humans; initial stage refers to the severity of genital ulceration as describe elsewhere [18]) were not normally distributed. The Kruskal-Wallis test using Dunn’s correction for multiple comparisons showed that antibody titers in the severe genital-ulcerated group of baboons were significantly higher when compared to clinically healthy animals (mean rank diff. = -30.04, p ≤ 0.0001) and baboons with an initial stage of genital ulceration (mean rank diff. = -17.56, p ≤ 0.05). The same was found for moderate genital-ulcerated baboons, which had significantly higher antibody titers against T. pallidum than animals without genital ulceration (mean rank diff. = -19.95, p ≤ 0.05). Fig. 1 provides an overview.
All serological tests used in this study, TTs as well as NTTs, were able to detect anti-T. pallidum antibodies in serum of infected baboons. While the presence of anti-T. pallidum IgG antibodies in all infected animals correlates to lifelong antibody titers in human infection [28], the absence of IgM type antibodies against the spirochete in infected baboons (only 6 animals were positive for anti-T. pallidum IgM antibodies, Mastablot IgM, S1 Table) could most likely be linked to the timing of sampling from two weeks to months post infection. This may not be fully consistent with human cases, in which IgM antibody titers are low to moderately high in primary syphilis, peak in secondary and sometimes tertiary syphilis, and are low in latent syphilis [29]; the majority of baboons, which were tested in this study were unlikely in a latent stage of infection. Although animals are known to be chronically infected and untreated over many months to years, genital ulceration in infected baboons rarely heals up. However, the expected lower amplitude of IgM titers [30] and the expected reduced half-life of IgM [31] may have contributed to the result. Results were excluded from multiple comparisons analysis and evaluation of performance characteristics, because of the uncertainty and low number of IgM positive samples. The finding of few chronically infected individuals positive for IgM antibodies against T. pallidum could be explained by persisting IgM titers, as described for other human spirochete infections (e.g., borreliosis) [32,33].
All TTs and NTTs used in this study are commercially available and licensed for use with human serum or plasma samples. Using TT as a screening test in NHPs (Fig. 2) is in accordance with international standards and EU Guidelines for the Management of Syphilis in humans [22,34,35] and is often referred to as “reverse testing algorithm” [36]. Although there were no data about efficiency, sensitivity, and specificity available from testing samples in baboons or other NHPs, it was reasonable to assume that a TP-PA would be a reliable standard in testing baboon serum samples both qualitatively and quantitatively (Table 3). A number of alternative TTs (Espline TP, Syphilitop Optima, Mastafluor FTA-ABS IgG, and Mastablot TP IgG) were included in the analysis to rule out uncertainties of TP-PA test performance in baboon sera. Mastafluor FTA-ABS IgG and Mastablot TP IgG were specifically added to cover different T. pallidum antigens than the rapid TTs Espline TP and Syphilitop Optima (TmpA; see Material and Methods). Interestingly, sensitivity, specificity, and the corresponding positive and negative predictive values were lower in Serodia TP-PA when compared to all other TTs, excluding the Syphilitop Optima, which had a higher sensitivity, but weaker specificity (Table 4). This was in contrary to the TP-PA manufacturer’s information about performance characteristics in human test sera (Table 1) and also when compared to FTA-ABS and immunoblot IgG results of human samples [21, 37]. In our experience the interpretation of the gelatin TP-PA requires a certain level of training and serum pre-absorption in baboon samples. This was achieved by incubating the test serum with non-sensitized particles so that unspecific binding factors were pre-absorbed. Another difficulty of the test was related to the endpoints, which can differ over time. Despite those difficulties, Serodia TP-PA has an advantage in that the readability was made by the naked eye and may be operated in resource poor laboratory settings. Also, it was the only TT that can be used for semi quantitative titer quantification. Validation of dried blood spots with a Serodia TP-PA assay for external quality assurance of T. pallidum serology as published elsewhere [38], provides an interesting outlook for the confirmation of screening test results, e.g., from the use of Espline TP in remote areas at the wildlife-human interface.
Both rapid TTs that could be used for screening, the Espline TP and Syphilitop Optima, were easy to perform. Results are reported within 15 min and especially the Espline TP comes in a handy cassette format. Both tests require only 25 μl of sample material. According to the manufacture’s description serum and plasma samples can be used for Espline TP, while Syphilitop Optima must not use plasma samples. Comparing Espline TP to the Serodia TP-PA in serum samples of baboons, Espline TP had higher sensitivity (97.7 vs. 91.3%) and nearly the same specificity (96.0 vs 95.5%; Table 3). When the test performances are compared to the consensus of infection status (Table 2), sensitivity is 100% in both screening tests, but Syphilitop Optima achieved only 88.0% specificity compared to the Espline TP with 100% (Table 4). The reason for these differences remains uncertain but might be promoted by protein variations of the coated filter membrane. Espline TP uses at least one additional antigen (Tp 15) that is not included into Syphilitop Optima. Also, proteins (Tp 47 and Tp15-17) are used to coat two different areas of the membrane in the Espline TP, but they are combined in one field on the reaction membrane of the Syphilitop Optima. Unfortunately, there was no information available about the quantity of protein coated to the membrane in Syphilitop Optima. The Espline TP uses a combination of alkaline and non-phosphatase labeled TP recombinant antigens. Due to the lack of information, it was not possible to compare the tests in that aspect, but differences may have an influence on the binding affinity of simian T. pallidum antibodies.
When compared to the consensus of infection status (Table 4) both Mastafluor FTA-ABS IgG and Mastablot TP IgG have higher sensitivity and specificity and thus can be recommended as a confirmatory test in baboons. However, it may be noted, that in human infection FTA tests are no longer recommended for the diagnosis of syphilis [22,39], which should make Mastablot TP IgG the preferred option.
In the context of the performance characteristics that are reported by the manufacturers (Table 1), Serodia TP-PA had a slightly weaker sensitivity (97.7 vs. 100%) and reduced specificity (92.0 vs. 100%, Table 4). The Espline TP rapid test had nearly identical sensitivity and specificity as indicated by the manufacturer, whereas Syphilitop Optima had an equal sensitivity (100%) with a reduced specificity value, 88% vs. 95% as communicated by the manufacturer. Future studies may benefit from heat pre-treatment of serum samples to reduce interference caused by complement proteins and unspecific binding of antibodies. Heat pre-treatment was not part of any manufacturer’s protocol of the tests that were used in this study.
Although the quantification of antibodies may not be of interest for disease prevalence studies in wild baboons, it may be an important tool for characterizing simian infection. The finding that severe genital-ulcerated baboons had significant higher anti-T. pallidum antibody titers than clinically non-affected animals (CNA vs. SEV, p < 0.0001; Fig. 1) or those with less severe genital ulceration (CNA vs. MOD and INI vs. SEV, both p ≤ 0.05; Fig. 1) is consistent with what can be expected from the course of infection. The rating of chronicity of infection in baboons at Lake Manyara National Park was based on gross-pathology and histological examination of skin tissue samples [18].
The use of a NTT for the initial screening in the traditional algorithm in human infection is to avoid the detection of previously treated and non-active cases [40]. NTTs are known to produce a higher percentage of false positives [41] and test performance data of the NTTs as they are reported in Table 4 need to be interpreted with caution since it is neither known when an individual was infected nor how long anti-cardiolipid antibodies can be found in the due course of infection in wild baboons.
The decision to use and recommend a TT as a screening test for T. pallidum infection in NHPs was based on the following three reasons and is in accordance with the current European Guidelines on the Management of Syphilis [23]. First, wild baboons are rarely treated and once infected, treponemal clearance may be an exception rather than the norm. Second, there is a paucity of data on cross-reactivity of proteins derived from human T. pallidum strains with antibodies against the simian strain in baboons. Lastly since the majority of baboons were chronically infected, we had reason to belief that a number of these chronically infected baboons were non-reactive in NTTs, as it was described in untreated human syphilis infection [25–27]. However, while a lifelong anti-T. pallidum antibody titer in baboons provides a most useful readout for the identification of a disease hot spot that offers the possibility for simian-human infection, therapeutic interventions in wild NHPs, as it is already conducted in baboons at Gombe Stream National Park in Tanzania (Collins et al. pers. communication) may benefit from the use of the traditional algorithm since NTTs may allow the differentiation of active and inactive infection.
It is generally believed that yaws has no animal reservoir. Until identical T. pallidum strains are found circulating in nonhuman primates and humans in their natural environment this understanding cannot change. Yet, to this end, more research is needed before nonhuman primates can definitely be ruled out to serve as a natural reservoir for human infection. We have only recently begun to explore the range of nonhuman primate infection in Africa. Because the human-livestock-wildlife interface is constantly growing, the potential for inter-species transmission increase significantly. It is also possible that simian strains do naturally infect humans but do not cause clinical manifestations, as it is the case in Guinea baboons (Papio papio) in Senegal; or it may be that at least the East African simian strains cause genital ulceration in humans and may therefore not be diagnosed as yaws based on their genetics. Clearly, further research is needed before any answers can be given and serological surveys are an important tool to support these investigations and to complete our picture of T. pallidum infection in humans and nonhuman primates.
Based on the outcome of this study we propose an algorithm for the screening of wild non-treated NHP populations (Fig. 2). The algorithm aims to identify T. pallidum infection in wild baboons and other NHPs and may complement the current yaws eradication campaign [42].
All tests used in this study provided reliable results to detect anti-T. pallidum antibodies in serum of baboons. We therefore favor hypothesis A, which suggests that commercially available serological tests are able to detect simian anti-T. pallidum IgM and IgG in serum samples of baboons, with the exception of IgM class anti-T. pallidum antibodies. It would be necessary to examine more animals in the initial stage of infection in order to test this part of the hypothesis, something that is difficult to achieve since the time of infection in wild baboons in general is not known. While NTTs may help to plan treatment and control of infections in baboons, TTs are most useful to screening non-treated baboon population for the presence of T. pallidum. Hypothesis B was partly rejected because some serological tests were not equally reliable in their sensitivity and specificity in baboon samples compared to human serum samples.
For screening purposes, the immunochromatography based Espline TP test provided the highest sensitivity and specificity values and in addition had the most suitable format for use in the field. For confirmation, the treponemal test Mastablot TP IgG had the best performance characteristics and is therefore recommended as a gold standard. Serodia TP-PA was able to quantify antibodies against T. pallidum in baboons and results were consistent with the chronicity of infection. Based on this study a testing algorithm for the screening of NHP populations for T. pallidum infection is proposed, which may help future yaws eradication campaigns or wildlife management to identify baboons as a potential reservoir for human yaws infection.
|
10.1371/journal.ppat.1007323 | Independent amplification of co-infected long incubation period low conversion efficiency prion strains | Prion diseases are caused by a misfolded isoform of the prion protein, PrPSc. Prion strains are hypothesized to be encoded by strain-specific conformations of PrPSc and prions can interfere with each other when a long-incubation period strain (i.e. blocking strain) inhibits the conversion of a short-incubation period strain (i.e. non-blocking). Prion strain interference influences prion strain dynamics and the emergence of a strain from a mixture; however, it is unknown if two long-incubation period strains can interfere with each other. Here, we show that co-infection of animals with combinations of long-incubation period strains failed to identify evidence of strain interference. To exclude the possibility that this inability of strains to interfere in vivo was due to a failure to infect common populations of neurons we used protein misfolding cyclic amplification strain interference (PMCAsi). Consistent with the animal bioassay studies, PMCAsi indicated that both co-infecting strains were amplifying independently, suggesting that the lack of strain interference is not due to a failure to target the same cells but is an inherent property of the strains involved. Importantly PMCA reactions seeded with long incubation-period strains contained relatively higher levels of remaining PrPC compared to reactions seeded with a short-incubation period strain. Mechanistically, we hypothesize the abundance of PrPC is not limiting in vivo or in vitro resulting in prion strains with relatively low prion conversion efficiency to amplify independently. Overall, this observation changes the paradigm of the interactions of prion strains and has implications for interspecies transmission and emergence of prion strains from a mixture.
| This is the first example of prion strains that replicate independently in vitro and in vivo. This observation changes the paradigm of the interactions of prion strains and provides further evidence that strains are a dynamic mixture of substrains. Strain interference is influenced by the abundance of PrPC that is convertible by the strains involved. These observations have implications for interspecies transmission and emergence of prion strains from a mixture.
| Prion diseases are a group of transmissible neurodegenerative diseases that affect animals, including humans. Animal prion diseases include scrapie in sheep and goats, transmissible mink encephalopathy (TME) in ranch-raised mink, chronic wasting disease (CWD) in cervids, and bovine spongiform encephalopathy [1–10]. The human prion diseases can be acquired, inherited, or can occur sporadically and include Creutzfeldt-Jakob disease (CJD), Gerstmann-Straussler-Scheinker disease, fatal familial insomnia, and kuru [11–15]. Prion diseases have long asymptomatic incubation periods ranging from months to decades followed by a short symptomatic phase characterized by progressive cognitive and/or motor deficits [16,17]. During the asymptomatic phase, prions can be detected in the central nervous system and extraneural locations [18]. Currently, effective treatment for prion diseases is not available, and they are inevitably fatal.
The prion agent is comprised mainly, if not entirely, of PrPSc which is an abnormal isoform of the host encoded prion protein, PrPC [19–23]. Prion propagation is thought to occur in a three-step process where PrPSc first binds to PrPC followed by a conformational conversion of PrPC to PrPSc. Next, fragmentation of the growing PrPSc polymer results in the generation of new PrPSc free ends for PrPC to bind. Repeated cycles of this process are thought to encompass prion formation in vivo and are recapitulated in vitro by protein misfolding cyclic amplification (PMCA) [24].
Prion strains are operationally defined as a phenotype of disease under a fixed set of agent and host parameters. Under experimental conditions where these parameters are precisely controlled, distinct phenotypes of disease correspond with prion strains [25–27] [28–36]. Differences in the distribution and relative intensity of spongiform degeneration in select areas of the central nervous system (CNS) are, currently, the most well-accepted criteria to distinguish strains [37,38]. Prion strains can differ in incubation period, clinical signs of disease, tissue tropism, and host range. Importantly, the strain specific phenotype is maintained upon serial passage and is, therefore, heritable.
Prion strain diversity may be encoded by distinct conformations of PrPSc. Studies using the hyper (HY) and drowsy (DY) strains of hamster-adapted TME were the first to show that PrPSc could have strain-specific differences in the proteinase K (PK) cleavage site, with the unglycosylated PrPSc polypeptide migrating at 21 and 19 kDa respectively, relative PK resistance and detergent insolubility [39–41]. Consistent with these findings, similar strain-specific differences in PrPSc migration properties of human prion isolates were preserved upon transmission to transgenic mice expressing chimeric mouse-human PrPC [42]. Structural studies of PrPSc using Fourier transform infrared spectroscopy indicated that strain-specific differences in PrPSc secondary structure may underlie strain-specific properties of PrPSc such as PK cleavage site [43,44]. The conformation dependent immunoassay (CDI) measures changes in immunoreactivity of PrPSc compared to immunoreactivity of PrPC under conditions of increasing denaturation. CDI has identified strain-specific differences in PrPSc conformation from several rodent prion strains [45]. The relationship between the strain-specific biochemical features of PrPSc and the outcome of disease are poorly understood.
Prion strains, when present in the same host, can interfere with each other. This was first observed in mice where inoculation of the long incubation period blocking strain, 22C, prior to superinfection with the short incubation period strain,22A, could extend the incubation period of 22A [46]. Short and long incubation period strains are categorized relative to the minimum and maximum incubation periods of known strains for a given host under experimental transmission parameters. As the interval between inoculation of 22C and 22A is increased, 22C can extend the incubation period or completely block 22A from causing disease [46]. This is consistent with subsequent studies indicating that the relative onset of conversion of the blocking and superinfecting strain influences strain emergence, and conversion of the blocking strain PrPSc corresponds with the occurrence of strain interference [47–49]. Additionally, the blocking and superinfecting strains must infect the same cells for strain interference to occur and potentially compete for PrPC [47,50]. Prion strain interference is a common property of prions and numerous host strain combinations have been identified where prion strains interfere with each other or completely block one strain from causing disease [16,46,51–55]. Most of these studies, however, investigated combinations of prion strains that have large differences in incubation period and/or prion agent conversion efficiencies. It is unknown if two relatively long-incubation period, low efficiency converting prion strains can interfere with each other. In this study we investigated the ability of relatively low prion conversion efficiency strains to interfere with each other in vivo and in vitro and the contribution of PrPC in this process.
Mixtures of DY and 139H or DY and ME7H have differentiable PrPSc Western blot migration profiles. To determine the strain-specific migration of PrPSc from mixtures of strains, Western blot analysis was performed on mixtures of DY and 139H or DY and ME7H brain homogenates at ratios of 10:1, 1:1 and 1:10 (Fig 1). Samples that contained an excess of one strain had unglycosylated PrPSc polypeptide migration of the excess strain at either 21 or 19 kDa (Fig 1). Samples that contained an equal ratio of DY and 139H or DY and ME7H resulted in the unglycosylated PrPSc polypeptide migrating at 21 kDa and 19kDa (Fig 1A) that PrPSc migration analysis confirms as a band migrating from 21 to 19 kDa (Fig 1B and 1C). Overall, a mixture of DY and 139H or DY and ME7H resulted in a dual unglycosylated PrPSc polypeptide pattern that was resolved by Western blot if the ratio of the two strains was within 10-fold of each other.
Strain interference does not occur between DY and 139H or DY and ME7H in vivo. To investigate strain interference in vivo, groups of hamsters (n = 5) were intracerebrally (i.c). inoculated with either an uninfected brain homogenate (negative control), DY-infected brain homogenate (positive control), 139H-infected brain homogenate (positive control), ME7H-infected brain homogenate (positive control), or co-infected with an equal ratio of DY and 139H (experimental group 1; Table 1) or DY and ME7H (experimental group 2; Table 1). Positive controls hamsters inoculated with DY (n = 5) all developed clinical signs of progressive lethargy at either 178±4 or 161±3 (Table 1) days p.i. with weight gain that did not significantly (p>0.05) differ from the uninfected negative control animals (Fig 2) and brain material contained an unglycosylated PrPSc polypeptide that migrated at 19 kDa (Fig 3A, lanes 1 & 7, B and C). All (n = 5) of the 139H-inoculated positive control animals developed clinical signs of ataxia at 125±3 days p.i. (Table 1) with a significant (p<0.05) gain in weight compared to age-matched mock-infected controls starting at 49 days p.i. that continued for the duration of the disease course (Table 1 and Fig 2). Western blot of brain material from the 139H-infected animals contained an unglycosylated PrPSc polypeptide that migrated at 21 kDa (Fig 3A, lane 2 and B). All of the negative control mock-infected animals (n = 5) remained asymptomatic for either 200 or 270 (Table 1) days post-infection (p.i.) and PrPSc was not detected in brains from these animals by Western blot (Fig 3A, lanes 4–5; cropped from a different blot performed concurrently with the rest of the figure). Hamsters (n = 5) co-infected with the DY and 139H agents developed clinical signs of ataxia at 125±3 days p.i. that did not significantly (p>0.05) differ from animals inoculated with 139H alone but did differ significantly (p<0.05) from animals inoculated with DY alone (Table 1). The DY and 139H co-infected animals had a statistically significant (p<0.05) weight gain starting at 56 days p.i. compared to age-matched mock infected animals but did not differ significantly (p>0.05) from animals inoculated with 139H (Fig 2). At 119 days p.i. an intercurrent death occurred in both the 139H and the DY/139H co-infected group resulting in a reduction of statistical power that may have contributed to the lack of statistical significance at this time point (Fig 2). Brain material from hamsters co-infected with DY and 139H agents had unglycosylated PrPSc that migrated at 21 kDa consistent with 139H infection (Fig 3A, lane 3 and B). All (n = 5) of the ME7H-inoculated positive controls developed clinical signs of prion infection at 250±3 days p.i. (Table 1), and Western blot of brain material from the ME7H-infected animals contained an unglycosylated PrPSc polypeptide that migrated at 21 kDa (Fig 3A, lane 6 and C). Hamsters co-infected with the DY and ME7H developed clinical signs of progressive lethargy at 161±3 days p.i. that did not significantly differ (p>0.05) from animals inoculated with DY alone but did statistically differ (p<0.05) from animals inoculated with ME7H alone (Table 1). Brain material from hamsters co-infected with DY and ME7H agents had unglycosylated PrPSc that migrated from 19 to 21 kDa consistent with a mixture of both DY and ME7H PrPSc (Fig 3A, lane 8 and C).
Similar PMCA conversion efficiency of DY, 139H, and ME7H PrPSc. Previous work indicated the PMCA conversion coefficients (PMCA-CC) of DY, 139H, and ME7H scrapie are similar based on 10-fold serial dilutions of brain homogenate [56]. The PMCA-CC of these strains was further refined by examining selected dilutions of infected brain homogenate that differed by less than 10-fold (Fig 4). The PMCA-CC was 1.80 ± 0.35 for DY, 1.89 ± 0.44 for 139H, and 1.33 ± 0.42 for ME7H. These PMCA-CC values did not differ statistically (p>0.05) when compared to DY. Mock-infected PMCA reactions did not amplify detectable PrPSc (Fig 4; DY sample was cropped from a separate blot that was performed concurrently)
Lack of strain interference between long-incubation period, low efficiency converting prion strains in vitro. PMCAsi was performed as previously described [50], and all PMCAsi reactions were performed in triplicate. Unseeded PMCA negative control reactions did not result in detectable PrPSc (Tables 2 and 3). Positive PMCA control reactions seeded with either DY, 139H or ME7H resulted in amplification of strain-specific PrPSc (Tables 2 and 3; Fig 5). For PMCAsi reactions, known ratios of DY and 139H (Table 2) or DY and ME7H (Table 3) were mixed and subjected to serial rounds of PMCA. The migration of PrPSc was determined by Western blot using the parameters described in Fig 1. In the co-infected PMCAsi reactions, the migration of PrPSc was similar to the strain that was initially seeded in excess (Tables 2 and 3; Fig 5). When an equal ratio of strains was used as the starting material we found a 19 to 21 kDa migration of unglycosylated PrPSc consistent with amplification of both strains that was maintained for 10 serial rounds of PMCA (Tables 2 and 3; Fig 5).
PMCAsi did not generate a new prion strain. PMCAsi reactions seeded with equal amounts of the long incubation period strains contain unglycosylated PrPSc that migrates from 19 to 21 kDa. This PrPSc species may not represent independent amplification of each strain but instead represent a new prion strain. To test this hypothesis, groups (n = 5) of hamsters were i.c. inoculated with round 10 PMCAsi material from either uninfected negative control PMCA reactions, DY or 139H seeded positive control PMCA reactions or reactions seeded with equal amounts of DY and 139H. The hamsters inoculated with the uninfected negative control PMCA samples remained asymptomatic for 275 p.i., weighed 152.4±5.0 g at 172 p.i., (Table 4) and PrPSc was not detected in brain material from these animals (Fig 6A). Hamsters inoculated with the round 10 PMCA DY positive control reaction developed clinical signs of progressive lethargy at 224±7 days p.i., weighed 170.6±5.9 g at 172 p.i. (Table 4), and brain material from all animals contained an unglycosylated PrPSc polypeptide that migrated at 19 kDa consistent with DY infection (Fig 6). Hamsters inoculated with the round 10 PMCA 139H seeded positive control reactions developed clinical signs of ataxia at 177±3 days p.i., weighed 203.4± 6.0 g at 172 p.i. (Table 4), and brain material from all animals contained an unglycosylated PrPSc polypeptide that migrated at 21 kDa consistent with 139H infection (Fig 6). Hamsters inoculated with round 10 PMCAsi DY and 139H seeded experimental group developed clinical signs of ataxia at 180±3 days p.i., weighed 198.0±5.5 g (Table 4) and brain material all of these animals contained an unglycosylated PrPSc polypeptide that migrated at 21 kDa consistent with 139H infection (Fig 6). The incubation period of animals inoculated with round 10 PMCA 139H and round 10 PMCAsi mixture of DY and 139H did not differ significantly (p>0.05) but did differ significantly (p<0.05) from animals inoculated with either round 10 PMCA DY or UN PMCA samples (Table 4). The weight of hamsters at the onset of clinical disease inoculated with round 10 PMCAsi DY and 139H mixture did not statistically (p>0.05) differ from the weights of hamsters inoculated with round 10 PMCA 139H reactions but did differ statistically (p<0.05) from animals inoculated with round 10 PMCA DY or UN PMCA reactions (Table 4). Overall, the hamsters inoculated with the round 10 PMCAsi DY and 139H mixture had the same incubation period, clinical signs, weight gain, and PrPSc properties of the round 10 PMCAsi 139H group. This is consistent with our findings that 139H and DY do not interfere (Tables 1 and 2) and suggest that a new prion strain was not produced.
Abundance of PrPC and PrPSc in PMCA strain interference reactions. The relative amounts of PrPC and PrPSc in a single PMCA reaction can be determined using the epitope accessibility immunoassay (EAI) (Fig 7). In the mock-infected seeded reactions, PrPC abundance was significantly (p <0.05) reduced (1.32±0.06 fold), and PrPSc abundance was not detected (0.06±0.21) compared to the unamplified control (Fig 7). In reactions seeded with HY, a positive control short incubation period high efficiency PrPSc converting strain, PrPC abundance was significantly (p<0.05) reduced (5.86±1.04 fold) and PrPSc abundance significantly (p<0.05) increased (11.16±1.01 fold) compared to the unamplified control (Fig 7). In reactions seeded with either DY, 139H or ME7H, PrPC abundance was significantly (p<0.05) reduced (1.62±0.13, 1.58±0.16 or 1.12±0.04 fold) and PrPSc abundance significantly (p<0.05) increased (4.02±0.61, 3.47±0.22 or 3.31±0.62 fold) respectively compared to the unamplified control (Fig 7). In reactions seeded with either 139H and DY, or ME7H and DY, PrPC abundance was significantly (p<0.05) reduced (1.63±0.07 or 0.81±0.07 fold) and PrPSc abundance significantly (p<0.05) increased (3.32±0.13 or 3.11±0.36 fold) respectively compared to the unamplified control (Fig 7). A significant (p<0.05) fold decrease of PrPC abundance was found for HY compared to DY, 139H, ME7H, 139H / DY mixture, and ME7H / DY mixture samples. A significantly (p<0.05) greater fold increase of PrPSc abundance was found for HY when compared to DY, 139H, ME7H, 139H / DY mixture, and ME7H / DY mixture (Fig 7). Overall, we found that prion strains with a relatively lower prion conversion efficiency consume correspondingly lower amounts of PrPC, compared to HY that has a relatively higher efficiency of prion conversion.
Previous strain interference studies have examined the capacity of a long-incubation period strain that have low prion conversion efficiencies to interfere with a short-incubation strains that have relatively higher prion conversion efficiencies [46,51–55]. To investigate if two long-incubation period low conversion efficiency prion strains can interfere with each other we co-infected hamsters with DY and 139H or ME7H. These three strains have similar relatively low prion conversion efficiencies (Fig 4). Additionally, 139H and ME7H were chosen because 139H has a shorter incubation period compared to DY and the incubation period of ME7H is much longer. In addition, all three of these strains have a longer incubation period compared with short-incubation period hamster strains such as HY. We found that co-infection of hamsters with DY in combination with either 139H or ME7H resulted in animals developing clinical signs of disease with an incubation period comparable to animals inoculated with the shorter incubation period strain alone (Table 1), suggesting that strain interference was not occurring. Additionally, in the animals co-infected with DY and ME7H, we found evidence that the PrPSc in these animals contained a mixture of DY and ME7H PrPSc (Fig 3), providing evidence of independent strain amplification. Based on these in vivo experiments, however, we cannot exclude the possibility that DY and ME7H are not infecting the same population of neurons that is needed for strain interference to occur [47]. For example, i.c. inoculation of hamsters with 139H scrapie prior to superinfection with Sc237 scrapie results in animals developing clinical signs, pathology and an incubation period similar to animals inoculated with Sc237 alone [57]. This data suggests that 139H is unable to interfere with Sc237 prions. Recent work using the sciatic nerve route of inoculation (i.sc.) indicates that 139H can block HY, a strain that is similar to Sc237, from causing disease [58]. HY and 139H initially replicate in ventral motor neurons (VMNs), following i.sc. inoculation, suggesting that the previous report of 139H and Sc237 failing to interfere with each other was not an inherent property of the strains, but rather was due to a failure to target both strains to the same location of prion conversion. We were unable to perform strain interference super-infection studies with these combinations of strains using i.sc. inoculation due to constraints on the relationship of incubation period of disease, the relative onset of PrPSc formation in VMNs and the lifespan of the host. To overcome this obstacle, we used PMCA to further examine strain interference between these long incubation period, low conversion efficiency prion strains since the relative onset of prion conversion of the strains governs which strain will emerge, independent if the strains were co-infected or super-infected [47].
Low prion conversion efficiency strains amplify independently when present together in PMCA. Strain interference can be recapitulated with PMCA similar to what is observed in vivo [50]. Since both prion strains are present in the same PMCA reaction, this in vitro system mimics when both strains are infecting the same cell in vivo. We found that in PMCAsi the strain that started in greater abundance maintained its strain properties through all 10 rounds of serial PMCAsi, suggesting the other strain was not able to interfere (Tables 2 and 3; Fig 5). Importantly, we found that when both strains started at an equal ratio, a mixture of the 19 and 21 kDa migrating unglycosylated PrPSc polypeptide was maintained for 10 rounds of PMCAsi (Tables 2 and 3; Fig 5). Based on the limitations of Western blot to resolve a mixture of PrPSc from both strains (Fig 1) and the strains used have similar PMCA-CC values (Fig 4) and a new prion strain is not generated (Table 4, Fig 6) we conclude that the strains are amplifying independently. This suggests that the failure of these strain combinations to interfere with each other in vivo is not due to a failure to target the same cell, but instead is an inherent property of the strain combinations. The concept that not all prion strain combinations result in interference suggests that strain diversity in naturally-infected animals may be greater than previously thought.
Prion strain interference occurs when PrPC abundance is limiting. Prion formation is dependent on PrPC, and evidence suggests that prion strains compete for PrPC [50,59]. By definition, therefore, if the amount of available PrPC for prion conversion is limiting, strain interference will occur. We hypothesized that PrPC abundance was not limiting with the long-incubation period, low prion conversion efficiency strains used in this study. To investigate this hypothesis, we used EAI to measure the PK resistant fraction of PrP and the fraction of PrP that binds to the monoclonal anti-PrP antibody 3F4 in the native state, which we interpret as PrPC (Fig 7). We cannot exclude the possibility that other forms of PrP (e.g. PK sensitive PrPSc) that bind to 3F4 in the native state, in addition to PrPC, contribute to the observed results, however, previous reports suggest this is unlikely [60,61]. Using this assay, we determined the abundance of PrPC and PrPSc in the PMCA reactions before and after amplification, and we found that PrPC levels following one round of PMCA were higher in reactions seeded with either DY, 139H or ME7H compared to PMCA reactions seeded with the relatively higher prion conversion efficiency strain HY (Fig 7). This is consistent with the observation that DY, 139H and ME7H have lower amounts of PrPSc produced after one round of PMCA and lower PMCA-CC (i.e. lower prion conversion efficiencies) compared to HY (Figs 4 and 7). When two low prion conversion efficiency strains are present in PMCA, after one round of PMCA, the level of PrPC is significantly higher compared to reactions seeded with HY (Fig 7). Overall, we conclude that the PrPC requirements of DY, 139H and ME7H are sufficiently low as to not result in PrPC abundance becoming limiting under the conditions tested; therefore, strain interference does not occur.
Not all forms of PrPC are equally suitable templates for prion conversion. Recent work suggests that the sialylation state of PrPC N-linked glycosylation can influence prion formation. Importantly, prion strains may convert a subset of PrPC, based on sialylation, more efficiently than other forms of PrPC [62–64]. Based on these findings, it is possible that the subset of PrPC that is convertible by both strains may have a greater influence on prion strain interference than the total PrPC abundance would suggest. In addition to post-translational modifications of PrPC, tissue specific ratios of the C1 and C2 cleavage fragments of PrPC may affect the abundance of convertible PrPC [65–67] and can affect strain emergence.
Prion strain interference can affect the emergence of a strain from a mixture. Interspecies transmission can generate new prion strains that, following intraspecies transmission, result in the emergence of a dominant strain [16,68–73]. Interference between prion strains that compete for limiting PrPC is thought to be an important parameter affecting this process [73]. Consistent with this observation, overexpression of PrPC can lead to the emergence of prion strains that are not identified when PrPC is expressed at physiological levels [74]. Additionally, extraneural prion inoculation can result in the emergence of novel strains, further suggesting that changes in the population of PrPC that is initially infected can result in profound differences in prion strain emergence [75]. This finding is consistent with previous work indicating that the initial population of cells infected by a mixture of strains has a large effect on prion strain emergence [47]. Overall, we hypothesize that the availability of strain-specific convertible PrPC can influence strain interference and alter the emergence of a strain from a mixture.
All procedures involving animals were approved and in compliance with the Guide for the Care and Use of Laboratory Animals (protocol numbers 811 and 880) by the Creighton University Institutional Animal Care and Use Committee.
Brains from terminally-ill hamsters inoculated with either the 139H (108.1 i.c. LD50/g), ME7H, HY (109.3 i.c. LD50/g) or DY (107.4 i.c. LD50/g) were homogenized to 10% w/v in Dulbecco’s phosphate buffered saline (DPBS) (Mediatech, Herndon, VA) or in PMCA conversion buffer (phosphate-buffered saline [pH 7.4] containing 5 mM EDTA, 1% [vol/vol] Triton X-100, and Complete protease inhibitor tablet [Roche Diagnostics, Mannheim, Germany] [56]. Uninfected hamster brain was homogenized to 10% w/v in PMCA conversion buffer. All homogenates were stored at -80°C.
Male Syrian hamsters (Harlan-Sprague-Dawley, Indianapolis, IN) were intracerebrally (i.c.) inoculated with 25 μl of an equal mixture of 1% w/v uninfected brain homogenate with either 1% w/v of 139H, ME7H, DY, or equal mixtures of a 1% w/v of DY and 139H or DY and ME7H-infected brain homogenates. PMCAsi generated uninfected brain homogenate, DY, 139H, or DY and 139H mixed material was diluted 1:100 and 25 μl was i.c. inoculated into hamsters. Hamsters were observed three times per week for the onset of clinical signs of prion disease and the incubation period was calculated as the number of days between inoculation and onset of clinical signs. Two tail Student’s T test (Prism Version 4.03, for windows; GraphPad Software Inc., La Jolla, CA) with a p value of 0.01 was used to compare incubation periods [55]. Animals were weighed once per week until the onset of clinical disease and weights were compared using ANCOVA analysis (Prism Version 4.03, for windows; GraphPad Software Inc., La Jolla, CA) with a p value of 0.05. Animals were sacrificed by CO2 asphyxiation and brain tissue were collected, flash frozen and stored at -80°C.
Protein misfolding cyclic amplification strain interference (PMCAsi) was adapted from a previously described protocol [50]. Briefly, samples (n = 3 per group) were placed in a Misonix Q700 sonicator (Farmingdale, NY). The sonicator output was set at amplitude 16 with an average output of 165W during each sonication cycle. The ratio of sonicated sample to uninfected brain homogenate was 1:20 for the reactions with either DY, 139H, or Me7H seeded reactions alone. For reactions with mixtures of 139H and DY or Me7H and DY seeded reactions, the ratio of sonicated to uninfected brain homogenate was 1:20 for the first round, 1:10 for the second round, and 1:2 for the remaining rounds of PMCA. Samples containing uninfected brain homogenate in conversion buffer were included in every round of PMCA as a negative control. The PMCA conversion coefficient is calculated as the reciprocal of the concentration of the highest dilution of prion-infected brain homogenate that resulted in detectable amplified PrPSc by Western blot following one round of PMCA [56]. PMCA conversion coefficients were compared using a two-tailed Student’s T test (Prism Version 4.03, for windows; GraphPad Software Inc., La Jolla, CA) with a p value of 0.05.
Western blot analysis was performed as previously described [76]. Briefly, samples were digested with 100 U/ml of proteinase K (PK) at 37°C for 30 minutes with constant agitation (Roche Diagnostics Corporation, Indianapolis, IN). The PK digestion was terminated by incubating the samples at 100°C for 10 minutes in gel loading buffer (4% w/v SDS, 2% v/v β- mercapto ethanol, 40% v/v glycerol, 0.004% w/v Bromophenol blue, and 0.5 M Tris buffer pH 6.8). Following size fractionation on 4–12% Bis-Tris gel, the proteins were transferred to immobilon P and Western blot analysis were performed using the anti-PrP antibody 3F4 (final concentration of 0.1 μg/ml; Chemicon; Billerica, MA) to recognize hamster prion protein. The Western blot was developed with Pierce supersignal west femto maximum sensitivity substrate according to manufacturer’s instructions (Pierce, Rockford, IL) and imaged on a Kodak 4000R imaging station (Kodak, Rochester, NY). The abundance and migration of PK resistant PrPSc was determined using the Kodak molecular imaging software v.5.0.1.27 (New Haven, CT). Cropped images are indicated by a vertical line and are from the same exposure of the same blot unless otherwise noted. The signal intensity of the unglycosylated PrPSc polypeptide as a function of migration distance was determined using the Kodak molecular imaging software v.5.0.1.27 (New Haven, CT).
The epitope of the monoclonal anti-PrP antibody 3F4 is accessible on PrPC in both native and denatured forms, and is largely unavailable in the native conformation of PrPSc that can be revealed following denaturation [61,77]. Using combinations of PK digestion and denaturation we investigated the relative abundance of various forms of PrP. Samples first digested with PK and then denatured are defined as PrPSc. Samples that have not been digested with PK or denatured are defined as PrPC. To accomplish this, samples were mixed with equal volumes of Dulbecco’s phosphate buffered saline (DPBS) or with PK (2.31 units/mL) and incubated at 37°C for 1 hour. The samples were examined for PrP as described previously with the following modifications [78]. A 96 well plate (Millipore, Billerica, MA) was activated with 150 μL methanol and then washed five times with 150 μL of tween tris buffered saline (TTBS) and centrifuged at 470 x g for 30 seconds adding 150 μL of TTBS after the first centrifugation. The samples were diluted into DPBS to a total volume of 150 μl and loaded onto the activated 96 well plate and centrifuged three times at 470 x g for 30 seconds adding 150 μl DPBS after the first centrifugation. The plate was incubated with to 0.3% H2O2 in MeOH for 20 mins and then centrifuged at 470 x g for 30 seconds two times adding 150 μL of TTBS after the first centrifugation. Next, one non-PK and one PK digested replicate was incubated with either DPBS or 3M guanidine thiocyanate (Sigma Aldrich, St. Louis, MO) for 10 mins and washed five times with 150 μL of TTBS and centrifuged at 470 x g for 30 seconds two times, adding 150 μL of TTBS after the first centrifugation. The wells were incubated with 5% w/v blotto in TTBS for 30 mins at 37°C. Blotto was removed and the 96 well was incubated for 1 hour at 37°C with mouse anti-hamster PrP antibody 3F4 (final concentration of 0.1 μg/ml; Chemicon; Billerica, MA) to recognize hamster prion protein and washed five times with 150 μL of TTBS and centrifuged at 470 x g for 30 seconds two times, adding 150 μL of TTBS after the first centrifugation. Next the wells were incubated with the secondary antibody HRP-conjugated goat anti-mouse antibody for 30 mins at 37°C (final concentration of 0.1 μg/ml; Thermo Scientific; Rockford, IL.) and washed five times with 150 μL of TTBS and centrifuged at 470 x g for 30 seconds rinsing with 150 μL of TTBS two times adding 150 μL of TTBS after the first centrifugation. The 96 well was developed with 40 μL per well of Pierce Supersignal West Femto Maximum Sensitivity Substrate according to manufacturer’s instructions (Pierce, Rockford, IL) and imaged on a Kodak 4000R Imaging Station. The abundance of PrP was determined using the Kodak molecular imaging software v.5.0.1.27 (New Haven, CT) and a two-tailed Student’s T test (Prism Version 4.03, for windows; GraphPad Software Inc., La Jolla, CA) with a p value of 0.05 was used to compare EAI values.
|
10.1371/journal.pntd.0007144 | Assessment of the new World Health Organization's dengue classification for predicting severity of illness and level of healthcare required | The objective of this study was to assess the validity of the new dengue classification proposed by the World Health Organization (WHO) in 2009 and to develop pragmatic guidelines for case triage and management. This retrospective study involved 357 laboratory-confirmed cases of dengue infection diagnosed at King Abdulaziz University Hospital, Jeddah, Saudi Arabia over a 4-year period from 2014 to 2017. The sensitivity of the new classification for identifying severe cases was limited (65%) but higher than the old one (30%). It had a higher sensitivity for identifying patients who needed advanced healthcare compared to the old one (72% versus 32%, respectively). We propose adding decompensation of chronic diseases and thrombocytopenia-related bleeding to the category of severe dengue in the new classification. This modification improves sensitivity from 72% to 98% for identifying patients who need advanced healthcare without altering specificity (97%). It also improves sensitivity in predicting severe outcomes from 32% to 88%. In conclusion, the new classification had a low sensitivity for identifying patients needing advanced care and for predicting morbidity and mortality. We propose to include decompensation of chronic diseases and thrombocytopenia-related bleeding to the category of severe dengue in the new classification to improve the sensitivity of predicting cases requiring advanced care.
| Dengue fever, the most prevalent arthropod-borne viral disease in humans, has been conventionally classified into four main categories: non-classical, classical, dengue hemorrhagic fever, and dengue shock syndrome. Several studies reported lack of correlation between the categories of the conventional classification and the disease severity. As a consequence, the World Health Organization proposed in 2008 a new classification that divides dengue into two categories: non-severe and severe dengue; the non-severe dengue is further divided into two categories: dengue with warning signs and dengue without warning signs. In this retrospective study we reviewed 357 cases of dengue diagnosed in our institution over a 4-year period to assess the validity of the new dengue classification in order to develop pragmatic guidelines for case triage and management in the Emergency Departments. We found that the sensitivity of the new classification for identifying severe cases was limited even though it had a higher sensitivity for identifying patients who needed advanced healthcare compared to the old one. We propose adding decompensation of chronic diseases and low platelets-related bleeding to the category of severe dengue in the new classification. This modification dramatically improves the sensitivity for identifying patients who need advanced healthcare and the sensitivity to predict severe outcomes.
| Dengue fever (DF) is the most prevalent arthropod-borne viral disease in humans and one of the major re-emerging communicable diseases. The World Health Organization (WHO) estimates that around 50 million dengue infections occur annually and approximately 2.5 billion of the world's population live in dengue-endemic areas [1].
Dengue virus is endemic in Saudi Arabia primarily in the western and southern provinces [2,3]. It is a well-recognized cause of seasonal outbreaks in Jeddah and Makkah [4–7]. Two large epidemics occurred in Jeddah and Makkah; the first in 2011, when 2569 cases were reported, and the second, in 2013 when 4411 cases including 8 deaths were reported [6]. Dengue was also reported in other regions of Saudi Arabia, including Al-Madinah (2009), and Aseer and Jizan (2013) [2,7].
Since the 1970s, dengue has been conventionally classified into four main categories (Table 1): non-classical DF, classical DF, dengue hemorrhagic fever (DHF), and dengue shock syndrome (DSS). DHF definition requires the presence of four criteria: fever, thrombocytopenia (<100,000 platelets/mm3), hemorrhagic manifestations, and plasma leakage manifesting as accumulation of fluids in the peritoneal, pleural, or pericardial spaces, lower limb edema, hypoalbuminemia, or hemoconcentration. Several studies reported lack of correlation between the categories of the conventional classification and the disease severity [8–10]. Despite high specificity of the DHF category, the sensitivity is unacceptably low in detecting severe cases of dengue that require specialized care and monitoring in a hospital setting [11–13]. As a consequence, a global expert consensus meeting at WHO in 2008 accorded on a new classification of DF. The revised classification divides dengue into two categories (Table 2): non-severe and severe dengue (SDF); the non-severe dengue is further divided into two categories: dengue with warning signs (D+W) and dengue without warning signs (D-W). The new classification was developed based on the level of clinical severity to establish management guidelines and to facilitate dengue reporting and surveillance. Warning signs were proposed to facilitate triage and early detection of potentially severe cases that need hospitalization, particularly in primary care settings and during outbreaks.
Several studies were conducted to assess the utility of the new dengue classification scheme in clinical practice [14–19]. After the release of the new WHO classification in 2009, the level of care required by patients with dengue was used by researchers as a gold standard to grade the severity of the illness to identify patients with severe disease that is likely to require high level of healthcare as in-patients and those with milder disease that can be managed as out-patients [11,14]. Such classification would conceivably improve outcome and reduce cost of healthcare. Additionally, the level of care is the most reliable indicator of severity in retrospective studies that were primarily used to assess the utility of the new classification in clinical practice. However, current data remain insufficient to establish the validity of this classification in predicting or identifying severe dengue cases that may need close observation or hospitalization for proper management.
This study aimed to assess the validity of the new dengue classification scheme based on data from Jeddah city as part of global evaluation of the new dengue classification.
This retrospective study included patients with dengue virus infection reported to the infection control unit at King Abdulaziz University Hospital (KAUH), Western Saudi Arabia over a 4-year period from January 2014 through December 2017. Only laboratory-confirmed cases presenting within 7 days of disease onset were included in the study. The diagnosis was confirmed if at least one of the following criteria was met in acute phase serum: (1) positive reverse transcription polymerase chain reaction (RT-PCR), (2) positive serology for dengue IgM, or (3) positive dengue-specific non-structural antigen-1 (NS1). Exclusion criteria included patients who presented 7 days after the onset of symptoms, or those who were transferred to other hospitals, or whose data were unavailable.
Persistent vomiting was defined as vomiting at least 5 times per day or vomiting everything the patient ingested. Shock was defined as tachycardia (pulse rate > 100 beats/minute) with either hypotension-for-age or narrow pulse pressure (<20 mmHg). Severe bleeding was defined as major bleeding (hematemesis, melena, or menorrhagia) associated with systolic hypotension, haemoglobin <8 g/dL, or a drop of haemoglobin of >2 g/dL, or bleeding that required blood transfusion. Renal impairment was defined as serum creatinine rise of at least 50% over the baseline that failed to improve after two days of re-hydration. The baseline creatinine was defined as the minimum value following two days of re-hydration. Peak creatinine was defined as the highest creatinine value recorded following two days of re-hydration. The baseline haematocrit (HCT) value was defined as the minimum HCT value following a minimum of two days of re-hydration provided that the patient had passed the 6th day of illness. If no baseline HCT could be defined for a given patient, the hospital's average value of normal HCT (41% for adults and 42% for pediatrics) was used as the estimated baseline value. Peak HCT was the maximum HCT value recorded during hospital stay. Patients' clinical outcomes were determined using data from the time of hospital presentation to the time of hospital discharge or demise, and from any subsequent follow up data when available.
Warning signs were only considered at patient’s presentation. The day patients developed severe dengue was documented. Lethargy was not included as a warning sign due to inadequate documentation of this symptom in the patients’ records.
Therapeutic interventions and healthcare required by patients were classified into three levels: level I, included patients who were treated on outpatient basis; level II, included hospitalized patients who received intravenous fluids for rehydration and/or those who received platelets due to thrombocytopenia that was not associated with major mucosal bleeding; level III, included hospitalized patients who required intravenous fluids for resuscitation, mechanical ventilation, blood transfusion, inotropic support, or specific treatment for organ failure.
The authors determined the level of care (the gold standard) for each patient based on the maximum healthcare required by the patient any time during his/her entire hospital stay. For instance, if a patient with dengue was stable at presentation but 2 days later the patient required ICU admission, level III of care (patients requiring intensive care) will be assigned to this patient. The aim of the classification was to determine the level of care the patient is most likely to require and to predict severe dengue before it happens. The gold standard was assessed during the phase of data collection (interventions needed were answered as yes or no) and phase of statistical analysis (patients were grouped as level I, II, III according to the pre-set definitions in the research proposal). Assignment of patients to the level of care they required was reviewed and confirmed by the senior author (TAM) who is a professor of medicine and infectious disease with a vast and long experience in managing dengue fever.
Data were analyzed using IBM SPSS, version 22. P values of <0.05 were considered statistically significant. Chi-square test was used to compare categorical variables. Diagnostic values including sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of the old and the new WHO classifications for identifying severe cases were analyzed in reference to the level of healthcare the patients required. Thus, the level of healthcare required by patients was the basis for determining severity of dengue in a retrospective manner. For the old classification, patients who were classified as DHF/DSS were considered to have severe dengue, while those classified as non-classical or classical DF were considered to have non-severe dengue. For the new classification, patients classified as SDF were considered to have severe dengue and those classified as dengue with or without warning signs were considered to have non-severe DF. With respect to the gold standard (management level), severe cases were defined as patients who were managed with level III care, whereas non-severe cases were patients who required level I or II care.
The study was approved by the Institutional Review Board of King Abdulaziz University Hospital, Jeddah, Saudi Arabia. All data analyzed were anonymized.
Of 471 laboratory-confirmed dengue cases, 357 met the inclusion criteria. Of those 357 cases, 244 (68%) were males; 53 (15%) were children (<15 years old) with a mean ± standard deviation (SD) age of 8 ± 4·4 years, and 304 (85%) were adults (≥15 years) with a mean age of 33 ± 14·1 years.
The clinical presentation and laboratory characteristics of the 357 eligible patients are summarized in Tables 3 and 4. The mean interval from the onset of illness to the hospital presentation was 4 ± 1·8 days (range 1–7, median 4 days). One of the 357 eligible patients met the inclusion criteria despite having no fever. The inclusion criteria were any patient with clinically suspected dengue that was laboratory-confirmed with positive IgM, PCR, or NS1 and presenting within 7 days after the onset of symptoms. This patient was a 25 y old Somalian woman with one day history of generalized body ache, vomiting, menorrhagia, and dyspnea Examination revealed jaundice and splenomegaly. Platelet count was 9.0 x103/mm3, WBC, 1.0 x103/mm3, and hemoglobin, 4.6 g/dl. Dengue was clinically suspected because of her thrombocytopenia and leukopenia and the fact that she presented during an outbreak of dengue. Dengue IgM was positive. She was managed with blood transfusion and supportive care and the patient fully recovered.
Since the old and the new classifications developed by the WHO were applicable to both adults and children, we decided to analyze and present the data for the whole group. Despite the small number of children (53 child or 15% of our study population), subgroup analysis was performed to see if there were any major differences between adults and children and there was none.
Of 357 eligible patients, 190 (53%) patients were hospitalized for a mean of 7·6 ± 9·8 days (range 1–95, median 5 days). Nineteen (5%) patients required admission to the intensive care unit (ICU) and 8 (2%) patients died. Of 357 eligible patients, 167 (47%) were assessed and managed in the emergency/ambulatory departments.
Cross-tabulations of the level of care (level I/II versus level III) with the old and the new classifications showed 86% and 93% proportional agreement (PA) of the old and the new classifications with the level of care, respectively (Tables 5 and 6).
There was a minimal agreement (Kappa standard error [SE] = 0·179 [0·075]; p = 0·01) between the two classifications, although proportional agreement (PA) was 85% (p = 0·003) (Table 6).
Diagnostic values including sensitivity, specificity, PPV and NPV of the old and the new classifications in identifying patients who required level III of care is presented in Table 7.
Of the 40 patients who needed level III of care, 14 (35%) patients were not classifiable as SDF using the WHO 2009 classification. Of these 14 cases, 8 had co-existing hematological conditions (sickle cell anemia and hereditary spherocytosis) that caused severe drop in hemoglobin level necessitating urgent blood transfusion, 5 patients had severe thrombocytopenia (<20,000 platelets/mm3) with evidence of bleeding, and one patient had a neurological deficit that was exacerbated by dengue. Thus, the 14 patients with SDF as per the gold standard (ie requiring level III of care) who were not identified by the new WHO classification were either patients with decompensated chronic medical diseases, hematological diseases presenting with aplastic crisis, or thrombocytopenia <20,000 platelets/mm3 and minor or major bleeding.
To improve the sensitivity of the clinical assessment in identifying patients who would require level III of care, we propose two revised versions of the new classification by integrating new criteria to the definition of SDF. The first proposed revision integrates patients with acutely decompensated chronic disease; e.g. patients with known cardiomyopathy who present with acute heart failure or patients with known hematological disease, such as sickle cell anemia, presenting with aplastic crisis. This first revision improves the sensitivity from 65% to 85%, the PPV from 74% to 79%, and the NPV from 96% to 98%, with no change in specificity (97%). The second proposed revision also integrates (besides decompensated chronic disease) patients with thrombocytopenia of <20,000 platelets/mm3 with any bleeding even if minor. This second revision results in further improvement of the sensitivity from 65% to 98%, the PPV from 74% to 80%, and the NPV from 96% to 99·7%, with without affecting specificity (97%). The diagnostic value of the first and second proposed revisions of SDF in the new classification is presented in Table 8 and the revised criteria are presented in Table 9.
Severe outcomes were defined as patients who recovered after having complications directly or indirectly related to dengue or those who died. The sensitivity of the new classification, the 1st proposed revision, and the 2nd proposed revision, for identifying patients with severe outcomes was 72%, 84%, and 88%, respectively, and the NPV was 97·7%, 99%, and 99%, respectively (Table 10).
Evaluation of the warning signs of the new classification showed that patients with hemoconcentration associated with concurrent drop in platelets count had approximately 5-fold increased risk of requiring level III healthcare (OR [95% CI] = 4·97 [2·35–10·50], p<0·001) and approximately 7-fold increased risk of severe outcome (OR [95% CI] = 6·89 [2·90–16·37], p<0·001). Other warning signs were not significant predictors of level III healthcare or severe outcome (Table 11). Correlation between the number of warning signs (<2 versus ≥2), the old and the new classifications, and the level of care is presented in Table 12.
This study compared the old and the new WHO dengue classifications as predictors of both levels of healthcare and patients' outcomes among 357 patients with confirmed dengue. It demonstrated that both classifications were inadequate in identifying patients who required advanced level of healthcare. Both classifications had a low-to-moderate sensitivity (30% to 65%), although the new classification had an acceptable sensitivity in predicting severe outcomes (72%). Consequently, we propose to revise the new classification by integrating 2 new criteria in the definition of SDF, namely, acute decompensation of chronic diseases and evidence of minor bleeding in association with thrombocytopenia <20,000 platelets/mm3. The proposed revision improved the sensitivity to 98% in detecting patients who need level III healthcare without significantly altering specificity, which remained high (97%). It also improved sensitivity of predicting severe outcomes to 88%. Furthermore, the proposed revisions had 99·7% and 99% NPVs for the level of care and outcome severity, respectively. This means that a patient who would not be classified as SDF according to our proposed criteria would have 0·3% probability to require level III of healthcare and 1% probability of having severe outcome including complications and mortality; versus 5% each for the original classification, respectively.
Pre-existing comorbidities have been considered as risk factors for progressing to severe dengue in the literature. In this study, patients with pre-existing morbidities that were clinically stable at presentation and throughout the duration of dengue illness did not need to be managed as severe cases (level III of care). On the other hand, patients known to have chronic diseases who presented with acute decompensation (unstable disease) at presentation or any time during their dengue illness needed to be managed as severe dengue and received level III of care; e.g. patients with known cardiomyopathy who presented with acute heart failure or patients with known hematological disease, such as sickle cell anemia, who presented with aplastic crisis.
Data from literature generally suggest that the new classification had a higher sensitivity as well as a higher or comparable specificity in identifying severe cases requiring higher level of healthcare in comparison with the old classification [14–19]. Several studies that compared the WHO classifications showed that a significant proportion of patients with SDF were misclassified as classical DF using the old classification, while they were readily identified by using SDF in the new classification [14–16]. Other studies compared the accuracy of the two classifications in certain clinical contexts. For example, a Brazilian study of 267 pediatric cases, showed that the old classification had a lower sensitivity (62% versus 87%) but a higher specificity (93% versus 73%) in discriminating severe cases compared to the new classification [20]. Another Brazilian retrospective study evaluating 121 autopsied individuals who died during 2011–2012 dengue epidemics, showed that the new classification had a higher sensitivity to discriminate dengue deaths and that the absence of plasma leakage and thrombocytopenia were the main reasons for failure of the old classification to discriminate DHF cases [21]. A review by Bandyopadhyay et al., analyzed 37 studies using the old WHO dengue classification, and demonstrated that this classification had a low sensitivity with frequent overlap between the different classes, especially in endemic areas [22]. Additionally, the new classification has been shown to be a better measure for case reporting and surveillance [21–23]. Conversely, other authors found no difference between the two classifications in term of sensitivity [24,25].
The new classification intended to develop a system that helps in directing patients' management and improving clinical outcome by reducing morbidity and mortality. Beyond the controversy over the usefulness of the old and the new WHO definition of SDF in identifying severe cases, the practicability of the new definition in epidemic contexts was contested, as it was perceived to be entailing heavier workload to healthcare personnel [24]. According to the new classification, all patients presenting with dengue warning signs should be admitted for observation. This would raise the hospitalization rate from 10% (DHF/DSS) to 75% (D+W/SDF) according to our data, resulting in unnecessary observation/admission, which might overwhelm healthcare personnel and resources particularly during outbreaks. As shown in our study, almost half of D+W patients were managed as outpatients, which makes the relevance of warning signs in terms of disease severity questionable [14–17]. Only hemoconcentration with concurrent drop in the platelet count was a significant predictor of severe outcome and the need for advanced healthcare. Most of the previous studies found that none of the suggested warning signs was a significant predictor of dengue severity [24,26,27], although one of them showed that the presence of five or more warning signs was a significant predictor of SDF [27]. A multi-center study across 18 countries, assessing user-friendliness and acceptance of the new classification from health professionals’ viewpoint, showed that 24% of health professionals had concerns with the new classification including: a possible increase of hospitalization rates, non-specificity of warning signs, a possible increase of cost if more patients were admitted, and the need for more training and dissemination of more concise clinical protocols [17]. These disadvantages called for a revision to improve its practicability and specificity in identifying severe cases.
In the present study, the new definition of SDF missed 14 (35%) patients who needed advanced (level III) medical care. Of these 14 cases, 8 had co-existing hematological conditions (sickle cell anemia and hereditary spherocytosis) that caused severe drop in hemoglobin level necessitating urgent blood transfusion, 5 patients had severe thrombocytopenia (<20,000 platelets/mm3) with evidence of bleeding, and one patient had a neurological deficit that was exacerbated by dengue. Adding two new criteria, namely, thrombocytopenia <20,0000 platelets/mm3 with evidence of bleeding, even though minor, and decompensated chronic illness, to the definition of SDF improved its sensitivity to identify patients who needed advanced level of care from 65% to 98% and its sensitivity to predict cases likely to have morbidity or mortality from 72% to 88%.
The first criterion suggested to be added to our proposed revision of the new classification as an indication of SDF is “acute decompensation of a preexisting comorbidity”. Hematological conditions constituted majority of the cases where 8/10 patients who had a co-existing sickle cell anemia or hereditary spherocytosis developed aplastic crisis. Dengue is a hemorrhagic virus and it’s known to cause thrombocytopenia and leukopenia but not anemia. However, a pre-existing defect along the line of red blood cells potentiated acute anemia.
The second criterion suggested to be included in our proposed revision of the new SDF definition is “thrombocytopenia<20,000 platelets/mm3 with any bleeding, even though minor”. Thrombocytopenia ≤100,000 platelets/mm3 constitutes only one of the warning signs in the original version of the new WHO classification (2009) but it is not included in the definition of SDF. In a French-Polynesian study, thrombocytopenia <20,000 platelets/mm3 was associated with a longer hospital stay, more frequent admission to ICUs, and higher mortality [28]. However, only two thirds of cases were classified as DHF/DSS and the remaining one third was classified as DF resulting in underestimation of the severity of illness [28]. This is consistent with our results demonstrating that severe thrombocytopenia (<20,000 platelets/mm3) is a strong indicator for severe dengue and the need for advanced level of care. Therefore, the new classification should be revised to include severe thrombocytopenia in the definition of SDF to improve identification of severe cases.
In our study, 9/35 patients classified as SDF received level II of healthcare and the remaining 26 patients were hospitalized for close observation and monitoring. Of these 26 patients, 3 patients had suspicion of liver damage (aminotransferases >1000 U/L), which proved to be a transient elevation of liver enzymes resolving spontaneously without complications. Among the patients who had neurological manifestations, impaired consciousness was the most frequent sign (5%), followed by meningeal signs (5%), and convulsion (3%) with one patient presenting with encephalitis. All cases with neurological manifestations benefited from conservative management. Similar to our findings, a Vietnamese study reported that impaired consciousness was the most frequent neurological manifestation [29]. Impaired consciousness is part of SDF criteria according to the new classification. A study by Gupta et al, demonstrated a greater risk for neurological complications among patients classified as classical DF in the old classification which did not include neurological symptoms or low level of consciousness at presentation as signs of SDF [30]. Recent reviews also highlighted a rising trend of neurological complications of dengue and that it is associated with more frequent hemorrhagic manifestations, higher prevalence of DSS, and increased hospital stay and mortality [31]. A study from Brazil reported 21% of neurological manifestations with confusion being the most frequent sign [32]. In Europe, an even higher prevalence (24%) of neurological manifestations was reported among imported cases of dengue infection, mainly including cases acquired in Asia and the Americas [33]. This variability in the prevalence of neurological manifestations may be explained by varying time of clinical assessment with respect to patient’s first presentation and symptoms onset. Furthermore, lack of clear definitions of organ damage that reflects the actual disease severity may lead to misplacement of patients under SDF category. In our cohort, only 61% of patients having organ damage needed level III care. Thus, standardized definition of organ damage is needed to improve the classification specificity.
Having a more sensitive classification will help capture severe cases and place patients in the right interventional care level they need to improve their quality of care and subsequently decrease morbidity and mortality associated with dengue. Having a firm classification will direct patient management as per guidelines instead of relying on individualized decisions and experiences in directing patient care. Additionally, having a more sensitive classification without altering specificity will help direct healthcare resources and avoid unnecessary hospitalization/observation which usually overwhelms health care systems especially during outbreaks.
Limitations of our study include the retrospective design yielding incomplete observations, especially for cases treated on outpatient basis, which may have compromised the robustness of the analysis. Moreover, most of the severe cases presented at a late stage, which limited the accuracy of warning signs assessment and analysis as predictors for severe dengue. Furthermore, this is a limited study in a specific area and transferability to other countries is questionable.
In conclusion, the new WHO dengue classification had low sensitivity for identifying patients in need of advanced level of care and for predicting morbidity and mortality. This classification needs to be revised to improve its sensitivity in predicting the required level of care. It is proposed to include 2 additional criteria to the definition of SDF, namely, decompensation of chronic disease and thrombocytopenia <20,000 platelets/mm3 with evidence of any bleeding, even though minor. Adding these two criteria substantially improves the sensitivity to predict cases requiring advanced level of care and to predict severe outcomes. Further prospective controlled two-arm study designs are needed from more than one region and different countries to confirm our results. Further, criteria such as organ damage and impaired consciousness should be distinctively defined to avoid overlap of the definitions and misclassification.
|
10.1371/journal.pntd.0000904 | Metabolomics to Unveil and Understand Phenotypic Diversity between Pathogen Populations | Leishmaniasis is a debilitating disease caused by the parasite Leishmania. There is extensive clinical polymorphism, including variable responsiveness to treatment. We study Leishmania donovani parasites isolated from visceral leishmaniasis patients in Nepal that responded differently to antimonial treatment due to differing intrinsic drug sensitivity of the parasites. Here, we present a proof-of-principle study in which we applied a metabolomics pipeline specifically developed for L. donovani to characterize the global metabolic differences between antimonial-sensitive and antimonial-resistant L. donovani isolates. Clones of drug-sensitive and drug-resistant parasite isolates from clinical samples were cultured in vitro and harvested for metabolomics analysis. The relative abundance of 340 metabolites was determined by ZIC-HILIC chromatography coupled to LTQ-Orbitrap mass spectrometry. Our measurements cover approximately 20% of the predicted core metabolome of Leishmania and additionally detected a large number of lipids. Drug-sensitive and drug-resistant parasites showed distinct metabolic profiles, and unsupervised clustering and principal component analysis clearly distinguished the two phenotypes. For 100 metabolites, the detected intensity differed more than three-fold between the 2 phenotypes. Many of these were in specific areas of lipid metabolism, suggesting that the membrane composition of the drug-resistant parasites is extensively modified. Untargeted metabolomics has been applied on clinical Leishmania isolates to uncover major metabolic differences between drug-sensitive and drug-resistant isolates. The identified major differences provide novel insights into the mechanisms involved in resistance to antimonial drugs, and facilitate investigations using targeted approaches to unravel the key changes mediating drug resistance.
| Visceral leishmaniasis is caused by a parasite called Leishmania donovani, which every year infects about half a million people and claims several thousand lives. Existing treatments are now becoming less effective due to the emergence of drug resistance. Improving our understanding of the mechanisms used by the parasite to adapt to drugs and achieve resistance is crucial for developing future treatment strategies. Unfortunately, the biological mechanism whereby Leishmania acquires drug resistance is poorly understood. Recent years have brought new technologies with the potential to increase greatly our understanding of drug resistance mechanisms. The latest mass spectrometry techniques allow the metabolome of parasites to be studied rapidly and in great detail. We have applied this approach to determine the metabolome of drug-sensitive and drug-resistant parasites isolated from patients with leishmaniasis. The data show that there are wholesale differences between the isolates and that the membrane composition has been drastically modified in drug-resistant parasites compared with drug-sensitive parasites. Our findings demonstrate that untargeted metabolomics has great potential to identify major metabolic differences between closely related parasite strains and thus should find many applications in distinguishing parasite phenotypes of clinical relevance.
| Health professionals are constantly challenged with the clinical polymorphism of infectious diseases. Pathogen diversity is known to play a major role in this clinically observed variability in disease manifestation, severity and drug response. However, to obtain a greater understanding of this relationship there is a need for in-depth characterisation of the diversity existing in endemic pathogen populations.
We believe that metabolomics is a powerful tool for studying such phenotypic diversity at the molecular level [1]. The advent of ultra-high mass accuracy mass-spectrometers heralded a new era in the analyses of metabolomes. This technology permits identification with a high level of confidence of low molecular weight analytes present in complex metabolite extracts [2] and thus has great potential in the unveiling of the metabolic fingerprints marking various pathogen phenotypes [1].
In this study we put our hypothesis to the test and applied a metabolomic approach to characterise clinical isolates of the parasite Leishmania donovani with different sensitivity to the antileishmanial drug sodium stibogluconate. Leishmania donovani is the causative agent of the infectious disease visceral leishmaniasis (also known as kala-azar), which is lethal if not treated [3]. Pentavalent antimonials such as sodium stibogluconate were for long used as the first-line treatment for leishmaniasis worldwide [4]. However, use of this drug was recently officially discontinued in the Indian subcontinent due to widespread resistance of the parasite to the antimonials, resulting in treatment failure in up to 60% of the patients [5], [6]. Clinical use of replacement drugs like Miltefosine could be less successful than anticipated, as their mode of action may be hampered or challenged by some of the unknown molecular adaptations present in antimonial resistant Leishmania populations [7]. Furthermore, screening for resistance to antimonials in endemic regions has been hindered as no molecular detection tools could be developed and validated [4], [8]. Hence there is an urgent need from a biological, clinical and epidemiological perspective to (i) characterise the molecular mechanisms underlying drug resistant phenotypes present in endemic parasite populations, and (ii) identify biomarkers of Leishmania drug-resistance.
We explored in this study if metabolomics is an adequate approach to address these research needs. This paper presents a proof-of-principle untargeted metabolome comparison of clinical L. donovani isolates with different antimonial sensitivity analysed with LTQ-Orbitrap mass spectrometry coupled to ZIC-HILIC chromatography. The untargeted nature of the study guarantees that we get a general overview of metabolic variability, rather than focusing on a preselected set of target metabolites. The results show that there are indeed numerous metabolic differences between the drug-sensitive and resistant isolates and thus illustrate how metabolomic approaches offer a unique potential to characterise diversity in a natural population of a major pathogen.
Written informed consent was obtained from the patients and in case of children from the parents or guardians. Ethical clearance was obtained from the institutional review boards of the Nepal Health Research Council, Kathmandu, Nepal and the Institute of Tropical Medicine, Antwerp, Belgium.
The L. donovani isolates MHOM/NP/02/BPK282/0 and MHOM/NP/03/BPK275/0 were obtained from bone marrow aspirates taken before treatment from confirmed visceral leishmaniasis patients recruited at the B.P. Koirala Institute of Health Sciences (BPKIHS), Dharan, Nepal, as described by Rijal et al. [9]. The patients received a full supervised course of Sodium Antimony Gluconate (SAG) (Albert David Ltd, Kolkata) treatment of 20 mg/kg/day i.m. for 30 days in the BPKIHS hospital. The patients were followed up for clinical and parasitological evaluation at the end of the 1-month drug course, as well as 3, 6 and 12 months after the start of treatment. Definite cure was defined as a patient with initial cure who showed no signs and symptoms of relapse at the 12-months follow-up visit. Non-responders were defined as patients with positive parasitology after a full 30-day SAG drug course.
Two clinical isolates, one antimonial-sensitive BPK282/0 and one antimonial-resistant BPK275/0, were selected for this study and were identified as L. donovani based on a CPB PCR-RFLP assay [10]. Both isolates belong to the same genomic subpopulation which is circulating in most leishmaniasis endemic regions in Nepal [11]. The two isolates were cloned using the micro-drop method [12], in order to obtain homogenous working parasite populations. Two sensitive (BPK282/0) and three resistant (BPK275/0) cloned parasite populations (further called clones) were obtained and used for further analysis. The in vitro antimonial susceptibility of the two parasite isolates and the corresponding five clonal populations was tested as described in our previous studies [9]. Although the derived clonal populations were found to have very similar drug sensitivity as the respective original parasite isolates (see Table 1), that does not preclude that the different clones of each parasite isolate differ in other characteristics.
Leishmania promastigotes were grown on modified Eagle's medium (Invitrogen) [13] supplemented with 20% (v/v) heat inactivated foetal calf serum (PAA Laboratories GmbH, Linz, Austria) pH 7.5 at 26°C. The cultures were initiated by inoculating day 3–4 stationary phase parasites in 20 mL culture medium to a final concentration of 5×105 parasites/mL; the resulting inoculated medium was equally distributed over 4 culture flasks. The four independently growing cultures of each parasite clone were further treated as biological replicates. The 5 different clones were grown synchronically with growth monitored by daily counting; the different clones were all harvested on day 3 of stationary growth phase for metabolite extraction. Day-3 stationary phase parasites were shown in pilot experiments to be the most reproducible source of metabolites, The differences in growth rate of the clones used in this study were relatively minor. The metabolite extraction protocol consists of (a) quenching (<20 sec) of L. donovani promastigotes in their culture flasks to 0°C in a bath containing a mixture of dry ice/ethanol, (b) aliquoting the necessary volume for harvesting 4×107 parasites, (c) triplicate washing of parasite cells in 1 ml of cold (0°C) phosphate buffered saline (PBS; pH 7.4 – Invitrogen) by centrifugation (20,800× g, 0°C, 3 min) and re-suspending cells using a vortex, (d) cell disruption and metabolite extraction of the washed cell pellet in 200 µl chloroform/methanol/water 20/60/20 (v/v/v) during one hour in a Thermomixer (1400 rpm, 4°C – Eppendorf AG, Hamburg, Germany), (e) separating the metabolite extract from cell debris by centrifugation (20,800× g, 0°C, 3 min) and (f) deoxygenating the extracts with a gentle stream of nitrogen gas for 1 min prior to tube/vial closure. Vials were stored at −70°C and analysed within 48 hrs.
Formic acid (ULC grade), acetonitrile (ULC grade), water (ULC grade), methanol (ULC grade) and chloroform (HPLC-S grade) were purchased from Biosolve (Valkenswaard, The Netherlands). The ZIC®-HILIC PEEK Fitting Guard column (15 mm×1.0 mm; 5 µm) and ZIC®-HILIC PEEK HPLC column (150 mm×2.1 mm; 3.5 µm) were obtained from HiChrom (Reading, UK). Gradient elution was performed using a Surveyor HPLC pump (Thermo Fisher Scientific Inc., Hemel Hempstead, UK). Elution of the ZIC-HILIC columns was carried out with a gradient of (A) 0.1% formic acid in acetonitrile; (B) 0.1% formic acid in water. The flow rate was 100 µl/min, with an injection volume of 5 µl. Gradient elution chromatography was always performed starting with 80% solvent A. Within a 6 min time interval, solvent B was increased to 40% and maintained for 12 min, followed by an increase to 90% within 4 min. This composition was maintained for 2 min, after which the system returned to the initial solvent composition in 2 min. The whole system was allowed to re-equilibrate under these conditions for 14 min.
High-resolution mass measurements were obtained with a Finnigan LTQ-Orbitrap mass spectrometer (Thermo Fisher Scientific Inc., Hemel Hempstead, UK). Optimal LTQ-Orbitrap parameters were based on previous results [14]–[16]. Briefly, the instrument was operated in both positive and negative ion electrospray mode. ESI source voltage was optimized to 4.0 kV and capillary voltage was set to 30 V. The source temperature was set to 250°C and the sheath and auxiliary gas flow rates were set respectively to 30 and 10 (machine-specific units). Full-scan spectra were acquired over an m/z-range of 50–1000 Da, with the mass resolution set to 30,000 FWHM. All spectra were collected in continuous single MS mode. The LC-MS system was controlled by Xcalibur version 2.0 (Thermo Fisher Scientific Inc., Hemel Hempstead, UK).
Raw data files acquired from analyzed samples were converted into the mzXML format by the readw.exe utility (a tool of the Trans-Proteomic Pipeline software collection, downloaded from http://tools.proteomecenter.org/wiki/index.php?title=Software:ReAdW). Further processing was handled by a flexible data processing pipeline mzMatch [17] (http://mzmatch.sourceforge.net/), performing signal detection [18], retention time alignment [19], blank removal, noise removal [20], and signal matching. In order to minimize the effects of biological and technical variation, the normalization procedure of Vandesompele et al. [21] was applied. This approach detects the signals of housekeeping metabolites, such as amino acids, and scales the data according to the variation found for those metabolites. Masses whose abundance was not reproducible for all biological replicates, as indicated by a Relative Standard Deviation (RSD) larger than 35%, were discarded, as quantification is expected to be at least 20% accurate over multiple runs [22]. Derivative signals (isotopes, adducts, dimers and fragments) were automatically annotated by correlation analysis on both signal shape and intensity pattern [23]. The derivative signals were removed before further statistical tests, as they would give excessive weight to abundant analytes with many derivatives. The selected mass chromatograms were putatively identified by matching the masses (mass accuracy <1 ppm) progressively to those from metabolite-specific databases. In a first round of identification, LeishCyc [24], LipidMAPS [25], and a contaminant database were used [26]. The latter allows removal of typical impurities and buffer components often detected in metabolomics experiments. The putative identifications for the lipids were manually annotated with the total number of carbons and double bonds in the side-chains. Only the remaining unidentified peak went through a second round of matching to KEGG [27] and a peptide database; and finally a third round was done with the Human Metabolome Database for any remaining unidentified analytes [28]. This iterative process was used in order to restrict the number of potential matches to the most likely [29]. Metabolite identification was aided by MS fragment interpretation and retention time matching to metabolite standards [15].
Statistical analysis and graphical routines were handled in R (http://www.R-project.org). Unsupervised hierarchical clustering analysis (HCA) and principal component analysis (PCA) are used to identify groups of samples that behave similarly or show similar characteristics. Hierarchical clustering algorithms build an entire tree of nested clusters out of objects in the dataset by an iterative clustering algorithm [30]. Principal component analysis (PCA) is an unsupervised multivariate analysis technique frequently used in metabolomics [31]. It implements a data dimensionality reduction of complex data matrices, so that clustering tendencies, trends and outliers can be visualized among samples. Rank products (Bioconductor RankProd Package [32]) is a non-parametric statistical method used to detect metabolites with significantly differential abundance in the two phenotypes studied [33], [34]. The R code consisting of reading and writing routines of data from/to PeakML file format (XML representation of processed data produced by the mzMatch pipeline) is available from the authors upon request.
Two parasite isolates were selected for this study; we derived two clones from the drug-sensitive clinical isolate and three clones from the drug-resistant clinical isolate for metabolic analysis (Table 1). The documented genetic homogeneity of the L. donovani population in the Indian subcontinent [35] indicates that the isolates are genetically very similar, maximizing the chances that any observed metabolic differences are related to the relative sensitivity to the antimonial drugs.
Mass spectrometry analysis of the metabolite extracts (4 biological replicates for each clone) yielded 71,000–73,000 regions of interest (mass spectrometry signals or potential peaks) per extract for positive electrospray ionisation (ESI) mode and 56,000–61,000 for negative ESI mode. Automatic detection of irreproducible and/or noise regions, as described in Materials and Methods, removed between 91–95% of the regions (i.e. non-reproducible and/or masses not producing a clear chromatographic peak), leaving a total of 4143 chromatographic peaks for positive mode and 4656 chromatographic peaks for negative mode as candidate biological analytes. Only 15–18% of these automatically extracted signals matched a compound of the selected metabolite databases (324 and 237 matches for positive and negative mode, respectively, using a mass accuracy <1 parts-per-million or ppm). The likelihood of the validity of the database hits was further assessed by manually verifying for each peak whether the retention time and mass spectrum fragment profile matched the chemical nature of the corresponding database hit. We accepted the metabolite identifications for 256 and 185 peaks from positive and negative mode respectively. Many of these metabolites (101) were present in both electrospray ionisation modes, in which case we selected the ionisation mode with the best quality signal (according to peak shape and signal intensity). Finally, a list of 340 compounds for which we had strong confidence of the identification being correct, was created. Table S1 gives this list of all the metabolites putatively identified together with the detected abundance in each sample and the Rank Product statistical analysis used to identify significant differential abundance of metabolites between the two isolates with differing drug sensitivities.
The largest class of metabolites identified is the lipids (116 glycerophospholipids, 18 sphingolipids, 9 glycerolipids, 9 sterol/prenol lipids), primarily eluting at an early chromatographic time-point as expected for HILIC chromatography. The next largest class is amino acids and their derivatives (40 amino acids, 49 amino acid derivatives subdivided in acylglycines, polypeptides and thiol compounds). Other metabolite classes detected include carbohydrates (21), fatty acyls (26), purines/pyrimidines and their conjugates (26), polyamines (3) vitamins and cofactors (10) and organic acids (9). Our total coverage is approximately 20% of the predicted core Leishmania metabolome (about 600 metabolites, excluding lipids; [36]), thus exceeding the number reported in previous untargeted metabolomic studies [37], [38]. The coverage over the various metabolic pathways is visualised on the L. donovani metabolic network in Figure 1, which shows 163 of the 340 identified compounds.
Unsupervised hierarchical clustering (Figure 2) of the samples (shown on x-axis) revealed that the metabolite abundance profiles of the drug-resistant and -sensitive clones differ sufficiently that they can be distinguished clearly and robustly. The 4 biological replicates from the individual clones are also correctly clustered together. Clustering of the metabolites (shown on the y-axis) reveals several large groups of metabolites that are either significantly higher or lower in the drug-resistant compared with the drug-sensitive clones. The results of the hierarchical clustering are confirmed in a principal component analysis as shown in Figure 3. Principal component analysis is a mathematical method to project a multidimensional dataset onto a smaller number of dimensions -principal components- which explain the maximum of variation in the data and thus enables the visualization of the major differences between samples. Clones of the drug-resistant and -sensitive isolate are clearly separated on the first principal component (explaining 61.8% of the total variance), while the second principal component separates the different clonal populations (explaining 8.9% of the total variance).
We only considered a metabolite to have a significantly differential profile in drug-sensitive and resistant clones when (i) there was a statistically significant differential abundance in the samples from the two phenotypes (Rank Product P-value <0.05), (ii) there was at least a 3-fold difference in average signal intensity between the two groups of samples, and (iii) the metabolite was consistently detected in all replicate samples of either all the drug-sensitive or all the drug-resistant clones. Using these criteria, we identified 100 (29.6% of those detected) compounds that differed between the samples of the two phenotypes. About half (51) of those compounds had a significant higher signal in drug-sensitive clones while the other half (49) had a higher signal in drug-resistant clones. The metabolites shown to differ in the two phenotypes participate in a variety of metabolic pathways, many related to sphingolipid, phospholipid, amino acid and purine/pyrimidine metabolism. Figure 4 shows the distribution of these 100 compounds; and 54 of those compounds have been mapped onto Figure 1. Full details are provided in Table S1. The detected compounds that are intermediates of the glycolytic pathway, the pentose phosphate pathway, and the TCA cycle, as well as growth factors and cofactors were found to be mostly similar between the two phenotypes (Figure 1, Table S1).
The most dramatic difference found between the two phenotypes is in phospholipid and sphingolipid metabolism. The heatmap in Figure 5 gives an overview of the full extent of the phospholipid/sphingolipid changes, the full details are given in Table S1. The significantly different sphingolipids (including 2 sphingomyelins) are 3.5–13 fold (median 4.1 fold) more abundant in drug-sensitive clones compared with drug-resistant clones. For the phospholipids the pattern was more complex, with 19 phosphatidylcholines (PC) and 2 phosphatidylethanolamines (PE) being significantly more abundant (3–61 fold; median 5.3 fold) in drug-sensitive clones and a different set of 10 PC and 12 PE being significantly more abundant (3–64.5 fold; median 5.7 fold) in drug-resistant clones. Scrutinizing the structural properties of the fatty acyl side chains of PE and PC lipids further revealed that the changes are of a different nature in PC lipids compared with PE lipids. Figure 6 shows that only diacyl PC lipids with highly unsaturated fatty acyl chains are enriched in drug-resistant compared with drug-sensitive clones; while all the diacyl PE lipids are more abundant in drug-resistant clones. However, the total intensity of all phospholipids (110) detected was almost identical in the 2 phenotypes.
A second major class of metabolites significantly modified in our drug-resistant parasites were the amino acids and amino acid derivatives. A total of 13 amino acids, including 9 proteinogenic amino acids (Figure 1), were 3–18 fold (median 4.4 fold) more abundant in the drug-resistant compared with the drug-sensitive clones (Figure 4). The remaining 11 proteinogenic amino acids were at similar abundance in the two phenotypes (Figure 1). In contrast to the amino acids, several purines (hypoxanthine, guanine, xanthine and adenosine) were more abundant (4–45.6 fold, median 8.7 fold) in drug-sensitive clones compared with drug-resistant clones (Figures 1 and 4). However, the related nucleotides that could be detected all were at similar levels in the 2 phenotypes (Figure 1).
In this proof-of-principle study, we set out to explore whether metabolomics is applicable as a global approach to elucidate the various phenotypes present in a pathogen population. We here studied L. donovani and used clones of an antimonial-sensitive clinical isolate and an antimonial-resistant clinical isolate. The two isolates are known to be genetically very similar [11], [35]. The molecular adaptations leading to antimonial resistance in natural Leishmania populations are still poorly understood; hypothesis-driven approaches have yielded fragmentary knowledge and suggest that antimonial resistance is multifactorial [39]. However, here we compared the global metabolomic profiles of the two phenotypes, and this has proved to be a method by which to clearly distinguish drug-sensitive and resistant isolates. Moreover, the data obtained highlights major metabolic differences between the two phenotypes which have not been reported before. The extraction procedure using chloroform/methanol/water 20/60/20 (v/v/v) leads to an enrichment of hydrophobic compounds in the metabolomic samples, which has revealed the notable differences in sphingolipid and phospholipid levels. However, other metabolites were also detected, with differences in amino acid and purine/pyrimidine metabolism also being observed (Figure 1 and 4).
Leishmania primarily utilize salvaged and de novo synthesized sphingolipids/sphingomyelins as a source of phosphorylethanolamine for phospholipid biosynthesis, particularly phosphatidylethanolamine (PE) [40], [41] (Figure 1). Our data on the steady-state lipid pools shows that there are clear differences in the metabolites of the pathways of both sphingolipid and phospholipid biosynthesis. Sphingolipids and sphingomyelins are less abundant in drug-resistant parasites, which could be consistent with their consumption at a higher rate to fuel PE biosynthesis which are more abundant in the resistant parasites (Figure 1). In contrast to PE, phosphatidylcholine (PC) profiles were changed in a more balanced manner; drug-sensitive clones had higher levels of PC with low fatty acyl unsaturation, while drug-resistant clones were enriched in PC with high fatty acyl unsaturation (Figure 6). This differential unsaturation profile in PC is unlikely to relate directly to the sphingolipid/PE pathway differences, but could point to another major metabolic difference between the 2 phenotypes. Although there are clear differences in the abundance of individual phospholipids, the total phospholipid content detected here appears to be similar in the 2 phenotypes. The total membranes (plasma and internal) of Leishmania contain 10–20% PE and approximately 40% PC [41], [42]. PE and PC are major components of all membrane types (e.g. plasma membrane comprises approximately 35% PE and 15% PC; mitochondrial membrane is approximately 10% PE, 25% PC) [41]–[43], hence it is not possible to know at present how the observed changes in phospholipid composition relate to functional changes in individual membranes. Nevertheless, the differences observed are strongly indicative that there are some functional differences too. High fatty acyl unsaturation, which is enhanced in the PC of drug-resistant parasites, is generally thought to decrease the ordered state of membranes and increase membrane fluidity [44], [45]. Changes in membrane fluidity due to modified lipid composition have also been reported for Leishmania parasites resistant to several other drugs including miltefosine [42], [46], amphotericin B [45], atovaquone [47] and pentamidine [48]. It was demonstrated that such changes in lipid metabolism affect (i) interaction between drug and plasma membrane and subsequent drug uptake [42], [45], [47], [49] and/or (ii) the membrane potential of the mitochondria [48]. Thus the major phospholipid changes we have identified here in antimonial resistant clones may also have some impact upon the transport of antimonials. Modified uptake, export or sequestration of antimonials (or a metabolite of it) could underlie the modified antimonial susceptibility of these parasites.
Leishmania are auxotrophic for many amino acids and must scavenge them from their environment. Additionally, they can also use amino acids, particularly proline, as a carbon source. Hence, free amino acids present in the environment are readily taken up by a large family of amino acid permeases [50], [51]. Purine biosynthetic enzymes are absent in Leishmania, and the parasite depends entirely on nucleobase and nucleoside transporters to salvage from their environment [52]. The large changes in membrane-associated phospholipids observed here in drug-resistant clones could also affect uptake of both amino acids and purines, and account for the detected differences in the intracellular abundance of these metabolites between the 2 phenotypes. A large set of amino acids including several essential amino acids (tryptophan, leucine, isoleucine, histidine) and some atypical amino acids (e.g. proline betaine and hydantoin-5-propionic acid, which are present in the culture medium and may simply be taken up by the parasite) were detected at significantly different levels in drug-resistant and drug-sensitive clones. Similar differences were detected for several purines, especially nucleobases taken up by the Leishmania transporter NT3 [52]. It has been reported previously that modified lipid metabolism in other drug-resistant Leishmania resulted in significant modifications in transport of some amino acids and purines/pyrimidines which were structurally unrelated to the respective drug [49], the changes being the indirect result of modifications in plasma membrane organisation [49], [53]. Our findings also support this notion that modified membrane composition might indirectly alter transport of metabolites.
The membrane changes we have identified in the antimonial-resistant parasites is concerning with regard to the newly installed drug policy in the Indian subcontinent. It is known that the two drugs in use, miltefosine and amphotericin B (the second-line treatment), rely on their interaction with lipids in the membrane of the parasites [46], [54]. Hence, a change in membrane composition of antimonial-resistant parasites may impact upon the efficacy of these drugs. Worryingly, there is a report of increased tolerance to all three drugs in some parasite isolates of the Indian subcontinent [7]. This demonstrates the importance of identifying the molecular mechanisms underpinning drug resistance in order to be prepared for using new drugs most effectively. Untargeted metabolomics has great potential to contribute to this much needed comprehensive characterisation of pathogens circulating in endemic regions.
Our study has exemplified how the application of metabolomic approaches could play an important role in the characterisation of clinical pathogens by identifying a fingerprint of metabolic differences between various clinical phenotypes. Further experiments are currently underway to compare a much larger number of isolates representing the entire parasite population of the Indian subcontinent, in order to document the phenotypic diversity that currently exists in the L. donovani population of this kala-azar endemic region. In parallel, we are also assessing the nature and extent of genomic diversity of this parasite population by applying new sequencing technologies to characterise the whole genome of the isolates characterised by metabolomics. The integration of genomic and metabolomic approaches will result in an unparalleled source of data and promises to yield a holistic insight into the impact of endemic pathogen diversity on clinical polymorphic treatment outcome. Future application of such integrated genomic/metabolomic approaches holds great promise to address the many challenging research questions related to pathogen diversity encountered in the field of infectious diseases.
|
10.1371/journal.pcbi.1003743 | No, There Is No 150 ms Lead of Visual Speech on Auditory Speech, but a Range of Audiovisual Asynchronies Varying from Small Audio Lead to Large Audio Lag | An increasing number of neuroscience papers capitalize on the assumption published in this journal that visual speech would be typically 150 ms ahead of auditory speech. It happens that the estimation of audiovisual asynchrony in the reference paper is valid only in very specific cases, for isolated consonant-vowel syllables or at the beginning of a speech utterance, in what we call “preparatory gestures”. However, when syllables are chained in sequences, as they are typically in most parts of a natural speech utterance, asynchrony should be defined in a different way. This is what we call “comodulatory gestures” providing auditory and visual events more or less in synchrony. We provide audiovisual data on sequences of plosive-vowel syllables (pa, ta, ka, ba, da, ga, ma, na) showing that audiovisual synchrony is actually rather precise, varying between 20 ms audio lead and 70 ms audio lag. We show how more complex speech material should result in a range typically varying between 40 ms audio lead and 200 ms audio lag, and we discuss how this natural coordination is reflected in the so-called temporal integration window for audiovisual speech perception. Finally we present a toy model of auditory and audiovisual predictive coding, showing that visual lead is actually not necessary for visual prediction.
| Since a paper was published in this journal, an increasing number of neuroscience papers capitalize on the assumption that visual speech would be typically 150 ms ahead of auditory speech. It happens that the estimation of audiovisual asynchrony in the mentioned paper is valid only in very specific cases, for isolated consonant-vowel syllables or at the beginning of a speech utterance. But the view that vision leads audition is globally oversimplified and often wrong. It should be replaced by the acknowledgement that the temporal relationship between auditory and visual cues is complex, including a range of configurations more or less reflected by the temporal integration window from 30 to 50 ms auditory lead to 170 to 200 ms visual lead. This has important consequences for computational models of audiovisual speech processing in the human brain.
| Sensory processing has long been conceived as modular and hierarchic, beginning by monosensory cue extraction in the primary sensory cortices before higher level multisensory interactions took place in associative areas, preparing the route for final decision and adequate behavioral answer. However, it is now firmly established that low-level multisensory interactions are much more pervasive than classical views assumed they were and affect brain regions and neural responses traditionally considered as modality specific [1], [2].
Restricting to audiovisual interactions in speech perception, direct connections have been displayed between primary auditory cortex and primary visual cortex (e.g. [3] on macaques), and electrophysiological data on speech perception display early influence of the visual component of speech stimuli on auditory evoked response potentials (ERPs). Indeed, there appears a decrease in amplitude and latency of the first negative peak N1 and the second positive peak P2, 100 to 200 ms after the acoustic onset, when the visual component is present [4], [5]. It is still under debate to determine the specific role of direct connections between primary sensory cortices vs. the role of associative cortex and particularly the superior temporal sulcus in these early interactions [6]–[8].
The computational nature of audiovisual interactions is now the focus of a large number of recent papers. Capitalizing on the natural rhythmicity of the auditory speech input, it has been suggested [9], [10] that the visual input could enhance neuronal oscillations thanks to a phase-resetting mechanism across sensory modalities. This has led to various experimental demonstrations that visual speech improves the tracking of audiovisual speech information in the auditory cortex by phase coupling of auditory and visual cortices [11], [12].
A number of these studies have proposed predictive coding as a possible unifying framework for dealing with audiovisual interactions. Predictive coding posits that neural processing exploits a differential coding between predicted and incoming signals, with decreased activity when a signal is correctly predicted [13], [14]. Visual prediction would be responsible for early modifications in auditory ERPs evoked by visual speech decreasing latency and amplitude of N1 and P2 (e.g. [5], [7]). This has led to recent proposals about the role of specific components in neural oscillations respectively conveying top-down predictions and bottom-up prediction errors in audiovisual speech processing [8], [15].
The previously mentioned studies capitalize on the underlying audiovisual structure of speech stimuli, that is the way sounds and sights provided by the speaker are comodulated in time (so that their phase can indeed be coupled) and more generally how one modality provides adequate information for partial prediction of the other modality.
It is actually known since long that the auditory and video streams are related by a high level of cross-predictability related to their common underlying motor cause. This is displayed in a number of studies about audio-visual correlations between various kinds of video (e.g. lip parameters, facial flesh points, video features extracted from the face) and audio (acoustic envelope, band-pass filter outputs, spectral features) parameters [16]–[20].
In a recent and influential paper published in this journal, Chandrasekaran et al. [21] present a number of analyses about the “natural statistics of audiovisual speech”, based on various databases in different languages (British and American English, and French), with four major results: firstly, there is a robust correlation in time between variations of mouth opening and variations of the acoustic envelope; secondly, focusing the acoustic envelope to narrow regions in the acoustic spectrum, correlation is maximum in two regions, one around 300–800 Hz, typically where is situated the first vocal tract resonance (formant) F1, and the other around 3000 Hz interpreted by the authors as corresponding to the second and third resonances F2 and F3; thirdly, temporal comodulations of the mouth and acoustic envelope appear in the 2–7 Hz frequency range, typically corresponding to the syllabic rhythm; last but not least in the context of the present paper, “the timing of mouth movements relative to the onset of the voice is consistently between 100 and 300 ms” (penultimate sentence of the paper abstract).
Since the publication of this paper and systematically referring to it, an increasing number of neuroscience papers – including some of those cited previously – capitalize on the assumption that visual speech would be typically 150 ms ahead of auditory speech. Let us mention a few quotations from these papers: “In most ecological settings, auditory input lags visual input, i.e., mouth movements and speech associated gestures, by ∼150 ms” [7], [8]; “there is a typical visual to auditory lag of 150–200 ms in face-to-face communication” [22]; “articulatory facial movements are also correlated with the speech envelope and precede it by ∼150 ms” [12].
The invoked natural audiovisual asynchrony is used in these papers in support to development on models and experiments assessing the predictive coding theory. The assumption that image leads sound plays two different roles in the above mentioned neuroscience papers. It is sometimes used as a trick to demonstrate that the visual stimulus plays a role in modulating the neural auditory response, rightly capitalizing on a situation where a consonant-vowel (CV) sequence (e.g. “pa” or “ta”) is produced after a pause. In this case, the preparatory movement of the mouth and lips is visible before any sound is produced, hence visual prediction can occur ahead of sound and results in visual modulation of auditory ERPs [4], [5], [7].
The second role is more problematic. Considering that there would be a systematic and more or less stable advance of vision on audition around 150 ms, it is proposed that this situation would play a role in the ability to use the visual input to predict the auditory one all along the time. Audiovisual asynchrony is implicitly incorporated in a number of models and proposals.
However, as we will see in the next section, the situation studied in [21] is very specific, characteristic of a CV sequence produced in isolation or at the beginning of an utterance after a pause. The objective of the present paper is to show that, while the method proposed by Chandrasekaran et al. to estimate audiovisual delays is adequate for the onset in preparatory sequences or the start of a speech utterance, in chained sequences which actually provide the most general case in speech communication, the method should be modified. Furthermore, if an appropriate method is used, delays actually vary in a different range from the one they propose – with the consequence that “there is no 150 ms lead of visual speech on auditory speech”.
The rationale in the measure of asynchrony proposed by Chandrasekaran et al. is based on the notion of preparatory gestures (Figure 1). This is also the case of the N1-P2 studies mentioned previously (e.g. [5], [8]). This can be related to a rather classical analogy, namely the movement of a hammer towards a table (Figure 1a). To produce a sound with a hammer, one must previously realize a downward stroke and the onset of this downward stroke is visible much before the hammer touches the table and makes a sound. Notice that in this scene, one could define actually two visible events, one at the onset of the downward stroke and one at the instant when the hammer touches the table; and only one auditory event, the sound onset, which is actually perfectly synchronous with the second visual event. The downward stroke may be called a “preparatory gesture” in that it prepares the sound and hence predicts something about it (its time of arrival, and also its acoustic content since a subject looking at the hammer going towards the table knows the kind of sound which will be produced soon).
It is exactly the same for preparatory lip gestures before “p” at the beginning of a speech utterance (Figure 1b): when the lips begin to close, a subject looking at the speaker knows that they will soon join together for a lip closure, and she/he can predict rather accurately when will sound occur and what will be its spectrum (the typical flat low-frequency spectrum of a bilabial burst [23]). Here again, there are two visual events, namely the onset of the lip closing gesture and the further onset of the lip opening gesture, and only one auditory event, the burst onset, quite synchronous with the second visual event. Notice that the analogy between the preparatory gestures for the hammer and for speech is not perfect. Indeed, the sound is produced by the hammer at the end of the downward stroke, while for speech the lips must open again. There is actually a complex coordination between larynx, lungs and lips to achieve the adequate aerodynamic strategy [24], which fixes rules about the duration of lip closure before lip opening. But the audiovisual asynchrony involved in preparatory gestures for both hammer and speech are similar: in both cases, audiovisual asynchrony is assessed by the duration between two different events, the onset of the preparatory gesture for the visual channel and its offset for the auditory channel.
Therefore it appears that the crucial aspect of preparatory gestures is that they are visible but produce no sound. This could be different, actually. Consider for example what happens if you replace the hammer by a whip or a flexible stick. Now the downward stroke produces a whistling sound (which also predicts the sound produced when the whip or stick touches the table). There are now two auditory events, just as there are two visual events, and for both pairs of audiovisual events (at the beginning and end of the visual stroke) the auditory and visual events are quite in synchrony.
This leads us towards another kind of gestures that we propose to call “comodulatory gestures” since these gestures produce both auditory and visual stimuli more or less in synchrony all along the time (Figure 2). Comodulatory gestures are actually by far the most common gestures in speech. Here we should move towards another analogy that is a balloon in which one adjusts the mouthpiece. When its size increases or decreases, shape, volume and pressure change leading to more or less synchronous auditory and visual events for both opening and closing phases (Figure 2a), just as opening and closing the lips while vocalizing produces auditory and visible events quite in synchrony (Figure 2b).
In the remaining of this paper we present simple audiovisual data on plosive-vowel syllables (pa, ta, ka, ba, da, ga, ma, na), produced either in isolation or in sequence. We show that when syllables are produced in isolation, preparatory gestures provide audiovisual asynchronies quite in line with those measured in [21]. However, when syllables are chained in sequences, they provide comodulatory gestures in which audiovisual synchrony is actually precise, contrary to the data provided on similar sequences in [21], just because the measure of audiovisual asynchrony is different. In such cases, there are actually auditory events that were not taken into account in the original paper, and these need to be taken into account if one is talking about asynchrony.
After presenting Methodology and Results, we discuss how natural coordination between sound and image can actually produce both cases of lead and lag of the visual input. We relate the range of leads and lags to the so-called temporal integration window for audiovisual speech perception [25]. We propose that the “visual lead” hypothesis, wrong in many cases, is actually not necessary to deal with audiovisual predictability, and we illustrate this by briefly introducing a simple audiovisual prediction model dealing with the speech sequences studied previously. We conclude by some methodological and theoretical remarks on neurophysiological developments about audiovisual predictability in the human brain.
In the experimental work we focus on audiovisual temporal relationships in CV sequences where C is a voiced, unvoiced or nasal stop consonant that is, for English or French (the two languages considered in [21]), one of the sounds /p t k b d g m n/, and V is the open vowel /a/. We consider both CV sequences produced in isolation and chained sequences VCVCVCV. This corpus is very simple though sufficient to illustrate the difference between preparatory gestures – for isolated syllables – and comodulatory gestures – for chained syllables. The /a/ context in which the plosives /p t k b d g m n/ are produced is selected because it provides a large gesture amplitude providing more salient trajectories both in the visual and auditory modality. We will consider more general phonetic material in the discussion.
We recorded a small database of 6 repetitions of 8 syllables /pa ta ka ba da ga ma na/ uttered by a French speaker either in isolation /Ca/ or in sequence /aCa/. The syllables were produced in a fixed order at a relatively slow rhythm (around 800 ms per syllable). In the “isolated syllables” condition, syllables were embedded in silence: /pa#ta#ka#ba#da#ga#ma#na/ where /#/ means a silence (typically 500 ms silence between two consecutive syllables). In the “chained syllables” condition, they were produced in the same order though with no silence between syllables: /apatakabadagamana/.
The recording was done with a PAL camera at 50 Hz. The recording set up was based on the classical paradigm we use in Grenoble since years [26], [27] with blue make up applied on the lips. For each image, we extracted automatically and precisely the lip contours by applying a Chroma Key process extracting blue areas on the face. The lips parameters were extracted every 20 ms, synchronously with the acoustic signal, which is sampled at 22.05 kHz.
We display on Figure 5 the data for isolated syllables. In this case, where there is no audible event for closure, we report the same measure as in [21] that is the delay between the first visible event, CVL, and the first audible event, OAI or OAF. There is a very large anticipation, which actually reaches values much larger than 150 ms here (and which may reach 400 ms in some cases). These values are compatible with the range 100-300 ms proposed in [21], the more so considering that the measure used by the authors for detecting visual events (half open point in the lip closing trajectory, while we used the onset of the closing phase) would produce values lower than the ones in Figure 5.
We display on Figure 6 typical audiovisual sequences for all types of chained syllables (with a zoom around the consonant). It clearly shows that there is comodulation of the auditory and visual information, with audible and visible events for both closing and opening phases. The event detection is sometimes not straightforward or not very precise in time (e.g. detection of CAI for /ata/ or /ada/), which is quite classical in this type of stimuli, and gross trends are more important that precise values in the following.
We display on Figure 7 the data about temporal coordination between audio and visual events for either closing (Figure 7a) or opening (Figure 7b) in the case of chained sequences. The mean delay between visual and acoustic events at the closure (in the /aC/ portion, Figure 7a) varies between −20 ms and −40 ms for intensity (CVL-CAI, in green) and reaches values from −40 to −80 ms for formants (CVL-CAF, in red). This means that there is a small lead of the visual channel compared to the audio channel (where information is available on intensity before formants). But this lead is much smaller than the 150 ms lead mentioned in [21], and there are actually cases where audio and video information are available more or less in synchrony, e.g. for /ad/, /ag/ or /ak/ where the tongue gesture towards the voiced plosive decreases intensity or formants while jaw may stay rather stable, and hence lip area does not decrease much – which prevents early video detection.
In the opening phase (Figure 7b) the synchrony is even larger. Concentrating on the delay between labial and intensity events (OVL-OAI, in green) we actually observe an almost perfect synchrony for labials (/p b m/). This is trivial: as soon as the lips begin to open, the sound drastically changes, from silence (for /p/) or prevoicing (for /b/) or nasal murmur (for /m/) to the plosive burst. For velars /k g/ there is actually a clear lead of the audio channel, since the first tongue movement producing the plosive release is done with no jaw movement at all and hence before any labial event is actually detectable: the audio lead may reach more than 20 ms (see examples in Figure 6). Notice that while the video sampling frequency at 50 Hz can make the detection of the opening event for bilabials a bit imprecise with a precision around 10 ms for very quick gestures, the variations of lip area for dentals or velars is smooth and hence imprecision in event detection cannot explain such an audio lead.
Therefore the discrepancy with [21] is clear for chained syllables, just because this corresponds to what we called comodulatory gestures, for which we argue that a different measure of the audiovisual asynchrony should be used.
The experimental results presented previously show that for isolated syllables associated with preparatory gestures, our measure of audiovisual asynchrony provides quantitative estimates from 200 ms to 400 ms of visual lead (Figure 5). This is in line with the 100 to 300 ms visual lead proposed in [21], the more so considering that the estimate of the visible onset for lip closure in [21] is done at the mid closing phase – while we prefer detecting the first visible event that is at the very beginning of the lip closure phase, typically 100 ms before. The coherence of both sets of measures was expected considering that the same definition of asynchrony for preparatory gestures is used in both papers, between the first visible event (onset of lip closing phase) and the first auditory event (plosive burst at labial release).
However the data are quite different for chained sequences associated with comodulatory gestures. In this case the range of asynchronies is much more restricted and more centered around 0, from 70 ms visual lead to 20 ms audio lead when auditory events are detected on intensity, auditory events detected on the formant trajectory being somewhat delayed in respect to intensity (Figure 7). Mean video lead amounts to 35 ms in the closing phase and 0 ms in the opening phase for intensity, 60 ms in the closing phase and less than 10 ms in the opening phase for formants. Therefore the departure between our data and those proposed in [21] is now important. This is not due to variability in the speech material, but to a difference in the measure proposed for assessing audiovisual asynchrony. As explained in Figure 4, the measures differ hence their results also differ. Speech gestures in chained sequences typically produce both auditory and visual events all along the time (see Figure 6) hence resulting in a rather precise audiovisual synchrony in most cases.
Preparatory gestures do exist in speech communication, and ERP studies rightly capitalized on this experimental situation in which the gap between the first visible and the first auditory event may be quite large and able to lead to significant influence of the visual input on the electrophysiological response in the auditory cortex, for both speech [5], [8] and non-speech stimuli [29], [30]. Notice that this may actually depend on the prephonatory configuration: if somebody keeps the lips closed while listening to the interlocutor, there will actually be no preparatory gesture before an initial bilabial sound such as /b/ or /m/, and hence there will be no visual lead at all in this case. One could even imagine a reverse situation in which a speaker keeps the lips closed and systematically signals her/his turn taking by a backchannel signal “hmm” (which is not rare): in this case the preparatory gesture would be actually audible and not visible, leading to an auditory lead in the preparatory phase.
However, most of the speech material is made of comodulatory gestures. Of course, speech utterances involve a range of phonetic configurations much larger than the /Ca/ sequences that were studied in this paper. This variety of configurations leads to a variety of situations in terms of audiovisual asynchronies. This is where the analogy we proposed previously with the deflating balloon being both audible and visible reaches some limits: actually, not every action realized on the vocal tract is always either audible or visible, which may lead to delays between perceivable auditory or visible cues for a given speech gesture.
A first general property of speech concerns anticipatory coarticulation – much more relevant and general than preparatory movements discussed in [21]. This relates to articulatory gestures towards a given phonetic target, which can begin within a previous phoneme. Anticipatory coarticulation generally capitalizes on a property of the articulatory-to-acoustic transform, in which an articulatory gesture has sometimes no or weak effect on the sound and hence can be prepared in advance without audible consequences.
A typical example concerns the rounding gesture from /i/ to /y/ or /u/ in sequences such as /iC1C2…Cny/ or /iC1C2…Cnu/ with a variable number of consonants C1…Cn not involving a specific labial control (e.g. /s t k r/) between the unrounded /i/ and the rounded /y/ or /u/. In this case the rounding gesture from /i/ towards /y/ or /u/ can begin within the sequence of consonants /C1C2…Cn/, and hence anticipate the vowel by 100 to 300 ms [31]. Various sets of data and various theoretical models of this anticipatory coarticulation process have been proposed in the literature [32]–[36]. In such cases the rounding gesture can hence be visible well before it is audible.
So there are cases where visible information is available before auditory information (e.g. in /iC1…Cnu/ sequences), others where vision and audition are quite synchronous (e.g. in /aCa/ sequences), and there are also cases where audition may actually lead vision as was shown e.g. in Figure 7. But the next question is to know if the auditory and visual systems are able to process the information efficiently as soon as it is available. This is actually not always the case, and in gating experiments on the visual vs. auditory identification of coarticulated sequences, Troille et al. [37] display in some configurations a lead of audition on vision which can reach up to 40 ms, because of the poor visibility of some articulatory gestures. This leads the authors to claim that they have discovered a case where “speech can be heard before it is seen”.
In summary, there are actually a variety of situations from audio lead (estimated to 40 ms in [37]) to visual lead (which can reach more than 200 ms). In their study of mutual information between audio and video parameters on speech sequences, Feldhoffer et al. [38] show that mutual information is maximal for some audio and video parameters when it incorporates a video lead up to 100 ms. In audiovisual speech recognition experiments, Czap [39] obtains a smaller value, recognition scores being higher with a small global video lead (20 ms). Altogether, these global estimations are concordant with the classical view that “in average, the visual stream may lead the auditory stream”, which is generally advocated by specialists of audiovisual speech perception (e.g. [40], [41]). However, the “average” view hides a large range of variations, typically inside a window between 40 ms audio lead to 200 ms visual lead in the phonetic content of normal speech communication.
A large number of recent studies have attempted to characterize the temporal integration window in various kinds of multisensory interactions. This typically involves two kinds of paradigms. Firstly, evaluation of intersensory synchrony may be based on either simultaneity or temporal order judgment tasks (see a recent review in [42]). Secondly, the “multisensory temporal binding window” describes the range of asynchronies between two modalities in which a fused percept may emerge [43].
The “audiovisual temporal integration window” is well described for speech perception (e.g. [44], [45]). Van Wassenhove et al. [25] compared estimates of audiovisual temporal integration window based on either simultaneity perceptual judgments or regions where the McGurk effect seems to stay at a maximal value. They show that these various estimates converge on an asymmetric window between about 30 ms audio lead and 170 ms audio lag.
This provides a set of values rather coherent with the range of possible asynchronies in the speech material itself. Small audio leads may occur because of the lack of visibility of certain audible gestures, as shown in Figure 7 or in gating experiments [37]. Large video leads are mostly due to labial anticipatory coarticulation and described in many studies [31]–[36]. A tentative interpretation is that the perceptual system has internalized this range through a learning process. This is in line with the so-called “unity assumption” [46] according to which subjects would naturally bind together multisensory stimuli referring to a common cause, which would lead to both fused percepts and decreased ability to detect temporal asynchronies [47]. We speculate that unity assumption is based on a statistical learning of the comodulation properties of the auditory and visual streams in the speech natural environment, naturally providing an asymmetrical window around the range [−30 ms, +170 ms].
The asymmetry of the temporal integration window has been the topic of much discussion – including assumptions about the difference between optic and acoustic wave speeds, which cannot however explain such a large asymmetry: a speaker 10 m apart from a listener would not provide more than 30 ms visual advance! We argue here that the psychophysical asymmetry just mirrors the natural phonetic asymmetry, according to which there are plenty of cases of large visual anticipation due to coarticulation – typically in the 100 to 200 ms range – and less cases of auditory anticipation, in a smaller range – typically less than 40 ms as displayed in our data in Figure 7 or in gating data [47]. But, once again, this does not mean that there is a constant visual lead, but rather a range of audiovisual asynchronies mirrored in the temporal integration window.
Recent data on the development of the audiovisual temporal integration window fit rather well with this proposal. Indeed, these data show that the window is initially quite large and then progressively refined by “perceptual narrowing” in the first months of life [48]. The window actually appears to stay rather wide and symmetrical until at least 11 years of age [49]. It is only after this age that the left part of the window (for auditory lead) refines from 200 ms to 100 ms, which is proposed by the authors as the typical value for adults (the fact that these values are larger than in [25] likely comes from the use of a different criterion to define binding windows from simultaneity curves). On the contrary, the right part of the window stays stable. The interpretation is that the large initial symmetric window [−200 ms, +200 ms] is progressively tuned to the window characteristic of the speech input, asymmetric in nature. The fact that learning the asymmetrical pattern occurs so late may appear surprising, but it is in fact compatible with data showing that the maturation of the McGurk effect is not complete before at least 8 years of age for native stimuli and even later for non-native stimuli [50].
There is also a rather large deal of variations of audiovisual temporal integration window from one subject to another [43]. These variations respect the asymmetry trend, though with large variations in quantitative values. The fact that these variations are correlated with the results of various fusion paradigms suggests that inter-individual differences could be related with specific weights attributed by subjects to one or the other modality [51], [52]. Interestingly, it also appears a large ability to tune and decrease the integration window with auditory or visual experience [53], [54], including the possibility to decrease the asymmetry and specifically decrease the large visual-lead part of the window, which suggests that the integration window actually combines stimulus-driven content with individually-tuned perceptual experience.
The data recalled in the previous section rule out over-simplistic claims about audiovisual predictability. Does it raise a problem for predictability in general? The answer is clearly no. The reason is that predictability does not require asynchrony. Actually, a pure auditory trajectory may provide predictions on its future stages, and the visual input may enhance these predictions, since it is naturally in advance on future auditory events, though not systematically in advance on present ones. This is illustrated on the toy model presented in [55] and sketchily introduced here under (see a detailed presentation in the Supplementary Text S1).
The model was developed for dealing with a corpus of repetitions of sequences /aba/, /ada/ and /aga/ uttered by a male French speaker. A predictive coding model was developed to provide guesses about the closure point of the acoustic trajectory /aC/ (with C one of the plosives /b, d, g/) from a given point of the trajectory. We implemented such a model within a Bayesian probabilistic framework, comparing predictions provided by audio-alone inputs with predictions provided by audiovisual inputs.
Importantly, audiovisual inputs were shown to produce better predictions, providing values closer to the actual endpoint than with audio-only inputs. This shows that the visual component provides information able to improve predictions. This toy model is of course highly oversimplified in respect to what should be a reliable system dealing with the whole complexity of speech. However it presents the interest to show that the visual input may strongly improve predictions, in spite of the close synchrony of basic temporal events in the auditory and visual streams, according to the data presented in the Results section. In a word, there is no theoretical requirement for visual lead to argue that visual predictive coding could be at work in the sensory processing of speech in the human brain.
The impressive advances of neurosciences on the processing of speech in the human brain, sometimes simplify the complexity of speech, and miss or forget a number of evidence and facts known from long by phoneticians – on the structure of phonetic information, on the auditory and visual cues, on some major principles of speech perception and production. In consequence, there is a serious risk that these advances oversimplify “much of the known complexity of speech as [it] is spoken and of speakers as they speak” [56].
This paper attempts to make clear that the view that vision leads audition is globally oversimplified and often wrong. It should be replaced by the acknowledgement that the temporal relationship between auditory and visual cues is complex, including a range of configurations more or less reflected by the temporal integration window from 30 to 50 ms auditory lead to 170 to 200 ms visual lead.
It is important to recall that fortunately, this caveat does not put in question the experimental studies that capitalized on the presumed “150-ms video lead” to assess audiovisual interactions in EEG or MEG data. Indeed, all these studies (e.g. [4], [5], [7]) used isolated plosive-vowel syllables for which the preparatory visual movement is actually realized without any audio counterpart, hence producing a clear visual anticipation (see Figure 5).
But the pervasive message linking visual lead and visual prediction within a predictive coding stance needs some refinement. Actually, as shown in the last part of this paper, audiovisual predictability does not require audiovisual asynchrony. The development of realistic computational proposals for assessing auditory and audiovisual prediction coding models in speech perception is a challenge for future work in cognitive neuroscience. For this perspective, precise knowledge of the natural statistics of audiovisual speech is a pre-requisite. A number of useful and important data and principles were provided in [21], though the last of its four conclusions needed some refinement. The present paper hopefully contributed to enhance the available knowledge about the complexity of human speech.
|
10.1371/journal.pgen.1000595 | The CUGBP2 Splicing Factor Regulates an Ensemble of Branchpoints from Perimeter Binding Sites with Implications for Autoregulation | Alternative pre-mRNA splicing adjusts the transcriptional output of the genome by generating related mRNAs from a single primary transcript, thereby expanding protein diversity. A fundamental unanswered question is how splicing factors achieve specificity in the selection of target substrates despite the recognition of information-poor sequence motifs. The CUGBP2 splicing regulator plays a key role in the brain region-specific silencing of the NI exon of the NMDA R1 receptor. However, the sequence motifs utilized by this factor for specific target exon selection and its role in splicing silencing are not understood. Here, we use chemical modification footprinting to map the contact sites of CUGBP2 to GU-rich motifs closely positioned at the boundaries of the branch sites of the NI exon, and we demonstrate a mechanistic role for this specific arrangement of motifs for the regulation of branchpoint formation. General support for a branch site-perimeter–binding model is indicated by the identification of a group of novel target exons with a similar configuration of motifs that are silenced by CUGBP2. These results reveal an autoregulatory role for CUGBP2 as indicated by its direct interaction with functionally significant RNA motifs surrounding the branch sites upstream of exon 6 of the CUGBP2 transcript itself. The perimeter-binding model explains how CUGBP2 can effectively embrace the branch site region to achieve the specificity needed for the selection of exon targets and the fine-tuning of alternative splicing patterns.
| Alternative splicing is a precisely controlled process that determines whether an exon will be included or skipped in the mature mRNA transcript. Factors that control alternative splicing bind to RNA sequence motifs in the exon or flanking introns and guide tissue and developmental specific splicing events. CUGBP2 is a dual functional regulator of alternative splicing that can cause inclusion or skipping of a target exon, depending on the context of its binding motifs. Previously, the mechanisms of regulation by this protein and the positional significance of its target motifs have not been characterized. In this study, the authors dissected the mechanism of exon skipping by CUGBP2 and demonstrate that a specific configuration of motifs at the perimeters of a functional reference point are intimately involved in this event. Furthermore, this mechanism of regulation is shown to have general significance because novel CUGBP2 target exons contain a similar arrangement of motifs. The most interesting of this group is an exon within the CUGBP2 transcript itself. This study underscores the importance of a functional reference point in the specificity of regulation by an alternative splicing factor and reveals a novel autoregulatory role for CUGBP2.
| Alternative pre-mRNA splicing is prevalent throughout vertebrate genomes where an individual gene can be diversified into hundreds or even thousands of related mRNA isoforms [1],[2]. Functional consequences of alternative splicing can involve changes to a subset of the protein's biochemical properties or subcellular localization. These are powerful mechanisms used to regulate protein functions across different cell types, during development, or in response to extracellular signals [3],[4]. One of the major challenges in postgenome biology is to understand how alternative splicing, which involves a high degree of inherent flexibility, can achieve the specificity needed to select the correct set of target transcripts for regulation.
The spliceosome is the functional context for regulation, since this is the macromolecular machinery that guides intron removal and exon joining. It is assembled from the dynamic associations of five small nuclear ribonucleoprotein particles (snRNPs) and hundreds of accessory factors [5],[6]. Initially, U1 snRNP and U2AF (U2 snRNP auxiliary factor) recognize the 5′ and 3′ splice sites of the exon, respectively, and U2 snRNP base pairs with the branch site region thereby designating the adenosine to be used as the branchpoint. The association of U456 tri-snRNP and various RNA rearrangements then activate the first step of catalysis, which generates the 5′ exon and lariat intron-3′ exon intermediate. As catalysis advances to the second step, the lariat intron is excised and the 5′ and 3′ exons are ligated. The overall pattern of exon inclusion/skipping depends on the ability of the spliceosome to recognize each splice site signal, which is a reflection of the inherent strength of the site as well as the regulatory effects of splicing factors acting from sequence motifs nearby [2],[7]. Exon definition, which involves the interactions of U1 snRNP bound to the 5′ splice site and U2AF bound to the 3′ splice site across an exon is a particularly sensitive mechanism to specify alternative splicing patterns [8],[9].
Two families of RNA binding proteins known to regulate alternative splicing by direct recognition of RNA sequence motifs include the arginine-serine rich (SR) and hnRNP splicing factors [10],[11]. SR splicing factors most commonly recognize exonic splicing enhancers, whereas hnRNP proteins recognize intronic or exonic splicing silencers and enhancers. These regulatory motifs are typically short and degenerate making it difficult to reliably predict the target exons of a splicing factor based upon sequence inspection alone. CUG Binding Protein 2, or CUGBP2 (also called NAPOR, CELF-2, ETR-3, or BRUNOL3) is a member of the larger family of CUGBP and ETR-3-like (CELF) RNA binding proteins, which have been shown to regulate alternative splicing through UG-rich motifs in accordance with their tissue-specific expression patterns [12],[13]. CELF proteins have been shown to play important roles in heart development, whereas their misregulation has been implicated in the pathogenesis of myotonic dystrophy [14]–[16].
We previously reported a close correlation between the distribution of protein expression patterns of CUGBP2 (called NAPOR in the previous study) and splicing patterns of the NI and CI cassette exons of the NMDA R1 receptor transcript (GRIN1 gene) in the rat brain [17]. In particular, high levels of CUGBP2 protein in the forebrain were associated with skipping of the NI exon and inclusion of the CI exon, and these splicing patterns reversed in the hindbrain where CUGBP2 was deficient. In vivo splicing reporter assays confirmed dual functional roles for CUGBP2 as a splicing silencer of the NI exon and as an enhancer of the CI exon. These dual roles are thought to be important in directing the brain region-specific distribution of GRIN1 mRNA isoforms for fine-tuning of receptor functions at the synapse [17]. Additional splicing factors have been shown to exhibit dual roles in enhancement and silencing depending upon the context of the target exons, but these mechanisms are, at present, poorly understood [18],[19].
In this study, we focus on the silencing face of CUGBP2's dual character to understand how it recognizes the NI target exon and the mechanism used for splicing silencing. A variety of intronic UG-rich motifs can be found within several hundred nucleotides of the NI and CI exons by sequence inspection, but functionally significant motifs in these regions have not yet been identified. We initially used a chemical-based RNA footprinting approach to detect RNA-protein interactions at nucleotide-level resolution. Here we identify the direct contact sites of CUGBP2 in the 3′ splice site region of the NI exon, and establish that this arrangement of binding sites plays a mechanistic role in silencing a group of branch sites in between. We show the significance of this mechanism by demonstrating its involvement in the regulation of other skipped exons, most notably exon 6 of the CUGBP2 transcript itself.
A silencing role for CUGBP2 was shown for the NI exon of the GRIN1 transcript in a previous study but the mechanism of silencing was not characterized [17]. To gain insight into the mechanism, we sought to extend this analysis to identify the sequence and spatial arrangement of motifs associated with direct binding of CUGBP2 and silencing of the NI exon. A nitrocellulose filter binding assay was used initially to locate the RNA region involved in stable binding by CUGBP2. Bound/total RNA was plotted as a function of increasing protein concentration and data were fit to a hyperbola to estimate the dissociation constants (Kd) for binding to individual RNA transcripts. These RNA substrates included the NI exon, the NI exon and flanking introns, and the upstream intron, substrates E5-8, E5-10, and E5-15, respectively (Figure 1A). CUGBP2 was found to bind to the E5-10 and E5-15 substrates containing the upstream intron region with apparent dissociation constants in the nanomolar range (96 and 92 nM Kd values), but not to the E5-8 substrate containing the exon alone (Figure 1B). Because the downstream intron region is not a common feature of the high affinity binding substrates, this region, as well as the exon, must be dispensable for high affinity binding.
To identify the specific nucleotides contacted by CUGBP2, we next carried out chemical modification footprinting with the E5-10 substrate, which contains the high affinity region identified by filter binding. CMCT modification at the N3 position of uracil and the N1 position of guanine causes termination of reverse transcription initiated at a downstream primer [20]. This chemical was chosen for footprinting because of the binding preference of human ETR-3 for (C)UG-rich motifs as indicated by an iterative selection procedure [13]. A representative footprint of the region upstream from the NI exon is shown (Figure 1C). These results reveal that a core (UGUGU) and upstream flanking (GU) motif are protected by CUGBP2 in a dose-dependent manner. Regions within the NI exon and in the downstream intron were also subjected to RNA footprinting with CUGBP2, but no additional binding sites were detected in agreement with the filter binding experiments.
To verify the specificity of the assay, additional footprinting reactions were carried out with purified splicing factors U2AF and PTB. Protected regions were distinct from those observed for CUGBP2 and consistent with the known RNA binding specificities of these factors (Figure S1, lanes 1–8). We also demonstrate that PTB can compete with CUGBP2 for binding to this region of RNA. That is, when PTB binds to its cognate sites which overlap with the core UGUGU motif, the pattern of CUGBP2 protection is lost from not only the core motif but also the upstream GU motif, suggesting that one CUGBP2 protein simultaneously contacts both of these sites (Figure S1, lanes 9–12). Notably, the core and flanking motifs protected by CUGBP2 are located at the boundaries of the predicted branch site region with the nucleotides protected by U2AF also within these boundaries. Thus, these results together with the high sequence conservation of the motifs (100%) across human, rat, mouse, fruit fly, and chicken genomes, support their involvement in the mechanism of silencing.
In order to determine whether nucleotides in contact with CUGBP2 upstream from the NI exon are necessary for its silencing role, we generated an in vivo splicing reporter with the NI exon and its immediate adjacent flanking introns inserted between β globin exons 1 and 2 (DUPNIwt) (Figure 2A). Mutant derivatives of this reporter contained site-directed mutations in the GU dinucleotide and UGUGU core motifs (m1 and m2 motifs, respectively) as identified by footprinting. A nearby UGUG motif (m3) was also mutated, since it showed weak protection in the footprinting assays (data not shown).
Splicing reporters were co-expressed in the presence and absence of CUGBP2 in C2C12 mouse myoblast cells or N18TG2 mouse neuroblastoma cells, which have little or no endogenous protein as shown by Western blotting (Figure S2). The change in exon inclusion value, ΔEI, was then calculated as the difference between the % exon inclusion±CUGBP2. The ΔEI value is used here as a convenient measure of the effectiveness of CUGBP2 to induce exon skipping (or silencing). While CUGBP2 expression in C2C12 cells induced exon skipping of the wild type substrate with a ΔEI value of −33% (Figure 2B, lanes wt) mutations in all three motifs eliminated silencing entirely as indicated by a ΔEI value of 0% (lanes m1,2,3). The double mutations also showed a significant reduction in the silencing effect of CUGBP2 (lanes m1,2, m2,3, m1,3). Of this group, combined mutations at positions m1 and m2 showed the smallest degree of silencing (ΔEI −5%), suggesting that these sites are intimately involved in the mechanism of action. Single mutations also showed a reduction in silencing indicative of additive effects (lanes m1, m2, m3). Similar results were observed in N18TG2 cells (Figure 2C). New to this cell type is higher basal levels of inclusion in the absence of CUGBP2 and reduced silencing by CUGBP2 on all substrates tested. This could be the consequence of the differential expression of splicing factors in these two cell lines. That is, an enhancer may act on the NI exon in N18TG2 cells and may be better able to compete with CUGBP2 when its binding sites are compromised. A good candidate enhancer is FOX because a perfect match to its enhancer element, (U)GCAUG, is located near the NI 5′ splice site in the downstream intron.
To further investigate the roles of the m1, m2, and m3 motifs for exon silencing by CUGBP2, we introduced a 39 nucleotide region containing the three motifs upstream of a constitutive exon in a different context (Figure 3A). Constitutive exon 3 of the DIP13β transcript was tested, since its splicing pattern is insensitive to CUGBP2 regulation, and unlike the NI exon is not under alternative splicing control. The introduction of the 39 nucleotide region conferred strong silencing by CUGBP2 (Figure 3B, lanes NIwt; ΔEI, −54%), in contrast to the parent plasmid, which was unregulated by CUGBP2 (lanes m93wt; ΔEI, ∼0%). Single and combined mutations in the m1, m2, and m3 motifs were also tested in this context (lanes m1, m2, m3, m1,2, m2,3, m1,3, and m1,2,3). Exon silencing by CUGBP2 was nearly eliminated when site-directed mutations were introduced into both the m1 and m2 positions (lanes m1,2; ΔEI, −6%). Mutations in both the m2 and m3 positions also led to a significant reduction in silencing (lanes m2,3). Thus, the general requirement for a pair of proximal CUGBP2 motifs, and the additive effects of the single mutations were verified in this context. This heterologous reporter was also tested in N18TG2 cells (Figure 3C). Here, the silencing effects of CUGBP2 were more consistent between cell lines across all mutations tested. Furthermore, compared to the NI exon, the DIP13β exon has stronger 5′ and 3′ splice sites, therefore mutations had less of an effect on the basal level of exon inclusion. Taken together, this reporter is a good system to study isolated effects of CUGBP2 on the m1, m2, and m3 motifs without indirect effects caused by other splicing factors or weak cis elements.
We also tested the silencing role of a closely related family member, CUGBP1, on wild type and mutant substrates since it is expressed in both cell lines tested (Figure S2). CUGBP1 silences the NI exon (Figure S3A, lanes DUPNI wt) with a dependence on the same motifs (lanes DUPNI m1,2,3). However, CUGBP1 silencing is much weaker in the DIPNI context (lanes DIPNI wt, ΔEI −16 compared to ΔEI −54 for CUGBP2) indicating that additional sites outside of the transferred region are necessary for strong silencing by CUGBP1. In support of this, exon 3 of the DIPNIwt reporter is included >99% of the time in N18TG2 cells (Figure 3C, lanes NIwt) despite high levels of endogenous CUGBP1 (Figure S2). Therefore, CUGBP1 shows a weaker silencing role compared to CUGBP2. As additional controls, we carried out similar experiments with overexpression of PTB or Nova, since both of these factors are known to silence the NI exon through distinct motifs (UCUU and YCAY, respectively; Y, pyrimidine). As expected, PTB and Nova were active in silencing the NI exon in the context of the DUPNIwt splicing reporter, and these effects were maintained in the presence of the m1,2,3 triple mutation (Figure S3B). Thus, the m1, m2, and m3 motifs are specific for silencing by CUGBP2.
We hypothesized that CUGBP2 may function to silence the NI exon by blocking branchpoint formation in the upstream intron. Based on complementarity to U2 snRNA, two candidate branch sites, A1 and A2, are located between the m1 and m2 motifs, and a third, weaker candidate, A3, resides just downstream between the m2 and m3 motifs (Figure 4A). As a test of this hypothesis, we measured branchpoint formation for the DUPNI substrate under in vitro splicing conditions in the presence and absence of recombinant CUGBP2. Note that endogenous CUGBP2 levels are not detectable in Hela nuclear extracts by Western blotting with the 1H2 antibody, which is highly specific for CUGBP2 [21]. According to our model, the addition of recombinant CUGBP2 to the extract should bind and preferentially occupy motifs m1 and m2 on the wild type substrate with the resulting inhibition of one or more of the branch sites in this neighborhood. The protein can also bind in an alternate register of lower affinity by contacting a GU at the m2 site and UGUG at the m3 site. Alternately, one protein may contact all three sites simultaneously.
Branchpoints were detected by primer extension as for the experiments in Figure 1C. The results for the wild type substrate verified the use of the predicted branchpoints with a preference for A1 and A2, compared to A3 (Figure 4B, lanes 2,3). The A1 and A2 branchpoints satisfied the criteria for ATP dependence (lane 1). Primer extension of reactions following debranching showed a loss of stops at A1, A2, and A3 providing confirmation that all three of these adenosines are used as branchpoints (data not shown). Notably, branchpoint formation was inhibited when the in vitro splicing reactions were supplemented with recombinant CUGBP2 (Figure 4B, lane 4). An analysis of the corresponding in vitro splicing reactions confirmed that CUGBP2 inhibited the formation of splicing intermediates of these reactions (Figure S4, lanes wt).
We next examined the effect of the single mutation in motif m1 as a test of whether branchpoint inhibition occurs between core and flanking motifs. That is, a single mutation in m1 should permit the binding of CUGBP2 to the remaining intact sites (m2 and m3) leading to preferential inhibition of branchpoint A3. Indeed, under conditions in which the m1 site was mutated, branchpoint A3 was preferentially inhibited as expected for a model involving flanking interaction motifs (Figure 4B, lanes 7,8). Consistent with this observation, the corresponding in vitro splicing gel showed that CUGBP2 inhibited the formation of one lariat intermediate but not the other (Figure S4, lanes m1). For the single mutant, m3, which should display the reciprocal pattern of inhibition by CUGBP2, branchpoints A1 and A2 were preferentially inhibited relative to A3 (Figure 4B, lanes 11,12). Finally, the triple mutant, m1,2,3, was tested. Here, the elimination of all three binding motifs neutralized the inhibitory effects of CUGBP2 on branchpoint formation (lanes 17,18). Again, these results were consistent with the splicing intermediates of these reactions (Figure S4, lanes m3 and m1,2,3). Adenovirus major late (Ad1) pre-mRNA was tested as a control, because Ad1 pre-mRNA lacks CUGBP2 motifs in the upstream intron. Both the in vitro splicing and branchpoint formation of Ad1 were unaffected by the addition of recombinant CUGBP2 (data not shown).
To determine which step before branchpoint formation is specifically affected by the addition of CUGBP2 to the splicing reaction, we analyzed complex assembly on the E5-10 RNA substrate in the presence or absence of recombinant CUGBP2. We demonstrate that CUGBP2 blocks U2 snRNP association because CUGBP2 inhibited complex A formation on the wild type substrate but not on the m1,2,3 mutant substrate (Figure S5A). The identity of the complex was verified as the U2 snRNP-containing complex A, since its assembly was inhibited by U2 snRNA cleavage (Figure S5A, S5B). In contrast, parallel samples assembled in the absence of ATP showed no effect of CUGBP2 on complex E assembly (Figure S5C).
The results shown above are consistent with a model in which site-specific binding of CUGBP2 surrounding the branch site region mediates exon skipping. Because exon definition could potentially affect branchpoint formation in the upstream intron by interactions involving U1 snRNP and U2AF across the exon, we asked whether strengthening the 5′ splice site of the NI exon would antagonize the silencing effect of CUGBP2. For this purpose, we increased the complementarity of the 5′ splice site to U1 snRNP and tested the ability of CUGBP2 to induce silencing in vivo. This mutation had no detectable effect on silencing by CUGBP2 (Figure S6). We also show that U2AF and CUGBP2 can contact the same RNA at the same time indicating that branchpoint inhibition occurs after U2AF but before U2 snRNP binding (Figure S7). This, together with the lack of effect of CUGBP2 on complex E, which contains U1 snRNP, U2AF, and SF1, is consistent with a mechanism involving inhibition at a step subsequent to exon definition. Thus, the inhibitory role of CUGBP2 is likely to involve direct antagonism of U2 snRNP binding at the NI branch site region.
A recent publication by Yeo, et al. (2007) used computational approaches to identify intronic splicing regulatory elements (ISREs) in the introns upstream and downstream from skipped exons [22]. One of the ISREs identified was a UGUGUU motif with the propensity to be found within 400 nucleotides of conserved skipped exons. The Yeo study identified 168 skipped exons with a UGUGUU motif in their upstream intron. We obtained this list for further analysis.
In order to extend our analysis to identify additional exons that are potentially silenced by CUGBP2, we searched the list of 168 exons for the following sequence features: (1) the presence of conserved pairs of UGUGU and GU motifs within 100 nucleotides of the 3′ splice site of the skipped exon with a spacing of 10–30 nucleotides between the motifs, and (2) the presence of potential branch site(s) between the motifs. Because the m1 and m2 motifs were sufficient to inhibit branchpoints A1 and A2 on the DUPNIwt substrate, we rationalized that two motifs flanking the branch site would be sufficient for the prediction of CUGBP2 regulation. Potential branch sites were required to match the human consensus sequence, YUNAY (Y, pyrimidine; N, any nucleotide) with one mismatch allowed [23]. From this analysis, we determined that 48 of the 168 exons (29%) fit these criteria. We chose 27 exons to test for regulation by CUGBP2.
To analyze the response of these endogenous exons to CUGBP2 overexpression, we optimized a calcium phosphate transfection method to obtain >90% transfection efficiency in HEK293T cells. HEK293T cells were chosen for these experiments, since there is no detectable expression of CUGBP2 (Figure 5A, Western blot). RNA was harvested from the cells and the test exon region was amplified by RT-PCR with primers specific for the flanking exons. Ten predicted exons showed an increase in exon skipping when CUGBP2 was overexpressed (Figure 5A, panels MAP4_E15, SORBS1_E5, PPF1BP1_E19, SMARCE1_E4, FOX2_E11, and CUGBP2_E6; not shown: NFAT_E2, CTBP1_E2, PTER_E3, and MLLT10_E13). Of the 17 exons that were not affected by CUGBP2, one was constitutively included and resistant to CUGBP2, 8 were not expressed in HEK293T cells, and 8 were always skipped, therefore, CUGBP2 could not induce additional skipping. Therefore, 10 out of 11 testable exons were regulated by CUGBP2 indicating that we have identified a specific code that can be used to accurately predict CUGBP2 regulation. For all of the confirmed target exons, the core and flanking motifs were separated by 14–29 nucleotides, and multiple branch site candidates were located between the CUGBP2 motifs (Figure 5B). Note that exon 6 of the SCAMP3 transcript, which contains two mismatches to the GU-rich motif pattern, was insensitive to silencing by CUGBP2. MAPT_E2 is shown as a positive control, because GRIN1 is not expressed in HEK293T cells [24].
An interesting observation was the appearance of an exon skipped product of the CUGBP2 transcript itself, which was specific for conditions in which CUGBP2 was overexpressed. However, the primers in this case also amplified mRNA expressed from transfected CUGBP2, thereby complicating interpretation. For this reason we designed a downstream primer specific for the junction between exons 5 and 7, since such a junction primer should amplify only the skipped product from the endogenous mRNA. The junction primer was used together with a forward primer specific for the first exon. The results with the junction primer clearly showed the accumulation of the exon 6 skipped version of the endogenous CUGBP2 transcript upon overexpression of CUGBP2 (Figure 5A, black box). Note that this primer did not amplify the CUGBP2 protein expression plasmid or mRNA from transfected CUGBP2 (data not shown). It is also important to note that although CUGBP2 protein is not detected by Western blotting, there are low levels of CUGBP2 RNA in HEK293T cells. This may indicate that trace amounts of the protein are present in these cells or that the mRNA is translationally repressed. Furthermore, because there is an enrichment of CUGBP2 protein in the rat cerebral cortex and a deficiency in the cerebellum [17], we predicted and confirmed that there would be more skipping of exon 6 in the cortex (data not shown). To establish the identity of the exon 6 skipped product, we cut the band out of the gel, cloned and sequenced it. The cloned product exactly matched the exon 5–7 junction sequence demonstrating its identity as the skipped product (data not shown).
We also tested an exon in the CUGBP1 transcript that is homologous to CUGBP2 exon 6 (CUGBP1_E6). Here, there is one mismatch to the core motif and although CUGBP2 can silence this exon, the effect is less than other target exons with perfect matches to the consensus motifs. This suggests that CUGBP2 regulation can be titrated depending on the sequence content and binding affinity to target motifs.
To determine whether CUGBP2 silences its own exon by a mechanism similar to that shown for the NI exon, exon 6 and the adjacent introns of CUGBP2 pre-mRNA were cloned into the DUP splicing reporter (Figure 6A). In this context, overexpression of CUGBP2 had a robust silencing effect changing the exon 6 splicing pattern from 100% to 18% inclusion in transfected HEK293T cells (Figure 6B, lanes Wild type). To address the functional significance of the CUGBP2 binding sites at the boundaries of the predicted branch sites, we tested site-directed mutations in the core (CORE) and downstream (DSM) motifs (Figure 6A). One perfect match to the branch site consensus (A2) and two additional candidates with a single mismatch (A1,A3) are the only plausible branch sites located within 100 nucleotides of the 3′ splice site of exon 6. Mutations in the CORE and DSM motifs resulted in a reduction of splicing silencing by CUGBP2 (Figure 6B, lanes CORE mt, DSM mt), and these effects were additive in the double mutant (lanes CORE/DSM mt). These are similar to the results shown for the NI exon, providing additional support for the perimeter-binding model. We note that although mutations did not completely eliminate CUGBP2 regulation of exon 6, footprinting experiments documented additional contact points extending from the CORE and DSM motifs suggesting that alternate binding registers might allow for some residual silencing (see below). Also note that we tested the possible role of an intronic UGUGU motif located 70 nucleotides downstream from exon 6. Mutation of this motif to UAUAU had a negligible effect on splicing silencing by CUGBP2, ruling out effects across the exon and further supporting our model (data not shown). Furthermore, we show that CUGBP1 is also a weak silencer of this exon, but does not act through the CORE and DSM motifs like CUGBP2 (Figure S3A, lanes CUGBP2_E6).
Finally, we used RNA footprinting to identify CUGBP2 contact sites in the neighborhood of the predicted branch sites upstream of exon 6 (Figure 6C and 6D). GU-rich motifs flanking A1, A2, and A3 were protected by the addition of purified CUGBP2 similar to that observed above for the NI 3′ splice site (Figure 6C, last three lanes at right). That is, two protected regions at the borders of the predicted branch sites overlap with the UGUGU core and UG flanking motifs in agreement with the perimeter-binding model. A difference in the pattern of protection, however, was the finding that two sets of motifs on either side of the branch site region extend outward, indicating variations in the mode of binding compared to the NI exon.
In this study we focused on the silencing face of the dual functional splicing factor, CUGBP2, to understand how it recognizes and silences the NI cassette exon of the NMDA R1 receptor. The first hint of how this exon target is recognized was revealed by chemical modification RNA footprinting of a high affinity binding region, which showed two contact sites—a core UGUGU and flanking GU—closely positioned in the neighborhood of the predicted branch sites. These contact sites and a third weaker footprinting site were shown to modulate alternative splicing of the NI exon in vivo. Furthermore, the positions of the branch sites were mapped between the core and flanking motifs. These branch sites were collectively inhibited by CUGBP2 with a dependence on the presence of flanking GU-rich binding motifs. Thus, guilt-by-association places CUGBP2 at the boundaries of the branch sites it regulates in support of the three-motif occupancy model illustrated in Figure 4A. The regulation of an ensemble of branchpoints by a perimeter-type binding model, and the discovery that an exon in the CUGBP2 transcript is itself silenced by a similar arrangement of binding motifs, are novel findings of this study.
We show additional support for this model by identifying novel skipped exons that are functionally silenced by CUGBP2 based on database searches for similarities to the configuration of NI regulatory motifs. These confirmed targets contained the characteristic pattern of candidate branch sites flanked by GU and UGUGU motifs, which were themselves separated by ∼20 nucleotides in the adjacent 3′ splice site region (Figure 5). Notably, exon 6 of the CUGBP2 transcript was the most interesting member of this group due to the implications for autoregulation. Thus, the specific arrangement of CUGBP2 binding motifs around the branch sites of the NI exon is a silencing code that can be generalized to have a functional impact on other skipped exons.
Our results support and extend those of a previous study, which reported the identification of a UGUGUU motif as an intronic splicing regulatory element (U17) enriched within 400 nucleotides upstream of conserved skipped exons [22]. This previous study reported the association of the U17 element with exon inclusion in brain tissue as indicated by microarray analysis. In contrast, our study shows that the UGUGU core of the U17 element is generally associated with exon silencing when the motif is paired with a flanking GU surrounding the functional branch sites. This is not necessarily a discrepancy, but more likely a reflection of a mechanism operating on a subset of exons containing a U17-related element. Here we demonstrate the value of the branch site as a functional reference point that can be used together with the precise binding interaction motifs of a splicing factor to computationally predict new splicing regulatory targets. Our results are consistent with the types of binding motifs identified for ETR-3 using a SELEX approach, although the relationship of the binding motifs to the branch site region and the autoregulatory role of CUGBP2 were not determined [13]. Furthermore, the types of motifs identified for Bruno-like proteins in the α-actinin transcript are in agreement with our results [25].
Branchpoint formation reflects a critical step in the catalysis of the splicing reaction, but its role in the regulation of alternative splicing across the transcriptome represents largely uncharted territory. Only a small number of branchpoints have been experimentally mapped, however, and there are often multiple candidate branch sites in the 3′ splice site region that match the consensus sequence [23]. Examples of alternative splicing regulation through the use of a suboptimal branchpoint include the calcitonin/calcitonin gene-related peptide exon 4 and fibroblast growth factor receptor 2 exon IIIc [26]–[29]. Branch site selection has also been implicated in the regulation of mutually exclusive exons of beta tropomyosin and in the processing of human growth hormone pre-mRNA [30],[31].
What advantages would a perimeter-binding model provide for the control of access to the branch site region? The pre-mRNA branch site and flanking sequences are sequentially contacted by several factors during spliceosome assembly [32],[33]. The branch site interacts with the RS domain of U2AF65 bound to the polypyrimidine tract of the 3′ splice site, followed by interactions with the RS domain of an SR splicing factor bound to an enhancer element in the adjacent exon [34]–[36]. Splicing Factor 1 (SF1) makes direct contacts with the branch site during complex E assembly [37],[38]. In complex A, SAP155 binds to sites flanking the branch site and replaces SF1 to recruit U2 snRNP [39],[40]. Here multiple contact sites may furnish CUGBP2 with the added stability to inhibit the association of U2 snRNP with the branch site region and/or may block conformational transitions of the spliceosome [41]. The perimeter-type binding model described here is significant in allowing for the coordinate regulation of multiple branchpoints to control alternative splicing of a cassette exon. Moreover, the distinctive pattern of RNA motif recognition by CUGBP2 may facilitate its enhancing roles in other contexts.
CUGBP2 is a modular protein containing three RNA recognition motifs (RRMs) in which a divergent domain of unknown function separates RRMs 2–3. The domain structure of the protein may be geared to facilitate binding of a monomer to a pair of core and flanking motifs forming a bridge between them as our model indicates. Alternatively, a single monomer might bind to all three GU-rich motifs. Both models would limit access to the branch sites by factors sliding along the RNA from upstream and downstream directions. Our footprinting results with CUGBP2 are in agreement with previous structural studies showing that a single RRM can contact ∼2–7 nucleotides of its bound RNA ligand, but additional studies will be required to understand the topology of binding associated with its silencing function [42]–[44]. Given the inherent flexibility of RNA binding proteins, it would not be surprising that breathing motions of CUGBP2 could adjust the relative conformations of the RRM domains to optimize recognition specificity in different sequence contexts.
Autoregulation has been shown for a growing number of splicing factors, including PTB, FOX-2, Nova-1, SRp20, SC-35, TIA1, and TIAR [45]–[48]. Here we dissect the mechanism of CUGBP2 autoregulation. CUGBP2 acts functionally through motifs surrounding the branch site region to silence exon 6 near the 5′ end of its own transcript. Conceptual translation reveals that skipping of this exon causes a shift in the reading frame, which introduces a premature termination codon in the exon 7 region of the transcript. In this way, the resulting transcript could be targeted for nonsense-mediated mRNA decay. Conversely, translation of the exon 6 skipped transcript could generate a truncated protein ending within RRM2. The advantage of a tight motif arrangement around the branch site region would be to dynamically adjust exon 6 inclusion based on fluctuations in the levels of CUGBP2 protein.
The observation that CUGBP2 can cross regulate exons in the CUGBP1 and FOX2 transcripts, and that CUGBP1 can silence CUGBP2 exon 6 to a lesser extent, implicates CUGBP2 in a network of splicing factor regulation (Figure 5 and Figure S3). We speculate that this may be important in specifying neural cell identity and for fine-tuning of neural exon splicing. In the future, it would be of interest to study the differences in binding specificities and target exon selection by CUGBP1 and CUGBP2.
The binding of CUGBP2 to the perimeters of the branch sites allows for sensitive gradations specifying the levels of NI exon inclusion. Because the peptide region encoded by the NI exon modulates sensitivity of the NMDA receptor to zinc ions, protons, and polyamines, such a mechanism would be advantageous for fine-tuning this modular property of receptor function in different regions of the brain or during development [49]. The biochemical functions of NMDA receptors are of fundamental importance in synaptic plasticity where deficits in this subunit are associated with altered brain function in the context of Alzheimer's disease. The CI cassette exon, which is regulated by the enhancing face of CUGBP2, encodes a functionally important region of the receptor involved in membrane trafficking and signaling to the nucleus. Our study provides the starting point to investigate the broader roles of CUGBP2 in regulation of the CI cassette and additional exons throughout the transcriptome. Insights from this study can also be applied to systematically examine the role of intronic mutations in the neighborhood of the branch site underlying mechanisms of human disease.
The mouse CUGBP2 (pcDNA4/NAPOR) and rat PTB (pcDNA4/PTB) protein expression vectors were described previously [17]. The Nova-1 protein expression vector was a gift of Robert Darnell [50] and the CUGBP1 expression vector was a gift of Thomas Cooper [51]. To generate the DUPNIwt and DUP-CUGBP2 splicing reporters, the cassette exon and flanking introns were amplified from rat genomic DNA and inserted between the ApaI and BglII restriction sites of the DUP4-1 splicing reporter [52]. The DIPm93wt splicing reporter [53] was used to generate DIPNIwt and mutant derivatives. DIPNIwt was generated by insertion of a 39 base pair fragment containing the 3′ splice site region of the NI exon at position −13 base pairs upstream from the test exon. pBSDUPNI wild type and mutant vectors for in vitro transcription were generated by PCR amplification from the DUPNIwt or mutant splicing reporters and insertion between the HindIII and EcoRI restriction sites of the pBS- phagemid vector (Stratagene). E5-8, E5-10, and E5-15 plasmids were generated in a similar manner. Plasmids were confirmed by restriction digestion and DNA sequencing.
RNA substrates were 32P-UTP-labeled by in vitro transcription and used at a final concentration of 1000 cpm/µl (20–100 nM). RNA was heated at 85°C for 5 min and then cooled to 37°C for 5 min to remove long-range secondary structures. Serial dilutions of CUGBP2 protein were prepared on ice in binding buffer (50 mM Tris pH 8.0, 150 mM NaCl, 0.1 mg/ml tRNA, 2 mM DTT, 20 units RNasin (Promega)) in a final volume of 199 µl. Protein samples were warmed to 37°C for 5 min before adding 1 µl of diluted RNA (final RNA concentration 100–500 pM). RNA-protein complexes were assembled in triplicate at 37°C for 30 min before filtration through 25 mm BA85 filters backed by DE81 filters in a Millipore 1225 vacuum manifold. Filters were separated and dried at room temperature overnight. The cpm retained on the BA85 filter corresponded to RNA bound to protein and cpms retained on the DE81 filter corresponded to free RNA. Bound/total RNA was plotted as a function of increasing protein concentration using KaleidaGraph Synergy Software and data were fit to a hyperbola to estimate the dissociation constant (Kd) according to the equation bound/total RNA = [CUGBP2]/([CUGBP2]+Kd).
RNA-protein complexes were assembled for 30 min at 37°C with 0.18 µM RNA substrate and purified recombinant CUGBP2 protein in a final volume of 50 µl. Each sample was then combined with an equal volume of 42 mg/ml 1-cyclohexyl-3-(2-morpholinoethyl)-carbodiimidemetho-p-toluene-sulfonate (CMCT) and chemical modification was carried out at 37°C for 7 minutes. Reactions were terminated by ethanol precipitation. Recovered RNA was treated with proteinase K followed by phenol chloroform extraction and ethanol precipitation. Primer extension was carried out with a 5′32P-labeled primer using Superscript II reverse transcriptase (Invitrogen). Sequencing ladders were generated using Thermo Sequenase Cycle Sequencing Kit (USB) according to manufacturers protocol. cDNA was separated on a 10% polyacrylamide gel and gel images were recorded on a BAS-2500 Phosphorimager (Fujifilm).
C2C12 and N18TG2 cells were grown in DMEM, 10% (v/v) fetal bovine serum (FBS). Twenty-four hours prior to transfection 1.5×105 C2C12 cells or 2×105 N18TG2 cells were seeded on 35-mm plates to achieve 60–80% confluency. For transient transfection, 1 µg pcDNA4 His/Max vector backbone or 1 µg pcDNA4/CUGBP2, pC1-Nova-1, pcDNA4/PTB, or pcDNA3.1/CUGBP1 expression vector and 0.25 µg splicing reporter were mixed with 250 µl Opti-MEM followed by addition of an equal volume of Opti-MEM mixed with 2.5 µl Lipofectamine 2000 (Invitrogen) and incubated at room temperature for 20 min. Media on cells was replaced with 1.5 ml DMEM, 10% (v/v) FBS prior to transfection. Total RNA was isolated 36 hours after transfection using TRIZOL reagent (Invitrogen). Two µg total RNA was reverse transcribed as described previously [17]. PCR was carried out in 10 µl reactions containing 1 µl of the reverse transcription reaction, 0.1 µM of each primer, 2 mM MgCl2, 0.2 mM dNTPs, and 2.5 units Taq DNA polymerase (Promega). Forward primers were 5′32P-labeled. Cycling parameters were adjusted to give amplification in the linear range. Conditions were as follows: denaturation 94°C, 1 min, annealing at 60°C, 1 min, and elongation at 72°C, 1 min for 22 cycles followed by a final elongation step at 72°C for 10 min. PCR samples were resolved on 6% polyacrylamide/5 M urea gels. Data were quantified using a BAS-2500 Phosphorimager system and Image Gauge software.
pBSDUPNIwt or mutant derivatives were digested with EcoRI for in vitro transcription in the presence of α32P-UTP. In vitro splicing assays were carried out for 45 min or 1 hour as previously described [54] except MgCl2 was at a final concentration of 2.2 mM. Branchpoints were detected by primer extension from parallel splicing reactions constituted with unlabeled pre-mRNA in a reaction containing 200 U MMLV reverse transcriptase (Invitrogen), 10 mM DTT, 1 mM dNTPs, 1× first strand buffer, and 50 nM 5′32P-labeled primer in a total reaction volume of 20 µl. Reactions were incubated at 37°C for 30 minutes and were terminated by ethanol precipitation. Samples were resuspended in 6 µl formamide loading buffer and 1/3 of the sample was separated on an 8% polyacrylamide/7 M urea gel next to a sequencing ladder. Gel images were recorded on a BAS-2500 Phosphoimager system.
Spliceosome complexes were assembled on the E5-10 wild type or mutant RNA in a 10 µl reaction containing 20 mM Hepes, pH 7.4, 44% Hela nuclear extract, 2.2 mM MgCl2, 60 mM KCl, 1.5 mM ATP, 5 mM creatine phosphate, and 0, 1.6, or 3.2 µM CUGBP2 at 30°C for 15 minutes. Spliceosome assembly was stopped by the addition of heparin at a final concentration of 2 mg/ml and incubation for an additional 3 minutes. Half of the reaction was separated on a 3.75% native polyacrylamide gel cast in 50 mM tris-glycine buffer and run at 4 watts at 4°C for 4 hours. Gels were dried under vacuum and visualized by phosphoimager. For assembly of the ATP-independent E complex, the nuclear extracts were preincubated at 30°C for 10 minutes to deplete ATP and complexes were assembled in the absence of ATP or creatine phosphate for 8 minutes. Oligonucleotide-directed cleavage of U1 and U2 snRNAs was carried out as described previously [55].
HEK293T cells were grown in DMEM, 10% (v/v) FBS. Twenty-four hours prior to transfection, 2×105 HEK293T cells were seeded in 35-mm dishes precoated with poly-L-lysine (Sigma). Cells were approximately 50% confluent at the time of transfection. Before transfection, all reagents were brought to room temperature. For one well, 1.6 µl of 1 µg/µl CUGBP2 protein expression vector or vector backbone was mixed with 16.1 µl 2.5 M CaCl2 by vortexing briefly. Next, 65.8 µl water was added and the CaCl2-DNA mixture was pipetted over 83.5 µl 2× BBS, pH 7.15 (50 mM N,N-bis(2-hydroxyethyl)-2-aminoethane sulfonic acid, 280 mM NaCl, 1.5 mM Na2HPO4) and vortexed for 3 seconds. Mixtures were incubated at room temperature for 10 minutes before 164 µl was added drop wise to each dish. After transfection, cells were incubated at 3% CO2 for 36 hours prior to RNA isolation using TRIZOL reagent (Invitrogen). RT-PCR was carried out as described above with unlabeled primers (Table S1). PCR samples were resolved on 2% agarose gels and quantified using Image Gauge software.
|
10.1371/journal.ppat.1006162 | HTLV-1 Tax Induces Formation of the Active Macromolecular IKK Complex by Generating Lys63- and Met1-Linked Hybrid Polyubiquitin Chains | The Tax protein of human T-cell leukemia virus type 1 (HTLV-1) is crucial for the development of adult T-cell leukemia (ATL), a highly malignant CD4+ T cell neoplasm. Among the multiple aberrant Tax-induced effects on cellular processes, persistent activation of transcription factor NF-κB, which is activated only transiently upon physiological stimulation, is essential for leukemogenesis. We and others have shown that Tax induces activation of the IκB kinase (IKK) complex, which is a critical step in NF-κB activation, by generating Lys63-linked polyubiquitin chains. However, the molecular mechanism underlying Tax-induced IKK activation is controversial and not fully understood. Here, we demonstrate that Tax recruits linear (Met1-linked) ubiquitin chain assembly complex (LUBAC) to the IKK complex and that Tax fails to induce IKK activation in cells that lack LUBAC activity. Mass spectrometric analyses revealed that both Lys63-linked and Met1-linked polyubiquitin chains are associated with the IKK complex. Furthermore, treatment of the IKK-associated polyubiquitin chains with Met1-linked-chain-specific deubiquitinase (OTULIN) resulted in the reduction of high molecular weight polyubiquitin chains and the generation of short Lys63-linked ubiquitin chains, indicating that Tax can induce the generation of Lys63- and Met1-linked hybrid polyubiquitin chains. We also demonstrate that Tax induces formation of the active macromolecular IKK complex and that the blocking of Tax-induced polyubiquitin chain synthesis inhibited formation of the macromolecular complex. Taken together, these results lead us to propose a novel model in which the hybrid-chain-dependent oligomerization of the IKK complex triggered by Tax leads to trans-autophosphorylation-mediated IKK activation.
| NF-κB is a key transcription factor that regulates many physiologically important cellular processes. However, persistent activation of NF-κB leads to chronic inflammation, autoimmunity and malignancy. Infection with the human retrovirus HTLV-1 causes adult T-cell leukemia, and HTLV-1 Tax-mediated persistent NF-κB activation is crucial for leukemogenesis. Therefore, a better understanding of the precise mechanism underlying aberrant NF-κB activation is essential to develop new therapeutic approaches. Ubiquitination is one of the major post-translational modifications that regulate various intracellular signaling pathways. We and others have shown that Tax activates NF-κB through activation of the IκB kinase (IKK) complex by generating Lys63-linked polyubiquitin chains. However, the molecular mechanism underlying Tax-induced IKK activation remains less well understood. Here, we demonstrate precisely how HTLV-1 Tax utilizes the ubiquitin system to activate the IKK complex. The IKK complex-associated Lys63/Met1-linked hybrid polyubiquitin chains are generated through the Tax-mediated recruitment of linear ubiquitin chain assembly complex (LUBAC) to the IKK complex. Furthermore, the hybrid chains are required for the Tax-induced formation of the active macromolecular IKK complex. Accordingly, we propose a novel model in which Tax triggers Lys63/Met1-linked hybrid-chain-dependent oligomerization of the IKK complex, leading to trans-autophosphorylation-mediated IKK activation.
| Human T-cell leukemia virus type 1 (HTLV-1) is etiologically associated with adult T-cell leukemia (ATL), an aggressive and lethal malignancy of CD4+ T cells, and with HTLV-1-associated myelopathy/tropical spastic paraparesis (HAM/TSP) [1, 2]. The HTLV-1 provirus genome encodes a transactivator protein (Tax), which is crucial for viral gene expression and the onset and development of ATL together with another viral protein, HBZ [3–5]. Tax aberrantly activates host cell transcription factors, including nuclear factor-κB (NF-κB), cyclic AMP response element-binding protein (CREB) and serum responsive factor (SRF), thereby perturbing transcriptional networks in host cells [6]. Among these factors, accumulating evidence indicates that persistent activation of NF-κB by Tax is crucial for T cell transformation and ATL development [7–9].
NF-κB plays critical roles in immune responses, inflammation, bone metabolism, cell proliferation and survival [10]. NF-κB is composed of five Rel/NF-κB family members—p50/p105, p52/p100, RelA, RelB and c-Rel—which form various combinations of homo- and heterodimers. NF-κB is sequestered in the cytoplasm with inhibitory proteins of the NF-κB family (IκBs) or NF-κB precursors. Two distinct pathways lead to activation of NF-κB. The canonical pathway is activated by cytokines, such as tumor necrosis factor (TNF)-α and interleukin (IL)-1, whose stimulation leads to activation of the IκB kinase (IKK) complex, which is composed of the catalytic subunits IKKα and IKKβ and the regulatory subunit NEMO [11]. The IKK complex then induces phosphorylation and subsequent degradation of IκBα, which allows the p50/RelA heterodimer to translocate into the nucleus and activate target genes. In the noncanonical pathway, stimulation of CD40, receptor activator of NF-κB (RANK) or lymphotoxin-β receptor results in the activation of IKKα in a NIK-dependent but IKKβ- and NEMO-independent manner. IKKα then phosphorylates the C-terminal ankyrin repeats of p100, which forms heterodimers with RelB in the cytoplasm, leading to the proteasome-dependent selective degradation of the p100 C-terminal end to generate p52 [12]. The resulting p52/RelB heterodimer translocates into the nucleus and activates target genes. Tax is able to activate both canonical and noncanonical pathways, which are thought to be coordinately involved in leukemogenesis [13].
The importance of ubiquitination in the regulation of NF-κB activity is well established [14]. Ubiquitination is catalyzed by three enzymes in a stepwise fashion [15]. Ubiquitin-activating enzyme E1 forms a thioester linkage with ubiquitin, and the activated ubiquitin is then transferred to the E2 ubiquitin-conjugating enzyme. E2 acts as an escort for ubiquitin to the subsequent enzyme, E3 ligase, which binds to both E2 and the substrate and catalyzes the formation of an isopeptide bond between carboxylic acid at the C-terminal end of ubiquitin and the epsilon amine of the lysine residue in the substrate. After the addition of a single ubiquitin to the substrate, more ubiquitin can be repeatedly added to the previously conjugated molecule, thereby yielding a polyubiquitin chain. Ubiquitin itself contains seven lysines (K6, K11, K27, K29, K33, K48 and K63), each of which can participate in the formation of the ubiquitin chain, allowing seven linkage types [16]. In addition, ubiquitin can also be attached to the N-terminus of the proximal ubiquitin to generate a linear ubiquitin chain or Met1-linked ubiquitin chain (M1 chain) [17]. In the TNFR signaling pathway, several types of polyubiquitin chains cooperatively regulate IKK activation. Upon TNF-α stimulation, TNF receptor-1 recruits the adaptor TRADD, TRAF2/5 and cIAPs. cIAPs have E3 ligase activity and conjugate Lys11- and Lys63-linked ubiquitin chains (K11 and K63 chains) to RIP1 [18–20]. These polyubiquitin chains conjugated to RIP1 act as a scaffold for the formation of an active signaling complex containing transforming growth factor-β-activated kinase (TAK)-1, TAK1-binding (TAB) 2/3 and linear ubiquitin chain assembly complex (LUBAC) [21]. LUBAC is composed of HOIL-1L, HOIP and Sharpin [22, 23]. HOIP is the catalytic subunit while HOIL-1L and Sharpin are also required for the enzymatic activity of this complex. LUBAC conjugates M1 chains to NEMO [24], which may induce oligomer formation or a conformational change of NEMO to activate the IKK complex. Although previous studies have shown that Tax binds to NEMO and induces constitutive activation of the IKK complex in a K63-chain-dependent manner [25–28], the involvement of other types of polyubiquitin chains in Tax-induced IKK activation is still controversial [29].
In this study, we show that Tax induces generation of hybrids of K63 and M1 chains by recruiting LUBAC to the IKK complex, leading to the formation of the active macromolecular IKK complex. Thus, we propose a previously unidentified mechanism by which K63 and M1 chains cooperate in Tax-induced IKK activation.
We previously established a cell-free assay to analyze Tax-induced IKK activation, in which the addition of recombinant Tax protein purified from E. coli into S-100 cytosolic extracts prepared from the Jurkat human T cell line, HEK293T cell line or mouse embryonic fibroblast (MEF) cells results in IKK activation [27]. To investigate which types of polyubiquitin linkages are required for Tax-induced IKK activation, we took advantage of a cell-free assay because the addition of dominant-negative (DN) ubiquitin mutants containing a single lysine-to-arginine substitution (K6R, K11R, K27R, K29R, K33R, K48R and K63R) or N-terminal HA-tagged ubiquitin results in linkage type-specific blockage of polyubiquitination. Immunoblots probed with anti-phospho-IKKα/β and phospho-IκBα antibodies revealed that the addition of K27R, K63R or HA-ubiquitin inhibited Tax-induced IKK activation (Fig 1A), suggesting that K27, K63 and M1 chains are required for IKK activation by Tax. Addition of K11R or K33R ubiquitin reproducibly enhanced Tax-induced IKK activation, probably because their addition could enhance the generation of K27, K63 or M1 chains. Note that phosphorylated IκBα is not degraded by proteasomes in a cell-free assay (S1 Fig), although the amount of IκBα was slightly reduced concomitantly with IκBα phosphorylation in some experiments in this paper. This could be due to the manufacturer-noted preferential binding of the anti-IκBα antibody used for immunoblotting to the non-phosphorylated form of IκBα. To identify the E2 ubiquitin-conjugating enzymes involved in Tax-induced IKK activation, a cell-free assay was performed using cytosolic extracts prepared from HEK293T cells expressing a series of E2 DN mutants, in which an active Cys residue was substituted with Ala. Expression of the Ubc13 DN mutant almost completely inhibited IKK activation, whereas other E2 DN mutants did not (Fig 1B). A cell-free assay using the extract from Ubc13−/− MEFs further confirmed that Ubc13 is required for Tax-induced IKK activation (Fig 1C), which is consistent with previous reports based on experiments using intact Ubc13−/− MEFs and a cell-free assay using the extract from Ubc13-knockdown cells [28, 30]. Low-level Tax-induced phosphorylation of IκBα was observed in Ubc13−/− MEFs, which could be due to residual Ubc13 attributable to incomplete gene disruption by the Cre/loxP system. The lack of candidate E2 enzymes other than Ubc13 suggests that several E2 enzymes may redundantly generate K27 and M1 chains or that E2 enzymes not tested here could be involved. Because it has been reported that RNF8, as an E3 ubiquitin ligase, is partially involved in the Tax-induced generation of K63 chains [28], we checked whether other E3 enzymes capable of generating K63 chains are involved [18, 31–33]. Cytosolic extracts were prepared from cIAP1/cIAP2-deficient (Birc2−/−/Birc3−/−), TRAF2/TRAF5-deficient (Traf2−/−/Traf5−/−), TRIM25-deficient (Trim25−/−) and Riplet-deficient (Rnf135−/−) MEFs [33–36] and were subjected to a cell-free assay. None of the extracts derived from the mutant cells showed reduced IKK activation (S2 Fig), suggesting that these E3 enzymes are not involved in Tax-induced IKK activation. In addition, TRAF6, another E3 enzyme, has been shown to be dispensable in Tax-induced IKK activation but instead can work together with Ubc13 to generate K63 chains for cytokine-induced IKK activation [27, 37]. We then hypothesized that Tax itself may be an E3 ligase as recently proposed [29], since Tax contains a putative zinc finger domain at its N-terminus (S3A Fig) [38] and the zinc finger domain may act as a catalytic domain of E3 ligase as previously shown in the zinc finger of A20 [39]. Some zinc finger mutants of Tax failed to activate the IKK complex and NF-κB (S3B and S3C Fig), indicating that the zinc finger of Tax is crucial for IKK activation. Although recombinant Tax purified from either E. coli or Sf9 cells can efficiently activate IKK (S3D Fig), neither of them induced polyubiquitination in the presence of E2 enzymes including UbcH5c, UbcH7 and Ubc13/Uev1A under conditions that allow TRAF6 to generate polyubiquitin chains together with Ubc13/Uev1A (S3E Fig). These results strongly suggest that Tax itself does not possess E3 ligase activity.
To further confirm the requirement for M1 chains, cytosolic extracts derived from MEFs that lack each component of LUBAC (the only known E3 ligase complex that catalyzes M1 chain generation) were tested. Tax failed to induce IKK activation when cytosolic extracts from HOIL-1L-deficient (Rbck1−/−) MEFs, Sharpin-deficient (cpdm) MEFs or MEFs in which the RING-IBR-RING region (the catalytic center) of HOIP was ablated (HOIPΔlinear) were used (Fig 2A–2C) [40]. To confirm the requirement for LUBAC for Tax-induced IKK activation in intact cells, Sharpin-deficient MEFs were infected with a Tax-expressing retrovirus, and subsequent phosphorylation of IKK and IκBα was detected by immunoblotting. Tax-induced IKK activation was significantly reduced in cells that lack LUBAC activity (Fig 2D). Taken together, these results clearly indicate that LUBAC is crucial for Tax-induced IKK activation.
To understand how LUBAC is involved in Tax-induced IKK activation, we first investigated whether LUBAC binds to Tax. When Tax was immunoprecipitated with an anti-Tax antibody after incubation with Jurkat cytosolic extracts, HOIP and Sharpin were co-immunoprecipitated with Tax (Fig 3A), indicating that Tax interacts with LUBAC. In addition, the Tax mutant M22, which is incapable of activating NF-κB due to a lack of binding ability to NEMO [30], also bound to HOIP and Sharpin (Fig 3A), indicating that the binding of Tax to LUBAC is not mediated by the IKK complex. This result led us to hypothesize that Tax acts as an adaptor in the formation of a multi-protein complex composed of LUBAC, Tax and the IKK complex. To test this hypothesis, the IKK complex was immunoprecipitated with an anti-Flag antibody from cytosolic extracts of Jurkat cells expressing Flag-NEMO. HOIP and Sharpin were recruited to the IKK complex in the presence of Tax, but M22 failed to recruit LUBAC to the IKK complex (Fig 3B), indicating that Tax functions as an adaptor to recruit LUBAC to the IKK complex. We then sought to determine whether Tax also acts as a bridge between LUBAC and the IKK complex in an intact Jurkat human T cell line. JPX-9, a Jurkat-derived cell line in which Tax expression is induced by Cd2+ treatment [41], was first cultured in the presence or absence of Cd2+, and cell lysates were subjected to immunoprecipitation using an anti-Tax antibody. HOIP and Sharpin were included in the immunoprecipitates only when Tax was induced (Fig 3C). These results indicate that Tax associates with LUBAC in intact T cells. Interestingly, the slower-migrating form of HOIP was observed only when the IKK complex was activated by Tax (Fig 3A). This band shift was due to the phosphorylation of HOIP because Phos-tag SDS-PAGE analysis identified slower-migrating bands (S4A Fig). Treatment of lysates with the IKKβ inhibitor TPCA-1 resulted in the disappearance of the slower-migrating bands in a dose-dependent manner (S4B Fig), suggesting that IKKβ phosphorylates HOIP during Tax-induced IKK activation. The significance of HOIP phosphorylation in IKK activation remains to be elucidated.
Given that HOIP binds to K63 chains but not M1 chains [42], we hypothesized that K63 chains are required for the binding of LUBAC to the IKK complex. To test this possibility, we first investigated whether the addition of DN ubiquitin mutants would inhibit the Tax-mediated binding of LUBAC to the IKK complex. The addition of K63R or HA-tagged ubiquitin inhibited Tax-induced IKK activation (Fig 1A), whereas the Tax-mediated binding of LUBAC to the IKK complex was not affected (Fig 3D). These results indicate that K63 and M1 chains are not required for the binding of Tax to LUBAC and the IKK complex. To determine which components of LUBAC bind to Tax and also whether the binding is direct, an in vitro binding assay was performed using purified recombinant proteins. Purified GST-HOIL-1L, GST-Sharpin or GST-HOIP was incubated with His6-Tax and subjected to GST pull-down assay. GST-HOIL-1L and GST-HOIP bound to His6-Tax, whereas GST-Sharpin did not (Fig 3E), indicating that HOIL-1L and HOIP directly bind to Tax. To elucidate the molecular basis of the binding of HOIL-1L or HOIP to Tax, a series of deletion mutants of HOIL-1L and those of HOIP were tested by co-immunoprecipitation assay. HOIL-1L ΔUBL and HOIP ΔRBR failed to bind to Tax, whereas the other mutants of HOIL-1L and HOIP proteins bound to Tax as efficiently as the full-length protein (Fig 3F and 3G). These results indicate that HOIL-1L and HOIP interact directly with Tax through their UBL and RBR domains, respectively.
To determine how the Tax-induced generation of polyubiquitin chains leads to IKK activation, Jurkat cytosolic extracts were incubated in the absence or presence of recombinant Tax, and the reaction mixtures were then subjected to immunoprecipitation with an anti-NEMO antibody. The resulting immunoprecipitates were immunoblotted with either an anti-ubiquitin (Ub) antibody that can recognize monoubiquitin and any type of polyubiquitin linkages or an anti-M1 chain-specific antibody. Both antibodies clearly detected smeared bands only when cytosolic extracts were incubated with Tax (Fig 4A, lane 2), indicating that M1 chains were associated with the IKK complex in a Tax-dependent manner. To further characterize the IKK complex-associated ubiquitin chains, the immunocomplexes precipitated with an anti-NEMO antibody were then treated with the following chain type-specific deubiquitinases (DUBs): Otubain-1 for K48 chains [43, 44], associated molecule with the SH3 domain of STAM (AMSH) for K63 chains [42, 45], OTULIN for M1 chains [46], and ubiquitin-specific protease 2 (USP2) for any type of polyubiquitin chain [47]. Otubain-1 treatment did not affect the smears detected by the anti-Ub or anti-M1 chain antibody (Fig 4A, lane 3 upper and lower), indicating that K48 chains were nearly nonexistent in the complex. In contrast, AMSH treatment almost completely abolished the smears detected by the anti-Ub antibody (Fig 4A, lane 4 upper), and OTULIN treatment completely abolished the smears detected by the anti-M1 chain antibody (Fig 4A, lane 5 lower). These results indicated that both K63 and M1 chains were associated with the IKK complex. These results were further supported by the quantification of different ubiquitin chain types associated with the IKK complex via the ubiquitin-AQUA method using mass spectrometry [48]. While residual amounts of K48 chains were detected irrespective of the presence of Tax, an approximately 2:1 ratio of K63 to M1 chain linkages was significantly associated with the IKK complex only when cytosolic extracts were incubated with Tax (Fig 4B).
Interestingly, when an immunoblot of the AMSH-treated IKK complex was probed with the anti-M1 chain antibody, extremely high-molecular-weight ubiquitin-containing complexes (EHUCs), which remained at the top of separating gels (arrows in Fig 4A and 4C), were degraded (Fig 4A, lane 4 lower), indicating that EHUCs are polyubiquitin chains that include both K63 and M1 linkages in a single chain. This notion was also supported by experiments showing that Tax failed to generate EHUCs when either the generation of K63 chains or that of M1 chains was blocked (Fig 4C, lanes 5 and 6). Furthermore, when an immunoblot of the OTULIN-treated IKK complex was probed with the anti-Ub antibody, the abundance of EHUCs was found to be significantly reduced while ladders between 17 and 75 kDa appeared (Fig 4A, lanes 5 and 7 upper, note that the same sample was applied in lanes 5 and 7). Importantly, among these ladders, four bands (dots in Fig 4A, lanes 5 and 7 upper) migrated almost identically to the bands corresponding to trimer (Ub3), tetramer (Ub4), pentamer (Ub5) and hexamer (Ub6) of recombinant K63 chains (arrowheads in Fig 4A, lane 8 upper). These results clearly indicate that K63/M1-linked hybrid chains are associated with the IKK complex activated by Tax.
To address whether these IKK complex-associated polyubiquitin chains are required for Tax-induced IKK activation, cell-free assays were performed in the presence of various DUBs. AMSH, OTULIN and USP2, but not Otubain-1, inhibited the phosphorylation of IκBα induced by Tax (Fig 4D), indicating that generation of both K63 and M1 chains is essential for Tax-induced IKK activation. Therefore, generation of IKK complex-associated K63/M1-linked hybrid chains is likely to be essential for Tax-induced IKK activation.
In the cytokine-induced NF-κB signaling pathway, IKK activation requires the formation of unanchored K63 chains or NEMO-conjugated M1 chains [24, 49]. To understand the roles of unanchored and substrate-conjugated (anchored) chains in Tax-induced IKK activation, cell-free assays were performed in the presence of Isopeptidase T (IsoT), a DUB specific for unanchored chains, or the OTU domain of the L protein of Crimean Congo hemorrhagic fever virus (viral OTU), a DUB specific for substrate-anchored chains. IsoT inhibited the Tax-induced phosphorylation of IKK and IκBα but not their polyubiquitination-independent phosphorylation by MEKK1 (Fig 5A). In addition, viral OTU, but not its catalytic inactive mutant (1A), inhibited the Tax-induced phosphorylation of IKK and IκBα (Fig 5B). These results indicate that both unanchored and substrate-anchored polyubiquitin chains are required for Tax-induced IKK activation. Several studies have shown that ubiquitination of Tax is required for IKK activation [50–52]. Among ten lysine residues present in Tax, ubiquitination of the C-terminal seven lysines (K4 to K10) are required for Tax-induced IKK activation in intact cells [50]. To examine whether Tax requires similar ubiquitination for IKK activation in our cell-free system, recombinant Tax mutants containing lysine-to-arginine mutations at the three N-terminal lysines (K1-3R) or at the seven C-terminal lysines (K4-10R) were generated. Both Tax-WT and the K1-3R mutant induced IKK activation equally well, whereas the K4-10R mutant did not (Fig 5C left). In addition, the K4-10R mutation significantly reduced Tax ubiquitination (Fig 5C right). These results strongly suggest that the polyubiquitin chains conjugated to Tax belong to the class of substrate-anchored polyubiquitin chains required for Tax-induced IKK activation.
Because LUBAC induces polyubiquitination at K285 and K309 of NEMO in cytokine-induced IKK activation [24], we next investigated whether these lysine residues are required for Tax-induced IKK activation. Human NEMO-WT or its mutant (K285R/K309R) was introduced into NEMO-deficient (Ikbkg−/−) MEFs, and cell-free assays were performed. Tax induced phosphorylation of IKK and IκBα equally well in NEMO-WT- and NEMO (K285R/K309R)-expressing cytosolic extracts (Fig 5D), indicating that polyubiquitination at K285 and K309 of NEMO is dispensable for Tax-induced IKK activation. It has been reported that the binding of NEMO to K63 and M1 chains is required for cytokine-induced IKK activation. NEMO binds to K63 chains through the C-terminal NZF domain and to M1 chains through the UBAN domain [53, 54]. To determine the requirement for the binding ability of NEMO to K63 or M1 chains, cytosolic extracts were prepared from NEMO-deficient MEFs expressing mouse NEMO-WT, its mutant (R309A/R312A/E313A) lacking the ability to bind to M1 chains or another NEMO mutant (H406A/C410A) lacking the ability to bind to K63 chains. Tax failed to induce IKK activation in cytosolic extracts expressing the NEMO (R309A/R312A/E313A) or NEMO (H406A/C410A) mutant (Fig 5E), indicating that the binding ability of NEMO to both K63 and M1 chains is required for Tax-induced IKK activation.
Because the ability of NEMO to bind to both K63 and M1 chains and the Tax-induced generation of the IKK complex-associated K63/M1-linked hybrid chains are required for the activation of IKK by Tax, we hypothesized that the macromolecular complex of IKK may be formed through multivalent interactions between polyubiquitin chains and NEMO, facilitating trans-autophosphorylation between the IKK complexes, thereby inducing IKK activation. To test this possibility, we performed cell-free assays in the presence or absence of DN ubiquitin mutants, and the reaction mixtures were subjected to Blue native-PAGE, which can be used to determine the size and composition of native protein complexes [55], followed by immunoblotting. Probing the immunoblots with an anti-NEMO antibody revealed that the NEMO-containing macromolecular complex (arrowheads in Fig 5F left) was formed only in the presence of Tax in addition to the regular complex of approximately 600 kDa (dots in Fig 5F), which is observed as an inactive IKK complex in the absence of Tax. Interestingly, the formation of the macromolecular complex was abrogated when the extract was incubated with DN ubiquitin mutants (K27R, K63R, or HA-Ub) that also inhibit Tax-induced IKK activation (Fig 5F left). Furthermore, probing the immunoblots with an anti-p-IKKα/β antibody revealed that activated IKK was observed only when the macromolecular complex was formed and that activated IKK was included in the macromolecular complex (Fig 5F right). These results strongly suggest that polyubiquitination-dependent formation of the macromolecular IKK complex triggers IKK activation.
To investigate the physiological significance of LUBAC in HTLV-1-infected cells, we first checked the interaction between Tax and LUBAC. Lysates prepared from the HTLV-1-infected human T cell line HUT102, which is derived from a mycosis fungoides patient [56], were subjected to immunoprecipitation using an anti-Tax or a control antibody. HOIP and Sharpin were precipitated only when Tax was precipitated by the anti-Tax antibody (Fig 6A). To further confirm the physiological interaction between Tax and LUBAC, MT-2 or MT-4 cell lines, T cell lines transformed by co-culture with HTLV-1-producing ATL cells [57, 58], were used. HOIP and Sharpin were co-precipitated with Tax using an anti-Tax antibody. However, neither HOIP nor Sharpin were precipitated when Tax was knockdown (Fig 6B). These results indicate that Tax interacts with endogenous LUBAC in three distinct HTLV-1-infected cell lines. To understand whether LUBAC is involved in IKK activation in HTLV-1-infected cells, effect of HOIP knockdown on IKK activation was analyzed. Immunoblotting with an anti-pIKKα/β antibody revealed that HOIP knockdown blocked IKK activation in MT-2 and MT-4 cells (Fig 6C). Consistently, expression levels of the NF-κB target genes were notably reduced in HOIP-knockdown MT-4 cells (Fig 6D). Moreover, HOIP knockdown significantly suppressed cell proliferation (Fig 6E). Taken together, these results show that LUBAC-mediated M1 chain formation is required for NF-κB activation leading to target gene expression and cell proliferation in HTLV-1-infected cells.
Extensive studies on the role of Tax in ATL development have demonstrated that Tax is involved in leukemogenesis largely through its ability to constitutively activate NF-κB [7–9]. It has been known for almost two decades that the binding of Tax to NEMO is required for IKK activation, but the precise molecular mechanisms by which this binding leads to IKK activation remain to be elucidated. We and other groups have shown that the Tax-induced generation of K63 chains is crucial for IKK activation [27, 28, 30]. Furthermore, Ho et al. [28] recently provided clear evidence that RNF8 acts as an E3 ligase to generate K63 chains for IKK activation. As an extension of these previous results, we propose a novel molecular model for Tax-induced IKK activation in which LUBAC, together with unidentified E3 ligases for K63 chains, generate K63/M1-linked hybrid chains to form the active macromolecular Taxisome, composed of LUBAC, Tax and the active IKK complex, thereby establishing persistent NF-κB activation (Fig 7). Cell-free experiments allowed us to address the effects of chain type-specific blocking of polyubiquitin synthesis on critical steps of Tax-induced IKK activation. Several lines of evidence presented here support our model. 1) In vitro binding and immunoprecipitation experiments revealed that Tax can bind to both the IKK complex and LUBAC to form an inactive pre-Taxisome without the generation of polyubiquitin chains, whereas activation of the IKK complex by Tax requires the synthesis of the K27, K63 and M1 chains. 2) Genetic evidence revealed that both Ubc13 (an E2 enzyme for K63 chain synthesis) and each component of LUBAC (the only E3 enzyme for M1 chain) are crucial for Tax-induced IKK activation. 3) Mass spectrometric analyses revealed that both K63 and M1 chains are associated with the IKK complex only in the presence of Tax. 4) Cell-free experiments with chain type-specific DUBs revealed that K63/M1-linked hybrid chains are associated with the active IKK complex. 5) Tax-induced IKK activation requires the ability of NEMO to bind to both K63 and M1 chains. 6) The formation of the active macromolecular IKK complex (active Taxisome) requires the synthesis of the K27, K63 and M1 chains. Based on 4), 5) and 6), a single hybrid chain may bind to multiple NEMO molecules, and a single NEMO may act as a bridge between the hybrid chains, which may explain why generation of the hybrid chain is required for the macromolecular IKK complex. Regarding how the formation of the macromolecular complex leads to IKK activation, there are two potential mechanisms. The first is that the formation of the macromolecular complex could induce an efficient physical interaction between IKK complexes, so that trans-autophosphorylation between IKK complexes results in full activation of IKK. The second possible mechanism is that the formation of the macromolecular complex could somehow recruit an IKK kinase (IKKK) such as TAK1, which phosphorylates and activates IKK in response to cytokine stimulation [59]. We previously reported that MAP3Ks, including MEKK1, MEKK3, NIK, TPL-2 and TAK1, are dispensable for Tax-induced IKK activation [60]. In addition, our extensive proteomics analysis of the Tax-activated IKK complex failed to identify any IKKK candidates [61], leading us to prefer the trans-autophosphorylation model.
In contrast to our model, conflicting results have been reported on the following three points. The first point is the involvement of IKKK. Yin et al. [62] demonstrated that a dominant negative MEKK1 mutant inhibits IKK activation induced by Tax. Ho et al. [28] and Wu et al. [63] reported that Tax fails to activate the IKK complex in TAK1-knockdown HeLa cells or TAK1-knockout MEFs. This discrepancy may due to the experimental conditions including cell types used. Extensive analysis of the macromolecular complex and genetic and biochemical experiments in T cells are required to determine the involvement of any IKKK in Tax-induced IKK activation. The second point concerns the E3 ligase for K63 chain formation. Wang et al. [29] recently reported that Tax acts as an E3 ubiquitin ligase for IKK activation through synthesis of mixed-linkage polyubiquitin chains and that K63 chains are dispensable for Tax-induced IKK activation. However, we show that K63 chains are essential and could not demonstrate the E3 activity of Tax under any conditions we tested. Further identification of factors involved in Tax-induced IKK activation may explain these discrepancies. The third point concerns the subcellular localization of the active Taxisome. Pujari et al. [64] reported that membrane-associated Cell adhesion molecule 1 (CADM1) functions as a scaffold for Tax and Ubc13 to activate the IKK complex in intact cells. Since the S-100 fraction used in the cell-free system does not contain membrane fractions, we believe that Tax can interact with Ubc13 in the absence of CADM1. However, our data do not exclude the possibility that, in the cell free assay, Tax induces IKK activation without molecules required for subcellular localization in intact T cells. Precise comparison of IKK activation in the cell-free system with that in intact cells will explain regulation of Taxisome formation in HTLV-1 infected cells
Inconsistent with our model, the regular complex (dots in Fig 5F), which was inactive in the absence of Tax, appeared to be activated in the presence of Tax in addition to the macromolecular complex. This could occur because the macromolecular complex is unstable, such that the activated macromolecular complex dissociates into the active regular-size complex, or because the active macromolecular complex is able to transiently associate with and activate the regular complex. However, we cannot completely rule out the possibility that the IKK complex can be activated without the formation of the macromolecular complex in the presence of Tax.
Interestingly, cell-free experiments using IsoT and viral OTU revealed that both unanchored and anchored polyubiquitin chains are required for Tax-induced IKK activation. Although the critical roles of unanchored and anchored polyubiquitin chains in Tax-induced IKK activation are still controversial, our experiments using Tax KR mutants support the idea that the polyubiquitin chains conjugated to Tax belong to the class of substrate-anchored polyubiquitin chains required for Tax-induced IKK activation. By taking advantage of specific DUBs that degrade only one type, we showed for the first time that unanchored and anchored polyubiquitin chains cooperate in Tax-induced IKK activation.
Although the addition of the DN ubiquitin mutant K27R inhibited IKK activation by Tax, we did not detect K27 chains associated with the active Taxisome via the ubiquitin-AQUA method. This may be because an undetectable amount of K27 chains is involved in the generation of the hybrid chains as their component or because K27 chains act as initial triggers of the hybrid chains synthesis but are dissociated from the Taxisome afterwards. Further studies are needed to identify the precise role of K27 chains in Tax-induced IKK activation.
Tax binds to LUBAC through the UBL domain of HOIL-1L and RBR domain of HOIP. The UBL domain of HOIL-1L is involved in the association with HOIP, while the RBR domain of HOIP is the catalytic active site [66]. Therefore, the association of Tax with LUBAC may activate the E3 activity of LUBAC upon HTLV-1 infection, which may be one of the critical events in the onset of ATL. In accord with this scenario, small compounds that specifically block the interaction between Tax and HOIL-1L or between Tax and HOIP could be used as novel therapeutic approaches for ATL and HAM.
Human cDNAs encoding dominant-negative mutants of E2 enzymes were generated via PCR and inserted into the pRK5 vector. pMRX-Cre was obtained from S. Akira (Osaka University). Expression vectors for HOIL-1L-HA, Myc-HOIP and their deletion mutants were generated as previously described [66]. Human and mouse cDNAs encoding NEMO and NEMO mutants were inserted into the retrovirus vector pMXs obtained from T. Kitamura (University of Tokyo). Viral OTU cDNA was obtained from A. García-Sastre (Icahn School of Medicine at Mount Sinai). 3xκB-luc was obtained from S. Miyamoto (University of Wisconsin-Madison). Tax cDNA was obtained from J. Fujisawa (Kansai Medical University). Tax mutants were generated via PCR and inserted into the pCG vector.
The following antibodies were used: anti-p-IκBα (9246), anti-IκBα (9242), anti-p-IKKα/β (2697), anti-NEMO (2695) and anti-ubiquitin (3936) (Cell Signaling Technology); anti-GST (sc-459), anti-Myc (sc-789) and anti-HA probe (sc-805) (Santa Cruz Biotechnology); anti-Flag M2 (F3165) (Sigma); anti-tubulin (CP06) (Calbiochem); anti-HOIL-1L (NBP-1-88301) (Novus Biologicals); anti-HOIP (ARP43241) (Aviva Systems Biology); anti-Sharpin (14626-1-AP) (Proteintech); anti-linear polyubiquitin-specific (AB130) (LifeSensors); and anti-Ubc13 (37–1100) (ThermoFisher Scientific). The anti-Tax antibody was generated as previously described [67].
HEK293T cells (purchased from ATCC), Plat-E cells (provided by T. Kitamura) and mouse embryonic fibroblasts (provided by J. Silke, The Walter and Eliza Hall Institute, or established by us) were maintained in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% heat-inactivated fetal bovine serum (FBS). Jurkat cells (purchased from ATCC), the Jurkat-derived cell line JPX-9 cells (provided by K. Ohtani, Kwansei Gakuin University), and HTLV-1-infected cell line HUT102, MT-2, and MT-4 cells (provided by J. Fujisawa) were maintained in RPMI1640 supplemented with 10% heat-inactivated FBS. Sf9 cells were maintained in Sf900IIISFM (Thermo Fisher Scientific) supplemented with 10% FBS. The transfection of HEK293T cells was performed by the calcium phosphate method. siRNAs were transfected using NEPA21 Super Electroporator (NEPAGENE). Control stealth siRNA and the following stealth siRNAs (Thermo Fisher Scientific) were used: HOIP-1 sense/anti-sense, 5′-GGUACUGGCGUGGUGUCAAGUUUAA-3′/5′-UUAAACUUGACACCACGCCAGUACC-3′; HOIP-2 sense/anti-sense, 5′-CACCACCCUCGAGACUGCCUCUUCU-3′/5′-AGAAGAGGCAGUCUCGAGGGUGGUG-3′.
For lentivirus production, HEK293T cells were transfected with the self-inactivating lentiviral vector construct, the packaging construct and the VSV-G- and Rev-expressing construct. After 48 h of incubation, culture supernatants were collected and centrifuged at 50,000 x g for 1 h at 20°C to concentrate lentivirus. MT-2 or MT-4 cells were infected with the lentivirus at 400 x g for 2 h at 20°C. After 48 h, puromycin (Wako) was added to the medium, and puromycin-resistant cell pools were used for further experiments. The following target sequences were used: Tax-1, 5′-GGCCTTCCTCACCAATGTTCC-3′; Tax-2, 5′-GGCAGATGACAATGACCATGA-3′; Tax-3, 5′-GCCTACATCGTCACGCCCTAC-3′; HOIP-1, 5′-GCTGCAGCTTTCAGAATTTGA-3′; HOIP-2, 5′-GCACTGCCCATCCTGTAAACA-3′; HOIP-3, 5′-GCTCCTTTGGCTTCATATATG-3′; Control, 5′-GATTTCGAGTCGTCTTAATGT-3′.
For retrovirus production, Plat-E cells were transfected with pMRX-Cre vector. After 24 h, culture supernatants were collected, and Ubc13+/+ or Ubc13fl/fl MEFs were incubated with the retrovirus containing polybrene (10 μg/ml; Sigma-Aldrich) for 8 h. After 48 h, puromycin was added to the medium, and puromycin-resistant cell pools were used for further experiments.
His6-Tax, His6-M22, GST-viral OTU (WT) and its catalytic inactive mutant 1A were expressed in E. coli and purified. His6-TRAF6 and His6-Tax were expressed in Sf9 cells using the Bac-to-Bac Baculovirus Expression System (Thermo Fisher Scientific) and purified. GST, GST-HOIL-1L, GST-HOIP and GST-Sharpin were generated using the wheat germ cell-free protein synthesis system and purified [68]. Ubiquitin (U-100H), ubiquitin mutants (K6R (UM-K6R), K11R (UM-K11R), K27R (UM-K27R), K29R (UM-K29R), K33R (UM-K33R), K48R (UM-K48R) and K63R (UM-K63R)), HA-ubiquitin (U-110), IsoT (E-320), Otubain-1 (E-522B), AMSH (E-548B), OTULIN (E-558), USP2 (E-504), His6-UBE1 (E-304), UbcH5c (E2-627), UbcH7 (E2-640) and His6-Ubc13/Uev1A (E2-664) were purchased from BostonBiochem. Recombinant K48- and K63-linked Ub2-Ub7 chains (UC-230, UC-330) were purchased from BostonBiochem. Recombinant M1-linked Ub2-Ub7 chains (BML-UW1010-0100) were purchased from Enzo Life Sciences.
Jurkat cells and MEFs were suspended in hypotonic buffer (10 mM Tris-HCl (pH 7.5), 1.5 mM MgCl2, 10 mM KCl, 0.5 mM dithiothreitol (DTT) and protease inhibitor cocktail (Roche)) and then lysed with a Dounce homogenizer. Cell debris was removed via ultracentrifugation at 100,000 x g for 1 h at 4°C to prepare the S-100 cytosolic extract. Cytosolic extracts (10 mg/ml) were incubated with recombinant His6-Tax in ATP buffer (50 mM Tris-HCl (pH 7.5), 5 mM MgCl2, 2 mM ATP, 5 mM NaF, 20 mM β-glycerophosphate, 1 mM Na3VO4, and protease inhibitor cocktail) in the presence or absence of various recombinant DUBs. After incubation at 30°C for 1 h, the reaction mixtures were subjected to immunoblotting or immunoprecipitation.
Cells were lysed in IP buffer (20 mM Tris-HCl (pH 7.5), 150 mM NaCl, 2 mM EDTA, 1 mM MgCl2, 10 mM NaF, 1% NP-40, 10 mM β-glycerophosphate, 1 mM Na3VO4, 1 mM DTT, 5 mM N-ethylmaleimide (NEM) and protease inhibitor cocktail), followed by centrifugation at 22,000 x g for 15 min at 4°C to remove the insoluble fraction. For detection of polyubiquitination of Tax, the reaction mixtures were boiled for 10 min in the presence of 1% SDS to remove noncovalently attached proteins. The mixtures were then diluted 10-fold in IP buffer to reduce the SDS concentration to 0.1%. The cell lysates or the cell-free reaction mixtures were subsequently incubated with the antibodies plus protein G-sepharose. The immunoprecipitates were washed five times and subjected to immunoblotting. For immunoblotting, immunoprecipitates or cell lysates were separated via SDS-PAGE and transferred to PVDF membranes (Immobilon P, Millipore). The membranes were then incubated with the primary antibodies. Immunoreactive proteins were visualized with anti-rabbit or anti-mouse IgG conjugated to horseradish peroxidase, followed by processing with an ECL detection system.
Glutathione sepharose was incubated with 300 ng of GST, GST-HOIL-1L, GST-HOIP or GST-Sharpin at 4°C for 1 h in binding buffer (50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1 mM EDTA, 1% Triton X-100, 1 mM DTT, 2.5 mg/ml BSA and protease inhibitor cocktail). The beads were then incubated with 500 ng of His6-Tax at 4°C for 1 h. After incubation, the beads were washed and subjected to immunoblotting.
Jurkat cytosolic extracts (10 mg/ml) were incubated with recombinant His6-Tax in ATP buffer. After incubation at 30°C for 1 h, the reaction mixtures were subjected to immunoprecipitation with an anti-NEMO antibody in IP buffer. The immunoprecipitates were washed three times with IP buffer without NEM and incubated with Otubain-1 (5 μM), AMSH (5 μM), OTULIN (5 μM) or USP2 (5 μM) at 37°C for 1 h in DUB buffer (50 mM HEPES-KOH (pH 7.5), 100 mM NaCl, 1 mM MnCl2, 0.01% Brij-35 and 2 mM DTT). After incubation, the reaction mixtures were subjected to immunoblotting.
Jurkat cytosolic extracts (10 mg/ml) were incubated with recombinant His6-Tax in ATP buffer. After incubation at 30°C for 1 h, the reaction mixtures were subjected to immunoprecipitation with an anti-NEMO antibody in IP buffer. The immunoprecipitates were analyzed via the ubiquitin-AQUA method as described previously [48]. The immunoprecipitates were separated through SDS-PAGE, and the gel region above 50 kDa was subjected to in-gel trypsinization. The extracted peptides were analyzed with a Q Exactive mass spectrometer in targeted MS/MS mode together with 10 fmol of ubiquitin AQUA peptides.
After the cell-free reaction, the reaction mixtures were mixed with NativePAGE Sample Buffer (Thermo Fisher Scientific). Electrophoresis was performed using NativePAGE Running Buffer (Thermo Fisher Scientific) containing 0.002% G-250. The gels were soaked in denaturation buffer (10 mM Tris-HCl (pH 6.8), 1% SDS and 0.006% 2-mercaptoethanol) for 30 min at 60°C, followed by transfer to PVDF membranes and immunoblotting.
The total RNA was isolated from MT-4 cells transfected with control or HOIP siRNA with Trizol reagent (Thermo Fisher Scientific). cDNA was synthesized from 2.0 μg of total RNA with Prime ScriptII (Takara). Quantitative real-time PCR analysis was performed on CFX Connect (Bio-Rad). The level of GAPDH expression was used to normalize the data. The following primers were used: IL-6 sense/anti-sense, 5′-CCTGAACCTTCCAAAGATGGC-3′/5′-TTCACCAGGCAAGTCTCCTCA-3′; IL-1β sense/anti-sense, 5′-TTCGACACATGGGATAACGAGG-3′/5′-TTTTTGCTGTGAGTCCCGGAG-3′; MMP-9 sense/anti-sense, 5′-ATGTACCGCTTCACTGAGGG-3′/5′-TCAGGGCGAGGACCATAGAG-3′; GAPDH sense/anti-sense, 5′-TGCACCACCAACTGCTTAGC-3′/5′-GGCATGGACTGTGGTCATGAG-3′.
HEK293T cells were transfected with the plasmids encoding wild type or various Tax mutants together with 20 ng of luciferase reporter (3xκB-luc) and 30 ng of β-actin-β-galactosidase plasmid. After 48 h, the luciferase activity was measured using the Luciferase Assay System (Toyo Ink). β-galactosidase activity was used to normalize the transfection efficiency.
The reactions were performed at 37°C for 1 h in the reaction buffer containing His6-UBE1 (0.1 μM), the indicated E2 (0.2 μM), ubiquitin (25 μM) and His6-TRAF6 or His6-Tax in the presence of ATP (2 mM). After incubation, the reaction mixtures were analyzed by immunoblotting.
After cell-free reactions, the reaction mixtures were separated using 6% polyacrylamide gels containing 20 μM Phos-tag acrylamide (Wako) and 40 μM MnCl2. After electrophoresis, the gels were washed with transfer buffer containing 10 mM EDTA for 15 min. The gels were further washed with transfer buffer without EDTA for 10 min, and the samples were transferred to PVDF membranes, followed by immunoblotting.
Statistically significant differences between mean values were determined using Student’s t-test (**P<0.01, *P<0.05). Data are presented as the means ± SD.
|
10.1371/journal.pntd.0003341 | Natural and Vaccine-Mediated Immunity to Salmonella Typhimurium is Impaired by the Helminth Nippostrongylus brasiliensis | The impact of exposure to multiple pathogens concurrently or consecutively on immune function is unclear. Here, immune responses induced by combinations of the bacterium Salmonella Typhimurium (STm) and the helminth Nippostrongylus brasiliensis (Nb), which causes a murine hookworm infection and an experimental porin protein vaccine against STm, were examined.
Mice infected with both STm and Nb induced similar numbers of Th1 and Th2 lymphocytes compared with singly infected mice, as determined by flow cytometry, although lower levels of secreted Th2, but not Th1 cytokines were detected by ELISA after re-stimulation of splenocytes. Furthermore, the density of FoxP3+ T cells in the T zone of co-infected mice was lower compared to mice that only received Nb, but was greater than those that received STm. This reflected the intermediate levels of IL-10 detected from splenocytes. Co-infection compromised clearance of both pathogens, with worms still detectable in mice weeks after they were cleared in the control group. Despite altered control of bacterial and helminth colonization in co-infected mice, robust extrafollicular Th1 and Th2-reflecting immunoglobulin-switching profiles were detected, with IgG2a, IgG1 and IgE plasma cells all detected in parallel. Whilst extrafollicular antibody responses were maintained in the first weeks after co-infection, the GC response was less than that in mice infected with Nb only. Nb infection resulted in some abrogation of the longer-term development of anti-STm IgG responses. This suggested that prior Nb infection may modulate the induction of protective antibody responses to vaccination. To assess this we immunized mice with porins, which confer protection in an antibody-dependent manner, before challenging with STm. Mice that had resolved a Nb infection prior to immunization induced less anti-porin IgG and had compromised protection against infection.
These findings demonstrate that co-infection can radically alter the development of protective immunity during natural infection and in response to immunization.
| Vaccination studies in animal models have focused on understanding responses in young, previously naïve mice. In reality, humans are vaccinated or respond to infection in the context of a life-time of accumulated exposure to multiple, systemic infections and other vaccines, some of which are themselves attenuated live organisms. This is even more pronounced in areas that are endemic for infectious diseases. We wished to examine the impact infectious history can have on the immune response against infection and the efficacy of vaccination. To do this, we used two classes of pathogens that model clinically important invasive infections. One pathogen is the bacterium, Salmonella Typhimurium against which we have also developed an experimental porin vaccine, and the second is an invasive helminth, Nippostrongylus brasiliensis, that models aspects of hookworm infections. Our studies indicate that exposure to a second, unrelated pathogen can reduce the efficiency of immunity generated during natural infection and immunity generated after vaccination. These results are important as they help to identify potential strategies for improving immune-mediated control of infection and the success of vaccination in infection-endemic regions.
| Models examining immunity to experimental infections primarily focus on responses to a single pathogen or vaccine in an immunologically naïve host. Such studies have shaped our understanding of how infections develop and are controlled. However, in reality individuals are exposed to multiple pathogens, often concurrently during their life-time [1]–[4]. Whether infectious history may influence the type of immune response mounted by the host to a new vaccine pathogen has not been extensively explored.
Of particular significance is that regions endemic for non-typhoidal Salmonella (NTS) serovars, such as Salmonella Typhimurium (STm) [5] are also endemic for parasitic nematode infections, such as hookworm [6]. This provides opportunities for concomitant STm and helminth infections to develop. In distinct forms both infections can be modelled in a murine system. Nippostrongylus brasiliensis (Nb), a natural parasite of rats is used as a model infection in mice of human hookworm disease. Nb induces Th2 features such as interleukin 4 (IL-4), IL-13, IgG1 and IgE [7]–[14]. Infection with Nb in mice is self-limiting, with worms cleared from BALB/c mice in a narrow period of 9–11 days post-infection when mice are infected with the common dose of 500–750 L3 larvae [7], [11]. Having these defined kinetics for clearance enables identification of factors that interfere with immunity. Exposure to an additional agent after resolution of Nb infection enables any lasting influence of helminth infection on host immunity to the second antigen to be identified.
Clearance of STm infections require Th1-mediated immunity, characterized by the induction of Interferon (IFN) γ and IgG2a in mice [15]–[18]. A mutation in the Slc11a1/Nramp gene renders mouse strains, such as BALB/c, hyper-susceptible to virulent strains of STm whilst attenuated strains are cleared gradually. For the latter strains, such as the AroA-deficient STm strain SL3261, clearance is achieved 1–2 months after infection with a typical dose of 5×105 bacteria administered systemically [19]–[21]. A striking component of host immunity to attenuated STm is a rapid and extensive extrafollicular (EF) antibody response with switching to IgG2a and IgG2b, which occurs without parallel germinal centre (GC) induction [19].
While B cells and antibody are wholly dispensable for controlling primary STm murine infections [20], [22], [23], the presence of antibody to STm prior to infection can be protective [19], [22], [24], [25]. Indeed, we have found that immunization with purified porins induced antibody sufficient to protect against subsequent STm infection [26], with IgG augmenting the protection afforded by IgM. Thus, factors that influence IgG responses are likely to affect protection and immunization with porin proteins.
Helminth infections may modulate responses to other pathogens [27]–[29] and to vaccination [30]–[33], although the nature of these influences have not been fully elucidated. Furthermore, such studies have often not addressed the impact of co-infection on the immunological response to each infection. In this study, we investigated the development and efficacy of immune responses after immunization with combinations of Nb, STm and porins. Our data shows that co-infection with Nb and STm impairs clearance of both pathogens. Whilst some changes in cytokine patterns were observed, the pathogen-associated pattern of isotype-switching was conserved so that specific IgG1, IgE and IgG2a responses all developed in parallel. Furthermore, prior Nb infection impaired the protective efficacy of porin immunization indicating a longer-term impact of helminth infection, suggesting these effects were not necessarily dependent upon an active infection. These data not only further our understanding of the relationship between host and pathogen and the mechanisms used to regulate immune function, but also identify the need to consider the impact of infectious history on the host's capacity to implement protective immunity.
Specific pathogen-free 6–8 week BALB/c mice were obtained from the animal facility at the University of Cape Town, South Africa. All animal procedures were carried out under Protocol 012-006 which was approved by the Animal Research Ethics Committee at the University of Cape Town. All procedures were also conducted in strict accordance with the South African code of practice for laboratory animal procedures.
STm SL3261 is an attenuated strain of STm SL1344 [34]. Nb was maintained through passage in rats. Outer membrane preparations from STm were generated by 2% (vol/vol) Triton X-100 extraction [19]. Purified porins from STm (strain ATCC 14028) were generated as described previously [26], [35] using SDS and FPLC and suspended in PBS in 0.1% (wt/vol) SDS. Nb total antigen preparations were generated by snap-freezing L3 stage larvae and homogenizing by sonication. Antigen was stored until use at −80°C.
Mice were infected intraperitoneally (i.p.) with 5×105 STm SL3261. Tissue bacterial burdens were determined by direct culturing. Mice were infected subcutaneously (s.c.) with 500 Nb L3 larvae. Adult worm burdens were determined by counting in the gut lumen under a dissecting microscope as previously described [9]. Where stated, mice were immunized i.p. with 20 µg porins in PBS. Opsonizing bacteria with antisera was performed as described previously [19], [26]. A single serum was used per mouse and sera were heat-inactivated at 56°C for 0.5 h to inactivate complement. Bacteria (2.5×106/mL) and sera (1∶100) were mixed for 1 h before infection. Bacterial viability and lack of agglutination were confirmed by plating.
Splenic single cell suspensions were prepared and red cells were lysed with ACK lysing buffer (Gibco Life Technologies). Cells were initially blocked prior to staining with anti-CD16/32 antibody before surface staining for 20 min at 4°C with CD3-FITC (Clone 145-2C11), CD4-PerCP Cy 5.5 (Clone RM4-5). Intracellular cytokine staining was performed by ex-vivo re-stimulation as described previously [15]. Briefly, 5×106 splenocytes were plated with 1 µg/ml anti-CD28 (clone 37.51) and re-stimulated in a pre-coated well with anti-CD3 (10 µg/ml) (clone 145-2C11). Cells were incubated for 2.5 h followed by 2.5 h with GolgiStop (BD Biosciences). Cells were then surface stained, fixed and permeabilised with Cytofix/Cytoperm Plus for 20 min at 4°C before intracellular cytokine staining using IL-13 PE (Clone JES10-5A2) or IFNγ-APC (Clone XMG1.2) or isotype controls (all BD Biosciences). Cells were acquired using a FACSCalibur (BD Biosciences) and analysed using FlowJo Software.
Immunohistology was performed on frozen sections as described previously [19], [36] with tissues frozen in liquid nitrogen. CD3, IgG2a, IgG1 (Clone LO-MG1-2), IgE (Clone LO-ME-2) and FoxP3 cells were detected using rat anti-mouse antibodies in conjunction with biotinylated rabbit anti-rat immunoglobulins. Signal was developed using streptavidin ABComplex alkaline phosphatase (DakoCytomation) with naphthol AS-MX phosphate with Fast Blue salt and levamisole. Sheep anti-mouse IgD binding was detected using horseradish peroxidase (HRP)-conjugated donkey anti-sheep immunoglobulins with Diaminobenzidine (Sigma Aldrich). Hamster anti-mouse CD3 binding was detected using goat anti-hamster IgG followed by HRP-conjugated donkey anti-sheep immunoglobulins with Diaminobenzidine. The area of the spleen occupied by germinal centres and cells per square millimeter were calculated using a point-counting technique as described previously [37].
Enzyme-linked immunosorbent assay (ELISA) was performed as described previously [19]. NUNC Maxisorp plates were coated overnight with antigen at 5 µg/ml in coating buffer. Plates were then blocked with 1% BSA before serum was added in PBS-0.05% Tween-20 and diluted stepwise. Bound antibodies were detected using alkaline-phosphatase conjugated, goat anti-mouse secondary antibodies (Southern Biotech) and Sigma-Fast p-nitrophenylphosphate (Sigma Aldrich). The absorbance at ODλ405 nm was determined using an Emax microplate spectrophotometer (Molecular Devices, Germany). Relative reciprocal titres were calculated by measuring the dilution at which the serum reached a defined ODλ405 nm.
Splenocytes (2×105) were plated for 48–72 h with 1 µg/ml anti-CD28 (Clone 37.51) and re-stimulated with either 10 µg/ml anti-CD3 (Clone 145-2C11) which was pre-coated overnight or 10 µg/ml heat-killed STm. Heat-killed STm was prepared by heat inactivation at 72°C for 1 hour. Control wells were stimulated with anti-CD28 and PBS. Cytokines secreted into the supernatants were then measured using the appropriate ELISA Ready-Set-GO kit (eBiosciences) as per manufacturers' instructions. Briefly, plates were coated overnight with capture antibody, blocked for 1 h at room temperature with 2% fat-free milk in PBS, after which samples and standards were added overnight at 4°C. Biotinylated secondary antibodies were then added and signal detected using streptavidin-HRP and 3,3′,5,5′-tetramethylbenzidine solution before stopping with 1 M H3PO4. The absorbance at ODλ450 nm (background at ODλ540 nm) was determined using a Versamax tunable microplate spectrophotometer (Molecular Devices, Germany).
The data is expressed as the mean plus one standard deviation. Significant differences were determined using the Mann-Whitney non-parametric two-tailed test using GraphPad Prism Version 5. P≤0.05 was accepted as significant.
To assess whether synchronous administration of STm and Nb altered the kinetics of clearance, we infected WT mice with either 5×105 attenuated STm i.p., 500 L3 Nb larvae s.c., or both pathogens for 5, 10, 18 or 32 days (Figure 1A). While equivalent bacterial numbers were found in the spleens and livers of STm and co-infected mice at day 5 post-infection, after this time bacterial numbers were consistently higher in co-infected mice (Figure 1B). As expected intestinal worm burdens in Nb-only infected mice were largely cleared by day 10. However, co-infected mice demonstrated persisting Nb infection up to 32 days. Thus, co-infection with STm and Nb impairs immunity to both pathogens.
The impact of co-infection on pathogen clearance suggested perturbed type-specific immunity to each pathogen. The proportion and numbers of T cells from co-infected mice that produced IFNγ or IL-13 after anti-CD3 re-stimulation were largely similar to that seen after single STm or Nb infection respectively, at all time-points (Figure 2A). As the capacity to induce pathogen-associated Th1 and Th2 cytokines was maintained, levels of secreted cytokines from splenocyte cultures were examined (Figure 3A). After re-stimulation with anti-CD3 in the presence of anti-CD28 secreted levels of IFNγ were similar between STm-only and co-infected mice at all time-points examined, reflecting the intracellular cytokine staining. In contrast, levels of IL-4 and IL-13 were greatly reduced at times after co-infection compared with Nb-only infected mice (Figure 3A). To examine if this reflected cytokine responses induced after re-stimulation with STm, splenocytes from day 10 infected mice were re-stimulated with heat-killed STm instead of anti-CD3 (Figure 3B). Levels of IFNγ were similar in both STm-infected groups but there was an increase in IL-4 and IL-13 in the co-immunized group. Therefore, co-infection has little impact on the development of Th1 and Th2 cytokine-producing T cells but can modulate the levels of cytokines secreted.
Cytokines from non-T cells can influence the functional response of T cells [38]. Therefore IFNγ or IL-13 expression in CD3−ve cells was assessed by flow cytometry (Figure 2B). This showed that at all times after infection <1% of cells were positive for IFNγ or IL-13. Furthermore, co-infection did not dramatically alter the cytokine pattern seen after single infection. In addition, we examined cytokine secretion by splenocytes from mice infected for 10 days which were cultured without stimulation (Figure 3C). Cytokines were detected at lower levels than after anti-CD3 stimulation. IFNγ levels were similar in all groups except for the group that only received Nb, where they were lower. IL-4, but not IL-13, levels were reduced in co-infected mice compared to Nb only infected mice.
Helminth infections are associated with the induction of T regulatory (Treg) cells [39]–[41]. Therefore, it is possible that co-infection could either augment or diminish Treg responses and the levels of associated IL-10 observed compared to each pathogen alone. Initially, levels of secreted IL-10 after re-stimulation with anti-CD3 were assessed by cytokine ELISA (Figure 3A). This revealed that IL-10 was readily detected after Nb infection, but after STm infection the levels were similar to those of non-infected cultures. When IL-10 was examined after co-infection it was found to be intermediate between the STm-only and Nb-only infected mice on days 10 and 18 (Figure 3A). Thus, the presence of STm was associated with a moderation in the levels of IL-10 detected in Nb-infected mice. Since Treg are significant sources of T cell-derived IL-10, the impact of co-infection on Tregs was assessed. To do this we used immunohistochemistry to examine the frequency of FoxP3+ T cells in the T zones of infected mice on day 5 after infection, when pathogen burdens were similar in mice that received one or both pathogens (Figure 1A). Reflecting the IL-10 results, the density of FoxP3+ cells in the T zones of co-infected mice was significantly lower relative to mice that were only infected with Nb, yet significantly higher than mice only infected with STm (Figure 4).
STm and Nb induces immunoglobulin-switching to the Th1-reflecting IgG2a isotype or the Th2-reflecting isotypes IgG1 and IgE respectively. Furthermore, in this model of STm infection GC are absent early in the response, only becoming detectable later, when the infection has largely cleared [19]. As the direction of immunoglobulin-switching in mice can be influenced, in part by the cytokine milieu, it was possible that the altered cytokine environment during co-infection could alter the immunoglobulin-switching profile. In mice infected only with Nb, robust IgG1 and IgE EF plasma cell responses were detected, with IgG2a barely detectable by day 10 post-infection. This response was further characterised by an extensive GC response (Figure 5A). Mice infected with STm alone developed a robust IgG2a response with few IgG1 and no IgE cells detected. This response developed in the near total absence of GC, which only developed late in the response (Figure 5A). Surprisingly, in co-infected mice at days 10 and 32 post-infection a mixed switching-pattern was observed with IgG1, IgG2a and IgE plasma cells all readily detectable in EF foci. Interestingly, in co-infected mice development of the robust Nb-associated GCs was abrogated, only becoming detectable at day 32 post-infection (Figure 5A). Thus, the direction of B cell switching is maintained during co-infection, with the features of the response to each individual pathogen conserved.
Antibody responses are dispensable for the control of primary STm infection in mice, although they play a central role in protecting against secondary infection and in vaccine responses [16], . Thus, we examined the impact of co-infection on serum antibody responses to both pathogens to identify if long-term protective immunity may be compromised by co-infection with Nb. Serum antibody responses against outer membrane antigens of STm at days 10, 18 and 32 post-infection were measured. Reflecting the early conserved EF plasma cell responses, IgM and IgG antibody titres were similar at days 10 and 18 post-infection in both groups that received STm (Figure 5B). In contrast, at day 32 post-infection when GC are detected, IgM, IgG and IgG2a titres were lower in co-infected mice relative to STm-only infected mice, despite the GC response being comparable between the two groups (Figure 5A). Measurement of specific antibody responses in Nb-only and co-infected mice revealed that serum anti-Nb IgM, IgG and IgG1 titres were reduced in co-infected mice relative to Nb-only infected animals at day 10 post-infection, when GC responses were diminished. However, by day 18 IgG responses were similar between the two groups as the GC start to become more established in co-infected animals (Figure 5A). Thus co-infection can impact serum antibody titres to each pathogen but does not necessarily alter the switching profile.
Since simultaneous infection with Nb and STm could impair host control towards each pathogen the influence of sequential exposure was assessed. WT mice were infected with Nb and at day 16 (6–7 days post worm-expulsion) mice were challenged with STm for 5 or 25 days (Figure 6). While early control of STm was comparable between non-Nb primed and Nb-primed mice, prior Nb infection impaired control of STm at day 25, reflecting our earlier observations (Figure 1B). The impact of prior Nb infection on serum antibody responses to STm was then examined. This showed that antecedent Nb infection had no influence on anti-STm IgM titres but impaired anti-STm IgG titres at day 25 post-STm infection, with both IgG2a and IgG2b titres lower in mice previously infected with Nb (Figure 6). Thus prior Nb-infection can impair antibody switching to STm and the late control of subsequent STm infection.
Antibody induced during infection can protect against secondary STm infection [16], [19], [20], [22], [23]. Previously, we demonstrated that immunization with the porin proteins OmpC, D and F (collectively called porins) was sufficient to protect mice from STm infection via an antibody-dependent mechanism [26]. This offered an opportunity to examine the impact of prior Nb infection on antibody-mediated control of STm infection. To do this, groups of mice either received no intervention before STm infection, or combinations of Nb and porins before STm challenge (Figure 7A). After 5 days of infection splenic bacterial burdens were assessed. This showed that both porin-immunized groups had significantly lower bacterial numbers relative to non-immunized mice. Nevertheless, porin-immunized mice that had first been infected with Nb had a greater bacterial load than mice that had only received porins before infection (Figure 7A), indicating that Nb-infection can impair the protection conferred by porin-immunization.
Prior Nb infection may impact upon the success of immunization through at least two routes. Firstly, it may alter the activity of antibody, possibly through altering macrophage populations and their opsono-phagocytic capacity. Secondly, reduced benefit from immunization may reflect lower levels of antibody induction. To test the former, bacteria were opsonized with complement-inactivated sera from mice that had either been infected with STm or immunized with porins (Figure 7B). Opsonized bacteria were then given to mice i.p. that had either received PBS or Nb 18 days previously and bacterial burdens were enumerated 5 days later. In each case bacterial numbers recovered from mice infected with opsonized bacteria were similar irrespective of whether they had previously been infected with Nb (Figure 7B). This suggests there was no intrinsic defect in antibody-mediated control of STm infection in Nb-infected mice.
Since ≥95% of the protection provided by anti-porin antibody is through the induction of IgG [26], anti-porin antibody titres in mice immunized with porins after Nb infection were assessed. After immunization, porin-immunized Nb-infected mice had lower total anti-porin IgG serum titres than non-Nb infected counterparts (Figure 7C). Analysis of the distinct IgG isotypes induced showed there was diminution in IgG1 titres, whereas there was a negligible effect on IgG2a (Figure 7C). Therefore, prior Nb infection influences the titre of anti-porin IgG induced, but does not necessarily affect the efficacy of killing bacteria pre-opsonized with antibody. Finally, we looked to see if boosting with porins in Nb-infected mice could restore anti-porin antibody titres (Figure 7D). WT mice given PBS or Nb were immunized 18 days later with porins and 18 days after this some mice received a second porin-immunization. Antibody responses were then assessed after 7 days. Anti-porin IgG titres were similar in both boosted groups, irrespective of whether they were previously infected with Nb. This suggests that the reduced antibody titres observed after porin immunization can be restored through engagement of B cell memory.
This work identifies the mutual impairment in immune regulation when infection with Nb and STm occurs concurrently, as marked by the delayed clearance of STm and expulsion of Nb. This impaired host control was not limited to synchronous challenge with both pathogens as prior infection with Nb also impacted on the host response to STm and impaired vaccine-mediated protection, despite adult worms having been cleared. This indicates that the persistence of viable adult worms is not necessary for this effect, as described previously [42]. This is important as it supports the concept that the impact of infectious history or co-infection may not always require direct physical association between the pathogens, as shown with bacterial microflora and Trichuris muris [43].
The delay in STm clearance after infection with Nb was only apparent at times when adaptive immunity controls infection. Nevertheless, an impairment in the induction of Th1 cells or secretion of IFNγ after anti-CD3 stimulation was not obvious, nor was there a change in the levels of IFNγ after culture of splenocytes without stimulation. This suggests the underlying reason for defective immunity is not one of a failure to mount an appropriate immune response but may relate to other factors, such as the inefficient migration of T cells or inappropriate interactions between T cells and macrophages. Otherwise, the elevated IL-10 production observed in co-infected mice may alter the kinetics of STm clearance. Relevant to this perhaps is the increase in FoxP3 cells detected in the T zone after co-infection compared to STm alone. This may alter the functionality of T cells and limit their ability to promote bacterial clearance. Furthermore, during co-infection diminished, but not absent, Th2 cytokine secretion was observed and IL-4 and IL-13 were detectable after stimulation of splenocytes with killed STm. Although Th1 and Th2-associated responses can co-develop [44], in vivo and in vitro Th1 and Th2 cytokines have been shown to have opposing and suppressing activities [45], [46]. In the context of this study, only lower Th2 cytokine production was observed and this was partial, suggesting some potential Th1 dominance here, possibly because STm directly colonizes the spleen. Furthermore, IL-4 and IL-13 were both detectable in the day 32 STm-only group, probably reflecting the function of these molecules in GC development [47]. Nevertheless, it may be the balance between Th1 and Th2-associated cytokines, rather than the absolute amounts of each cytokine considered in isolation, which is the important factor. Such a consideration is relevant in other systems such as experimental Leishmania major infection [48]. Alternatively, this may simply reflect this specific combination of pathogens.
Other reasons may help account for the delayed control of STm infection. Levels of IL-4 and IL-13 were higher in non-stimulated splenocyte cultures from co-infected mice relative to mice only infected with STm. This may indicate other non-T cells contribute or impair clearance of STm through collaboration with T cells. Obvious candidates are innate lymphoid cells. Group 2 innate lymphoid cells (ILC2s) have been shown to release IL-13 in response to helminth infection [49] and recently the importance of ILC2s for the efficient development of Th2 cell responses during a Nb infection was demonstrated [38]. Therefore, in the same way that ILC2s can contribute positively to clearance of helminth infection they may impede the functioning of Th1 immunity. Many of the factors identified that potentially explain the failure to properly control STm infection in co-infected animals may also explain the delayed clearance of Nb. The cytokine most associated with efficient clearance of helminth infection is IL-13. Therefore, the diminished IL-13 cytokine production detected, in combination with the elevated levels of IFNγ, may inhibit the rate of worm expulsion. Other reasons that could help account for the delayed clearance of Nb include reduced levels of IL-4 production or a reduced expression of the respective receptors for IL-4 and IL-13 on cells such as smooth muscle cells [11], [50] or B cells [14]. The intermediate levels of FoxP3 T cells observed during co-infection may paradoxically have a negative effect on Nb clearance through enhancing Th1 inflammation and thus restricting the limited Th2 response induced from functioning.
Furthermore, in responses to other helminths loss of MyD88 in mice can enhance protection [51]. Therefore it may be that strong engagement of this molecule, for instance through the multiple TLRs triggered by STm, inhibits immunity. These factors could collaborate to limit the efficacy of the Th2 response induced and diminish the efficiency of worm clearance. One possibility to consider is if the addition of exogenous Th2 cytokines would recapitulate the protective immunity to Nb seen in the absence of STm co-infection. We would expect not for two reasons. First, the presence of Nb during STm infection has virtually no impact on IFNγ production, suggesting that the pro-inflammatory cytokine profile and possibly its anti-Th2 activities would be retained. Second, relates to the technical complexity of delivering IL-4 or IL-13 sufficient within the host to overcome this inhibition. This can be achieved by delivering these cytokines through a pump or as a complex with antibodies [52], although being able to provide this continuously and throughout infection would be challenging and prohibitive.
Antibody plays an important role in preventing re-infection with STm and the appearance of antibody to the pathogen correlates with reduced risk of bacteraemia in infants, but in the mouse it is not required for the control of primary infection [53]. Furthermore, the Vi capsular polysaccharide vaccine against typhoid works via the induction of antibody [54] and provides equivalent protection in the first few years after administration as the live, attenuated vaccine. Thus understanding how optimal levels of antibody to STm are induced is important to understand the mechanisms of control to this pathogen. STm alone failed to induce GC in the first weeks of infection, whereas Nb-infection induced pronounced GC responses and co-infection resulted in the abrogation of this response to Nb. Therefore, whilst the direction of EF switching in the spleen is largely independent of the presence of a second pathogen, the development of GC responses is not. In vitro and in vivo IL-4 is essential for directing B cell switching to IgE [55], but is dispensable for IgG1 switching [56]. Unexpectedly, EF IgG1 and IgE switching in the spleen was detectable at similar levels in both co-infected and Nb-only infected mice, despite reduced levels of IL-4 after co-infection. This implies that whilst IL-4 is essential for IgE switching, it may only be required at low levels. Furthermore, the augmented levels of Th2 cytokines during co-infection did not moderate the induction of IgG2a to STm. Therefore, both Th1 and Th2 cell priming and the characteristic class-switching profile is conserved and co-developed in the same responding secondary lymphoid tissue during co-infection. This is compatible with our earlier observations immunizing with soluble flagellin and flagellated bacteria where the direction of antibody-switching was conserved relative to the direction of T cell differentiation [15], [57]. This is important as it indicates that only selective elements of immunity are influenced by the presence of infecting organisms.
Despite EF switched plasma cell numbers being similar between co-infected mice and mice challenged with either STm or Nb there were some effects of co-infection on antibody titres. The anti-STm antibody response was similar between both STm-infected groups at day 18, yet at day 32, a time when antibody would largely originate from the GC, there was a clear reduction in IgM and switched antibody titres despite no difference in the splenic area occupied by GC. One possibility is that although the total number of GC may be similar between STm-only and co-infected mice at day 32, some of the GC in co-infected mice are Nb-specific and others STm-specific. Alternatively, it may relate to the higher bacterial burdens seen on day 32 in co-infected mice, which can alter the kinetics of GC induction [19] or other factors may be involved. Such influences may also explain why there was a lasting influence of Nb infection on anti-STm IgG antibody titres when Nb infection preceded STm infection. This impact on antibody titres was not restricted to live STm as the antibody response to STm porins was also lower when administered after Nb infection. Lower IgG titres were associated with diminished protection from infection, whilst the capacity of Nb-infected mice to control infection with antibody-opsonized STm was similar to non-Nb infected controls. This suggests that the capacity of cells to phagocytose and kill STm is not influenced by Nb-infection since antibody does not kill STm via cell-free complement-mediated mechanisms in mice [58]. Anti-NTS IgG strongly correlates with lower risk of invasive NTS infection in humans [53], and our study implies that the level of anti-porin IgG titres may influence protection. Whether co-infection with STm and helminths in humans is associated with altered IgG titres to STm and risk of infection needs to be addressed.
Helminth infections in humans are associated with lower vaccine efficacy to subunit and live vaccines [30]–[33], [59]. For instance, helminth infections are associated with diminished IgG and IgA antibody responses to cholera toxin B subunit [60] and to a live-attenuated oral cholera vaccine strain [61]. Interestingly, while treatment for helminth infection prior to vaccination can improve vaccine responses [61] our results indicate that prior infection could continue to have a detrimental effect on efficacy, although this that can be circumvented by antigen boosting.
In summary, helminth infections can influence antibody responses to STm and subunit vaccines and this should be considered when translating findings generated in animal models into humans, particularly in regions endemic for helminths. Understanding how helminths influence antibody induction will help us identify how best to employ vital life-saving vaccines. As antibody titres to porins post-Nb infection reached normal levels after boosting it would suggest that exploiting memory B cell responses would be important for the efficacy of subunit vaccines in helminth-endemic regions.
|
10.1371/journal.pgen.1005335 | Separable Crossover-Promoting and Crossover-Constraining Aspects of Zip1 Activity during Budding Yeast Meiosis | Accurate chromosome segregation during meiosis relies on the presence of crossover events distributed among all chromosomes. MutSγ and MutLγ homologs (Msh4/5 and Mlh1/3) facilitate the formation of a prominent group of meiotic crossovers that mature within the context of an elaborate chromosomal structure called the synaptonemal complex (SC). SC proteins are required for intermediate steps in the formation of MutSγ-MutLγ crossovers, but whether the assembled SC structure per se is required for MutSγ-MutLγ-dependent crossover recombination events is unknown. Here we describe an interspecies complementation experiment that reveals that the mature SC is dispensable for the formation of Mlh3-dependent crossovers in budding yeast. Zip1 forms a major structural component of the budding yeast SC, and is also required for MutSγ and MutLγ-dependent crossover formation. Kluyveromyces lactis ZIP1 expressed in place of Saccharomyces cerevisiae ZIP1 in S. cerevisiae cells fails to support SC assembly (synapsis) but promotes wild-type crossover levels in those nuclei that progress to form spores. While stable, full-length SC does not assemble in S. cerevisiae cells expressing K. lactis ZIP1, aggregates of K. lactis Zip1 displayed by S. cerevisiae meiotic nuclei are decorated with SC-associated proteins, and K. lactis Zip1 promotes the SUMOylation of the SC central element protein Ecm11, suggesting that K. lactis Zip1 functionally interfaces with components of the S. cerevisiae synapsis machinery. Moreover, K. lactis Zip1-mediated crossovers rely on S. cerevisiae synapsis initiation proteins Zip3, Zip4, Spo16, as well as the Mlh3 protein, as do the crossovers mediated by S. cerevisiae Zip1. Surprisingly, however, K. lactis Zip1-mediated crossovers are largely Msh4/Msh5 (MutSγ)-independent. This separation-of-function version of Zip1 thus reveals that neither assembled SC nor MutSγ is required for Mlh3-dependent crossover formation per se in budding yeast. Our data suggest that features of S. cerevisiae Zip1 or of the assembled SC in S. cerevisiae normally constrain MutLγ to preferentially promote resolution of MutSγ-associated recombination intermediates.
| At the heart of reproductive cell formation is a nuclear division process (meiosis) whereby homologous chromosomes segregate from one another. Meiotic partner chromosomes establish exclusive associations via a patterned distribution of crossover recombination events. During the maturation of recombination intermediates into crossovers, homologous axes are aligned in the context of a striking proteinaceous structure, the synaptonemal complex (SC). While genetic data link the SC with crossovers, it is unclear whether the mature SC structure facilitates crossover formation. Here we describe an interspecies complementation experiment in which we replace the S. cerevisiae version of an SC structural protein with an ancestrally related version from K. lactis. Our experiment reveals that, while SC proteins are required, mature full-length SC is dispensable for the formation of SC-associated crossovers in budding yeast. We furthermore discovered that most, but not all, members of a conserved meiotic crossover pathway are required for the crossovers that form in this interspecies context. Our findings strengthen the notion that a primary function of many SC proteins is to facilitate crossover recombination, independent of a role in building the larger SC structure. Furthermore, these data suggest that during normal meiosis in S. cerevisiae the assembled SC may act to functionally couple key crossover recombination proteins to one another.
| The segregation of homologous chromosomes at meiosis I is essential for the formation of haploid reproductive cells. Accurate segregation is dependent on the establishment of one or more associations between homologous chromosomes [1,2]. For most organisms, crossover recombination events in conjunction with sister chromatid cohesion provide the temporary associations needed between homologous chromosomes for their proper alignment and segregation on the meiosis I spindle. Interhomolog crossovers arise via the resolution of joint molecule (JM) intermediates, such as double Holliday junctions (dHJs), that form between homologous partner chromosomes during the repair of programmed, double-stranded DNA breaks (DSBs). The formation of interhomolog crossovers during meiosis depends on meiosis-specific proteins and, in a number of organisms, is temporally and functionally linked to a conserved meiotic chromosomal structure called the synaptonemal complex (SC).
Recombination-based associations between homologs can be cytologically detected and are referred to as chiasmata [1,3–5]. During the maturation of recombination intermediates into crossovers, however, such sites are often obscured by the presence of SC, a prominent, proteinaceous structure assembled along the entire lengthwise interface of aligned homologous chromosomes. The SC has a tripartite organization. One component of the larger structure is established via the multimeric assembly of coiled-coil containing proteins that form transverse filaments [6–8]. Transverse filaments are oriented perpendicular to the long axis of an aligned homolog pair and span the width of the SC, bridging the proteinaceous axes of each chromosome. Chromosome axes are referred to as lateral elements within the context of the mature SC. Additional proteins that make up the mature SC’s “central element” substructure assemble at the midline of the SC’s central region, apparently associating with and perhaps organizing transverse filament proteins. Zip1 is a coiled-coil protein component of the transverse filaments of the budding yeast SC [9,10], while Ecm11, SUMO and Gmc2 are proteins that are incorporated into the central element substructure [11–13].
Several additional proteins that are critical for the elaboration of SC along chromosomes in budding yeast do not appear to form structural components of the complex. These so-called “Synapsis Initiation Complex” (SIC) proteins [14], which include Zip2, Zip3, Zip4 and Spo16, localize at SC assembly (synapsis initiation) sites on meiotic chromosomes, many of which are thought to correspond to sites of ongoing recombination, and remain predominantly distributed as foci on full-length SCs after synapsis is complete [11–13,15–18].
The SC structure is established downstream of initial homology recognition and mediates the close apposition of homologous chromosomes (synapsis) during mid-meiotic prophase; the SC thus forms the context in which the majority of meiotic crossovers mature. The characterization of meiotic mutants has revealed a tight correlation between the presence of SC and the establishment of a proper number and distribution of interhomolog crossover recombination events, raising the possibility that SC structure itself plays a functional role in promoting crossover formation [1,7,19,20]. However, the molecular relationship between SC proteins, SC structure and the processing of recombination intermediates remains uncertain. The SC has also been linked to meiotic checkpoint signaling during meiosis, which can delay or arrest meiotic progression [21,22].
In many species, mutants defective in SC assembly (synapsis) exhibit a deficit in a genetically defined subset of crossovers, sometimes referred to as “class I” events [7,23–29]. SC-associated crossovers rely on SC proteins (SIC proteins and SC structural proteins in budding yeast) and often also rely on specific eukaryotic homologs of the bacterial MutS and MutL mismatch repair proteins (the Msh4/Msh5 and Mlh1/Mlh3 heterodimers, which comprise MutSγ and MutLγ, respectively) to promote the formation, maturation and resolution of the majority of dHJ intermediates that arise during meiosis [23,27,30–39]. The Msh4/Msh5 heterodimer (MutSγ) is capable of forming a “clamp” on double-stranded DNA and can recognize HJ structures [36]; these observations in conjunction with other data have led to the idea that Msh4/Msh5 acts to protect a dHJ intermediate from the anti-crossover activity of helicases such as Sgs1 [40,41]. Alternatively, or in addition, Msh4/Msh5 might promote the formation of a JM structure that can be recognized by a MutLγ-associated resolvase complex (in budding yeast this resolvase complex appears to involve MutLγ and Exo1 [23]), or may directly recruit MutLγ complexes to dHJs [32]. Once targeted, the MutLγ-Exo1 complex presumably resolves dHJ intermediates through its endonuclease activity [23,33,38,42]. The MutSγ complex can be detected cytologically at chromosomal sites where SIC proteins (Zip2, Zip3, Zip4) localize, and although MutSγ is dispensable per se for Zip1 elaboration along chromosomes, mutants missing MSH4 have been reported to exhibit delayed SC formation [30], suggesting the possibility of a complex interplay between the SC assembly process and discrete steps in the processing of DNA intermediates at recombination sites. Precisely how MutSγ and MutLγ complexes collaborate with one another and with SC-associated proteins to process recombination intermediates into interhomolog crossover products is not well understood.
On the other hand, MutSγ-MutLγ-independent crossovers can be detected in many organisms, including budding yeast [26]. Such so-called “class II” crossovers, in budding yeast, are genetically unlinked to SC protein activity and resolution of recombination intermediates associated with this class rely on the Mus81-Mms4, Slx1-Slx4, and/or Yen1 structure-selective endonuclease complexes [23,24,26,28,43]. While these observations suggest a conserved and perhaps functional relationship between the SC and MutSγ-MutLγ, it should be noted that SC-associated crossovers, MutSγ and MutLγ might not be strictly linked in all organisms. C. elegans, for example, relies on SC proteins and Msh4/Msh5 (MutSγ) for processing recombination intermediates toward an interhomolog crossover fate [44–47] but apparently employs predominantly MUS-81 and XPF-1 endonuclease complexes (presumably instead of MutLγ) to resolve such intermediates [48–50]. Drosophila also relies on SC proteins for crossover formation [51] but does not appear to have msh4 nor msh5 homologs, and instead uses a meiosis-specific version of mini-chromosome maintenance proteins to perform at least some of roles of MutSγ in processing recombination intermediates [52]. For any of these scenarios, how SC proteins and/or the SC structure might interface with DNA repair enzymes to facilitate a crossover event is poorly understood.
At least in budding yeast, several SC-associated proteins appear to facilitate early steps in MutSγ-MutL-associated crossover formation, prior to the elaboration of full-length SC. Mutant meiotic cells missing either ZIP2 or ZIP3 exhibit the same deficit as msh5 mutants in the accumulation of single end invasion (SEI) and dHJ intermediates, early crossover recombination intermediates that occur largely prior to full-length SC formation [27,35,53]. Although at one recombination hotspot, zip1 mutants showed a distinctly weaker defect in the accumulation of SEI and dHJ intermediates relative to zip2, zip3 and msh5 mutants [27], the altered kinetics of SEI and dHJ formation observed in zip1 mutants has been used to argue that even the SC transverse filament protein Zip1 acts early, prior to its elaboration along chromosomes, to facilitate the formation of qualitatively normal dHJ structures [27,54,55]. Consistent with the idea that SC proteins are involved in recombination independent of their role in elaborating an SC structure along the chromosome, SIC proteins localize to chromosomal sites that are correlated with class I crossover-designated recombination events, even in the absence of full-length SC [14,15,56]. How SC proteins functionally interface with the processes mediated by MutSγ, MutLγ and/or other recombination proteins during crossover formation is unknown: Are SC proteins, particularly SC structural proteins like Zip1, merely forming a scaffold upon which recombination enzymes dock, or do these proteins have a more specialized role in the processing of recombination intermediates?
Furthermore, is there any role for the fully assembled SC structure per se in MutSγ-MutLγ-associated crossover formation? Later steps in the maturation of crossovers occur in the context of full-length SC, and MutLγ-mediated resolution occurs concomitant with SC disassembly in budding yeast; the latter two events are triggered by Cdc5 activity at a mid-late prophase transition marked by elevated Ndt80 activity [3,19]. As the relevant protein targets of Cdc5 with respect to these events remain unknown, it is unclear whether the process of SC disassembly is normally mechanistically linked to the resolution of recombination intermediates into crossovers.
As noted above, mutants lacking ZIP1 appear to have a weaker defect in accumulating JM structures (presumed to be dHJs) relative to mutants missing ZIP2, ZIP3 or MSH5 [27], but zip1 mutants nevertheless lack MutSγ-MutLγ-associated crossovers [23,25,27,30,34]. This observation is consistent with a role for the mature SC in facilitating later steps in the successful maturation of MutSγ-MutLγ interhomolog recombination events. Other studies support the possibility that SC is dispensable for generating meiotic crossovers in budding yeast. These studies describe mutant situations in which SC assembly is disrupted yet crossovers form (i.e. in the absence of normal SUMOylation [11,13] and in the absence of the meiosis-specific chromosomal axis protein Red1 [55]. However, these prior investigations did not explore whether the apparently SC-independent crossovers form through a canonical SIC protein/ MutSγ-MutLγ–dependent mechanism. Thus the question remains: Is the assembled, full-length SC structure required for MutSγ-MutLγ-dependent crossover formation in budding yeast?
Here we describe an interspecies complementation experiment that reveals intriguing features about the relationship between the full-length SC structure, the Zip1 transverse filament protein, and MutLγ-mediated crossover events in S. cerevisiae. We generated S. cerevisiae strains that express Kluyveromyces lactis ZIP1 in place of Saccharomyces cerevisiae ZIP1, and observed that spores from K. l. ZIP1-expressing budding yeast display wild-type crossover levels despite a failure in SC formation. Our in-depth analysis of this separation-of-function version of Zip1 reveals several interesting findings. In particular, our study demonstrates that the full-length SC structure is dispensable for Zip-protein mediated and Mlh3-dependent crossover formation in budding yeast. Our data strongly suggest that crossover recombination activity, independent of SC elaboration, is sufficient to overcome a checkpoint-induced block to meiotic progression. Furthermore, we describe the surprising result that K. lactis Zip1 promotes crossovers in S. cerevisiae cells that are SC protein- and Mlh3-dependent, but largely independent of the MutSγ proteins Msh4 and Msh5. MutLγ activities are thus uncoupled from MutSγ in the context of K. l. ZIP1, suggesting that at least one aspect of S. c. Zip1 normally mediates a constraint that couples MutLγ-dependent resolvase activity to MutSγ-associated crossover intermediates. We discuss the idea that SC assembly itself could be involved in establishing such a constraint.
Kluyveromyces lactis ZIP1 encodes a protein that shares 25% identity and 16.7% homology with S. cerevisiae’s Zip1, and the two versions of Zip1 are predicted to share overall structural characteristics including an extended central coiled-coil domain flanked by non-coiled coil segments (Fig 1A). To determine whether K.l. Zip1 can rescue the meiotic functions of S. c. Zip1, we created an S. cerevisiae strain (CO9) in which the S. c. ZIP1 ORF is replaced by the K. l. ZIP1 ORF (Fig 1B).
The success of homologous chromosome segregation at meiosis I in budding yeast correlates with the viability of the haploid spore products formed. Accordingly, when we assessed spore viability among S. cerevisiae strains, greater than 90% of spores from meiotic cells carrying S. c. ZIP1 (YAM1252) were viable while only 56% of spores were viable from diploids missing ZIP1 (and therefore missing class I crossovers; Table 1). We found that 77% of spores from diploids expressing K. l. ZIP1 were viable. Thus, K. l. Zip1 is able to promote successful meiotic chromosome segregation to some extent, even in an S. cerevisiae cell context.
S. cerevisiae cells expressing K. l. ZIP1 as the sole source of Zip1 also display an intermediate sporulation efficiency. About half of sporulating diploid cells from wild-type S. cerevisiae of the BR1919-8B background progress to form spores (44% in the experiment shown in Table 1). Due to a Pch2-mediated checkpoint [57], only ~5% of sporulating diploids from zip1 null S. cerevisiae strains form spores (Table 1). We found that K.l. ZIP1-expressing cells exhibit ~16% sporulation efficiency. PCH2 removal from K. l. ZIP1-expressing S. cerevisiae cells resulted in a nearly wild-type (45%, n = 3002) sporulation efficiency, indicating that the Pch2-mediated prophase checkpoint is responsible for the diminished spore formation by K. l. ZIP1-expressing S. cerevisiae cells.
We investigated the localization of K. l. Zip1 on S. cerevisiae meiotic chromosomes using antisera raised against K. l. Zip1 (kindly provided by Abby Dernburg) as well as antisera raised against S. c. Zip1 [10,58]. Both sets of antisera gave similar results, but because anti—S. c. Zip1 antisera gave a more robust and consistent signal, this latter antibody was used for the analyses presented here.
To assess the distribution of S. c. or K. l. Zip1 on meiotic prophase chromosomes, meiotic nuclei from S. c. cells expressing either S. c. ZIP1 (control) or K. l. ZIP1 were harvested at two-hour time points between 12 and 24 hours after transfer to sporulation medium, and surface-spread on glass slides for examination by immunofluorescence. Each strain expressed a single copy of ECM11-MYC; Ecm11 localizes uniformly along the length of the budding yeast SC central element substructure, and a fraction of Ecm11 protein in the SC is SUMOylated [12,13]. Strains were additionally missing NDT80 activity, which is required for meiotic nuclei to progress beyond the pachytene stage of meiotic prophase [59]. Chromosome spreads from control and experimental nuclei were stained with anti-Zip1, anti-SUMO and anti-MYC antisera in order to assess whether SC structure is properly established.
At the 22 and 24 hour time points, a majority (> 80%) of chromosome spreads from ndt80Δ mutants appeared to be at the pachytene stage where homologous chromosomes are aligned and exhibit nearly full synapsis (Figs 2 and 3A). The DAPI-stained DNA morphology of surface-spread pachytene chromosomes from wild-type S. cerevisiae reveals distinct, individualized chromosome pairs with Zip1, Ecm11-MYC, and SUMO coinciding as linear structures at the interface of each chromosome pair (Figs 2 and 3A, [11–13]). In contrast, while the DAPI-stained DNA morphology of our K.l. ZIP1 meiotic time course nuclei suggested normal progression into the pachytene stage, none of the meiotic chromosome spreads from cells expressing K. l. ZIP1 at any of the seven time points (n = 560) exhibited full-length linear structures of Zip1, SUMO, or Ecm11-MYC. Across time points, the vast majority of nuclei displayed either no detectable K. l. Zip1 or a handful of K. l. Zip1 foci dispersed along meiotic chromosomes (Figs 2 and 3), accompanied by a punctate distribution of SUMO and Ecm11 on chromosomes. The number of K. l. Zip1 chromosome-associated foci exhibited by these nuclei ranged from 1–36, with an average of 11 K. l. Zip1 foci per nucleus. The most prominent K.l. Zip1 structure found associated with S. cerevisiae meiotic nuclei was an aggregate of K. l. Zip1, Ecm11-MYC and SUMO proteins (examples in Figs 2 and 3A and S1). K. l. Zip1 polycomplexes also contained the SIC protein, Zip3 (Fig 4). Such “polycomplex” aggregates of synapsis proteins are a characteristic feature of meiotic nuclei in S. cerevisiae mutants that fail to assemble SC [12,15,18,20]. K. l. Zip1 polycomplexes were exhibited by over half (358/560) of the surface-spread nuclei, and were observed at both early and later meiotic time points, regardless of whether they displayed detectable chromosomal K. l. Zip1 foci.
An additional Zip1 staining pattern was rarely observed, in which a single or a small number of short linear Zip1 structures appear on chromosomes (examples in Figs 2 and 3). Such short linear structures may result from bona fide but aborted elaborations of an SC precursor, or could be the result of several K. l. Zip1 foci assembled side-by-side on the chromosome. Interestingly, especially in those nuclei that showed robust Zip1 foci or short linear stretches, Ecm11 and SUMO often appeared as short linear assemblies that encompass but surpass the Zip1 structures in length (Figs 2 and 3). Short linear Zip1, Ecm11 and/or SUMO assemblies were rarely found in any nuclei among all time points examined, indicating that these structures are not stable; we observed an apparently linear Zip1, Ecm11 or SUMO structure in 0/75 nuclei at 12 hours, 3/91 nuclei at 14 hours, 4/92 nuclei at 16 hours, 10/85 nuclei at 18 hours, 6/89 nuclei at 20 hours, 1/84 nuclei at 22 hours and 3/44 nuclei at 24 hours. Taken together, our data for three readouts of SC structure (Zip1, Ecm11, SUMO), across a 12-hour meiotic prophase time course, indicate that K. l. Zip1 fails to assemble mature SC in S. cerevisiae cells.
To guard against the possibility that our antibody recognizes only a subset of potentially detectable K. l. Zip1 protein, we examined the distribution of an epitope-tagged version of K. l. Zip1, which retains function. Insertion of a V5 epitope tag just after asparagine at position 647 in the K. l. Zip1 protein failed to rescue the spore viability defect of zip1 null S. cerevisiae cells, despite the fact that YFP, inserted at the equivalent position (amino acid 700) of S. c. Zip1, creates a functional S. c. Zip1-YFP protein [60]. However, insertion of the V5 epitope tag just after arginine at position 472 generated a K. l. Zip1 protein that rescues the spore viability defect of S. c. zip1 null diploids to the same extent as untagged K. l. Zip1: Diploids carrying the K. l. ZIP1-V5 cassette in place of the S. c. ZIP1 ORF exhibited 20.6% sporulation efficiency and the spore products exhibited 79.9% viability (307 viable out of 384 spores dissected). Immunolocalization of the V5 tag in S. cerevisiae meiotic nuclei expressing K. l. ZIP1-V5 in conjunction with ECM11-MYC revealed a distribution of K. l. Zip1 on S. cerevisiae meiotic chromosomes indistinguishable from that observed using anti-Zip1 antisera (S1 Fig). Importantly, linear V5 structures were never observed among the meiotic pachytene nuclei we screened. Instead, V5 staining most often appeared as a small polycomplex structure, which typically contained Ecm11-MYC (S1 Fig). Occasionally, meiotic chromosomes displayed a limited number of faint K. l. Zip1-V5 foci on chromosomes; these V5 foci often, but not always, co-localized with Ecm11-MYC foci.
Further support for the conclusion that K. l. Zip1 fails to assemble SC in S. cerevisiae came from staining of the axial element protein, Red1, on surface-spread meiotic chromosomes. Red1 labels the axes of meiotic prophase chromosomes [61]; because the SC structure brings homolog axes into intimate alignment along their entire lengths, the closely apposed Red1-labeled axes of partner homologs in wild-type meiotic pachytene bivalents appear as a single linear structure along their full-lengths (Fig 5, top left). In contrast, meiotic pachytene chromosomes from zip1 null cells exhibit loosely-associated chromosome axes labeled by Red1 (Fig 5, bottom left) [10]. The Red1-labeled “loops” apparent in such synapsis-defective mutants correspond to homolog axes joined in intimate alignment only at sporadic positions along the chromosomes (these “axial associations” are presumably where a crossover event has been established) [62]. The Red1-stained chromosome axis patterns exhibited by surface-spread meiotic chromosomes from K. l. ZIP1-expressing S. cerevisiae cells appeared indistinguishable from those seen in zip1 null cells, consistent with an absence of mature SC structure (Fig 5, middle left).
As described above, proteins that appear to have a structural role in building SC (such as Ecm11 and SUMO) fail to assemble normal linear structures in S. cerevisiae cells expressing K. l. ZIP1. However, evidence that K. l. Zip1 is able to interface, at least to some extent, with components of the SC in S. cerevisiae cells was revealed by an examination of SUMOylated forms of Ecm11-MYC in wild-type, zip1 null, and K. l. ZIP1-expressing cells. Humphryes et al. reported that the SUMOylation of Ecm11-MYC during meiosis is largely dependent on Zip1 [12]. Consistent with their report, we found that levels of mono- and poly-SUMOylated forms of Ecm11-MYC were severely diminished in meiotic cell extracts from zip1 null cells, relative to wild-type meiotic cell extracts (Fig 3B). In meiotic cell extracts from S. cerevisiae cells expressing K. l. ZIP1, mono- and poly-SUMOylated Ecm11-MYC rose to near wild-type levels between the 0 and 12 hour time points, and appeared intermediate between wild-type and the zip1 null at the 18 and 24 hour time points (Fig 3B). These data demonstrate that K. l. Zip1 can support partial levels of Ecm11-MYC SUMOylation in S. cerevisiae meiotic cells. K. l. Zip1 might promote the SUMOylation of Ecm11 within complexes on chromosomes and/or within the polycomplex structure (where K. l. Zip1, Ecm11-MYC, and SUMO co-localize) [11,12].
We also examined the distribution of SIC proteins on meiotic chromosomes in K.l. ZIP1-expressing cells. SIC proteins, such as Zip2, Zip3 and Zip4, are required for SC assembly, but localize as multiple foci along the length of SCs instead of displaying a linear, Zip1-like distribution ([14–16,18] and Figs 4 and 5). We first examined Zip3-MYC and Zip1 on surface-spread meiotic chromosomes from S. cerevisiae cells expressing K. l. ZIP1 (Fig 4). Nuclei were harvested every two hours from 12 to 24 hours after transfer to sporulation medium. At each time point, the number of Zip3-MYC foci on surface-spread chromosomes from K.l. ZIP1 expressing cells ranged from ~10–40, which is diminished relative to the range of foci (50–70) observed on wild-type pachytene chromosomes (S2 Fig). A fraction of K. l. Zip1 foci in each nucleus (arrows in Fig 4) appeared to overlap or localize adjacent to a Zip3-MYC focus. Taking nuclei from all time points into account, 50% (1059/2108, n = 184 nuclei) of Zip1 foci overlapped or localized adjacent to a Zip3-MYC focus. However, the low number of Zip1 relative to Zip3-MYC foci exhibited by each nucleus prevents a rigorous assessment of whether the apparent adjacency events are significantly different from what one would observe from a random distribution of Zip3-MYC and K. l. Zip1.
Next we analyzed the co-localization of Zip3-MYC and Zip4-HA on pachytene stage meiotic chromosomes from S. cerevisiae cells expressing S. c. ZIP1 or K.l. ZIP1, or in cells missing ZIP1 altogether (Fig 5). As has been previously reported [14–16,18,58], we observed that the number of Zip3 and Zip4 foci on wild-type pachytene chromosomes ranged between 50–70, and well over 90% of Zip3 and Zip4 foci co-localize with one another (Figs 5 and S2). The number of Zip3-MYC and Zip4-HA foci observed on pachytene chromosomes from zip1 null cells was diminished, relative to wild type, ranging from 8–31 with an average of 19 +/- 0.95 Zip3-MYC and from 4–35 with an average of 16 +/- 1.00 Zip4-HA foci per nucleus (n = 39 nuclei). This observation is in contrast to a prior report stating that normal numbers of Zip3 foci are observed on meiotic chromosomes in zip1 null cells [15] but is consistent with the lower number of Zip3 foci that were observed on meiotic pachytene chromosomes from zip1 null cells in other studies [63,64]. The diminished number of SIC foci on chromosomes from zip1 null cells indicates a role for the SC or Zip1 in either the formation or persistence of SIC complexes during meiotic prophase.
As previously reported [18], the co-localization between Zip3-MYC and Zip4-HA on meiotic chromosomes from zip1 null strains is high, although in our experiments not as high as that observed on wild-type pachytene chromosomes (S2B Fig). In the 39 zip1 null pachytene chromosome spreads examined, the number of Zip3-Zip4 coincident localization events ranged from 38%-100% with an average of 70 +/- 3% (S2 Fig).
We found that the number of Zip3-MYC and Zip4-HA foci observed on surface-spread meiotic chromosomes from cells expressing K. l. Zip1 resembled the levels observed on pachytene-stage chromosomes from zip1 null cells. We counted between 5–40, with a mean of 16 +/- 0.83 Zip3-MYC foci, and between 3–35, with a mean of 15 +/- 0.90 Zip4-HA foci on meiotic pachytene stage chromosomes from K. l. ZIP1-expressing cells (n = 47). Apparent co-localization events observed between Zip3-MYC and Zip4-HA on meiotic chromosomes from S. cerevisiae expressing K. l. Zip1 ranged from 33%-100% with an average of 63 +/- 3% (S2B Fig). The percent Zip3-Zip4 co-localization values for zip1 null and for K. l. ZIP1-expressing cells are not significantly different from one another, as evaluated by an unpaired t test using Welch’s correction (two-tailed P = 0.1). Our data indicate that expression of K. l. ZIP1 is not sufficient to restore a wild-type number of cytologically-detectable SIC foci to pachytene chromosomes in S. cerevisiae meiotic cells missing S. c. ZIP1.
Zip1 has been found to associate with the centromere regions of meiotic prophase chromosomes and centromeres mark sites where many of the earliest SC assembly events occur in S. cerevisiae [63,65]. Furthermore, S. c. Zip1 promotes pairwise associations between centromeres outside of the context of the SC, during early and late meiotic prophase [65–67]. To investigate whether K. l. Zip1 plays a role at centromeres in S. cerevisiae meiotic nuclei, we monitored an epitope-tagged version of the kinetochore protein, Ctf19-MYC, which localizes to the centromere regions on meiotic prophase chromosomes [65,68].
In order to assess co-localization between K. l. Zip1 and meiotic centromeres, we harvested sporulating cells at two-hour intervals that spanned 12 to 24 hours following transfer to sporulation medium. Across all time points, surface-spread meiotic chromosomes from cells expressing K. l. ZIP1 and CTF19-MYC exhibited an average of 10 K. l. Zip1 foci and 22 Ctf19-MYC foci (n = 120 nuclei). Despite the fact that centromere foci typically far outnumbered detectable K.l. Zip1 foci, K. l. Zip1 foci appeared co-localized or adjacent to Ctf19-MYC foci only 46% of the time (539/1163 Zip1 foci) (S3 Fig). From an analysis of exclusively pachytene stage nuclei (classified based on DAPI-stained DNA morphology) we measured an average of eight K. l. Zip1 foci and 21 Ctf19-MYC foci (n = 68 nuclei); in this subgroup, K. l. Zip1 foci appeared co-localized or adjacent to Ctf19-MYC foci 59% of the time (325/555 Zip1 foci). Thus, while K. l. Zip1 and centromeres do not exhibit a strong co-localization pattern, these data do not rule out the possibility that K. l. Zip1 may have some preferential affinity for centromere sites on S. cerevisiae meiotic chromosomes.
We additionally explored the relationship between K. l. Zip1 and centromeres in S. cerevisiae meiotic cells through a functional assay. Zip1 facilitates two-by-two associations between meiotic prophase centromeres, independent of SC formation [65–67]. For example, spo11 mutant meiotic cells fail to initiate recombination and also fail to assemble SC, but centromeres nevertheless tend to associate in pairs. Thus, surface-spread meiotic prophase nuclei from spo11 strains exhibit fewer than 32, and often an average of 16, centromere groups. In contrast, surface spread nuclei from spo11 meiotic cells that are also missing ZIP1 exhibit closer to 32 centromere foci, demonstrating that Zip1 is required for the observed Spo11-independent centromere associations. Zip1-dependent centromere associations can also be observed outside of the context of SC, in haploid cells capable of entry into meiosis. In the haploid cell context, Zip1-dependent centromere associations are found in both spo11 null and SPO11 contexts (neither of which supports extensive SC formation); the mechanisms used for Zip1-dependent centromere associations in spo11 null versus SPO11 cells may involve distinct (yet overlapping) mechanisms since only the latter is dependent on the Pph3 phosphatase [65–67].
We assessed the capacity of K. l. Zip1 to facilitate centromere associations in both diploid and haploid spo11 null meiotic cells as well as in SPO11 haploid meiotic cells expressing CTF19-MYC. Haploids capable of progressing through meiotic prophase were created by targeting a MATa locus cassette to an ectopic location in the genome (the THR1 locus) in MATα haploids [69]. We compared the number of Ctf19-MYC foci observed on surface-spread meiotic chromosomes when such strains carried S. c. ZIP1, K. l. ZIP1, or the zip1 null genotype.
Consistent with the observations described in the initial report on “centromere coupling” [65], spo11 diploid meiotic cells expressing S. c. ZIP1 exhibited a variable number of Ctf19-MYC foci per nucleus, ranging from 4–27 with an average of 17 (n = 245 total nuclei over 5 experiments; S4 Fig), while spo11 haploid meiotic cells exhibited from 4–14, with an average of 9 Ctf19-MYC foci per nucleus (n = 143 total nuclei over 3 experiments). In contrast, spo11 diploid cells missing ZIP1 exhibited between 16–35 with an average of 26 Ctf19-MYC foci (n = 258 total nuclei over 5 experiments), and spo11 haploid cells missing ZIP1 exhibited between 9–22 with an average of 15 Ctf19-MYC foci (n = 168 total nuclei over 3 experiments; S4 Fig).
In spo11 diploid meiotic cells expressing K. l. ZIP1, we observed between 5–36 with an average of 20 Ctf19-MYC foci (n = 288 total nuclei over 5 experiments, S4 Fig), suggesting that K. l. Zip1 may weakly restore the centromere association function of S. c. Zip1 in the context of a diploid spo11 cell. In contrast, however, spo11 null haploid meiotic cells expressing K. l. ZIP1 displayed no capacity for centromere association: spo11 null haploid meiotic cells expressing K. l. ZIP1 exhibited an average of 15 Ctf19-MYC foci (n = 152 total nuclei over 3 experiments).
As reported in [66], SPO11 haploid meiotic cells exhibited between 6–12 with an average of 8 Ctf19-MYC foci, while SPO11 zip1 null haploid meiotic cells exhibited between 8–22 with an average of 14 Ctf19-MYC foci. We found that SPO11 K. l. ZIP1-expressing haploid meiotic cells exhibited between 7–22 with an average of 14 Ctf19-MYC foci.
Taken together, our findings suggest that while K. l. Zip1 may maintain a weak capacity to mediate SPO11-independent centromere associations in diploid spo11 meiotic cells, K. l. Zip1 fails to facilitate persistent centromere associations in a haploid meiotic cell context (with or without Spo11 activity). The basis for the difference observed between diploid and haploid cell contexts may reflect a sensitivity (on the part of centromeres) to the dosage of K. l. Zip1.
Since crossover recombination events are critical for the formation of the stable connections between homologs that ensure proper chromosome disjunction at meiosis I, it is reasonable to speculate that the basis for the diminished viability of spore products from K. l. Zip1-expressing S. cerevisiae strains lies in a failure of K. l. Zip1 to rescue S. c. Zip1’s crossover function. We therefore assessed crossover formation in four consecutive intervals on chromosome III, one interval on chromosome VIII and one interval on chromosome XI in S. cerevisiae cells expressing K. l. ZIP1 (Fig 6B and Table 2).
To our surprise, crossover recombination levels measured using genetic marker segregation analysis on spores from S. cerevisiae cells expressing K. l. ZIP1 were nearly indistinguishable from wild-type levels (Table 2). Crossovers are typically reduced by 30–60% in mutant budding yeast strains that are missing a “class I” crossover pathway protein [20,53]. Accordingly, in our experiments cells missing the MutS component, Msh4, displayed 30%-73% (depending on the interval) of the wild-type level of crossovers (Fig 6B and Table 2). On the other hand, the map distances derived from four-spore viable tetrads of K.l. ZIP1-expressing strains were found to be within 90–105% of wild-type values. Two exceptions to this general finding existed in a pair of adjacent intervals on chromosome III: one of the two intervals exhibited 147% of the wild type map distance and the adjacent interval showed 69% of the wild-type map distance (Fig 6B and Table 2). The addition of these exceptional intervals thus gives a map distance that is 108% of our control (S.c. ZIP1-expressing) meiotic cells. Overall our data indicate that, for meioses resulting in four-spore viable tetrads (~8%, n = 7714, Table 1) the crossover recombination deficit of a zip1 null [55,70] is completely rescued by expression of K. l. ZIP1.
To ask whether the rescue in crossover formation observed for K.l. ZIP1-expressing cells is specific to four-spore viable tetrads, we used random spore analysis to assess crossing over in the three-spore viable, two-spore viable, and one-spore viable tetrads that arose in the same crossover experiment described above. Like the four-spore viable tetrads, analysis of spores from three-, two- and one-spore viable K. l. ZIP1-expressing cells gave wild-type map distances (S1 Table). Furthermore, the frequency of chromosomes III displaying zero, single, double, triple and quadruple crossovers is similar between meiotic cells expressing S. c. ZIP1 and cells expressing K. l. ZIP1 (S2 Table). Thus, in meioses that are productive for spore formation, regardless of whether four-spore viable tetrads are produced, K. l. Zip1 rescues the crossover function of S. c. Zip1.
A question that our genetic data raises is why meioses in K. l. ZIP1-expressing cells with a wild-type crossover map (at least in the intervals measured) nevertheless result in reduced spore viability (Tables 1 and S3). One explanation for reduced spore viability despite wild-type crossover levels in K. l. ZIP1-expressing S. cerevisiae cells is that the K. l. Zip1 protein fails to provide a function at centromeres that normally supports proper MI segregation; such a function could provide centromere associations between the rare chromosome pairs that fail to sustain a crossover, or alternatively could ensure that crossover events do not occur within centromeric regions [67,71,72]. An additional or alternative possibility involves the distribution of crossovers on meiotic chromosomes, which normally exhibits measureable positive interference. The interfering distribution displayed by meiotic crossovers in wild type means that two crossover events are less likely to occur close to one another than expected from a random distribution of crossover events. In the case of weakened interference, some chromosomes (especially small chromosomes) will more frequently fail to establish stable chiasmata, relative to when strong interference is imposed [73]. Consistent with reduced interference, K. l. ZIP1-expressing strains exhibited a significantly elevated frequency of viable spores carrying a chromosome III with zero interhomolog crossovers among the intervals measured (P = 0.0004) (S2 Table).
We assessed interference among the crossovers detected in S. c. ZIP1 and K. l. ZIP1-expressing strains in two distinct ways. First, we measured an “interference ratio” [74,75] by comparing the map distances of an interval when an adjacent interval had, or had not, experienced crossover recombination. To do this for intervals along chromosome III, we parsed tetrads that showed no evidence of recombination in a “reference” interval (Parental Ditype (PD) tetrads) from those tetrads containing a single or double crossover in that reference interval (Tetratype (TT) and Non-Parental Ditype (NPD) tetrads). Next we compared the distributions of tetrad types and map distances for an adjacent, “test” interval between the parsed groups—those associated with a non-recombinant reference interval versus those associated with a recombinant reference interval. The “interference ratio” is derived from the ratio of two map distances associated with the same test interval: the map distance calculated from tetrads in which the adjacent reference interval is recombinant (contains NPD or TTs) divided by the map distance calculated from tetrads that are PD for the reference interval. Since the two map distance values should approximate 1 in the case that a recombination event in an adjacent reference interval has no interfering effect on the frequency of crossing over in a test interval, the interference ratio gives an estimate of the strength of interference; a ratio of less than one can signify positive interference. The significance of differences between map lengths calculated for an interval in either the case of the recombinant or the non-recombinant reference interval was determined using Stahl Online Tools (http://molbio.uoregon.edu/~fstahl/), and a chi-square test was employed to determine if the distribution of tetrad types is considered significantly different in test intervals associated with the recombinant versus the non-recombinant reference interval (S5 Table). When both 1) the P value associated with comparing the distribution of tetrad types between test intervals and 2) the difference in the calculated map lengths were found to reflect statistical significance, we associated the interference ratio with positive interference (dark arrows in Fig 6C).
By this “interference ratio” method, positive interference was observed between two sets of genetic intervals on the right arm of chromosome III in wild-type strains (Fig 6C, top row, S5 Table). In contrast to previously obtained measurements of interference for msh4Δ mutant strains [25,30], this method did not indicate a strong diminishment in interference over these intervals in msh4Δ mutant strains. However, the method identified a uniform loss in positive interference for the two intervals examined in S. cerevisiae cells expressing K. l. ZIP1 (Fig 6C, third line, S5 Table). The interference ratio values associated with K. l. ZIP1-expressing msh4Δ cells (Fig 6C, fourth line, S5 Table) appeared broadly similar to K. l. ZIP1-expressing, MSH4 cells. Thus, according to this method for estimating the strength of interference, the wild-type levels of Msh4-independent crossovers promoted by K. l. Zip1 exhibit little interference, while (unexpectedly) the Msh4-independent crossovers observed in S. c. ZIP1-expressing meiotic cells exhibit significant levels of positive interference. The reason that interference among crossovers in S. c. ZIP1 msh4Δ strains was detected by the “interference ratio” method is unknown.
Interference can also be detected by a lower-than-expected incidence of NPDs, which normally arise from a double crossover within a single interval. The observed number of NPDs is compared to the number expected in the case of a random distribution of crossovers (i.e. no interference), using the equation of Papazian (1952) [76]. Using this latter method for analyzing interference we found that, compared with MSH4 S. c. ZIP1-expressing strains, crossover interference in msh4Δ mutants is nearly ablated in all intervals assessed, while crossover interference appears reduced (although not ablated) for every interval assessed in S. cerevisiae cells expressing K. l. ZIP1 (Table 2 and Fig 6D).
In summary, both measurements of interference identified a defect in crossover patterning in K. l. ZIP1-expressing cells. The basis for why the interference defect (for both msh4Δ mutant, and K. l. ZIP1-expressing cells) appears stronger in one versus the other measurement remain unclear.
Our genetic analysis of interhomolog recombination in spores from K. l. ZIP1-expressing cells uncovered one additional deviation from wild-type: The frequency of gene conversion events in K. l. ZIP1-expressing S. cerevisiae meiotic cells that were productive in spore formation was elevated at eight out of nine loci (S4 Table). This result, in conjunction with the absence of SCs in K. l. ZIP1-expressing cells, is consistent with the idea that the SC structure prevents additional interhomolog recombination events in budding yeast, perhaps through a mechanism involving a downregulation of DSBs [77]. It is also possible that an altered gene conversion tract length for K. l. Zip1-mediated recombination events contributes to the elevated gene conversion frequency observed. We note that the 2–3 fold elevated gene conversion frequencies in K. l. ZIP1-expressing cells is not accompanied by an increase in the frequency interhomolog crossover events (over wild-type levels).
The MutSγ heterodimer Msh4/Msh5 is required for the class I crossovers mediated by S. c. Zip1 [27,30,34,53,78]. Consistent with prior reports, we observed that the loss of MSH4 in wild-type cells resulted in 30–70% reductions in crossover levels (Table 2). In contrast, our genetic analysis revealed that the bulk of the crossovers mediated by K. l. Zip1 in S. cerevisiae cells occur in a Msh4-independent manner. In K. l. ZIP1 msh4Δ strains, map distances are reduced, relative to K. l. ZIP1 MSH4 strains, by less than seven percent in every interval measured with two exceptions: a 22% reduction in the MAT-RAD18 interval in chromosome III and a 33% reduction in the iLEU2-iTHR1 interval on chromosome XI (Fig 6 and Table 2). Overall, the crossover reductions observed when Msh4 is removed from K. l. ZIP1-expressing strains are dramatically less pronounced than the crossover reductions resulting from the removal of Msh4 in S. c. ZIP1–expressing strains. These data indicate that K. l. Zip1 rescues crossover formation in S. cerevisiae cells through a mechanism that does not rely heavily on the MutSγ component, Msh4.
Perhaps not surprisingly given its dispensability in crossover formation, the abundance of Msh4 on mid-meiotic prophase chromosomes in K. l. ZIP1-expressing cells is severely diminished relative to Msh4’s abundance on meiotic chromosomes in S. c. ZIP1-expressing cells (S5 Fig). Consistent with previous reports, we observed ~40–65 Msh4-HA foci co-localized with Zip3-MYC protein on mid-meiotic prophase chromosomes from S. c. ZIP1-expressing meiotic cells (at a stage when chromosomes normally exhibit full-length SC). In contrast, only 0–20 Msh4-HA foci were observed on similarly staged meiotic chromosomes from K. l. ZIP1-expressing cells; such low levels of Msh4 on meiotic chromosomes resembled the level detected in a zip1 null mutant (S5 Fig).
In order to measure crossover recombination among all meiotic cells regardless of their capacity to successfully form spores, we turned to a physical assay for recombination on chromosome III. In this “circle-linear” assay, meiotic nuclei harboring one linear and one circular chromosome III are subjected to pulsed-field gel electrophoresis followed by a Southern blot to detect the position of chromosome III on the gel [79]. The circular chromosome III fails to enter the gel and thus is not detectable. However, the non-recombinant and recombinant forms of linear chromosome III fall into three size categories that are detectable on these gels: A single crossover between linear and circular chromosomes III runs at twice the molecular weight of the parental linear chromosome III, whereas a double crossover involving three chromatids runs at three times the molecular weight of the parental chromosome III. The proportion of trimer and dimer chromatids relative to the total (detectable) chromatids can be used to generate a relative measure of crossing over on chromosomes III in the population.
In wild-type strains, crossover recombination values estimated using this physical assay for crossovers on chromosome III were at nearly 100% by 40 and 70 hours of sporulation (Fig 7A). In zip1 null mutants, on the other hand, approximately 20% and 30% recombination was measured at 40 and 70 hours after transfer to sporulation medium, respectively. In strains expressing K.l. ZIP1, approximately 50% and 65% crossover recombination was measured at 40 and 70 hours of sporulation, respectively (Fig 7A). Because K. l. ZIP1-expressing meiocytes that go on to form spores display wild-type levels of crossing over on chromosome III, the intermediate level of crossing over measured by this physical assay indicates that K. l. ZIP1-expressing meiocytes that fail to form spores are crossover-deficient.
We used this assay to explore whether K.l. Zip1-mediated crossovers are dependent on synapsis-associated proteins and MutLγ, or on the so-called “class II” crossover pathway components (Fig 7B). In control strains expressing S.c. ZIP1, removal of MMS4 or YEN1 (which encode proteins that have been genetically linked to the “class II” crossover pathway) resulted in a modest decrease (~10%) in the percentage of recombinant chromosomes III. In contrast, the individual removal of ZIP3, ZIP4, SPO16, MSH4 or MLH3 (each encoding a protein that has been linked to a discrete “class I” pathway for meiotic crossovers) resulted in a larger (50%-70%) reduction in crossover formation on chromosome III. zip1 null strains missing MMS4, YEN1, or any of the “class I” crossover genes tested displayed similarly low levels of crossover recombination on chromosome III.
Analysis of K. l. ZIP1-expressing strains missing these crossover-associated genes revealed strong evidence that K. l. Zip1 functionally interfaces with a canonical Zip1/SC–associated crossover pathway in S. cerevisiae cells. Crossover recombination in K.l. ZIP1-expressing strains is strongly reduced (to nearly zip1 null levels) in the absence of ZIP3, ZIP4, SPO16, or the MutLγ protein-encoding gene, MLH3 (Fig 7B). On the other hand, crossover recombination on chromosome III was reduced only modestly (by ~10%) in K.l. ZIP1-expressing strains missing either MMS4 or YEN1 (Fig 7B), indicating that these DNA repair-associated factors are dispensable for the bulk of meiotic interhomolog crossovers in both wild-type and K. l. ZIP1-expressing strains.
Consistent with our genetic analysis, recombination on chromosome III was reduced only modestly (by ~10%) in K. l. ZIP1-expressing strains missing MSH4, and a similar result was obtained for K. l. ZIP1-expressing strains missing both MSH4 and MSH5 (Fig 7B). We did not find evidence that class II crossover pathway components rescue Msh4 function when it is absent from K. l. ZIP1-expressing cells, as the small reduction in crossovers on chromosome III measured in K.l. ZIP1-expressing strains missing both MSH4 and MMS4 was similar to that observed in either the msh4Δ or mms4Δ single mutant (Fig 7B).
Taken together, our data clearly indicate that K.l. Zip1, like S.c. Zip1, functionally interfaces with other synapsis-associated proteins in order to facilitate the maturation of MutLγ-associated crossovers in budding yeast, but that K. l. Zip1-mediated crossovers can largely bypass a requirement for MutSγ.
We examined the capacity for K. l. Zip1 to facilitate MutSγ-independent recombination in greater detail by asking whether K. l. Zip1 can rescue the JM deficit reported for cells missing MutSγ complex function [27]. We analyzed six strains, each carrying S.c. ZIP1, K.l. ZIP1 or a zip1 null allele, in either a MSH4 or a msh4 null background. As our strains (BR1919-8B-derived [69]) progress through meiosis in an asynchronous manner, we reasoned that we would be more likely to detect JMs if we prevent their resolution. Thus, each of our strains is also missing NDT80 activity, which is normally required to promote the molecular pathways that resolve JMs into crossovers in S. cerevisiae [3,19], and is indeed required for crossover formation in K. l. ZIP1-expressing cells (S6 Fig). Cells were harvested at 0, 24, 32 and 40 hours after being introduced into sporulation medium, then subjected to psoralen crosslinking to preserve JM structures. Crosslinked DNA was extracted, digested with HindIII, and DNA fragments were separated by two-dimensional (2D) electrophoresis. The branched nature of crosslinked JMs causes them to migrate to a position on the 2D gel which is displaced from the arc of the bulk of crosslinked genomic DNA [5,35] (see cartoon in Fig 8). The positions of all DNA fragments that correspond to the ERG1 and YCR047c loci, which are associated with DSB hotspots [35,77,80,81], were analyzed by Southern blot hybridization.
Signals representing JM structures were undetectable at the t = 0 time points in any of our strains. However, JMs were detectable at both ERG1 and YCR047c sites in all strains at 24 hours after introduction into sporulation medium (Figs 8 and S7). Quantification of the percentage of DNA that was present in the JM spot in either MSH4 or msh4 strains is shown for the ERG1 locus in Fig 8. In MSH4 strains, we observed that both K. l. ZIP1-expressing samples and zip1 null samples exhibited a diminished JM signal relative to S. c. ZIP1 samples at the 24 hour time point. However, at the 40 hour time point the JM signal in K. l. ZIP1 samples appeared closer to that of S. c. ZIP1, and elevated above the JM level exhibited by the zip1 null. These data are consistent with the crossover data we obtained with the circle-linear chromosome III assay (Fig 7) and suggest that K. l. Zip1 has an (albeit diminished) capacity to facilitate stable JM formation, and that K. l. Zip1-dependent JMs accumulate over time in an ndt80Δ mutant background.
Support for the idea that MutSγ is critical for the bulk of JM formation in S. cerevisiae meiosis comes from the observation of strongly diminished JMs at the HIS4-LEU2 artificial hotspot in msh5 mutants (using the SK1 strain background) [27]. As Mlh3-dependent crossovers form in K. l. ZIP1-expressing cells despite the absence of Msh4, we wondered whether K. l. Zip1 rescues the deficit in JM formation presumed to occur in the absence of Msh4. We asked this question by analyzing JM formation in S. c. ZIP1, K.l. ZIP1 and zip1 null strains that were also missing MSH4.
In light of the strong reduction in JMs at the HIS4-LEU2 artificial hotspot in msh5 mutants, we were surprised to observe a robust JM signal at both ERG1 and YCR047c in our S. c. ZIP1 msh4Δ ndt80Δ strains (Fig 8B, blue line, and S7 Fig). Furthermore, the JM signal at the ERG1 site appeared to increase between 24 and 40 hours of sporulation in the ndt80-arrested, msh4 mutants. We presume that the extensive period of late prophase arrest performed for our analysis facilitated the slow but steady accumulation of JMs even in the absence of Msh4. Consistent with this possibility, a prior report demonstrated that msh5Δ ndt80Δ mutants of the SK1 background exhibit JM accumulation over time, ultimately achieving ~1/3 of the peak wild-type JM level by a late (8 hour) time point [41].
Our analysis thus reveals the existence of Msh4-independent JMs that accumulate in an ndt80Δ, meiotic prophase-arrested cell population at ERG1 and YCR047c sites in the BR1919 strain. Interestingly, we observed that a substantial fraction of the Msh4-independent JM signal at ERG1 and YCR047c sites in S. cerevisiae is dependent on Zip1. In zip1 msh4 double mutants (Fig 8B, red line), JMs do not accumulate to the same high levels as seen in the ZIP1 msh4 strain. Thus, our data indicate that S. c. Zip1 can promote Msh4-independent JM formation.
Since msh4 mutants are missing the same set of crossovers as zip1 mutants [30], the bulk of the Msh4-independent JMs promoted by S. c. Zip1 (the set of JMs present in S. c. ZIP1-expressing cells but not present in zip1 null cells) are not likely resolved to form interhomolog crossovers. This observation raises the possibility that the crossover defects of S. c. ZIP1 msh4 mutants may not be solely the result of a deficit in JM formation per se, but rather could be in part the result of a function for Msh4 in channeling SIC protein-associated recombination intermediates into an interhomolog JM pathway that is resolved by Mlh1/Mlh3.
Finally, our analysis revealed that, like S. c. Zip1, K. l. Zip1 promotes Msh4-independent JM formation in S. cerevisiae cells, albeit with a reduced capacity relative to S. c. Zip1 (Fig 8B, green line, and S7 Fig). While the Msh4-independent JMs in S. c. ZIP1-expressing cells presumably do not resolve to give interhomolog crossovers, in light of our genetic and physical crossover data (Figs 6 and 7) we propose that a substantial fraction of the Msh4-independent JMs in K. l. ZIP1-expressing cells are successfully resolved into interhomolog crossovers via an Mlh3-dependent mechanism.
Although experiments aimed at understanding the molecular nature of Msh4-independent JMs are outside the scope of the current work and will be the subject of a future study, we note that the shape of the JM signal in msh4 mutants appears elongated relative to that of the JMs observed in MSH4 strains (illustrations in Fig 8A and 8B). The elongated shape of the observed signal for Msh4-independent JMs in our strains suggest the presence of JM species with a similar molecular mass but different branched pattern, possibly the result of an altered dHJ structure or perhaps from junction migration in the msh4 mutants. An alteration in the structure of dHJs in the absence of Msh4 is consistent with the finding that hMSH4-hMSH5 recognizes Holliday Junctions and can potentially form a clamp which “embraces” partner DNA molecules of homologous chromosomes [36,82].
Here we report on the capacity of the Kluyveromyces lactis Zip1 protein to carry out S. cerevisiae Zip1 functions in the S. cerevisiae meiotic cell context. Kluyveromyces lactis and S. cerevisiae last shared a common ancestor well over 100 million years ago, prior to the fungal lineage’s whole genome duplication event [83]. The K. lactis genome encodes apparent homologs of most if not all synapsis-related proteins that have been thus far characterized in S. cerevisiae (including SUMO, Hop1, Red1, Ecm11, Gmc2, Zip2, Zip3, Zip4, Spo16, and Pch2), as well as the Msh4, Msh5, Mlh1 and Mlh3 proteins (http://www.genome.jp/kegg-bin/show_organism?org=kla). Whether K. lactis meiotic cells assemble an SC is unknown.
K. l. Zip1 exhibits ~ 40% overall homology with S. c. Zip1 at the primary amino acid level, and K. l. Zip1 and S. c. Zip1 share predicted structural characteristics. In particular, both K. l. Zip1 and S. c. Zip1 have a ~550 residue, centrally located group of amino acids that have a high probability of forming coiled-coil. The N- and C- terminal, non-coiled-coil regions of S. c. Zip1 are ~30–40% larger than the corresponding regions of K. l. Zip1. Several ~5–20 residue blocks of conserved sequence identity exist between the two ancestrally related proteins (Fig 1).
K. l. Zip1 fails to assemble mature SC structures in S. cerevisiae cells, as indicated by the absence of full-length linear Zip1, Ecm11 or SUMO assemblies on meiotic prophase chromosomes at any time point during meiotic prophase, and by the asynapsis phenotype of Red1-labeled chromosome axes (Figs 2, 3 and 5). Apart from a distinct polycomplex aggregate, little K. l. Zip1 was detectable on S. cerevisiae meiotic prophase chromosomes in our experiments, including those that assessed the distribution of an epitope-tagged version of K. l. Zip1 (Figs 2–5 and S1). Moreover, levels of the SC- and/or crossover-associated Zip3, Zip4, and Msh4 proteins on S. cerevisiae prophase chromosomes appeared similar to the levels of these proteins in zip1 null cells (Figs 4 and S2 and S5).
However, evidence that K. l. Zip1 can interface, at least to some extent, with S. cerevisiae SC-associated proteins stems from the observation that K. l. Zip1 polycomplex structures are decorated by S. c. Ecm11, SUMO and Zip3 proteins (Figs 2–5 and S1). Furthermore, K. l. Zip1 promotes SUMOylation of the S. cerevisiae Ecm11 protein (Fig 3), an activity that normally also largely relies on the function of synapsis proteins Zip2 and Zip4 [12]. Finally, the interhomolog crossover events that are promoted by K. l. Zip1 in S. cerevisiae cells are dependent on other so-called SIC proteins, namely Zip3, Zip4 and Spo16 (Fig 7). These observations suggest that at least some molecular features of S. c. Zip1 responsible for interfacing with SC-associated proteins are preserved in the K. l. Zip1 protein. Such molecular features could be represented by the short segments of identical sequence shared by the two Zip1 proteins, and/or may be based in a shared secondary structure.
The sparse distribution of detectable K. l. Zip1 and the reduced number of Zip3 and Zip4 proteins observed on meiotic prophase chromosomes in S. cerevisiae expressing K. l. ZIP1 suggests that SC precursor structures and/or their associations with chromosomes are unstable in this context. Because zip1 loss-of-function mutants and other S. cerevisiae mutant meiotic cells with such a dramatic asynapsis phenotype typically also exhibit a deficit in crossovers [16,18,27,70], we were surprised to measure wild-type levels of crossover recombination in spores derived from K. l. ZIP1-expressing S. cerevisiae meiotic cells (Fig 6). A combination of genetic and physical assays to measure crossing over revealed that K. l. Zip1-mediated crossovers are dependent on the SC-associated proteins Zip3, Zip4 and Spo16, and are dependent on the MutLγ protein Mlh3, but are relatively unaffected by the loss of Mms4 and Yen1, which is as expected for Zip1-mediated (SC-associated) crossover events (Fig 7). Furthermore, the resolution of K. l. Zip1-mediated repair intermediates into crossovers is, like most if not all meiotic crossovers in S. cerevisiae, dependent on the Ndt80 transcription factor (S6 Fig). Our data strengthen the notion that at least one pro-crossover function of Zip1 is separate from its role in assembling SC, a possibility previously raised by an analysis of the red1 mutant in the presence and absence of Zip1 and to a certain extent by analysis of the recombination phenotype of zip1 mutants [27,55]. K. l. Zip1’s behavior in S. cerevisiae cells demonstrates that these independent activities of Zip1 can be uncoupled at the protein level.
If K. l. ZIP1-expressing cells rely on other SC-associated proteins to promote crossing over, why is the level of Zip3 and Zip4 on meiotic chromosomes in K. l. ZIP1-expressing cells at the low level seen in the zip1 null? One possibility is that nascent SC-initiation structures are dynamic in the absence of elaborated SC, and thus only a subset of the so-called SIC complexes are detectable on meiotic chromosomes at a given time in the zip1 null or the K. l. ZIP1 context. The discrepancy between the low observed level of K. l. Zip1 protein on S. cerevisiae meiotic chromosomes and the high level of crossovers observed in at least a subset of S. cerevisiae meiotic cells expressing K. l. ZIP1 raises the important point that the abundance and spatial distribution of a protein that is minimally sufficient to provide crossover function may not necessarily be detectable by immunostaining.
A discrepancy exists between our genetic and physical analyses of crossing over in S. cerevisiae cells expressing K. l. ZIP1. When measured genetically in spores, K.l. ZIP1-expressing strains exhibit wild-type map distances within intervals across chromosome III, and within intervals on two additional chromosomes (VIII and XI; Fig 6 and Table 2). On the other hand, by our physical assay we observed an intermediate crossover level across chromosome III in strains expressing K.l. ZIP1, relative to the levels exhibited by S.c. ZIP1 and zip1 null strains (Fig 7). Similarly, K.l. ZIP1 msh4Δ double mutants exhibit significantly higher crossover levels relative to S.c. ZIP1 msh4Δ strains when measured genetically, but crossover levels across chromosome III are at comparable levels in K.l. ZIP1 msh4Δ and S.c. ZIP1 msh4Δ strains by our physical assessment.
The discrepancy between crossover levels measured genetically versus a physical assay is likely due to a Pch2-mediated, prophase surveillance system that blocks spore formation in the majority of K. l. ZIP1-expressing meiotic cells. The triggers that activate a Pch2-mediated checkpoint have been associated with defects in both synapsis and in DSB repair, and can be modulated by environmental factors in budding yeast [27,57,84–87]. Our data suggest that this meiotic prophase checkpoint activity is more robust in K.l. ZIP1 msh4Δ than in S. c. ZIP1 msh4Δ cells as the sporulation efficiency of K.l. ZIP1 msh4 strains is lower than the sporulation efficiency of S.c. ZIP1 msh4 strains (16.6% for K.l. ZIP1 msh4 versus 30.0% for S.c. ZIP1 msh4; n > 1000).
Overcoming a block to meiotic progression could occur either by removing or by bypassing the insult that triggered the checkpoint. The phenotype observed in K.l. ZIP1-expressing, S. cerevisiae strains, where only those meiocytes with a nearly wild-type interhomolog crossover level progress to form spores, appears to underscore the strong influence that SC protein-associated, MutLγ-mediated recombination can have on overcoming the Pch2-associated checkpoint (regardless of how the checkpoint is triggered in these cells). Our data indicate that a capacity to overcome the prophase checkpoint in K. l. ZIP1-expressing cells tightly correlates with crossover recombination outcomes: Wild-type crossover levels are measured for K. l. ZIP1-expressing meiotic nuclei that succeed in forming spores, but crossover recombination is lower (intermediate between the levels exhibited by zip1 null and wild-type) when examined by a physical assay, an analysis that includes meiotic cells that are blocked from progressing to form spores. Since mature SC is absent in S. cerevisiae cells expressing K. l. ZIP1, these data suggest that crossover levels alone, independent of the SC, may be sufficient to overcome the Pch2-mediated prophase checkpoint block to spore formation.
On the other hand, the S. c. ZIP1 msh4 mutant phenotype is difficult to explain with the simple model that crossover levels overcome the prophase checkpoint to allow spore formation, since msh4 meiotic cells with strongly diminished crossovers can nevertheless successfully form spores. Perhaps an absence of MutLγ-associated recombination intermediates in S. c. ZIP1 msh4 cells results in a less stringent checkpoint activity than that observed in K. l. ZIP1-expressing cells. Alternatively, perhaps the SC structure is capable of modulating the prophase checkpoint [21,88]. Previous reports indicate that SC formation is delayed but ultimately occurs to some extent in msh4 mutants [17,25,30]; thus the increased capacity of cells deficient in MutSγ-associated crossovers to complete spore formation could be a consequence of signals from the SC structure itself that overcome the checkpoint.
While prior studies indicated that Zip1 might play a role in recombination separate from its role in SC assembly, the question remained whether the SC structure has a mechanistic role in the formation of a set of crossovers that are normally associated with synapsis. The presence of K. l. Zip1 as the sole source of Zip1 in S. cerevisiae cells fails to support SC assembly but promotes the formation of a set of crossovers that are Mlh3-dependent. Thus, this unique separation-of-function version of Zip1 demonstrates that an elaborated SC structure is not required per se for the formation of Mlh3-dependent crossovers in S. cerevisiae.
By what mechanism is Zip1 (including K. l. Zip1) involved in MutLγ-associated crossover recombination? The notion that Zip1 acts early in the pathway leading to stable crossover recombination intermediates could account for Zip1’s entire role in promoting MutLγ-dependent events, via a function in shaping or stabilizing proper JM structures that are recognizable and/or accessible to MutLγ and its companion resolvase-promoting factors. On the other hand, current data does not rule out the idea that Zip1 protein acts at later stages in the JM maturation process to facilitate the targeting of MutLγ proteins to MutSγ-associated crossover intermediates. Our analysis of JM formation in S. c. ZIP1, K. l. ZIP1 and zip1 null cells (Figs 8 and S7) supports either model: K.l. Zip1 appears to promote some stable JM formation above the level seen in the zip1 null, consistent with the idea that K. l. Zip1 might act early to promote the formation of a stable JM structure. Our observation of S. c. or K. l. Zip1-dependent JMs in msh4 mutants also supports a role for Zip1 in establishing a stable JM structure. On the other hand, the fact that (in the S. c. ZIP1 context) S. c. Zip1-dependent JMs form in the absence of Msh4 but do not resolve properly highlights the possibility that an S. c. Zip1-mediated constraint linking MutLγ resolvase activity to MutSγ-associated recombination intermediates may act downstream of or in parallel to stable JM formation, (see below).
Importantly, while spores from K. l. ZIP1-expressing cells exhibit nearly wild-type crossover levels over multiple genetic intervals, this result is influenced heavily by a stringent meiotic checkpoint; it is certainly not the case that all meiotic cells enjoy wild-type crossover levels in K. l. ZIP1-expressing S. cerevisiae strains (as indicated by our physical assays of crossover recombination on chromosome III). We imagine that the capacity of K. l. Zip1 to promote crossover recombination in S. cerevisiae cells is diminished relative to S. c. Zip1 because of suboptimal protein function and/or diminished protein levels.
Our physical and genetic crossover analyses indicate that a small subset of S. cerevisiae meiotic cells exhibit nearly wild-type levels of crossing over on chromosome III. If K. l. Zip1 only partially rescues S. c. Zip1’s crossover function, why do some cells experience a wild-type level of crossing over in the context of K. l. Zip1? Thacker et al. (2014) reported that zip1 mutants fail to properly down-regulate DSBs at later meiotic prophase stages [77]. With the idea in mind that the presence of SC could participate in down-regulating recombination-based interhomolog interactions, we propose that ongoing DSB-initiated interhomolog interactions allowed because of the absence of SC in K. l. ZIP1-expressing cells may be critical for the gradual establishment of a class of K. l. ZIP1-expressing cells that achieve wild-type interhomolog crossover levels.
Data presented in this study indicate that K. l. Zip1 retains a robust pro-crossover activity that functionally interfaces with several canonical Zip1 crossover pathway factors, including the so-called SIC proteins Zip3, Zip4, Spo16, and the MutLγ protein Mlh3 in S. cerevisiae cells. However, K. l. Zip1 crossovers in the context of the S. cerevisiae cell are different from S. c. Zip1-associated crossovers in two ways: First, K. l. Zip1 crossovers are unassociated with SC formation. Second, in the context of K. l. Zip1, MutLγ-mediated crossovers bypass a requirement for the MutSγ complex.
In K. l. ZIP1-expressing cells, Mlh3 promotes the resolution of recombination intermediates into crossovers even in the absence of MutSγ complex proteins Msh4/Msh5. While evidence from both Tetrahymena and C. elegans indicates that eukaryotic versions of bacterial MutS may not always be functionally linked to MutL proteins [48–50], K. l. ZIP1-expressing S. cerevisiae meiotic cells reflect the first example, to the authors’ knowledge, of MutLγ-dependent crossover formation that does not rely on MutSγ. The result indicates that MutLγ is not intrinsically constrained to act on MutSγ-associated DNA structures in S. cerevisiae nuclei, but that a constraint is normally active in the context of S. cerevisiae ZIP1 that normally couples MutLγ to MutSγ-associated recombination intermediates. Our study furthermore demonstrates that K. l. Zip1 can bypass this constraint. Understanding how K. l. Zip1 bypasses the requirement for Msh4/Msh5 in generating MutLγ-associated crossovers will provide a useful framework for understanding the molecular mechanism normally used by budding yeast to couple MutLγ-associated resolvase activity to MutSγ-associated intermediates.
It is noteworthy that Zip1 appears to be central to the mechanism that normally links MutLγ-associated resolvase activity to MutSγ-associated intermediates in budding yeast. As raised in the Introduction, one explanation for the role of SC structural proteins such as Zip1 in meiotic interhomolog recombination is that SC proteins or the SC itself act as a recruitment platform upon which specialized recombination enzymes can dock. While the data presented here do not rule out this model, the fact that an alternate version of Zip1 can bypass the requirement for MutSγ in MutLγ-mediated crossover formation raises the possibility that Zip1 may play a more specialized role in the processing of joint molecule intermediates.
Does K. l. Zip1 promote Msh4-independent crossovers by replacing Msh4/Msh5’s function in JM formation? The Msh4/Msh5 heterodimer is thought to form a sliding clamp on DNA and thus could recognize and stabilize both SEI and dHJ structures [27,36] in order to protect them from disassembly by helicases [40] and/or to facilitate their resolution by MutLγ-Exo1 [23,33,38,40–42]. Interestingly, the pro-crossover function(s) of Msh4/Msh5 in stabilizing meiotic crossover recombination intermediates are replaced by novel minichromosome maintenance protein complex in Drosophila [52]. Furthermore, proteins that promote SC formation have been implicated in antagonizing the anti-crossover activity of the Sgs1 helicase, consistent with an the idea that these proteins may share functionality with Msh4-5 in protecting JM structures from dissolution by helicases [40]. Prior observations of the DNA intermediates that accumulate at the HIS4-LEU2 artificial DSB hotspot in msh5 mutant strains (of the SK1 background) are consistent with the idea that Msh4/Msh5 activity is required for the accumulation of the bulk of stable dHJ recombination intermediates [27,41], although Msh5-independent JMs were found to accumulate over time in an ndt80Δ mutant background [41]. If K. l. Zip1 can bypass the need for Msh4 through rescuing a function of Msh4 in promoting stable JMs, we expected to observe a larger abundance of JMs in K. l. ZIP1 msh4 mutants, relative to S. c. ZIP1 msh4. Surprisingly, our analysis of JM formation at two natural hotspots (ERG1 and YCR047c) in MSH4 and msh4 mutant strains of the BR1919 background indicate that S. c. Zip1 and K. l. Zip1 (to a lesser extent) both promote the formation of a population of Msh4-independent JMs, although the elongated shape of the observed signal suggests the possibility that JMs that form in the absence of Msh4 in our strains have altered structure.
While these data reveal the interesting result that Msh4 and Zip1 may indeed share overlapping roles upstream of JM formation, JM formation activity per se is not likely to be the reason that Msh4 is dispensable for MutLγ-dependent crossovers in K. l. ZIP1-expressing strains, since both S. c. Zip1 and K. l. Zip1 promote Msh4-independent JM formation. We presume that only in the context of K. l. ZIP1 can such Msh4-independent JMs resolve via MutLγ.
Perhaps the critical crossover function of MutSγ in S. cerevisiae, instead of JM formation per se, is in ensuring that MutLγ-associated resolvase activity is successfully targeted to SC-associated JMs. When K. l. Zip1 is present, MutLγ is targeted to SIC protein-dependent crossover intermediates independently of MutSγ. Evidence that mammalian MutSγ and MutLγ components can directly interact [89] raises the possibility that Msh4/Msh5 might directly recruit Mlh1/Mlh3 complexes to JM structures. If the major mechanism for targeting MutLγ to MutSγ-associated JMs involves a direct protein-protein interaction between MutSγ and MutLγ components, one might propose that K. l. Zip1 bypasses the normal requirement for Msh4/Msh5 via a capacity to directly interact with Mlh1 or Mlh3 in S. cerevisiae cells. On the other hand, S. cerevisiae MutLγ complex can recognize and bind preferentially to JM structures in vitro [38]. Thus, perhaps the critical role of Msh4/Msh5 in coupling SC-associated recombination intermediates with MutLγ-associated resolvase activity is not through its potential capacity to interact directly with Mlh1/Mlh3, but through a capacity to promote the formation of a JM structure that is recognizable and/or accessible to the S. cerevisiae MutLγ complex. In this case, a simple explanation for the bypass of MutSγ provided by K. l. Zip1 is that K. l. Zip1 activity is functionally redundant with Msh4/Msh5 in S. cerevisiae meiotic nuclei (this idea is reminiscent of the functional redundancy with MutSγ proposed for the minichromosome maintenance protein complex in Drosophila [52]) and can facilitate the processing of JM intermediates in a manner that allows their resolution by a MutLγ-mediated mechanism.
On the other hand, Fig 9 presents an alternative model to explain both the mechanism that normally constrains MutLγ activity to target MutSγ-associated recombination intermediates in S. cerevisiae and how K. l. Zip1 bypasses this constraint. In our alternative model, we propose that S. c. Zip1 is normally associated with both pro-crossover and anti-crossover activities, and that Msh4/Msh5 counters the anti-crossover activity of Zip1 at JMs. A possible anti-crossover aspect of Zip1 activity could be an action that destabilizes dHJ structures, or one that prevents the accessibility of dHJ structures to MutLγ-associated resolvase activity. The presence of MutSγ at JMs might protect them or directly counter Zip1’s anti-crossover activity. In the context of S. cerevisiae cells expressing K. l. ZIP1, K. l. Zip1 retains S. c. Zip1’s pro-crossover activity but lacks its anti-crossover activity, thus rendering MutSγ dispensable for the MutLγ-mediated resolution of Zip1/SC protein-associated recombination intermediates.
The proposed antagonistic relationship between Msh4/Msh5 and S. c. Zip1 that this model proposes would effectively constrain MutLγ-mediated resolvase activity to act exclusively on MutSγ-associated recombination intermediates: Although S. c. Zip1 may be able to promote the formation of JM structures in the absence of MutSγ, only MutSγ-associated crossover intermediates are processed during meiosis in a manner that protects them from S. c. Zip1’s anti-crossover activity and allows their resolution by a MutLγ-mediated mechanism.
It is tantalizing to suggest that the putative “crossover constraining” activity of S. c. Zip1 proposed by this model is the process of SC assembly or the assembled SC itself. Perhaps the Msh4/Msh5 complex is required to protect the integrity of SEI and dHJ recombination intermediates during the process of SC elaboration, or to maintain the accessibility of such intermediates to resolvases in the context of full length SC later on. Under this scenario, the absence of SC in K. l. ZIP1-expressing cells renders MutSγ complexes dispensable for MutLγ-dependent crossover formation.
In budding yeast, Zip1, Msh4/Msh5, and Mlh1/Mlh3 are associated with the successful resolution of crossover recombination intermediates that exhibit interference [25,27,30,34,70,90]. However, SC proteins likely play no role in the initial establishment of an interfering distribution pattern of SC-associated (MutSγ-MutLγ-mediated) crossover recombination events. Such a conclusion is supported by the fact that early crossover-correlated recombination intermediates form at a stage of meiotic prophase that is prior to when full-length SCs are present [27]. Furthermore, Zip2 and Zip3 chromosomal foci, which co-localize at SIC structures and are presumably cytological manifestations of SC-associated crossovers, exhibit an interfering distribution on meiotic pachytene chromosomes even when Zip1 is absent [14,91]. However, while the initial establishment of interfering crossover events may not require SC-associated proteins, the fact that Mlh3-dependent crossovers exhibit diminished interference in K. l. ZIP1-expressing S. cerevisiae cells indicates that SC protein activity may well influence not only the resolution of, but the ultimate pattern of MutLγ-associated crossover events.
Indeed, taken to the extreme, a model postulating that SC proteins and/or SC structures play no role in crossover interference predicts that if MutSγ-MutLγ-dependent crossovers could successfully mature in the absence of SC, these crossovers would exhibit normal interference. Our data, however, show that K. l. Zip1 can promote wild-type levels of Mlh3-dependent crossovers in a subset of S. cerevisiae meiotic cells, but these crossovers exhibit a substantially weakened interference pattern (Table 2 and Fig 6). As the cytological manifestation of MutLγ-associated crossovers (Zip2 or Zip3 foci) exhibit an interfering distribution pattern even in strains missing Zip1 altogether [14,91], we assume that K. l. Zip1-mediated crossovers are designated with proper interference in K.l. ZIP1–expressing S. cerevisiae meiotic cells. If the designation of interfering crossovers is intact in strains expressing K. l. ZIP1, then the preservation of an interfering distribution pattern of crossover-designated recombination events apparently requires a Zip1 activity that is separable from its crossover promoting function per se.
The influence of Zip1 on crossover patterning could be explained if an interfering pattern of designated crossover events is different depending on the stage in meiotic prophase when those events are initiated. Perhaps the presence of S. c. Zip1 (and/or the SC) preserves the integrity of a discrete set of “earliest designated” crossover intermediates, which exhibit a robust interference pattern. When K. l. Zip1 is present, perhaps fewer of these earliest-designated intermediates undergo successful maturation into stable JM intermediates. Since the SC structure may possibly be a barrier to the formation of ongoing interhomolog recombination-based interactions [77], it is reasonable to speculate that SC protein-mediated interhomolog recombination events may initiate in an ongoing manner and occur at later meiotic prophase stages in a K. l. ZIP1 (synapsis-defective) context, relative to normal S. cerevisiae meiosis. Thus, that subset of K. l. ZIP1-expressing meiotic cells carrying a wild-type crossover level may ultimately exhibit a crossover landscape that includes both early- and late-designated crossover intermediates. If the distribution of later prophase-designated recombination intermediates is less subject to interference, the result would be an overall weakening of the interfering distribution pattern of Mlh3-resolved crossovers in K. l. ZIP1-expressing cells.
All strains used in this study (S6 Table) are isogenic to BR1919-8B [69]. Strain variants were created by standard genetic crosses and transformation procedures. Every strain in which the K. l. ZIP1 open reading frame replaced the S. c. ZIP1 open reading frame was derived from the same parent, CO1. CO1 was created by first inserting URA3 in place of S. c. ZIP1 open reading frame sequences. Next, a PCR product containing the K. l. ZIP1 open reading frame (amplified off of genomic DNA extracted from K. lactis cells) flanked by ~50 bp of homology to the 5’ and 3’ sequences of the S. c. ZIP1 endogenous locus was transformed into the zip1::URA3 strain, in order to replace the URA3 sequences at the ZIP1 locus with K. l. ZIP1 sequences. Primers used for this step were: 5’TTCTTTGAGATTCGGAAGTAAAATACCCTCGGCGGCTAAATTTTTAGAGAATGTCTAACTTCTTCAGAGACAACTCG 3’ and 5’ACAAAATGAAATGTATTCGCACAAAACGATTTCAAATTTTCCATTATCCTTTATCTGAATCTTTTGGTCTTTTTTAATCGAGG 3’ (underlined regions correspond to K. l. ZIP1 sequences). Counterselection against Ura+ was carried out using 5-FOA medium.
The K.lactis ZIP1-V5 fusion cassette was created by first inserting URA3 between the codons for amino acids 472 and 473 of K. lactis Zip1. Next, a PCR product with flanking homology to K.lactis ZIP1 but carrying an in-frame V5 sequence was used to counterselect against Ura+ cells on 5-FOA medium. DNA sequencing confirmed the position of V5 coding sequences in frame with the codon for amino acid 472 in an otherwise complete K.lactis ZIP1 gene.
To construct a haploid strain capable of sporulation, MATa was integrated at the THR1 locus in a haploid MATα strain, using the B211 plasmid from Beth Rockmill [69].
Strains used for crossover analysis in spores carry a hphMX cassette inserted near the chromosome III centromere, ADE2 inserted upstream of the RAD18 locus, a natMX cassette inserted near the HMR locus, TRP1MX4 was inserted just downstream of the SPO11 locus, and LEU2 and THR1 were inserted on chromosome XI at 152kb, and at 193,424bp, respectively.
Chromosome III circular MATα strains as well as TY521 and TY522 [18] were received from the Roeder lab.
Meiotic nuclei were surface spread on glass slides and imaged as described in [13]. The following antibodies were used: rabbit anti-Zip1 (created as described in [10]), rabbit anti-Red1 [61], guinea pig anti-SUMO [11], chicken anti-HA (Abcam), mouse anti-MYC (clone 9E10, Invitrogen), rabbit anti-V5 (Abcam). Secondary antibodies conjugated with Alexa Fluor dyes were purchased from Life Technologies and used at a 1:200 dilution.
Genetic crossover data was compiled and processed using an Excel Linkage Macro program, created by Jonathan Greene (Rhona Borts, pers. comm.) and donated by Eva Hoffmann (University of Sussex, UK). Final crossover and interference values (and their standard errors) were obtained using the Stahl lab online tools (http://molbio.uoregon.edu/~fstahl/), with the method of Perkins [92]. All other statistical analyses were carried out using Graphpad Prism or Graphpad InStat (www.graphpad.com).
Agarose plugs were prepared from meiotic cultures at 0, 40 and 70 hours of sporulation and subjected to pulsed-field gel analysis [18,79]. For Southern blotting, a 1 kb probe from the THR4 region of chromosome III was prepared using a DIG High Prime DNA Labeling and Detection Kit (Roche). A Syngene “G:Box” was used to detect chemiluminescence and the Syngene “Gene-Tools” program was used to analyze the data. A value for % recombination (Fig 7) was calculated by summing twice the intensity of the trimer band (a double crossover product) plus the dimer band (product of a single crossover) over the total intensity of the three bands (trimer, dimer and monomer). Note that circular chromosome III chromatids do not enter the gel, and thus are not included in the calculation to estimate recombination. The average of two experiments is presented.
Western blotting was performed as described previously [13].
2D gel electrophoresis followed by Southern analysis to assay JMs was performed as previously described [35,66,93] Probes for detection of JMs at the ERG1 locus [77,81] were amplified from yeast genomic DNA with primers- 5’-GGCAGCAACATATCTCAAGGCC-3’ and 5’-TCAATGTAGCCTGAGATTGTGGCG-3’. Probes for detection of JMs at YCR047c [35,80] were amplified from yeast genomic DNA using primers 5’-GGAATTCCGAGAGAATCGACTTGCTAA-3’ and 5’-GGAATTCCAGCCACCAGTGGGCTTTTC-3’. Hybridization signal was detected and quantified using a Typhoon FLA 9000 (GE) and the ImageJ software (http://imagej.nih.gov/ij/).
|
10.1371/journal.ppat.1004781 | Macrocyclic Lactones Differ in Interaction with Recombinant P-Glycoprotein 9 of the Parasitic Nematode Cylicocylus elongatus and Ketoconazole in a Yeast Growth Assay | Macrocyclic lactones (MLs) are widely used parasiticides against nematodes and arthropods, but resistance is frequently observed in parasitic nematodes of horses and livestock. Reports claiming resistance or decreased susceptibility in human nematodes are increasing. Since no target site directed ML resistance mechanisms have been identified, non-specific mechanisms were frequently implicated in ML resistance, including P-glycoproteins (Pgps, designated ABCB1 in vertebrates). Nematode genomes encode many different Pgps (e.g. 10 in the sheep parasite Haemonchus contortus). ML transport was shown for mammalian Pgps, Pgps on nematode egg shells, and very recently for Pgp-2 of H. contortus. Here, Pgp-9 from the equine parasite Cylicocyclus elongatus (Cyathostominae) was expressed in a Saccharomyces cerevisiae strain lacking seven endogenous efflux transporters. Pgp was detected on these yeasts by flow cytometry and chemiluminescence using the monoclonal antibody UIC2, which is specific for the active Pgp conformation. In a growth assay, Pgp-9 increased resistance to the fungicides ketoconazole, actinomycin D, valinomycin and daunorubicin, but not to the anthelmintic fungicide thiabendazole. Since no fungicidal activity has been described for MLs, their interaction with Pgp-9 was investigated in an assay involving two drugs: Yeasts were incubated with the highest ketoconazole concentration not affecting growth plus increasing concentrations of MLs to determine competition between or modulation of transport of both drugs. Already equimolar concentrations of ivermectin and eprinomectin inhibited growth, and at fourfold higher ML concentrations growth was virtually abolished. Selamectin and doramectin did not increase susceptibility to ketoconazole at all, although doramectin has been shown previously to strongly interact with human and canine Pgp. An intermediate interaction was observed for moxidectin. This was substantiated by increased binding of UIC2 antibodies in the presence of ivermectin, moxidectin, daunorubicin and ketoconazole but not selamectin. These results demonstrate direct effects of MLs on a recombinant nematode Pgp in an ML-specific manner.
| Macrocyclic lactones (MLs) are widely used drugs against parasitic nematodes, but drug resistance is rapidly increasing in prevalence and spatial distribution in parasites of ruminants and horses, and is suspected in human nematodes after mass drug applications. Changes in expression levels or the amino acid sequences of P-glycoprotein (Pgp) transporters have frequently been implicated in ML resistance, but direct evidence for transport of MLs by nematode Pgps is still missing. Here, cloning of pgp-9 of the equine parasite Cylicocyclus elongatus and its functional recombinant expression in a Saccharomyces cerevisiae yeast strain deficient in seven endogenous ABC transporters is described. Expression decreased susceptibility to several fungicidal mammalian Pgp substrates including e.g. actinomycin D and ketoconazole, but had no influence on susceptibility to the benzimidazole thiabendazole, which is active against both, yeasts and nematodes. Addition of some MLs strongly increased ketoconazole susceptibility in yeasts expressing C. elongatus Pgp-9, while other MLs had no effect. These interactions are a strong hint that some MLs act as substrates or at least as inhibitors of Pgp-9 mediated drug transport.
| Due to their broad-spectrum antiparasitic activity with effects against both, nematodes and arthropods (endectocides), macrocyclic lactones (MLs) are among the most important antiparasitic drugs in veterinary and human medicine [1]. However, resistance to MLs is widespread in nematodes of small ruminants and currently increasing in prevalence and spatial distribution in nematodes of cattle, horses and humans [2–9] although for the latter the number of reports describing unresponsiveness of parasites to drugs is still only low and future investigations are required to formally prove resistance. In equines, ML resistance was initially observed in Parascaris equorum [10] but recently reports of ML resistant cyathostominae have also emerged [11–15].
In nematodes, the most important ML targets are glutamate-gated chloride channels (GluCl-Rs) whereas ionotropic γ-amino-butyric acid receptors respond only at higher drug concentrations [9].
Specific resistance mechanisms involving single nucleotide polymorphisms (SNPs) in the β-tubulin isotype 1 gene of nematodes from the order Strongylida are well known to be responsible for or at least strongly correlate with resistance to benzimidazoles (BZs) in ruminants and equines [16–20]. For levamisole resistance decreased density and open probability of nicotinic acetylcholine receptors, which are activated by levamisole, and certain splice variants encoding only truncated subunits of the receptors [21–23] have been described in resistant isolates. In contrast, no genotypes have been clearly involved in ML resistance except for a single report describing a SNP in a Cooperia oncophora GluCL-R subunit [24] but this specific change has never been observed in ML resistant nematodes in the field [25]. In the recent past, ABC (ATP-binding cassette) transporters and in particular P-glycoproteins (Pgps, i.e. orthologs of the mammalian ABCB1) have frequently been implicated in ML resistance mechanisms [26]. First hints that Pgps are involved in resistance were obtained by comparison of pgp-2 alleles between ML susceptible and resistant isolates of Haemonchus contortus [27]. Using antibodies against a highly conserved epitope, larger amounts of active Pgps were detected on the egg shell of ML resistant H. contortus [26] and C. oncophora [28]. Moreover, the competitive Pgp inhibitor verapamil was shown to strongly sensitize non-parasitic stages of C. oncophora, whereas the verapamil effects on inhibition of development of the model nematode Caenorhabditis elegans in the presence of ivermectin (IVM) were only small [29]. Moderately increased efficacy of MLs in C. elegans strains which are deficient in individual Pgps was described in several assays [29–31]. Using H. contortus eggs it was also demonstrated that Pgps are activated by MLs [32]. Despite this large amount of work, direct evidence that an individual nematode Pgp is able to interact with MLs was missing for a long time. In the present study a Saccharomyces cerevisiae yeast strain deficient in seven endogenous ABC transporters [33] was used to express a recombinant Pgp-9 cloned from the equine parasitic nematode Cylicocyclus elongatus (Cyathostominae) and to compare interaction of several fungicidal Pgp substrates with this CegPgp-9 in a simple and cheap yeast growth assay. Effects of different MLs were compared in the presence or absence of ketoconazole (Ket) to identify any potential competitive or enhancing effects of the drug combinations. Binding assays using the monoclonal antibody UIC2, which binds to a Pgp epitope only present during active transport, demonstrated activation of Pgp by MLs and fungicidal substances in the absence of a second drug.
Adult C. elongatus were collected from euthanized naturally infected horses, which were bought from their owners. These animal experiments were in accordance with the “Tierschutzgesetz” in Germany and with the European Union directive 2010/63/EU. Experiments were approved by the Landesamt für Verbraucherschutz und Lebensmittelsicherheit (LAVES) in Hannover (Germany) under the reference number 06A435 and by the Landesamt für Gesundheit und Soziales (LaGeSo) in Berlin (Germany) under the reference number L 0088/10.
The full-length cDNA of C. elongatus pgp-9 was obtained using a strategy described previously [34]. Amplification of small fragments with degenerated primers was followed by rapid amplification of cDNA ends (RACE) PCR and amplification of full-length cDNAs. Initially, degenerated primers were designed based on sequence alignments of orthologous Pgp sequences of C. elegans, Caenorhabditis briggsae, C. oncophora and H. contortus. Oligonucleotide sequences are listed in S1 Table. RNA was extracted from adult nematodes and approximately 100 ng were reverse transcribed to cDNA according to the manufacturer’s protocol using random hexamer primers (Revert Aid First Strand cDNA Synthesis Kit, Thermo fisher Scientific). The PCR contained 16 μl H2O, 10 μM each forward and reverse primer, 2.5 μl Accu Prime buffer 1 (with dNTPs), 0.5 μl AccuPrime Taq DNA polymerase (Life Technologies) and 1 μl cDNA. PCR protocols were carried out as follows: After 2 min at 94°C for initial denaturation, 10 cycles of 94°C for 15 s, 50°C for 30 s and 68°C for 1 min were followed by 30 cycles with equal settings but an increased annealing temperature at 60°C. PCR fragments were gel-purified and cloned into pCR4-TOPO vector and sequenced by GATC Biotech (Konstanz).
RACE-PCR was carried out according to manufacturer’s instructions (3'/5'-RACE 2nd generation Kit, Roche) as described recently [34] with primers and temperature profiles as listed in S1 Table. For 5'-RACE, cDNA synthesis was started with a gene-specific primer (S1 Table) and purified cDNAs were tailed with dATP to allow annealing of the oligo-dT anchor primer. PCR mixtures consisted of 18.75 μl H2O, 2.5 μl AccuPrime Buffer 1, 0.5 μM of each primer, 0.5 μl AccuPrime Taq DNA polymerase and 1.0 μl cDNA. PCR protocols for amplification were set as follows: Initial denaturation at 94°C for 2 min, followed by 40 cycles of 94°C for 15 s, 55°C for 30 s and 72°C for 1 min, and a terminal elongation at 72°C for 10 min.
For amplification of a full-length product, cDNA synthesis was carried out using oligo-(dT) primers. PCR mixture for amplification of Cegpgp-9 contained 10 μl Q-solution (Qiagen), 0.5 μM each forward and reverse primer, 10 nM dNTPs, 0.5 μl Phusion II Hot Start Polymerase (Thermo fisher Scientific) and 4 μl cDNA in 50 μl 1×HF buffer. After initial denaturation at 98°C for 30 s, 40 cycles consisting of 98°C for 10 s, 68°C for 30 s and 72°C for 2 min were carried out followed by a terminal elongation at 72°C for 10 min.
Pgp protein sequences from nematodes as well as representative sequences from vertebrates, insects, mollusks and platyhelminthes were aligned using ClustalX2 [35]. The optimal amino acid substitution model was identified using Prottest 3.0 [36] with the number of evolution rate categories set to 8. Phylogenetic trees were calculated with PhyML 3.01 [37,38] assuming the same number of rate categories and the LG+I+F+G model [39] using both, nearest neighbor interchange (NNI) and subtree pruning and regraftment (SPR) moves. To avoid trapping of the iterative optimization process in a local maximum of the likelihood function, calculations started with one neighbor joining and five random trees. Branch support was obtained by conducting the Shimodaira-Hasegawa [SH] approximate likelihood ratio test and the Bayesian transformation of the approximate likelihood ratio test. Finally, the best tree was visualized using MEGA5 [40].
Based on the monoclonal antibody UIC2, which is specific for a conserved epitope present in active Pgps, expression levels of CegPgp-9 were determined by fluorescence activated cell scanning. Briefly, yeast cells were collected from overnight cultures by centrifugation and washed twice with PBS. After re-suspension in 3 ml PBS, three aliquots of 10 μl were transferred to 1.5 ml tubes and incubated with 1 ml blocking buffer (50 mg bovine serum albumin in 25 ml PBS) for saturation of non-specific binding sites. Samples were centrifuged and washed before incubation with the monoclonal antibody specific for active Pgps (UIC2, Drako), PBS or the isotypic antibody (mouse IgG 2aλ, clone HOPC-1, Beckman Coulter). Before flow cytometry analysis cells were again washed and filtered to remove cell agglomerates (30 μM mesh size). The cellular fluorescence intensities were measured with a MoFlo cytometer (Beckman Coulter). Instrument settings corresponded to the protocol as described elsewhere [42].
After amplification of a small RT-PCR fragment using degenerated primers followed by 5' and 3' RACE PCR, amplicons containing the entire ORFs were obtained for C. elongatus. The fragments showed high similarity to C. elegans Pgp-9 as determined by Blast analyses. The cDNA sequence was deposited in GenBank under the accession no. KJ701410. The deduced amino acid sequence reveals the typical domain arrangement of Pgps with two similar halves, each consisting of an ABC transporter transmembrane domain (CDD accession number cd03249) containing six transmembrane helices followed by a nucleotide-binding domain (CDD accession number cl005249) containing the typical, highly conserved Walker A and B motifs as well as Q, D and H loops. Results of phylogenetic analysis using all Pgp proteins encoded in the genomes of C. elegans and C. briggsae as well as many previously published Pgps from other parasitic nematodes confirmed that the protein was a clear ortholog of C. elegans Pgp-9 and was therefore designated C. elongatus Pgp-9 (Fig 1).
The ORF of C. elongatus pgp-9 was amplified with and without stop codon to allow expression without and with a COOH-terminal tag (V5/6×His). PCR products were cloned into the pYes2.1 TOPO vector allowing galactose-induced expression. Plasmids were transformed into the S. cerevisiae strain AD1-7 which is deficient in seven major endogenous ABC transporters [33]. Initially, expression of Pgps in yeast was analyzed by RT-PCR and transcription of pgp-9 mRNAs could be confirmed (S1 Fig). Disappointingly, expression of C. elongatus Pgps in induced yeast cultures using Western blotting was not successful neither using the monoclonal antibody C219 (known to detect in Western blotting a highly conserved Pgp epitope which is present in CegPgp-9) nor with the anti-V5 antibody although detection of β-galactosidase in the control strain was successful using the anti-V5 antibody (S2 Fig). FACS analysis using the monoclonal antibody UIC2, which is specific for a highly conserved epitope present on active Pgp transporters, revealed specific binding of the UIC2 antibody compared to the isotype control. Results of a representative experiment for cells expressing CegPgp-9 are shown in Fig 2. According to forward (FS) and side (SS) scatter, two major populations of yeast cell were defined, i.e. small, non-granular (region R1) vs. large, granular (region R2) cells (Fig 2A). No obvious differences were found between cells expressing CegPgp-9 with or without V5/6×His tag and the following data summarize experiments irrespective of the presence of a tag: In region R1, only a small fraction of the cells (4.4–8.9%) was positive in terms of increased binding of UIC2 in comparison to the isotype control (Fig 2B–2D). Despite much higher background fluorescence in region R2, a much higher number of the yeast cells in this region, i.e. 16.3–20.4%, had detectable amounts of active Pgp on their surface (Fig 2E–2G). No specific binding of UIC2 was detected for AD1-7lacZ yeasts, i.e. percentage of positive cells was between 0.23 and 0.6% (n = 3) in region R1 and 0.31 and 1.3% in region R2 (n = 3).
In order to demonstrate that the expression of recombinant Pgp-9 influences drug susceptibility of the yeast cells, the substrates of mammalian Pgps actinomycin D, daunorubicin, valinomycin and Ket were used [46,47]. For selected concentrations of TBZ as well as for a vehicle control typical growth curves are shown in S3 Fig as examples. Initial experiments were carried out using Ket and cells expressing Pgp-9 without tag (AD1-7CegPgp-9) or with V5/6×His tag (AD1-7CegPgp-9V5His), which were compared with cells expressing lacZ from the same vector backbone (AD1-7lacZ). CegPgp-9 expression decreased susceptibility to Ket irrespectively of the presence of a V5/6×His tag (Fig 3A). The concentration response curves were extremely steep and even less than twofold differences in concentrations led to changes from ≥95% to ≤10% growth, in particular in the strain expressing CegPgp-9 without a tag. Nevertheless, due to very small dilution steps, the EC50 values and confidence intervals of all three strains could be determined (Table 1). The EC50 values of the strains expressing CegPgp-9 were approximately 2.2 (without tag) and 3.0 (with tag) fold increased. Although the difference between EC50 values of both strains was statistically not significant (p = 0.19), all further experiments were conducted using the AD1-7CegPgp-9V5His strain. For actinomycin D, valinomycin and daunorubicin significantly higher EC50 values were observed when comparing AD1-7CegPgp-9V5His strain with the control strain AD1-7lacZ (Fig 3B–3D, Table 1) with increases between 2.1 and 2.3 fold.
Since the anthelmintic TBZ is also an effective fungicide—although at relatively high concentrations—the ability of CegPgp-9 to reduce susceptibility of AD1-7 yeast cells to TBZ could also be analyzed directly. Concentration response curves (Fig 3E) revealed only minimal differences between the control strain and the strain expressing CegPgp-9 and EC50 values were virtually identical (Table 1).
Since MLs were not known to exert any significant fungicidal effects, an indirect method was used to quantify ML interactions with CegPgp-9. For this purpose, a critical concentration (i.e. the highest concentration showing no significant effect on yeast growth) of the fungicidal Pgp substrate Ket was used. For AD1-7lacZ this was 0.18 μM while AD1-7CegPgp-9V5His still grew normally at 0.72 μM. MLs were then used in a concentration range starting with equimolar concentrations of Ket and ML and ending with up to 16 fold higher ML concentrations (Fig 4). These assays were conducted with the aim to test whether MLs can inhibit transport of Ket by CegPgp-9, which would lead to increased susceptibility to Ket.
In the absences of Ket, high concentrations of IVM and eprinomectin (EPM) surprisingly showed a direct effect of the ML on growth of AD1-7CegPgp-9V5His (Fig 4A and 4B). In the same strain, no significant direct concentration dependent effects of moxidectin (MOX), selamectin (SLM) and doramectin (DRM) were observed (Fig 4D and 4E and S4A Fig). With the exception of EPM, which significantly reduced growth at the highest concentration of 2.94 μM, there were no directs effects on growth of AD1-7lacZ (Fig 4E–4H and S4B Fig), which can be explained by the fact that ML concentrations used in this strain were generally lower.
When used in combination with Ket, even equal concentrations of EPM, IVM and MOX (but not of SLM and DRM) with Ket were able to significantly enhance the fungicidal effects of Ket in CegPgp-9 expressing yeast cells (Fig 4A–4C, S4A Fig). Effects ranged between 31 and 79% inhibition of relative growth with the strongest effects exerted by EPM and the least pronounced effects by MOX. Twofold higher ML concentrations virtually abolished relative growth in the presence of EPM, allowing only minimal growth in the presence of IVM and reducing growth in the presence of MOX by approximately 59%. Remarkably, even at higher concentrations MOX only inhibited growth of AD1-7CegPgp-9V5His by about 92% whereas maximal inhibition of growth in the presence of IVM and EPM was approximately 95% and 98%. In contrast to EPM, IVM and MOX, no effects were observed for SLM and DRM even at the highest concentrations used (Fig 4D, S4A Fig).
In contrast, effects of MLs on growth of AD1-7lacZ cells were much smaller and at eightfold higher concentrations of IVM, EPM or DRM than those of Ket, inhibition rates in the rage of 24–34% were observed (Fig 4E–4H, S4B Fig). Even 16-fold excess of MLs over Ket caused only 47–78% growth inhibition.
Since UIC2 is known to bind only to active Pgps, binding of the antibody should increase in the presence of substrates. AD1-7lacZ and AD1-7CegPgp-9V5His cells were incubated with drugs followed by addition of UIC2 or an isotype control and detection via a HRP-labeled secondary antibody and chemiluminescence. Since this type of assay is much more complicated in terms of sample handling than the growth assay, only selected drugs and two concentrations per drug were analyzed. Drug concentrations were chosen to have one concentration with a strong and one with no or only a minimal effect in the growth assay. For comparisons of MLs, one with a strong effect (IVM), one with a moderate effect (MOX) and one without effect (SLM) on growth in the presence of Ket were chosen. The same ML concentrations were used for all MLs and concentration with small and strong effects of IVM were selected.
AD1-7CegPgp-9V5His cells showed significantly higher binding of UIC2 than of the isotype control antibody (Fig 5). Binding of either antibody to AD1-7lacZ was comparable to binding of the isotype control to AD1-7Pgp-9V5His and significantly lower than the binding of UIC2 to AD1-7CegPgp-9V5His (p<0.05). In the presence of drugs, changes were observed only for the combination of AD1-7CegPgp-9V5His with UIC2 but not with any of the combinations involving the isotype control antibody or the AD1-7lacZ yeasts. The higher concentrations of IVM, MOX, daunorubicin (4 μM) and Ket (2 μM) clearly increased binding of UIC2 to AD1-7CegPgp-9V5His as demonstrated in terms of chemiluminescence intensity by roughly 5–6 fold (p<0.05), although a high variation was observed between samples (Fig 5). In contrast, the higher concentration of SLM (4 μM) had no significant effect on binding of UIC2 to AD1-7CegPgp-9V5His cells. The lower concentrations of the drugs resulted in 0.9–1.8 fold changes in median relative light units and none of these changes was significant.
MLs represent the most important antiparasitic drug class with numerous fields of application in human and veterinary parasitology. Unfortunately, during the past decade their unprecedented efficacy is facing increasing resistance problems most often observed in parasitic nematodes of ruminants and horses. Non-specific mechanisms such as drug metabolism and in particular drug extrusion have been frequently implicated in anthelmintic resistance, in particular for the MLs [48] which are well established substrates of mammalian Pgps [49].
Several studies have suggested correlation between Pgp expression or genotype and ML resistance [9,50]. Particularly, Pgp-2, Pgp-9 and Pgp-11 have been implicated in ML resistance in H. contortus, Teladorsagia circumcincta and P. equorum [27,51–54]. In C. elegans, loss of function mutants of any Pgp gene increase IVM susceptibility significantly but effects of these mutations were in general modest to low. This is presumably due to substantial redundancy caused by 14 Pgp transporters encoded in the C. elegans genome which probably have substantial overlaps in their substrate specificity [29–31]. Effects of a putative loss of function mutation in C. elegans pgp-9 were moderate in a development assay and increased IVM susceptibility by approximately 2.8-fold in terms of the EC50 [29].
Despite the fact that many studies dealt with the role of Pgps in anthelmintic resistance, to the best knowledge of the authors recombinant expression of only one C. elegans Pgp (Pgp-1) [55] and one parasitic nematode Pgp have been reported yet [54]. The latter, very recently published study was conducted largely in parallel with the current study published here [56] and describes expression of H. contortus Pgp-2 in porcine LLC-PK1 cells [54] known to have low endogenous transporter activity.
Detection of CegPgp-9 by Western blotting failed although the β-galactosidase in the control strain was detectable. This could suggest that no or only very low amounts of CegPgp-9 were produced by the cells. To evaluate CegPgp-9 expression further, FACS analysis employing the monoclonal antibody UIC2 was initially used as an alternative detection method. This antibody was frequently used to characterize mammalian Pgps and is known to bind to a non-continuous, conformation-sensitive epitope that is only present on Pgps during certain stages of active transport/ATP hydrolysis [57,58]. Binding of UIC2 to Pgp—in particular in the presence of low concentrations of Pgp substrates—is well known to inhibit Pgp-mediated transport and the presence of Pgp substrates increases binding of UIC2 [57,59]. This antibody has also several times been used to characterize nematode Pgps by flow cytometry [32,42,60,61] and it was suggested that the epitope recognized is conserved between mammals and nematodes. The results presented here using yeasts expressing transgenic CegPgp-9 show for the first time formally that UIC2 binds to an epitope that is conserved between nematode and vertebrate Pgps and therefore corroborate the previously published reports regarding nematode Pgps. They also clearly proof that expression of CegPgp-9 was successful, which is further confirmed by consistent shifts in the drug concentration response curves towards higher drug concentrations.
In the Pgp-9 transgenic yeasts it was not surprising that only a minor percentage of yeast cells exposed the UIC2 epitope. It has been found previously that in the absence of any Pgp substrate—as in the FACS experiments described here—only a small fraction of mammalian cancer cells expressing Pgp appeared to be positive for UIC2 binding [59]. The frequency of positive cells was described to increase markedly in the presence of Pgp substrates. This probably explains the low frequency of UIC2-positive yeast cells observed herein, despite the usage of a clonal expression system. Nevertheless, it remains unclear why expression of the UIC2 epitope was much higher in the larger cell population. Not unexpectedly, background fluorescence (i.e. autofluorescence plus binding of the isotype control) was much higher in the larger cells from region R2. Since increased background should in principle decrease the ability to detect specific binding, the observed higher frequency of UIC2 binding despite increased background is most likely specific. The authors currently have no experimentally substantiated explanation why the expression of an epitope restricted to active Pgps is preferentially observed on large yeast cells. There are no hints in the literature that the gal-1 promoter used in the expression vector displays a cell cycle specific expression pattern. In fact, this promoter is widely used for high-level expression and cell-cycle dependency would probably counteract this aim. Increased production of endogenous Pgp substrates in the later parts of the cell cycle resulting in more active Pgp on the cell surface would be a possible explanation for the data but further experiments are required to fully understand this observation. In the chemiluminescence assay, increased binding of UIC2 but not of the isotype control antibody was observed in the presence of drugs for AD1-7Pgp-9 but not for AD1-7lacZ. Results were overall very similar to those observed in the growth assay but variation between samples was much larger. Increased binding was concentration dependent with the lower drug concentrations showing no significant effects.
The results obtained with various well characterized fungicidal substances known to be substrates of mammalian Pgps clearly show that the direct growth assay is able to evaluate the substrate specificity of transporters as previously shown by others [46] who tested interaction of cytotoxic drugs with variants of human ABCB1 in similar yeast assays. The anthelmintic drug TBZ also has fungicidal activity and could therefore be tested in the direct growth assay. Despite the fact that Pgps were recently implicated in resistance to TBZ [62], no protective effects of CegPgp-9 were detected here. This might be explained by the large number of Pgp genes encoded in the genomes of nematodes and maybe also by species or even isolate-specific differences. Although the CegPgp-9 analyzed in the present study does not have the potential to confer TBZ resistance, other members of the gene family might still have this capability. Moreover, the ability of Pgp-9 from other parasite species or even other C. elongatus isolates can also not be excluded from the negative data obtained in the present study. Recombinant expression of C. elegans Pgp-1 increased resistance of Spodoptera frugiperda Sf9 cells to the cytotoxic drugs actinomycin D and paclitaxel. ATPase activity was also stimulated by both drugs and additionally by valinomycin, progesterone and dipyridamole. In contrast many other drugs stimulated ATPase activity only marginally (e.g. vinblastine) or not at all (including verapamil and daunorubicin) [55]. Comparison with the result of the direct yeast growth assay for CegPgp-9 suggest partially overlapping substrate spectra but also important discrepancies. In order to obtain a more complete overview of nematode Pgp substrate specificity, comparisons of all paralogs from a single nematode species with the same method would be extremely valuable.
In contrast to the BZs, most other anthelmintic and in particular most nematocidal drug classes target neuronal receptors and therefore any additionally observed fungicidal activities most likely are due to completely different modes of action. For MLs, activity against the microorganisms Mycobacterium tuberculosis and Chlamydia trachomatis has previously been described [63,64], although IVM effects against M. tuberculosis could not be confirmed by another group [65]. In Candida spp., milbemycin oximes have been shown to not only potentiate azole effects by inhibition of azole export but also to cause formation of reactive oxygen species and activate transcription of several genes involved in stress response pathways [66]. However, activity of MLs against AD1-7 yeasts was only observed at high concentrations and only for IVM and EPM. Since the direct growth assays can only be conducted for drugs exhibiting considerable fungicidal activity, an indirect growth assay was developed to test for interaction of MLs with susceptibility to Ket in the presence or absence of CegPgp-9. In this assay, IVM and EPM showed strong, MOX intermediate and SLM and DRM no interaction with Ket in CegPgp-9 transgenic AD1-7 yeasts. A similar ranking was also observed in the chemiluminescence assay were strongly increased binding of UIC2 was detected for IVM and MOX but no increase occurred in the presence of SLM. Although the growth assay was designed similarly to competitive inhibition assays with a fixed Ket concentration together with increasing ML concentrations, the observed effects do not necessarily reflect competitive inhibition. Interaction depends on the presence of CegPgp-9 and therefore most likely on interference with Ket transport, but other types of inhibition mechanisms than competitive cannot be excluded currently. Despite this ambiguity, the assay is clearly able to identify differences between MLs regarding their interaction with CegPgp-9 in the context of Ket. These positive interactions for example substantially corroborate previous correlative support for involvement of Pgp-9 in IVM resistance in T. circumcincta [52].
Mammalian Pgps are well known to transport several MLs and the distribution and bioavailability of MLs in different organs and tissues of the hosts are strongly influenced by the host Pgps in vivo [67]. Particularly IVM and DRM are known to exert strong neurotoxic effects in mdr-1 (ABCB1a) knock-out mice and in certain dogs which have a loss of function mutation in the mdr-1 gene. Uptake of IVM, DRM and SLM in the brain of mice is strongly decreased by Pgp transport, whereas differences are smaller for MOX and EPM [68,69]. Using cells expressing human or canine Pgp, strong interaction of IVM, EPM and DRM and weak interaction of MOX were reported [70,71]. For SLM, results of these studies were inconsistent with one [71] describing strong and the other [70] only weak interaction. In the present study, strong interaction for CegPgp-9 with IVM, EPM and slightly weaker interactions with MOX were found to be similar to the above-mentioned observation for the mammalian orthologs. In contrast, results for DRM and SLM showed profound discrepancies between mammalian Pgp and CegPgp-9. Future experiments comparing these transporters in the same environment are required to ultimately identify these differences as related to the transporter and not to the experimental system. Previously published results of functional analyses of nematode Pgps surprisingly describe stimulatory effects of anthelmintic drugs on extrusion of the Pgp substrate rhodamine 123 from H. contortus eggs [32]. This study also found no significant interaction of TBZ with Pgps on the nematode egg surface which is in agreement with previous results regarding mammalian Pgps [72]. For the MLs, strong effects were observed for emamectin and DRM, medium effects for abamectin, EPM and MOX, only small effects of SLM and surprisingly no significant effects of IVM, which is a very good substrate of mammalian Pgps [72] and also showed strong interaction with CegPgp-9 here and H. contortus Pgp-2 [54]. However, the experimental approaches used here differ significantly from those used with H. contortus eggs. In contrast to stimulation of rhodamine transport by MLs for nematode Pgps [32], MLs have been consistently described to inhibit rhodamine transport by mammalian Pgps (ABCB1) [49,70–74] or by ABC transporters in azole-resistant Candida spp. fungi [66,75]. The effects of MLs on rhodamine transport in H. contortus eggs require further investigations to confirm these results and identify potential reasons for the observed differences. Using transfected LLC-PK1 cells, Godoy et al. [54] could recently show that MLs inhibit transport of the Pgp substrates rhodamine 123 and calcein-AM by recombinant H. contortus Pgp-2 in a concentration dependent manner. Effects of MLs differed between the assays. In the presence of rhodamine abamectin showed the strongest, IVM an intermediate and MOX the smallest influence. In contrast, effects of the MLs on transport of calcein-AM were very similar. One possible reason for the very unusual stimulation of rhodamine transport by MLs in the eggs might be the localization of Pgps in or on the eggshell in contrast to its normal localization in a plasma membrane. The drug binding site of Pgps is very complex and for mouse ABCB1b and C. elegans Pgp-1 crystal structures have been obtained by x-ray diffraction [55,76,77]. Both structures reveal that drug entry apparently occurs at the level of the inner leaflet of the plasma membrane. The localization of the substrate binding pockets in the membrane causes a significant impact of the local lipid environment on binding affinity. Indeed, binding affinity of drugs to Pgps is much higher in a lipid bilayer than in detergent micelles [55]. Lipid dependency of substrate affinity might result in changes in substrate affinity in recombinant expression systems in comparison to the natural situation. This potential systematic error can currently not be excluded for any of the systems used to express Pgps. Lipid composition of mammalian [54,70,72], insect [55] and yeast ([46] and this study) expression systems for Pgps is probably always significantly different from that of parasitic nematodes. One could propose that C. elegans would be a better expression system due to its closer phylogenetic relationship but evidence for this hypothesis is lacking. This is probably valid for soluble proteins but regarding lipid composition of plasma membranes, considerable differences between C. elegans and parasitic nematodes must be assumed, since optimal environmental temperatures of 20°C and 37°C surely imply different lipid composition to achieve similar membrane fluidities.
The mouse Pgp crystal structures in the presence of different drugs indicate drug-dependent interaction with specific amino acid residues suggesting that different, only partially overlapping binding sites for different drugs are used [76,77]. Earlier, much simpler models of mammalian Pgps considered three distinct substrate binding sites. Transport of substances binding to the R site (e.g. rhodamine 123, daunorubicin and vinblastine) is stimulated by drugs binding to the H site (e.g. Hoechst 33342 and colchicine) [78]. In contrast, Pgp substrates such as Ket and verapamil are well known to inhibit transport of other substrates in a more or less competitive manner [79].
Although nematode Pgps have been implicated not only in resistance to MLs but also to BZs [62] and have been shown to be activated by levamisole [32], current knowledge about substrate specificity of individual nematode and in particular parasitic nematode Pgps regarding anthelmintics is still very sparse. Only very recently, the brain penetration of emodepside has been shown to be increased in mdr-1 (ABCB1a) deficient mice [80] suggesting that the cyclooctadepsipeptides [81] might also be Pgp substrates. The study presented here aims to initiate more systematic approaches, leading to detailed comparison of substrate spectra. Such comparisons could be made between (i) Pgp paralogs within a single species, (ii) Pgp orthologs from different species and (iii) alleles of individual Pgp genes from independent isolates of the same parasite species with different resistance status. This aim is currently hampered by several problems. First, at least 10 Pgp genes can be suspected to be encoded in the genomes of most parasitic nematodes with no complete collection of full-length sequences available for any species. Secondly, a large number of anthelmintic drugs must be considered to be potential Pgp substrates. Finally, there are huge differences between drugs belonging to the same drug class regarding interaction with the same Pgp. Therefore, research aiming to provide a detailed picture of interaction of anthelmintic drugs with individual Pgps is obviously still in its infancy.
In conclusion, this report shows functional interaction of MLs with a Pgp from a target species. At least for Pgp-9 of C. elongatus it was clearly shown, that there are marked differences regarding interaction of individual MLs with this particular transporter. The growth assay described herein is a relatively simple and inexpensive method for the investigation of the functional analysis of transport by Pgps of parasitic nematodes but potentially also arthropods. In comparison to the chemiluminescence-based UIC2 binding assay, which suffers from very high hands-on times and only limited numbers of samples that can be analyzed in parallel, and to the assay using stably transfected mammalian cells [54], the yeast assay described here is cheap, largely automatically performed and even has the potential to be up-scaled to 384 well plates. The chemiluminescent UIC-binding assay offers the advantage that it does not rely on interaction of two drugs. The transport assays using mammalian cells described by Godoy et al. [54] has revealed that MLs differ in interaction with H. contortus Pgp-2 depending on the co-substrate that is used to show increased intracellular accumulation (rhodamine 123 vs. calcein-AM). The same does most likely also apply to the growth assay, i.e. using another fungicidal drug than Ket might result in another ranking of MLs. Since numerous Pgp genes as well as alleles from the same Pgp gene have been described for different parasitic nematodes, this new assay provides the opportunity to directly compare their substrate specificity and their ability to interact with anthelmintics. Therefore, this approach provides a powerful tool to compare substrate specificity of different Pgps, enabling pharmacological and functional research in the future, including various anthelmintic drug classes and compounds. Due to the episomal expression system, Pgps differing only in single amino acid changes can be expected to be expressed at equal levels as previously exploited to analyze human Pgp variants [46]. In mammalian systems, targeted integration would be required for such comparisons. Usage of targeted integration systems such as homologous or targeted integration is also much easier in the yeast than in the mammalian system. This should also allow determination of effects of non-synonymous SNPs in Pgps that were described to be associated with ML resistance. Finally, the specific functional analysis of Pgps as modulators of anthelmintic activity will contribute to the understanding of AR mechanisms, foster the development of tools for early detection of resistance and increase our ability to prevent or postpone the development of AR.
|
10.1371/journal.pntd.0004836 | Strong-LAMP: A LAMP Assay for Strongyloides spp. Detection in Stool and Urine Samples. Towards the Diagnosis of Human Strongyloidiasis Starting from a Rodent Model | Strongyloides stercoralis, the chief causative agent of human strongyloidiasis, is a nematode globally distributed but mainly endemic in tropical and subtropical regions. Chronic infection is often clinically asymptomatic but it can result in severe hyperinfection syndrome or disseminated strongyloidiasis in immunocompromised patients. There is a great diversity of techniques used in diagnosing the disease, but definitive diagnosis is accomplished by parasitological examination of stool samples for morphological identification of parasite. Until now, no molecular method has been tested in urine samples as an alternative to stool samples for diagnosing strongyloidiasis. This study aimed to evaluate the use of a new molecular LAMP assay in a well-established Wistar rat experimental infection model using both stool and, for the first time, urine samples. The LAMP assay was also clinically evaluated in patients´ stool samples.
Stool and urine samples were obtained daily during a 28-day period from rats infected subcutaneously with different infective third-stage larvae doses of S. venezuelensis. The dynamics of parasite infection was determined by daily counting the number of eggs per gram of feces from day 1 to 28 post-infection. A set of primers for LAMP assay based on a DNA partial sequence in the 18S rRNA gene from S. venezuelensis was designed. The set up LAMP assay (namely, Strong-LAMP) allowed the sensitive detection of S. venezuelensis DNA in both stool and urine samples obtained from each infection group of rats and was also effective in S. stercoralis DNA amplification in patients´ stool samples with previously confirmed strongyloidiasis by parasitological and real-time PCR tests.
Our Strong-LAMP assay is an useful molecular tool in research of a strongyloidiasis experimental infection model in both stool and urine samples. After further validation, the Strong-LAMP could also be potentially applied for effective diagnosis of strongyloidiasis in a clinical setting.
| Human strongyloidiasis, a soil-transmitted infection mainly caused by Strongyloides stercoralis, is one of the most neglected among the so-called neglected tropical diseases (NTDs). The difficult diagnosis lead to an underreporting of infection rates. Strongyloidiasis can easily be misdiagnosed because many infections remain asymptomatic and the lack of sensitivity of the conventional fecal-based techniques for morphologically identification of infective larvae in feces. Although serologic tests are useful, a limitation in standardization to avoid cross-reactions still remains. There is an urgent need to improve more sensitive and specific diagnostic tests, particularly in immunocompromised patients or candidates to immunosuppressive treatments. Several molecular approaches for Strongyloides spp. DNA detection have already been assayed, but they have a very limited use in routine diagnostic, particularly in endemic areas. In addition, all molecular approaches for Strongyloides spp. DNA detection have always been mainly assayed for stool samples and no other more advantageous biological samples, such as urine, have been investigated for molecular purposes. In this study we have developed, for the first time, a molecular assay using LAMP methodology as a simple, sensible and robust method for the detection of S. venezuelensis DNA in a well-established Wistar rats experimental infection in both stool and urine samples. The LAMP assay was also successfully evaluated in patients´ stool samples. Our LAMP assay (Strong-LAMP) is an useful molecular tool in a strongyloidiasis experimental infection model and could be a potential field-friendly diagnostic test in a clinical setting, following further validation.
| Strongyloidiasis, a soil-transmitted helminth human infection, is considered by World Health Organization (WHO) as a neglected condition affecting an estimated 30–100 million people worldwide [1]. The accuracy of these estimates remains actually uncertain due to lack of efficient guidelines for screening the population in epidemiological surveys [2, 3]. At least, two species of nematodes of the genus Strongyloides, namely Strongyloides stercoralis (the most common human pathogen species) and S. fuelleborni, are known to infect humans causing strongyloidiasis [4, 5, 6]. Human infection is primarily acquired by the filariform larvae (the infective third-stage larvae, iL3) penetrating the skin or mucous membranes through unprotected contact with contaminated soil [7]. S. stercoralis biology is complex involving two separate life cycles, the free-living heterogonic cycle and a parasitic cycle [8, 9]. The exceptional ability of this parasite to replicate in the human host permits ongoing cycles of autoinfection thus resulting in a chronic strongyloidiasis that can therefore persist for several decades without further exposure to a new exogenous infection [6]. Inmunocompetent patients with uncomplicated strongyloidiasis usually develop an asymptomatic, mildly symptomatic or chronic infection, which are typically associated with low intestinal worm burdens and intermittent larval excretion [10, 2]. However, a deregulation of the host´s immune response during the latent infection may lead in an uncontrolled multiplication of the parasites (hyperinfection syndrome) which can be life-threatening [6, 2, 10, 11, 7] with mortality rates of up 87% [12, 13]. Thus, detecting latent cases of S. stercoralis is crucial to decrease morbidity and mortality of the infection.
The diagnosis of strongyloidiasis is suspected when clinical signs and symptoms, eosinophilia or serologic findings are observed [14, 6], but definitive diagnosis is accomplished by parasitological examination of stool samples allowing the morphological identification of S. stercoralis, including direct smear in saline, the spontaneous sedimentation method [15], centrifugation [16], the Baermann´s technique [17], the agar plate culture [18] and the Harada-Mori´s filter paper culture method [19]. These methods have classically low sensitivity because of the low and irregular load of larvae in the feces [6]; the collection of a larger number of stool samples on alternate days instead of a single one and the combination of several diagnostic methods may increase the sensitivity [20]. On the other hand, several immunological methods have been described for diagnosing strongyloidiasis showing a high sensitivity when compared with parasitological methods but a limitation in the standardization of more specific serological tests in order to avoid the possibility of cross-reaction with other helminths still remains [21, 22]. Several DNA-based techniques (i.e. single-PCR, nested-PCR, PCR-RFLP, real-time PCR) have provided useful alternatives not only for identification of Strongyloides species [23, 24] but also for S. stercoralis DNA detection in feces with high accuracy in the diagnosis of strongyloidiasis [22, 25, 26, 27]. Nevertheless, such molecular methods have a very limited use in routine diagnostic, particularly under field conditions in endemic areas requiring special equipment manipulated by trained personnel. Thus, the development of new, simple, applicable and cost-effective alternative molecular assays is necessary to diagnose human strongyloidiasis, mainly in those immunocompromised individuals in which the infection can be fatal.
At present, there is a nucleic acid amplification method named loop-mediated isothermal amplification (LAMP) [28]. Compared to PCR, the simplicity of the LAMP method makes it suitable for field testing in developing countries [29, 30], and many LAMP reactions have already been developed for molecular detection and diagnostics of infectious diseases, including parasitic diseases [31]. In this sense, a first LAMP assay for the detection of S. stercoralis in feces has been recently developed and preliminary evaluated with human stool samples [32]. To date, all new successfully approaches for molecular methods to be used for Strongyloides spp. DNA detection have been focused in analyzing mainly stool samples and no other type of sample, such as urine, has been considered for the detection of parasite DNA. Urine is a biological sample that would have a number of advantages in diagnosis of strongyloidiasis over the stool samples since it has less inconvenience to obtain from patients as well as it is easier in handling and storing. It has been demonstrated that small amounts of cell-free circulating DNA are able to pass the kidney barrier and end up in urine [33, 34, 35]; this circulating DNA from the bloodstream that passes into the urine can be isolated and used in diagnostic applications.
This study aimed to assess the diagnostic utility of a new designed LAMP assay in an active experimental rodent strongyloidiasis in parallel with parasitological method by direct fecal examination. We used as a source for Strongyloides spp. DNA detection both stool and, for the first time, urine samples from rats experimentally infected with different doses of S. venezuelensis iL3. The LAMP developed in this work (namely, Strong-LAMP) was shown to be sensitive and specific in detecting Strongyloides spp. DNA. The potential diagnostic applicability of the Strong-LAMP could be also demonstrated on a number of human clinical stool samples with previously parasitological and real-time PCR confirmed strongiloidiasis.
The study protocol was approved by the institutional research commission of the University of Salamanca. Ethical approval was obtained from the Ethics Committee of the University of Salamanca (protocol approval number 48531), which approved the animal protocol. Animal procedures in this study complied with the Spanish (Real Decreto RD53/2013) and the European Union (European Directive 2010/63/EU) guidelines on animal experimentation for the protection and humane use of laboratory animals and were conducted at the accredited Animal Experimentation Facility of the University of Salamanca (Register number: PAE/SA/001). The human stool samples used in this study were obtained as part of public health diagnostic activities at Severo Ochoa and Gregorio Marañón Hospitals, Madrid, Spain. A standardized epidemiological questionnaire and clinical information were obtained from each participant included in the study. Participants were given detailed explanations about the aims, procedures and possible benefit of the study. Written informed consent was obtained from all subjects and samples were coded and treated anonymously. The study received the approval of the Committee of Research Ethics and Animal Welfare from the Instituto de Salud Carlos III (PI number: CEI PI06_2012-v2).
Twelve six-week-old male Wistar rats weighing 150–175 g (Charles River Laboratories, Barcelona, Spain) were used in our study as the source for stool and urine samples. Animals were housed at the accredited Animal Experimentation Facility of the University of Salamanca in individual metabolic polycarbonate cages and placed in humidity and temperature controlled environment with a 12 hour photoperiod and received sterilized food and water ad libitum. Animals were monitored regularly by qualified members in animal welfare at the Animal Experimentation Facility of the University of Salamanca.
Strongyloides venezuelensis used in this study was obtained from a strain originally used in the Department of Parasitology, University of Minas Gerais, Belo Horizonte, Brazil. This strain has been maintained by serial passages in laboratory rats routinely infected in the Laboratory of Parasitic and Molecular Immunology, CIETUS, University of Salamanca. Feces from infected rats were cultured using vermiculite mixed with distilled water at 28°C for 3–7 days and infective third-stage larvae (iL3) which came out of the feces were then collected and concentrated by using the Baermann extraction method as described elsewhere [36]. Recovered larvae were washed in phosphate-buffered saline (PBS) and their viability was checked using a light microscope prior to infection. The number of viable iL3 was determined and the animals were afterwards infected subcutaneously with different iL3 doses of S. venezuelensis to ensure a potential range of low, middle and high fecal egg production during the development of infection [37], as follows: group one (n = 3; each rat infected with 40 iL3), group 2 (n = 3; each rat infected with 400 iL3), group 3 (n = 3; each rat infected with 4,000 iL3) and group 4 (n = 3; non infected, as control group). During the 28-day infection, the animals were housed individually in metabolic cages, thus allowing separate collection of urine and feces from rodents and also to eliminate the possibility of rats re-infecting themselves from fecal sources during the experimental period. Infected rats were euthanatized in a CO2 gas chamber 29 days after the infection.
Human stool samples (n = 12) were obtained from outpatients (including Spanish nationals, immigrants, tourists and aid workers) attending Severo Ochoa and Gregorio Marañón Hospitals in Madrid, Spain, during June 2010 to June 2012 as a part of a collaborative research study on human strongyloidiasis. Those patients showed significant levels of IgE, eosinophilia or other symptoms suggestive of disease. Stool samples were examined after arrival by qualified laboratory technicians. Eleven of these 12 stool samples (nos. 030, 140, 231, 232, 338, 339, 069, 259, 331, 468 and 126) were subjected to different parasitological methods as screening tests for strongyloidiasis, including microscopic examination (MOE) for the presence of rabditiform larvae in direct fecal smears, agar plate culture (APC) or Harada-Mori´s filter paper culture method (HMM). Unfortunately, for one sample (no. 496) was not possible to perform any of the parasitological methods. Strongyloidiasis was confirmed in 7/11 stool samples by one or more parasitological tests applied. Four samples (nos. 259, 331, 468 and 126) were found to be negative; however, in two of these negative samples (nos. 468 and 126) eggs from Taenia saginata and "hookworm", respectively, could be observed upon microscopic inspection. All patients´ samples were obtained before treatment with ivermectin. Thereafter, patients´ stool samples were sent to the Instituto de Salud Carlos III (ISCIII), Madrid, Spain, for further DNA extraction and molecular analyses by real-time PCR (RT-PCR) as described below. Table 1 shows the patients´ stool sample numbers, the parasitological tests applied at Hospitals as well as results obtained in parasitological and molecular tests performed.
Patients´ stool samples included in this study were firstly tested by a RT-PCR optimized at ISCIII, Madrid, Spain, as described by Saugar et al. [27]. Briefly, the RT-PCR was standardized in laboratory settings using the specific primers Stro18S-1530F (5′-GAATTCCAAGTAAACGTAAGTCATTAGC-3′) and Stro18S-1630R (5′-TGCCTCTGGATATTGCTCAGTTC-3′) to amplified a 101 base pair (bp) region of S. stercoralis 18S rRNA (Gene Bank accession no. AF279916.2) as previously described by Verweij et al. [25]. The amplification was performed with a 25 μL reaction mix containing 5 μL of DNA extracted from stool samples, 1X Quantimix Easy Master Mix (Biotools B&M Laboratories), 0.2 μM of each Stro18S-1530F and Stro18S-1630R primer and 0.5 μL of 50X SYBR Green I (Invitrogen). The program consisted of an initial step of 15 min at 95°C followed by 50 cycles of 10 s at 95°C, 10 s at 60°C and 30 s at 72°C. The reaction and fluorescence detection were performed on the Corbett Rotor-Gene 6000 real-time PCR System (QIAGEN, Hilden, Germany) and The Rotor Gene 6000 Series software v.1.7 was used for data analysis. In each RT-PCR run both negative (DNA from uninfected stool sample) and positive (DNA from stool samples artificially infected with different amounts of S. venezuelensis iL3 DNA) controls were routinely included.
After searching on literature reports to identify potential sequences of DNA to be used in detection of Strongyloides spp., a 329 nucleotide bp corresponding to a linear genomic DNA partial sequence in the 18S rRNA gene from S. venezuelensis was selected and retrieved from GenBank (Accession No. AJ417026.1) for the design of specific primers [39]. A BLASTN search and alignment analysis [40] indicated that the sequence had 94–99% similarity with other sequences reported for Strongyloides spp. and no regions of similarity between this sequence and other sequences reported for possible human pathogens were detected. The 329 bp sequence selected was also tested in silico for similarity in the currently available genome databases for S. stercoralis at NemBase4 (www.nematodes.org) and a 94% identity with a partial sequence in contig SSC06134_1 annotated for the parasite was obtained. Forward and backward outer primers (F3 and B3) and forward and backward inner primers (FIP: F1c-F2 and BIP: B1c-B2, respectively) were designed using the online primer design utility, Primer Explorer v.4 (Eiken Chemical Co., Ltd., Japan; http://primerexplorer.jp/e/). Several LAMP primer sets were suggested by the software and further refinement in primer design was developed manually based on the criteria described in “A Guide to LAMP primer designing” (http://primerexplorer.jp/e/v4_manual/index.html). LAMP primers sequences finally selected are indicated in Table 2 and their positions relative to the 329 bp target sequence of S. venezuelensis compared with the 94% similarly partial sequence of S. stercoralis in contig SSC06134_1 are shown in Fig 1. All the primers were of HPLC grade (Thermo Fisher Scientific Inc., Madrid, Spain); the lyophilized primers were resuspended in ultrapure water to a final concentration of 100 pmol/μL and stored at -20°C until use.
The outer LAMP primer pair, designated F3 and B3 (Table 2), was firstly tested for S. venezuelensis specificity by a touchdown-PCR to verify whether the correct target was amplified. The PCR assay was conducted in 25 μL reaction mixture containing 2.5 μL of 10x buffer, 1.5 μL of 25 mmol/L MgCl2, 2.5 μL of 2.5 mmol/L dNTPs, 0.5 μL of 100 pmol/L F3 and B3, 2 U Taq-polymerase and 2 μL (10 ng) of DNA template. Initial denaturation was conducted at 94°C for 1 min, followed by a touchdown program for 15 cycles with successive annealing temperature decrements of 1.0°C (from 57°C to 52°C) every 2 cycles. The specificity of PCR was also tested with a panel of 22 heterogeneous DNA samples from other parasites included in the study. Besides, the sensitivity of the PCR was also assayed to establish the detection limit of S. venezuelensis DNA with 10-fold serial dilutions prepared as mentioned above.
We tried to evaluate the LAMP primer set designed by using different in house reaction mixtures each containing a different Bst polymerase (namely, Bst DNA polymerase Large Fragment, Bst DNA polymerase 2.0 and Bst DNA polymerase 2.0 WarmStart; New England Biolabs, UK) as well as varying concentration of betaine (Sigma, USA) and supplementary MgSO4 (New England Biolabs, UK) to compare results in S. venezuelensis DNA amplification.
Thus, LAMP reactions mixtures (25 μL) contained 40 pmol of each FIP and BIP primers, 5 pmol of each F3 and B3 primers, 1.4 mM of each dNTP (Bioron), 1x ThermoPol Reaction Buffer -20 mM Tris-HCl (pH 8.8), 10 mM KCl, 10 mM (NH4)2SO4, 2 mM MgSO4, 0.1% Triton X-100; New England Biolabs, UK- (when using Bst Polymerase Large Fragment) or 1x Isothermal Amplification Buffer -20 mM Tris-HCl (pH 8.8), 50 mM KCl, 10 mM (NH4)2SO4, 2 mM MgSO4, 0.1% Tween20; New England Biolabs, UK- (when using either Bst DNA polymerase 2.0 or Bst DNA polymerase 2.0 WarmStart), betaine (ranging 0.8, 1, 1.2, 1.4 or 1.6 M), supplementary MgSO4 (ranging 2, 4, 6 or 8 mM) and 8 U of the tested Bst polymerase in each case with 2 μL of template DNA.
All LAMP reactions mixtures were performed in 0.5-mL micro centrifuge tubes that were incubated in a heating block (K Dry-Bath) at a range of temperatures (61, 63 and 65°C) for 60 min to optimize the reaction conditions and then heated at 80°C for 5–10 min to inactivate the enzyme and thus to terminate the reaction. In each case, the optimal temperature was determined and used in the subsequent tests. As the LAMP reaction is highly sensitive, possible DNA contamination and carry-over of amplified products were prevented by using sterile tools at all times, performing each step of the analysis in separate work areas, minimizing manipulation of the reaction tubes and even closing them with a plastic paraffin film. Template DNA was replaced by ultrapure water as negative control in each LAMP reaction.
Amplified DNA in the LAMP reaction causes turbidity due to the accumulation of magnesium pyrophosphate, a by-product of the reaction. Once the reaction was finished and following a brief spin of the reaction tubes, the turbidity of reaction mixture was visually inspected by naked eyes. The LAMP amplification results could also be visually inspected by adding 2 μL of 1:10 diluted 10,000X concentration SYBR Green I (Invitrogen) to the reaction tubes. To avoid as much as possible the potential risk of cross-contamination with amplified products, all tubes were briefly centrifuged and carefully opened before adding the fluorescent dye. Green fluorescence was clearly observed in successful LAMP reaction, whereas it remained original orange in the negative reaction. The LAMP products (3–5 μL) were also monitored using 2% agarose gel electrophoresis stained with ethidium bromide, visualized under UV light and then photographed using an ultraviolet image system (Gel documentation system, UVItec, UK).
The specificity of the LAMP assay to amplify only S. venezuelensis DNA was tested against 22 DNA samples obtained from other parasites used as controls as mentioned above. To determine the lower detection limit of the LAMP assay, genomic DNA from S. venezuelensis 10-fold serially diluted as mentioned above was subjected to amplification in comparison with the PCR using outer primers F3 and B3.
To evaluate the ability of the LAMP assay designed to amplify S. venezuelensis DNA in real samples, we used DNA extracted from the pooled feces and urine samples taken daily from each experimentally infected group of rats with different iL3 doses. To check whether LAMP assay designed was also able to amplify DNA from S. stercoralis in clinical samples, we used the patients´ stool samples included in the study. In all amplification assays, positive (S. venezuelensis DNA) and negative (DNA mix from non-infected rats stool or urine samples or ultrapure water) controls were always included.
In the three infected groups, parasite eggs were detected for the first time in feces on the 6th day p.i. regardless of the initial infecting doses (Fig 2). The maximal fecal egg count was 3,921 EPG on day 10 p.i in group 1 (Fig 2A), 12,092 EPG on day 9 p.i. in group 2 (Fig 2B) and 116,016 EPG on day 8 p.i. in group 3 (Fig 2C).
When a PCR verification reaction was performed using primers F3 and B3 to amplify S. venezuelensis DNA a 215 bp PCR product was successful amplified; the minimum amount of DNA detectable by PCR was 0.01 ng. When a panel of 22 DNA samples from other parasites were subjected to this PCR assay, amplicons were never obtained (S1 Fig).
Considering the most consistent color change by adding SYBR Green I into the tubes, the intensity of the ladder-like pattern on agarose gel electrophoresis as well as reproducibility of tests, the best amplification results were always obtained when the reaction mixtures contained Bst DNA polymerase 2.0 or Bst DNA polymerase 2.0 WarmStart combined with 1 M of betaine and supplementary 6 mM of MgSO4 and the reaction tubes were incubated at 63°C for 60 min. We also obtained amplification results when using Bst polymerase LF in such conditions, but the color change in reaction tubes as well as the intensity of the ladder-like pattern on agarose were always less evident compared to that obtained when using the other two enzymes; additionally, we did not get a good reproducibility of amplification trials so we discarded to use Bst polymerase LF in the following applications. When we evaluated the sensitivity of both LAMP reaction mixtures containing Bst polymerase 2.0 and Bst polymerase 2.0 WarmStart, the limit of detection in S. venezuelensis DNA amplification was 0.1 ng and 0.01 ng, respectively. As sensitivity was tenfold higher when using Bst polymerase 2.0 WarmStart, the reaction mixture containing this enzyme was used in assessing the specificity of the LAMP assay. Then, the LAMP assay was positive only for S. venezuelensis and no positive DNA products were observed when other parasites species were used as templates (S2 Fig). Thereby, the LAMP reaction mixture containing Bst polymerase 2.0 WarmStart was set up as the most suitable to analyze all the samples included in the study and thenceforth was namely Strong-LAMP.
In addition, all non-template controls were negative for each batch of LAMP reactions, thus indicating that there was no cross contamination and that with the primers set used there was no template free amplification [41, 42].
We tested by LAMP each daily pool of stool samples obtained from each infection group of animals during a 28-day period (Fig 3). To avoid possible cross-contamination LAMP assays were performed into two batches of 14 samples each. When testing stool samples from infected rats with 40 iL3 (group 1) we obtained LAMP positive results continuously from day 6 p.i. -when parasite eggs were detected in feces for the first time- until the end of infection at day 28 (Fig 3A). When testing stool samples from infected rats with 400 iL3 (group 2) and 4,000 iL3 (group 3) we obtained in both groups LAMP positive results continuously from day 5 p.i. -one day before the onset of parasite eggs in feces- until the end of infection at day 28 (Fig 3B and 3C). Negative controls (pooled DNA samples from feces from non-infected rats; group 4) were never amplified and in all LAMP positive reactions a green fluorescence was clearly visualized under natural light.
LAMP assay was also performed in each daily pool of urine samples obtained from each infection group of animals during a 28-day period (Fig 4). The 28 urine samples obtained from each infection group were tested in two batches of 14 samples each. Analyses of urine samples from rats infected with 40 iL3 (group 1) showed LAMP positive results on days 6, 11–14, 16–23 and 26 p.i. (Fig 4A). Analyses of urine samples from rats infected with 400 iL3 (group 2) showed LAMP positive results on days 3, 7–8, 10–23 and 25 p.i. (Fig 4B). Finally, analyses of urine samples from rats infected with 4,000 iL3 (group 3) showed LAMP positive results on days 3, 6–7 and continuously from day 9 until the end of infection on day 28 (Fig 4C). Four urine samples considered as positive results (including those obtained on day 19 from group 1, on day 20 from group 2 and on days 12 and 21 from group 3) did not show a color change to green fluorescent as appreciable as other LAMP positive results, but a faint ladder-like pattern could be observed on agarose gel electrophoresis. We obtained DNA amplification in pooled urine sample from group 1 on the 6th day (the same day that parasite eggs were detected in feces for the first time), but regarding group 2 and group 3, we detected DNA amplification in pooled urine sample on the 3rd day (two days before than onset of parasite eggs in feces). Curiously, LAMP positive results were not obtained on those days that the maximal fecal egg count was observed in each infection group (i.e., day 10 for group 1, day 9 for group 2 and day 8 for group 3, respectively).
All patients´ stool samples were tested by RT-PCR and LAMP to compare results. The RT-PCR resulted positive in 6/7 patients´ stool samples with confirmed strongyloidiasis by parasitological tests previously applied. In addition, a positive result was obtained in a sample to which none parasitological test could be performed; negative results were obtained in negative parasitological samples for S. stercoralis (Table 1). All patients´ stool samples with confirmed strongyloidiasis by parasitological tests could be detected by LAMP, including the sample which resulted negative in previous RT-PCR analyses. In addition, as for RT-PCR, we also obtained a positive result in the sample to which parasitological tests were not available. Negative parasitological patients´ stool samples for strongyloidiasis resulted in a negative LAMP amplification, including those positive samples for Taenia saginata and "hookworn", respectively (Table 1; S3 Fig).
There are many difficulties in correctly diagnosing strongyloidiasis because most patients are asymptomatic and the lack of sensitivity and specificity of the commonly used parasitological and serological diagnostic methods, respectively [43]. Several PCR-based molecular methods offering high sensitivity and specificity have been recently proposed in diagnosing strongyloidiasis [25, 22, 26, 27]. A LAMP method could be an economic, simple and applicable alternative to PCR-based methods in field conditions for diagnostic assays [44]. On the other hand, all new PCR-based approaches for Strongyloides spp. DNA detection have been always mainly assayed for stool samples from both experimentally infected animals and clinical stool samples [45, 46, 22, 25, 26, 27] but no other biological samples, such as urine, have been investigated for molecular diagnostic purposes. In our work, we used a S. venezuelensis rodent model in order to test a new LAMP assay for diagnosing strongyloidiasis both in stool and, for the first time, urine samples. We used an experimental infection with S. venezuelensis since this parasitic nematode has been widely used as a tool and laboratory model for human and animal strongyloidiasis research [47, 48]. The use of a S. venezuelensis rodent model allowed us to collect well-defined stool and urine samples that would otherwise have been very difficult to obtain from human patients, including samples from recently acquired infections and samples with low parasite load resembling to those likely obtained in chronic human infections. Additionally, a classical parasitological diagnostic method, such as direct faecal examination by counting EPG was used for monitoring infection as well as to compare results in parallel with molecular assays.
Results obtained by counting EPG showed a similar dynamics of S. venezuelensis infection to that previously reported not only by this parasite or by S. ratti in Wistar rats [49] but also in Lewis rats [46] and in male Sprague-Dawley rats [50].
To design specific primers for our LAMP assay, a 329 nucleotide bp from the 5´ end of a linear DNA partial sequence in the 18S ribosomal RNA gene from S. venezuelensis was selected [39]. For Strongyloides species, 18S ribosomal RNA gene (rDNA) has been analyzed [51, 52, 39, 53]. It is considered that small subunit ribosomal RNA (SSU rDNA) sequences within Strongyloides species are all very similar making the resolution of their phylogeny problematic as many branch lengths are inferred to be very short when distance and likelihood methods are applied [39, 53]. Closer analysis of the SSU rDNA sequences from a number of Strongyloides species have been shown to identify a putative molecular synapomorphy (comprising 8 to 10 nucleotides) within the E9-2 stem-loop of the V2 variable region, thus allowing to distinguish two clades within Strongyloides genus: one containing Strongyloides spp. ex snake, S. stercoralis and S. fuelleborni (namely "stercoralis" clade, with a 10 nucleotides sequence: ATTTTATATT), and another containing S. ratti, S. suis, S. venezuelensis, S. cebus, S. fuelleborni kelleyi and S. papillosus (namely "cebus" clade, with a 8 nucleotides sequence: ATT—TTTTC) [39]. Among the set of primers automatically generated when designing LAMP for specific amplification of S. venezuelensis, the F2 primer was finally manually selected to be used since its sequence at 3´ end -which location serve as the replication starting point after annealing- would allow not only a specific annealing in the 8 nuleotides sequence of "cebus" clade but also, theoretically, in the 10 nucleotides sequence of "stercoralis" clade if present in samples. Thus, the LAMP assay may be employed for simultaneous detection of several Strongyloides species. At present, only S. stercoralis and S. fuelleborni are known to cause infection in humans but infection with other species might be possible. Besides, the designed LAMP can also be use in the S. venezuelensis experimental infection rodent model.
After verifying the operation and specificity of PCR F3-B3, we attempted to establish the most suitable reaction mixture for the set of primers operation in the LAMP assay. The limit of detection of the LAMP assay resulted tenfold higher when Bst polymerase 2.0 WarmStart was used in comparison with Bst polymerase 2.0 (corresponding to 0.01 ng vs. 0.1 ng, respectively). It has been previously reported a number of advantages of Bst polymerase WarmStart version compared to other commercially available versions, such as faster amplification [54], increased stability at room temperature [55] and also greater sensitivity [56, 57]. We emphasize the importance of setting up the best conditions and molecular components for primers set operation in a LAMP assay. When analyzing the stool samples, the Strong-LAMP resulted more sensitive than microscopy, at least in moderate and high levels of infection. A similar result has been also reported for RT-PCR in comparison to microscopy in detecting first-stage larvae of S. ratti in a rodent model infected subcutaneously with 2,500 iL3 [58]. When analyzing the urine samples daily collected, we obtained Strong-LAMP positive results during the course of infection depending on infection dose.
In a work carried out by Marra et al. [46], in which the migration route of S. venezuelensis was evaluated by PCR and histological analysis in Lewis rats infected subcutaneously with 4,000 iL3, it was noted that the appearance of larvae in alveoli was already clear at 48 h p.i.. It was also observed that at 72 h p.i. all infected animals had larvae in the lungs and no larvae were found in any other organs that were examined. It was at this time, at 72 h p.i., when we obtained Strong-LAMP positive results in urine samples from groups 2 and 3, thus indicating the possibility of detecting S. venezuelensis free circulating DNA as a consequence of destroying larvae passing through the lungs and ending up in urine. Also according to that study, at 120 h p.i. larvae begin to disappear from the lungs and were found inserted in the small intestine villosities at 48–72 h later. Interestingly, it is also at this time in our study (approximately on the 9th day p.i.) when urine samples from group 3 (infected with 4,000 iL3) resulted Strong-LAMP positive every day until the end of infection.
In group 2 (400 iL3) a first positive result was also obtained by Strong-LAMP in the pool of urine samples at 72 h p.i.; however, a time lag in the appearance of positive results until the end of infection in comparison to group 3 (4,000 iL3) was detected, possibly related to the lower initial infective dose. Such time lag in the appearance of positive results was much more apparent when testing urine samples from group 1; since group 1 was infected with the lowest infective dose of larvae (40 iL3), a first positive result obtained on day 6 p.i. would suggest that parasites reached the lungs later and consequently it would take longer to settle them in the small intestine villosities.
Unexpectedly, we did not obtained Strong-LAMP positive results in urine samples on days in which the maximal fecal egg count was observed in each infection group. The absence of information on this event or similar in already published data does not allow us to compare our results. We can only speculate on the possibility of some features related to the dynamics of the biological cycle of the parasite.
The potential clinical applicability of the Strong-LAMP could be demonstrated on a number of human clinical stool samples. We obtained positive results in those stool samples with both parasitological demonstration and confirmed detection by RT-PCR of S. stercoralis. In addition, the analysis of one sample (no. 496) with no parasitological test applied but positive by RT-PCR, resulted positive by Strong-LAMP. Moreover, another sample (no. 069) with parasitological demonstration of S. stercoralis but negative by RT-PCR resulted also positive by Strong-LAMP; however, considering the limited number of human samples tested it is difficult at this time to suggest a potential greater sensitivity of LAMP assay than RT-PCR in detecting S. stercoralis DNA in stool samples. Furthermore, confirmed negative stool samples for S. stercoralis both by parasitological and RT-PCR methods resulted negative by Strong-LAMP as well, even those samples infected with T. saginata and hookworm, thus corroborating once again the specificity of our designed LAMP assay for exclusively detection of Strongyloides spp. DNA. Although specificity was also previously determined in silico by using a thoroughly BLASTN search and alignment analysis in online databases and no cross-reaction of other sequences reported for possible human pathogens were detected, it is important to note that, considering the absence of a single gold standard for strongyloides diagnosis, and because LAMP products cannot be routinely sequenced to confirm identity, other micro-organisms that may be commonly found in stool samples (e.g. bacteria and fungi, such as Candida spp.) should be investigated in order to further validate the LAMP assay for human diagnosis.
In this work, we report for the first time, on the development of a new LAMP assay (Strong-LAMP) for sensitive detection of S. venezuelensis DNA in both stool and urine samples in a well-established Wistar rats experimental infection model. In addition, this Strong-LAMP assay can be also applied effectively for the detection of S. stercoralis DNA in patients´ stool samples. Clearly, in terms of potential human diagnostic, this assay requires additional validation using a greater number of clinical stool samples.
The successfully amplification of Strongyloides spp. DNA in infected urine samples by LAMP assay as well as the advantages that urine would have in collection, storage and processing in comparison to patients´ stool samples, should make us consider the possibility of starting to use urine specimens in diagnosing human strongyloidiasis.
However, it will be convenient to further consider the difference between the samples from a S. venezuelensis rodent model in acute disease and chronic human S. stercoralis infections. Since urine is actually an unusual requested biological sample from patients to detect S. stercoralis, further studies using clinical urine samples for human diagnostics of strongyloidiasis are strongly-(LAMP) recommended.
|
10.1371/journal.ppat.1003708 | Regulation of Innate Responses during Pre-patent Schistosome Infection Provides an Immune Environment Permissive for Parasite Development | Blood flukes of the genus Schistosoma infect over 200 million people, causing granulomatous pathology with accompanying morbidity and mortality. As a consequence of extensive host-parasite co-evolution, schistosomes exhibit a complex relationship with their hosts, in which immunological factors are intimately linked with parasite development. Schistosomes fail to develop normally in immunodeficient mice, an outcome specifically dependent on the absence of CD4+ T cells. The role of CD4+ T cells in parasite development is indirect and mediated by interaction with innate cells, as repeated toll-like receptor 4 stimulation is sufficient to restore parasite development in immunodeficient mice in the absence of CD4+ T cells. Here we show that repeated stimulation of innate immunity by an endogenous danger signal can also restore parasite development and that both these stimuli, when administered repeatedly, lead to the regulation of innate responses. Supporting a role for regulation of innate responses in parasite development, we show that regulation of inflammation by neutralizing anti-TNF antibodies also restores parasite development in immunodeficient mice. Finally, we show that administration of IL-4 to immunodeficient mice to regulate inflammation by induction of type 2 responses also restores parasite development. These findings suggest that the type 2 response driven by CD4+ T cells during pre-patent infection of immunocompetent hosts is exploited by schistosomes to complete their development to reproductively mature adult parasites.
| Schistosomiasis is a devastating disease caused by Schistosoma blood flukes and is a leading parasitic cause of morbidity and mortality in the Developing World. The regulation of inflammatory responses to schistosome eggs trapped in tissues is critical for host survival and is established before egg deposition begins, with the production of the cytokine IL-4 being a hallmark of this process. Here we show that regulation of inflammatory responses also contributes to the development of schistosomes into egg-laying adult parasites. We demonstrate that failure of schistosome development in immunodeficient mice correlates with the absence of the chronic liver inflammation and subsequent immune regulation found in infected wild type mice. Restoration of liver inflammation in immunodeficient mice by repeated administration of liver toxins restored parasite development. Repeated administration of an endogenous inflammatory stimulus also restored parasite development, and also restored aspects of the immune regulation found in wild type mice. Finally, administration of IL-4 alone to immunodeficient animals also restored parasite development and the regulation of inflammation. We propose that schistosomes require immune regulation of inflammation to develop in the hostile immune environment within their hosts. Hence, targeting regulation of inflammation may represent a novel approach to treating or preventing schistosome infections.
| As a result of extensive host-parasite co-evolution, helminths exploit resources within their hosts to complete their development and ensure transmission to new hosts. Indeed, most helminths are obligate parasites, requiring the intra-host environment for successful life cycle completion. However, for the most part, the precise host factors that helminths require or utilize, in terms of host cells or molecules, are poorly defined. Previously, CD4+ T cells were shown to play a fundamental role in schistosome development [1]–[3], as significant impairment of parasite growth and reproductive activity occurred in mice that lack CD4+ T cells. While the precise mechanism by which CD4+ T cells mediate this effect is unclear, the mechanism is indirect, as chronic stimulation of innate immune responses with lipopolysaccharide (LPS), a toll-like receptor 4 (TLR4) agonist, during pre-patent infection was able to restore parasite development in the absence of CD4+ T cells [4]. Thus, all the host factors necessary for schistosome development are present, or at least can be induced, independently of CD4+ T cells. However, whether the mechanisms by which CD4+ T cells and chronic LPS stimulation restore schistosome development share any common elements has remained an open question.
Regulation of pro-inflammatory responses is critical for host survival of S. mansoni infection [5], and in response to schistosomes and other helminths, the immune system establishes robust T helper 2 (TH2) responses that modulate pro-inflammatory processes [6], [7]. In schistosomaisis, TH2 responses against parasite antigens are required for the formation of protective granulomas around parasite eggs [8], [9]. TH2 responses to worm antigens develop even before the onset of egg production [10], [11] and there is evidence that this immune priming by the developing worms is necessary to ensure proper TH2 granuloma formation [12]. TH2 responses are also critical for host survival after egg production begins, as lack of IL-4 signaling leads to severe disease and early mortality as a result of excessive pro-inflammatory processes [8], [9], [13]–[15]. Thus, in schistosomiasis, TH2 responses serve a dual purpose, to mediate granuloma formation and to regulate inflammation.
Here, we present evidence to suggest that, while fundamentally different, chronic innate responses in immunodeficient mice and adaptive responses in immunocompetent mice ultimately promote parasite development by resulting in a similar outcome, namely the establishment of an immunological milieu where inflammatory processes are regulated. These findings provide insights into the developmental requirements of schistosomes and may identify host dependencies that could be exploited to disrupt schistosome infection.
Although chronic LPS stimulation, administered twice weekly during the first six weeks of infection [4], can restore schistosome development in recombination activating gene-deficient (RAG−/−) mice, this stimulus is unlikely to be present at high concentrations in mouse plasma during pre-patent infection. We therefore sought to identify other inflammatory processes occurring during the pre-patent stage of schistosome infection in normal mice, as these might be candidates for a physiological stimulus for parasite development. Before the deposition of eggs, worm development in the portal vasculature is associated with both liver inflammation and hepatocellular necrosis of an unknown etiology [16], [17]. To determine whether this pathology also occurs in RAG−/− mice, we compared liver tissue sections from 4 week-infected RAG−/− and wild type mice. As previously reported [16], [17], wild type mice exhibited areas of coagulative necrosis and infiltration with inflammatory cells (Figure 1A). Inflammatory infiltrates, consisting of lymphocytes, eosinophils and mononuclear cells, were located in periportal areas and in the parenchyma, and also surrounded necrotic areas (Figure 1A). In contrast, areas of coagulative necrosis were not seen in the 4 week-infected RAG−/− mice (Figure 1B), with the exception of two animals where small foci of necrosis were detected (data points included in Figure 1C). Furthermore, we observed very little liver inflammation in 4 week-infected RAG−/− mice (Figure 1B), which, when present, was restricted to areas around or near vessel walls (figure 1B). The mean percentage area occupied by necrotic liver tissue was 1.6% for wild type animals, while RAG−/− mice exhibited almost none (Figure 1C). Likewise, the mean percentage area occupied by inflammatory infiltrates was 8.7% in wild type mice, while RAG−/− mice exhibited almost none (Figure 1D). These data show that failure of parasite development in RAG−/− mice correlates with a lack of liver necrosis and inflammation. Furthermore, while the etiology of the liver necrosis in wild type mice is unknown, these data suggest that death of hepatocytes during pre-patent schistosome infection requires an intact adaptive immune system.
In view of the fact that a failure of parasite development in RAG−/− mice correlates with a lack of liver necrosis and inflammation, we hypothesized that induction of liver necrosis and inflammation would restore parasite development in these animals. To test this hypothesis, we administered two well-characterized hepatotoxins, acetaminophen (AAP) or D-galactosamine (GalN), to RAG−/− mice throughout pre-patent infection, at doses sufficient to result in hepatocellular death and inflammation [18]–[21], in an attempt to simulate the cell death observed in wild type mice. While chronic hepatotoxin treatment did not recapitulate the coagulative necrosis seen in wild type mice, both treatments induced histological evidence of widespread hepatocellular damage and inflammation (supplementary Figure S1). Furthermore, by six weeks post infection (p.i.), both treatments partially restored parasite growth when compared to control RAG−/− animals that received vehicle alone (Figure 2A and 2B), while parasite egg production was unaffected (essentially none, data not shown). These results suggested that liver necrosis and inflammation played a role in modulating schistosome development.
Cellular injury results in the release of uric acid into the extracellular environment that crystallizes to form monosodium urate (MSU) [22], an endogenous danger-associated molecular pattern (DAMP) that activates the NALP3 inflammasome [23]. Since chronic hepatotoxin treatment partially restored parasite development in RAG−/− mice, we hypothesized that DAMP-mediated inflammatory processes would also restore parasite development in these animals. To test this hypothesis, we administered MSU to infected RAG−/− mice throughout the first six weeks of infection and compared worm development to that in control RAG−/− mice that received vehicle alone. Treatment with MSU resulted in robust restoration of parasite growth (Figure 2C) and partial restoration of egg production (Figure 2D) in RAG−/− mice suggesting that, like LPS, chronic DAMP-mediated inflammation can also stimulate parasite development. In further support of a role for inflammasome-mediated inflammation in stimulating parasite development, we also found that treatment with alum, an exogenous NALP3 inflammasome agonist [24], [25], also resulted in robust restoration of parasite growth (Figure 2E) and partial restoration of egg production (Figure 2F) in RAG−/− mice. Taken together, these data suggest that, like the exogenous danger signal LPS, endogenous danger signals that stimulate inflammation via inflammasomes can also stimulate schistosome development.
By serving as a molecular platform for caspase 1 activation, inflammasomes drive IL-1β-mediated inflammation by catalyzing the conversion of inactive pro-IL-1β to the bioactive form [26]. As two different inflammasome agonists restored parasite development in RAG−/− mice, we hypothesized that IL-1β-mediated inflammation may be implicated in parasite development. To address this issue, we first examined IL-1β mRNA levels during pre-patent infection of wild type mice. Unexpectedly, we found that steady-state splenic mRNA levels of IL-1β in wild type mice at 3 and 4 weeks p.i. were down-regulated compared to the baseline levels found in non-infected control mice (Figure 3A). In contrast, IL-1β mRNA levels remained unchanged in the spleens of 4 week-infected RAG−/− mice when compared to non-infected controls (Figure 3B). Thus, normal parasite development correlated with down-regulation of steady-state IL-1β transcription. To test whether a failure to down-regulate IL-1β signaling in RAG−/− mice is the cause of impaired schistosome development in these animals, we infected RAG−/− IL-1R−/− knockout mice, predicting that, if this were the case, ablation of IL-1 signaling would restore parasite development in a RAG-deficient context. However, worms recovered from RAG−/− IL-1R−/− mice 6 weeks p.i. did not differ significantly from those obtained from RAG−/− mice, being small in size (Figure 3C) and reproductively inactive (Figure 3D). Therefore, parasite development correlates with regulation of IL-1β transcription, but IL-1R signaling is not directly responsible for inhibiting parasite development in RAG−/− mice.
As our examination of IL-1β transcription in wild type and RAG−/− mice revealed that normal schistosome development correlated with down-regulation of IL-1β transcription in wild type mice, we next examined the effect of the LPS and MSU treatment regimens on IL-1β transcription in RAG−/− mice, as both treatments restore parasite development in these animals. Repeated treatment with LPS (Figure 4A) or MSU (Figure 4B) throughout the first six weeks of infection resulted in down-regulation of splenic IL-1β mRNA levels in infected RAG−/− mice by week six p.i., similar to the down-regulation seen in infected wild type mice (Figure 3A). Thus, while down-regulation of IL-1β transcription is mediated by the adaptive immune system in wild type mice, chronic administration of LPS or MSU to RAG−/− mice can also result in IL-1β down-regulation, in the absence of an adaptive immune system. Furthermore, chronic LPS and MSU treatments resulted in down-regulation of other pro-inflammatory signals, as evidenced by reduced splenic mRNA levels for TNF (Figure 4C and 4D) and CCL2 (Figure 4E and 4F), a chemokine important for inflammatory macrophage recruitment [27]–[29]. Finally, we found that transcriptional down-regulation of pro-inflammatory genes to levels lower than those in control animals required chronic exposure to MSU (supplementary Figure S2 A–C), as transcription of pro-inflammatory genes peaked rapidly following a single injection of MSU (data not shown) and then returned to the levels observed in non-treated mice by 18 hours post injection (supplemental Figure S2 D–F).
Our analysis of wild type, RAG−/− and MSU- and LPS-treated RAG−/− mice showed that normal parasite development correlates with the overall down-regulation of pro-inflammatory gene transcription. However, we considered the possibility that elevated pro-inflammatory gene transcription early in the course of LPS or MSU treatment, before regulation was induced, could be the factor important for stimulating parasite development, rather than the ultimate down-regulation of these genes. To explore this possibility, we sought to identify comparable treatments where chronic administration of an inflammatory stimulus did not result in down-regulated pro-inflammatory gene transcription. To this end, we found that administration of poly I:C [30], a TLR3 ligand, to RAG−/− mice throughout the first six weeks of infection, resulted in up-regulation of splenic mRNA for IL-1β, TNF, and CCL2 (Figure 5 A–C), rather than their down-regulation. Consistent with a role for pro-inflammatory gene down-regulation in permitting schistosome development, chronic administration of poly I:C also failed to restore parasite development in RAG−/− mice, as worms recovered from treated animals did not differ significantly in size (figure 5D–E) or egg output (essentially none, data not shown) from vehicle-treated controls at either of the two doses tested.
As restoration of parasite development in RAG−/− mice correlated with down-regulation of pro-inflammatory gene transcription (in LPS- and MSU-treated RAG−/− mice) and was unaltered when pro-inflammatory gene transcription was sustained (in poly I:C-treated RAG−/− mice), we hypothesized that it was the down-regulation of pro-inflammatory genes that permits parasite development to proceed. To test this hypothesis, we attempted to suppress pro-inflammatory gene activity in RAG−/− mice by blocking TNF signaling with a neutralizing antibody, which has been shown to decrease IL-1β production in models of sepsis [31] and arthritis [32]. Administration of the anti-TNF antibody did not cause acute increases in pro-inflammatory gene transcription (data not shown), and its administration throughout pre-patent infection led to the down-regulation of IL-1β, TNF, and CCL2 transcription in the spleens of RAG−/− mice by four weeks p.i. (Figure 6A–C), similar to that observed in wild type mice and with chronic LPS or MSU treatment in RAG−/− mice. Furthermore, anti-TNF treatment resulted in significant increases in parasite size (Figure 6D) and reproductive activity (Figure 6E) when compared to control RAG−/− mice. Thus, these data supported our hypothesis that schistosome development requires the down-regulation of pro-inflammatory gene transcription during pre-patent infection.
In S. mansoni-infected wild type mice, down-regulation of IL-1β transcription occurs via an adaptive immune mechanism, as there is a failure of IL-1β mRNA down-regulation when the adaptive immune system is ablated (Figure 3). In previous studies, we showed that, prior to the onset of egg production, pre-patent schistosome infection results in the rapid establishment of a TH2 response [10], [11], where CD4+ T cells produce significant quantities of IL-4 in response to worm antigens [10]. As IL-4 is a type 2 cytokine that regulates pro-inflammatory signals, including IL-1β [33]–[35], we hypothesized that IL-4 may represent the adaptive mechanism by which IL-1β is regulated in wild type mice. To test whether IL-4 was sufficient to regulate pro-inflammatory gene transcription in RAG−/− mice, we administered IL-4 complex (IL-4c) to infected RAG−/− mice during pre-patent infection and examined pro-inflammatory gene transcription in the spleen at week six p.i. Administration of IL-4c resulted in down-regulation of IL-1β and TNF transcription in the spleens of treated RAG−/− mice (Figure 7A and 7B), similar to that observed in wild type mice and RAG−/− mice treated with LPS, MSU or anti-TNF. Transcription of CCL2 was also reduced by IL-4c treatment, although the difference between treated and control animals was not significant (Figure 7C). The transcription of RELM-α and YM1, both markers of alternative macrophage activation [36], was dramatically up-regulated in the spleens (data not shown) and the livers of IL-4c-treated animals (Figure 7D and 7E), suggesting that IL-4c treatment induced an innate type 2 response and the accumulation of alternatively activated (M2) macrophages in livers of RAG−/− mice. Indeed, administration of IL-4c also restored the accumulation of mononuclear cells in the livers (Figure 7F) of infected RAG−/− mice and induced giant cell formation (Figure 7G), a previously reported hallmark of alternatively activated macrophage responses [37]. Finally, administration of IL-4c resulted in the restoration of parasite growth (Figure 7H) and reproductive activity (Figure 7I). Our data demonstrate that IL-4, a type 2 cytokine produced as part of the adaptive immune response to pre-patent schistosome infection, is sufficient to regulate pro-inflammatory signals and restore schistosome development and may represent the mechanism by which CD4+ T cells permit normal parasite development in wild type mice.
Our finding that IL-4 administration restored parasite development in RAG−/− mice, in a manner similar to chronic LPS or MSU administration and TNF blockade, prompted us to examine whether these other treatments, like IL-4, promoted M2 conditioning of macrophages, perhaps by stimulating release of endogenous IL-4 from innate immune cells. However, levels of RELM-α transcript in the liver were unchanged by either chronic PAMP (LPS, poly I:C) or DAMP (MSU) treatment, or by TNF blockade (Figure 8A–D), suggesting these treatments do not induce M2 responses and that parasite development does not require M2 conditioning of macrophages per se. To further test the potential involvement of innate sources of IL-4 in restoring parasite development during chronic immune stimulation, we tested whether LPS could restore schistosome development in RAG−/− mice that are also deficient for the common γ chain (γc) [38], a critical component of the IL-4 receptor complex. Chronic LPS treatment of infected RAG−/−/γc−/− mice restored parasite growth to levels comparable with those observed in LPS-treated RAG−/− mice (Figure 8E), indicating that IL-4 signaling is not required for restoration of schistosome development by chronic LPS administration. Together, these findings suggest it is the regulation of pro-inflammatory processes by IL-4, rather than IL-4-driven M2 macrophage responses, that is critical in permitting schistosome development to proceed.
Numerous lines of evidence indicate that type 2 responses are beneficial in schistosome infection, not because these responses mediate immunity against schistosomes but because they limit potentially damaging pro-inflammatory responses. For example, in mice deficient in IL-4, IL-4 and IL-10, IL-4 and IL-13 or IL-4 receptor, decreased host survival is observed during acute schistosome infection due to excessive pro-inflammatory cytokine expression and increased liver and intestinal pathology [8], [9], [14], [39]. Likewise, in schistosomiasis patients, severe disease is correlated with decreased production of type 2 cytokines and elevated levels of IFN-γ, TNF and nitric oxide [40]. However, there is also evidence that type 2 responses ultimately benefit schistosomes, and that this benefit extends beyond the obvious relationship between extended host survival and the increased likelihood of transmission to snail intermediate hosts. For example, it has long been recognized that egress of schistosome eggs across the bowel wall is immune-dependent [41]. Subsequent macrophage-specific ablation of IL-4R expression showed that IL-4/IL-13-responsive macrophages are specifically required for egg passage into the intestinal lumen [39]. These observations suggest that host-parasite co-evolution has not only selected for immune responses that prolong host survival, but also for parasites that are able to take advantage of the resulting immunological milieu. The data we present here support the hypothesis that control of pro-inflammatory signals may also be intimately linked to parasite development before the onset of egg production.
Our finding that chronic stimulation with LPS could restore schistosome development in RAG−/− mice presented a paradox, as there is no obvious parallel between the inflammatory response to LPS and the response induced by pre-patent schistosome infection in wild type mice. However, inflammation, albeit in response to necrotic hepatocytes, is a feature of pre-patent schistosome infection in immune-competent, but not RAG−/− mice, suggesting there is a link between inflammation and normal parasite development. In support of this hypothesis, we show that restoration of DAMP-mediated inflammation in RAG−/− mice also restored parasite development. While endogenous DAMPs stimulate inflammation by pathways distinct from exogenous PAMPs such as LPS [42], our finding that both can restore parasite development suggests it is the inflammation itself rather than the inciting cause that is relevant to parasite development. The contribution of necrosis-induced inflammation to promoting parasite development in wild type mice is an interesting and unresolved question. One way to address this question may be to inhibit necrosis in wild type mice and examine for effects on parasite development. Identification of the mechanism leading to hepatocellular necrosis in wild type mice may make this approach possible. The absence of necrosis in RAG−/− mice suggests that adaptive responses are involved in necrosis induction. Alternatively, the lack of necrosis in these animals may be a result of diminished parasite growth, rather than a cause. However, we have not observed hepatocellular necrosis in RAG−/− mice where parasite development is restored by LPS treatment (data not shown), suggesting that developing parasites do not directly cause liver necrosis.
Our observation that steady state transcription of the pro-inflammatory cytokine IL-1β is down-regulated in infected wild type mice, but not in RAG−/− animals, further suggested a role for inflammatory processes in schistosome development, but in an inhibitory capacity. However, the non-permissiveness of RAG−/− mice for parasite development is not specifically due to a failure to down-regulate IL-1 signaling, as ablation of IL-1R activity in a RAG−/− context did not restore parasite development. This result led us to hypothesize that parasite development may require a more global regulation of pro-inflammatory processes and that decreased IL-1β transcription was simply a correlate of this regulation. If this was correct, we predicted that restoration of parasite development in RAG−/− mice by chronic LPS and MSU administration would be associated with transcriptional regulation of IL-1β and other pro-inflammatory genes. The induction of LPS tolerance in response to repeated LPS exposure is a well-recognized negative feedback mechanism thought to be mediated by a variety of mechanisms, including regulation of downstream protein kinases, resulting in down-regulation of inflammatory cytokine transcription [43]. We show here that chronic MSU exposure also results in down-regulation of pro-inflammatory gene transcription. While MSU signals via pathways distinct from LPS [44], the existence of negative feedback mechanisms that regulate persistent MSU signaling is not unexpected and evidence for the induction of regulatory mechanisms by endogenous DAMPs exists. For example, toxicological injury to the liver by AAP first results in an early pro-inflammatory response dominated by classically activated (M1) macrophages, but this initial response is followed by suppression of the initial pro-inflammatory response and promotion of wound healing by immunoregulatory, alternatively activated (M2) macrophages [45], [46]. Second, both alum and uric acid have been shown to promote TH2 immunity and suppression of pro-inflammatory processes through NALP3 independent mechanisms [47], [48]. Similarly, LPS tolerance is associated with induction of immunoregulatory M2 macrophages [49]. Based on our findings, we suggest that LPS and MSU restore parasite development in RAG−/− mice by virtue of their ability to induce regulation of pro-inflammatory signals when administered chronically.
To further test whether inflammation per se, or the regulation that results from the inflammation was required for parasite development, we sought to identify inflammatory stimuli that did not lead to regulation. We found that the TLR3 ligand poly I:C, even when administered repeatedly under the same regimen as LPS or MSU, failed to reduce baseline levels of inflammatory gene transcription, resulting instead in overall elevated levels of transcription, even 18 hours post administration of the final dose. Unlike other TLR ligands like LPS, poly I:C does not stimulate MyD88-dependent signaling, utilizing instead a TRIF-dependent pathway that appears not to be subject to the same negative feedback regulation [50]. Chronic poly I:C administration therefore afforded us the opportunity to examine schistosome development in the context of persistent inflammation without the associated regulation. Consistent with a role for regulation in parasite development rather than inflammation, chronic administration of poly I:C, at two different doses, failed to enhance parasite development.
Because poly I:C signaling is mediated via a distinct receptor and adapter molecule [50], we cannot exclude the possibility that some essential component of the response induced by MSU or LPS is absent from the response to poly I:C. However, we reasoned that if regulation of inflammation was the critical element in permitting parasite development, then direct regulation of inflammation in the absence of exogenous inflammatory stimuli would also be able to restore schistosome development. The success of anti-TNF neutralization therapy in controlling inflammatory disorders is due to the ability of this intervention to broadly control inflammation, mediated by TNF and associated signals, including IL-1β [32]. In the absence of any additional inflammatory stimuli, administration of anti-TNF antibodies to infected RAG−/− mice recapitulated the regulation of inflammatory gene transcription observed after chronic LPS or MSU administration and also restored schistosome development, lending further support to our conclusion that regulation of inflammation is required for normal parasite development.
In an immunocompetent host, pre-patent schistosome infection induces a TH2 response, characterized by production of IL-4 by CD4+ T cells in response to schistosome antigens [10]. In addition to driving TH2 effector mechanisms such as antibody isotype class switch recombination in B cells and M2 macrophage development, IL-4 also regulates pro-inflammatory signals [33] and is therefore a likely regulator of pro-inflammatory processes during pre-patent schistosome infection. Consistent with this role, IL-4 administration to RAG−/− mice was sufficient to restore regulation of IL-1β and TNF transcription and induce type 2 responses, as evidenced by the dramatic up-regulation of RELM-α and YM1 transcripts. Significantly, IL-4 treatment also resulted in increased schistosome growth and egg production, demonstrating that a single type 2 cytokine was sufficient to significantly augment parasite development. From these results, it is tempting to speculate that regulation mediated by IL-4 may be the principle contribution of the adaptive response in wild type mice to permitting parasite development to proceed. However, IL-4 is unlikely to be the only adaptive immune factor to promote schistosome development, as parasite development proceeds normally when IL-4 signaling is blocked in otherwise immunologically intact mice, whether by anti-IL-4 antibody or through genetic disruption [1]. Furthermore, we previously showed that CD4+ T cells that lack specificity for schistosome antigens and cannot respond to schistosome infection can still positively influence parasite development [4]. There is no evidence this non-cognate effect requires IL-4, but is likely a consequence of homeostatic interactions between CD4+ T cells and innate antigen-presenting cells that modulate myeloid cell function [4]. Thus, there is likely considerable latitude in the requirement of schistosomes for regulation, as illustrated by the diversity of mechanisms by which parasite development can be restored in RAG−/− mice (chronic administration of LPS or MSU, or administration to anti-TNF antibody or recombinant IL-4). However, it remains a possibility that the host elements required for schistosome development may be common to the responses induced by each of these mediators.
As there is evidence that alternatively activated M2 macrophage rather than classically activated M1 macrophage responses are favored under conditions of chronic immune stimulation [49], [51]–[55], it is tempting to hypothesize that M2 macrophages are critical for parasite development. However, we could find no evidence that our repeated treatment of RAG−/− mice with LPS, MSU or anti-TNF antibody led to M2 induction. Thus, while the type 2 response and possibly M2 macrophages may represent host factors that schistosomes co-opt to complete development in immunologically intact mice, our data suggest there are other more fundamental aspects of innate responses that schistosomes exploit, rather than M2 macrophages per se. For example, LPS tolerization and alternative macrophage activation induce overlapping changes in macrophage function, including the down-regulation of pro-inflammatory mediator expression and production of toxic reactive oxygen intermediates, that in both scenarios is mediated by a common regulatory pathway that involves NF-κB p50 [49]. Thus, it is possible that schistosomes have evolved to take advantage of the regulatory aspects of type 2 responses that modulate pro-inflammatory responses.
Why regulation of pro-inflammatory signals would influence schistosome development remains to be determined. Allen and Wynn recently suggested that TH2 immunity evolved in order to supply a rapid response that repairs the tissue damage generated by helminths [56], highlighting that the need for regulation of innate immune responses is not limited to schistosome infections, but is a common feature of immune responses directed towards tissue-penetrating helminths. For example, M2 macrophages and TH2 responses are critical for limiting lung tissue damage after experimental Nippostrongylus brasiliensis infection [57], by controlling initial IL-17-driven inflammatory responses and promoting resolution of tissue damage [58]. M2 macrophages have also been shown to limit brain tissue pathology in a murine model of neurocysticercosis [59] [60], where decreased numbers of brain M2 macrophages in Mesocestoides corti-infected mice was shown to result in increased disease severity. The induction of systemic TH2 responses has been shown to occur early in S. mansoni infection [10], [11]. Indeed, multiple exposure of skin to invading cercariae, as likely occurs under field conditions, has been shown to be sufficient to induce M2 conditioning of macrophages at the site of infection [61]. Thus, type 2 responses are induced sufficiently early during infection to exert an effect on the developing schistosomes. There is already evidence that schistosomes require M2 macrophages later in infection, as these cells are critical for the egress of schistosome eggs from the body of the host [39], and are required for host survival after egg production begins [8], [13]–[15]. M2 macrophages in particular are of critical importance in the regulation of excessive egg-induced inflammation and the lack of M2 macrophages during acute infection is lethal, as shown by the macrophage-specific ablation of IL-4Rα expression [39]. As M2 macrophages are a specific hallmark of the host response to schistosomes and other helminths [56], the hypothesis that schistosomes have evolved to specifically exploit this aspect of the host response is an attractive one.
In addition to regulating inflammatory processes that may damage developing schistosomes, there are other mechanisms by which type 2 or regulatory responses might contribute to schistosome development. As mediators of tissue repair and remodeling, one possibility is that M2 or immunomodulatory macrophages mediate critical niche remodeling in portal venules where the rapidly growing schistosomes reside, akin to the lymphatic vascular remodeling induced by filarial nematodes [62]. Alternatively, the presence of M2 or immunomodulatory macrophages may alter the availability of host-derived nutrients or other molecules that the developing parasites require [63]. Macrophage activation status is associated with profound changes in cell metabolism [64] that could influence the concentrations of host factors in the immediate vicinity of larval schistosomes. Finally, molecules associated with immunoregulation and type 2 responses may be directly recognized by schistosomes and utilized as signals of an environment that is appropriate for parasite development. A somewhat similar relationship was recently proposed to influence the development of Litomosoides sigmodontis, a filarial nematode, which accelerates its larval development and produces greater numbers of microfilaria in response to IL-5 and eosinophils [65]. It was suggested that these parasites utilize IL-5 as a predictor for future survival and altered life expectancy [65]. Studies to examine these and other possibilities in the context of schistosome infection are currently underway.
Here we presented evidence that regulation of pro-inflammatory processes is a contributor to determining the developmental fate of schistosomes in their definitive mammalian host. It remains to be determined how schistosomes might recognize a regulated immune environment and how this environment influences parasite development. However, these findings suggest that inflammation and its regulation are key components of a host environment permissive to schistosome infection. Thus, modulation of inflammatory processes may be a useful approach to disrupting schistosome infection and could lead to new insights for improved treatments or vaccines for schistosomiasis.
All animal studies were performed in strict accordance with the recommendations of the Office of Laboratory Animal Welfare at the National Institutes of Health and the USUHS Institutional Animal Care and Use Committee. All animal protocols were reviewed and approved by the USUHS Institutional Animal Care and Use Committee, permit number A3448-01.
RAG-1−/− mice on a C57BL/6 background [66] were originally purchased from Jackson laboratory (Bar Harbor, ME) and then bred in-house for experimental use. Wild type C57BL/6 mice were purchased from the National cancer institute (NCI, Frederick, MD). RAG-1−/− IL-1R−/− were generated by crossing C57BL/6 RAG-1−/− to C57BL/6 IL-1R−/− mice [67] purchased from Jackson laboratory (Bar Harbor, ME). The RAG-1−/− IL-1R−/− genotype was confirmed via PCR. All mice used in experiments were age matched.
Mice were infected percutaneously via tail exposure to water containing 160 S. mansoni cercariae (Puerto Rican strain) shed from infected Biomphalaria glabrata snails. Infections were terminated at 6 weeks post infection (p.i.). Worms were perfused from the portal system and immediately fixed in 4% neutral buffered formaldehyde. Male and female worms were counted and photographed at 20×magnification using a Nikon D80 10.0 megapixel digital camera attached to a Zeiss trinocular dissecting microscope. Worm growth was assessed by measuring the length of male worms from digital micrographs using Image J software (http://rsb.info.nih.gov/ij), as described previously [4]. Only male worms were measured as female growth is dependent upon receiving developmental cues from pairing with maturing males [68]. Length of male worms was compared to that of worms recovered from wild type C57BL/6 mice at 6 weeks p.i., as reported in previous publications [1], [4], [69] and in unpublished data. Fecundity of the parasites was assessed by calculating egg production per worm pair from liver egg burdens, as described previously [4].
C57BL/6 wild type and RAG-1−/− mice were infected with cercariae as described above. At 4 weeks p.i., mice were sacrificed and their livers removed and immediately fixed in 35 ml of 4% neutral buffered formaldehyde. Liver sections were cut and stained with hematoxylin and eosin stain (H&E stain, Histoserv INC.,Germantown, MD). Slides were digitally scanned using the Hamamatsu Nanozoomer 2.0RS (Hamamatsu City, Japan). Tissue sections were analyzed using the Nanozoomer digital pathology (NDP) software. At low magnification, areas of liver tissue measuring 20 mm2 total area were randomly selected. The selected areas were then scanned at 5×magnification and the area occupied by inflammatory infiltrate was measured. For each liver, sections were obtained at 3 different levels and measurements were taken for at least 3 different tissue sections. The percentage area of inflammation was determined by summing up the area occupied by inflammatory infiltrate and dividing it by the total area examined. The percentage area occupied by coagulative necrosis was determined by the same method.
RAG-1−/− mice were infected with cercariae as described above. Mice received weekly intraperitoneal (i.p.) injections of AAP (Sigma-Aldrich, St Louis, MO) at a dose of 5 mg/mouse dissolved in 100 µl 30% DMSO for the first 3 weeks. Mice then received AAP at a dose of 10 mg/mouse dissolved in 30% DMSO for the remaining 3 weeks. Control mice received weekly i.p. injections of 30% DMSO. D-GalN (MP Biomedicals, Solon, OH) -treated mice received biweekly i.p. injections at a dose of 10 mg/mouse for 6 weeks, using PBS without calcium and magnesium as a vehicle. Control mice received biweekly i.p. injections of PBS (Mediatech, Manassas, VA) alone. At 6 weeks p.i. mice were euthanized, H&E staining of livers was performed to confirm liver inflammation in treated mice, and parasite parameters were determined as described above.
RAG-1−/− mice were infected with cercariae as described above. Mice received biweekly i.p. injections of MSU (Invivogen, San Diego, CA) at a dose of 500 µg/mouse, Imject Alum (Thermo Scientific, Rockford, IL) at a dose of 1 mg/mouse, ultrapure LPS, E.coli 0111:B4 (Invivogen) at a dose of 20 µg/mouse, or poly I:C 20 µg or 40 µg/mouse. Control mice received biweekly i.p. injections of PBS without calcium and magnesium. At 6 weeks p.i., mice were euthanized and parasite parameters were determined as described above.
RAG-1−/− mice were infected with cercariae as described above. Mice received weekly i.p. injections of Adalimumab (Abbott, Chicago, IL) at a dose of 100 µg/mouse, using PBS without calcium and magnesium as a vehicle. Control mice received weekly i.p. injections of PBS alone. At 6 weeks p.i., mice were euthanized and parasite parameters were determined as described above.
RAG-1−/− mice were infected with cercariae as described above. Mice received weekly i.p. injections of 5 µg IL-4 (Peprotech, Rocky Hill, New Jersey) complexed to 25 µg anti-IL-4 antibody 11B11 (BioXCell, West New Lebanon, New Hampshire) [70]. Control mice received weekly i.p. injections of the isotype control antibody HRPN (BioXCell). At 6 weeks p.i., mice were euthanized, livers were removed for histology, and parasite parameters were determined as described above.
RNA was isolated from the spleens and/or livers of wild type or RAG-1−/− mice. After removal, tissues were immediately placed in 1 ml RNA-BEE (Tel-Test, Friendswood, Texas), homogenized, snap-frozen in liquid nitrogen and stored at −80°C until isolation of total RNA, following manufacturer's instructions. RNA was further purified following the RNeasy mini protocol for RNA cleanup (Qiagen, Valencia, California). Purified RNA was quantified using a ND-1000 spectrophotometer (Nanodrop, Wilmington, DE). 2 µg of RNA was used for cDNA preparation using a high capacity RNA to cDNA kit (Invitrogen, Grand Island, New York) following manufacturer's instructions. Real time PCR was performed with a MJ Research Chromo4 PTC-200 thermocycler unit (Bio-Rad, Hercules, CA) using Taqman gene expression assays and TaqMan gene expression master mix (Invitrogen) following manufacturer's instructions. Assays for the following mRNAs were performed: rsp29, GAPDH, IL-1β, TNF-α, CCL2, Relm-α, and YM1. Expression of genes of interest was normalized to the expression of GAPDH or rsp29 and fold changes in expression were calculated following the 2−ΔΔCT method [71].
All statistical analyses were performed using GraphPad Prism Inc. version 4.0 software (San Diego, California). Significant differences between two groups were determined using a student's unpaired T-test with Welch's correction or a Mann-Whitney test. Significant differences between 3 or more groups were determined using a Kruskal-Wallis test followed by a Dunn's multiple comparison test. P values of less than 0.05 were considered significant. 3–5 mice were used per experimental group and all experiments were performed at least twice.
|
10.1371/journal.pntd.0003477 | Identification of the Mycobacterium ulcerans Protein MUL_3720 as a Promising Target for the Development of a Diagnostic Test for Buruli Ulcer | Buruli ulcer (BU) caused by Mycobacterium ulcerans is a devastating skin disease, occurring mainly in remote West African communities with poor access to health care. Early case detection and subsequent antibiotic treatment are essential to counteract the progression of the characteristic chronic ulcerative lesions. Since the accuracy of clinical BU diagnosis is limited, laboratory reconfirmation is crucial. However, currently available diagnostic techniques with sufficient sensitivity and specificity require infrastructure and resources only accessible at a few reference centres in the African endemic countries. Hence, the development of a simple, rapid, sensitive and specific point-of-care diagnostic tool is one of the major research priorities for BU. In this study, we have identified a previously unknown M. ulcerans protein, MUL_3720, as a promising target for antigen capture-based detection assays. We show that MUL_3720 is highly expressed by M. ulcerans and has no orthologs in other prevalent pathogenic mycobacteria. We generated a panel of anti-MUL_3720 antibodies and used them to confirm a cell wall location for MUL_3720. These antibodies could also specifically detect M. ulcerans in infected human tissue samples as well as in lysates of infected mouse footpads. A bacterial 2-hybrid screen suggested a potential role for MUL_3720 in cell wall biosynthesis pathways. Finally, we demonstrate that a combination of MUL_3720 specific antibody reagents in a sandwich-ELISA format has sufficient sensitivity to make them suitable for the development of antigen capture-based diagnostic tests for BU.
| According to the recommendations of the World Health Organization, the clinical diagnosis of BU should be reconfirmed by at least two laboratory techniques. However, out of the four currently available tests, three (PCR, histopathology and cultivation of M. ulcerans) can only be performed at centralized reference laboratories; the fourth (microscopic detection of acid fast bacilli) lacks the required sensitivity and specificity. Therefore, a simple tool for early diagnosis of the disease, which can be implemented in rural health care facilities of the endemic countries, is of urgent need. In this study we aimed at the identification of M. ulcerans proteins as potential targets for the development of a simple and rapid diagnostic antigen detection assay. Among 36 proteins, MUL_3720 best met the predefined criteria of being highly expressed by M. ulcerans and not having orthologs in other pathogenic mycobacterial species prevalent in the endemic regions. Here we generated monoclonal and polyclonal antibodies against this protein and carried out pilot studies for the development of an antigen capture-based diagnostic test.
| Buruli ulcer (BU) is a neglected mycobacterial skin disease, reported from tropical and subtropical countries world-wide with highest incidence rates in Western Africa [1]. Populations in rural areas with limited access to health facilities are most affected and often seek medical advice at late disease stages [2]. Advances in the clinical management of BU have shifted options for treatment from surgical resection to combination antibiotic therapy [1]. While PCR analysis targeting the insertion sequence IS2404 has evolved into the gold standard for laboratory diagnosis of BU, this test is only available at a few reference centres. Therefore, the diagnosis of BU is currently often based on clinical findings and antibiotic therapy is started before laboratory diagnostic results can be obtained. BU has a wide range of clinical manifestations including non-ulcerative forms such as subcutaneous nodules or papules, plaques and oedema, which may progress to chronic ulcerative lesions. Due to this diversity of disease presentations the accuracy of clinical diagnosis is limited [1, 3–5] and thus a significant proportion of patients reporting with skin lesions may not receive adequate treatment. This includes cases of cutaneous tuberculosis which may be misdiagnosed as BU and thus receive the recommended eight week course of Streptomycin/Rifampicin combination chemotherapy for BU [5], which is much too short for the treatment of tuberculosis. As for IS2404 PCR, two of the other three currently applied methods for laboratory reconfirmation of BU—histopathology and cultivation of the extremely slow-growing mycobacteria—equally require expensive equipment and expertise [4, 6–8] not accessible at peripheral health facilities. The only available point-of-care diagnostic test, direct-smear examination by microscopy for the detection of acid fast bacilli (AFB), has limited sensitivity and specificity [6]. Hence, one of the major research priorities for BU is the development of a fast, low-tech, sensitive and specific point-of-care diagnostic test, which can be directly implemented at peripheral health centres.
The development of a specific point-of-care diagnostic test for the detection of M. ulcerans is complicated by the broad antigenic cross-reactivity among the various mycobacterial species. Serological approaches targeting the few M. ulcerans-specific antigens identified, turned out to be not suitable for differentiation between BU patients and exposed control individuals, as both groups may or may not exhibit serum IgG titers against these antigens [9–11].
In recent years, point-of-care tests in the form of antigen capture assays have successfully been developed for tropical infectious diseases [12]. Extensive studies focussing on rapid diagnostic tests for malaria [13–17] have paved the way for the development of antigen capture based assays for other diseases such as dengue fever [18, 19], hepatitis C [20, 21] or visceral leishmaniasis [22] to name but a few.
In the present work we aimed at the identification of targets for the development of an antigen capture test for the diagnosis of M. ulcerans infection by using a proteomics approach.
Ethical clearance for the analysis of clinical specimens was obtained from the Cameroon National Ethics Committee (N°172/CNE/SE/201) and the Ethics Committee of Basel (EKBB, reference no. 53/11). Immunization of mice for the generation of monoclonal antibodies was performed in strict accordance with the rules and regulations for the protection of animal rights (“Tierschutzverordnung”) of the Swiss “Bundesamt für Veterinärwesen”.
All animal infection experiments performed were approved by the animal welfare committee of the Canton of Vaud (authorization number 2261) and were conducted in compliance with the Swiss animal protection law under BSL-3 conditions.
In this study we analyzed M. ulcerans isolates from Ghana (NM20/02), Côte d’Ivoire (ITM 940511), Togo (ITM 970680), China (ITM 98912), Japan (ITM 8756) and Australia (JS5147) as well as additional mycobacterial strains including M. abscessus (ATCC 19977), M. avium (MAC101), M. chelonae (DSM 43804), M. fortuitum (ATCC 49403), M. gordonae (Pasteur 14021.001), M. haemophilum (ATCC 29548), M. intracellulare (clinical isolate), M. kansasii (NCTC 10268), M. lentiflavum (clinical isolate), M. malmoense (NCTC 11298), M. marinum (ATCC 927), M. scrofulaceum (Pasteur 14022.0031), M. simiae (clinical isolate) M. smegmatis (Pasteur 14133.0001), M. terrae (clinical isolate), M. xenopi, M. bovis (ATCC 35734) and M. tuberculosis (Pasteur 14001.0001). M. ulcerans strains were grown in BacT/Alert culture bottles supplemented with enrichment medium according to the manufacturer’s protocol (bioMérieux).
For the preparation of M. ulcerans protein lysates, bacteria (5 ml of culture, OD600~1) were washed in PBS, heat-inactivated at 95°C for 35 min, centrifuged at 10′000 × g for 10 min and resuspended in 400 μl lysis buffer (PBS containing 5% SDS, 1 mM phenylmethylsulfonyl fluoride (PMSF) and a protease inhibitor cocktail (complete mini, Roche)). The mix was transferred to lysing tubes (Precellys) and homogenized using a mechanical bead beater device (Precellys 24, Bertin Technologies) twice at 6′800 rpm for 30 s. Beads and non-lysed cells were removed by centrifugation at 10′000 × g for 10 min. The preparation of lysates of other mycobacterial species was described previously [9].
90 μg of trichloroacetic acid (TCA) precipitated M. ulcerans (NM20/02) protein lysate was resuspended in rehydration buffer (8 M urea, 2% 3-[(3-Cholamidopropyl)-dimethylammonio]-1-propanesulfonate (CHAPS), 0.5% (v/v) ZOOM Carrier Ampholytes (Invitrogen), 0.002% bromophenol blue and 0.4% dithioerythritol (DTE)). The mix was incubated with a 3–10 pH gradient strip (ZOOM Strip; Invitrogen) over night (ON) at room temperature (RT). First-dimension isoelectric focusing (IEF) was performed on a ZOOM IPG runner (Invitrogen) using a step voltage protocol (175 V for 15 min, 175–2000 V for 45 min, 2000 V for 2 h). After IEF, the strips were incubated for 15 min with equilibration buffer (6 M urea, 50 mM Tris pH 8.8, 30% glycerol, 2% SDS, 30 mM DTE) followed by a 15 min incubation period with alkylating solution (6 M urea, 50 mM Tris (pH 8.8), 30% glycerol, 2% SDS, 0.23 M iodacetamide). Second-dimension gel electrophoresis was performed at 200 V for 35 min using a 10% NuPAGE Novex Bis-Tris ZOOM Gel (Invitrogen). The gel was stained with Coomassie blue (Invitrogen).
All Coomassie stained protein spots were selected for mass spectrometry analysis. Spots were excised from the 2D gel, placed in a low-binding microcentrifuge tube and destained in 0.1 M ammonium bicarbonate / 30% acetonitrile at 30°C. Gel spots were dried in a SpeedVac concentrator and digested with 4 μl of 10 μg/ml trypsin (trypsin porcine, Roche Applied Science) ON at 37°C. Peptides were extracted from gel pieces with 4 μl of 0.3% trifluoroacetic adic (TFA) / 50% acetonitrile. The samples were desalted and concentrated using ZipTipC18 tips (Millipore). Eluted peptides were loaded onto a MALDI target.
MS analysis was performed using a MALDI-TOF mass spectrometer (Bruker ultraflex III TOF/TOF, Bruker Daltonics Inc.) in the reflector mode. 1 μl of tryptic digest and 1 μl of matrix (5 mg/ml α-cyano-4-hydroxycinnamic acid, 50% acetonitrile, 0.1% TFA) were spotted onto a MALDI target (MTP AnchorChip 600/384, Bruker Daltonics) and allowed to co-crystallize at room temperature. Data were processed using FlexAnalysis software (Bruker Daltonics flexAnalysis 2.4). Spectra were smoothed (Sawitzgy Golay algorithm, 0.2 m/z width, 1 cycle), baseline subtracted (median algorithm, 0.8 flatness) and calibrated using trypsin autocleavage or internal standard peptide mass peaks. A monoisotopic peak list was generated from the spectrum using SNAP algorithm and analyzed with BioTools (Bruker Daltonics BioTools 3.0). Peptide mass fingerprinting searches were performed using the Aldente search engine on the Expasy server.
The full length MUL_3720 (aa 1–207) and a truncated version of this protein lacking the lectin domain (aa 115–207) were recombinantly expressed in Escherichia coli BL21 Star (DE3, Invitrogen) as N-terminal hexahistidin-tagged fusion proteins. Briefly, PCR was performed on a pUC57 vector containing the DNA sequence of MUL_3720 generated by gene synthesis (Genscript), including NdeI and NotI restriction sites. The amplified sequences were inserted into a TOPO-TA cloning vector using the TOPO Cloning Kit and introduced into E. coli (Top 10, Invitrogen). The vector was digested with NdeI and NotI (New England Biolabs) and the sequence was ligated into a pET28a expression vector using the Rapid DNA Ligation Kit (Roche). E. coli BL21 Star (DE3, Invitrogen) were grown in Luria-Bertani (LB) medium until an OD600 of ~0.5. Protein expression was induced by addition of isopropyl thiogalactoside (IPTG) to a final concentration of 1 mM and subsequent incubation for 3 h at 37°C. Bacteria were lysed by sonication and His-tagged proteins were purified by nickel-nitrilotriacetic acid (Ni-NTA) chromatography.
MUL_3720 was amplified from genomic M. ulcerans DNA and cloned into TOPO vector using NdeI and ScaI restriction sites. Electrocompetent E. coli TOP10 cells (Invitrogen) were transformed with TOPO::MUL_3720 vectors and spread onto LB-Ampicillin (50 μg/ml) agar. Plasmid DNA of the mutant colonies was prepared and inserts with correct size and sequence were excised from TOPO by NdeI/ScaI and ligated into the mycobacterial vector pSD5. Chemically competent E. coli TOP10 were transformed with pSD5::MUL_3720 and grown on LB-Kanamycin (50 μg/ml) plates. Plasmid DNA was prepared and the presence of the insert was confirmed.
M. ulcerans strain NM20/02 was grown in BacT bottles (bioMérieux) containing enrichment medium (bioMérieux). Bacteria were harvested and washed twice with 10% glycerol or distilled water. Competent M. ulcerans bacteria were electroporated (2.5 kV, 1000 Ohm, 25 μF) with varying amounts (50–1000 ng) of DNA, transferred to MGIT-OADC medium (BD) and grown under non-selective conditions for 36 hours at 30°C. After the recovery phase, bacteria were spread on 7H10-Kanamycin (25 μg/ml) agar and incubated for several months at 30°C. Colonies were picked and regrown on selective agar and in BacT bottles (bioMérieux) in order to prepare lysates and stocks.
For the preparation of mAbs, mice were immunized two times intraperitoneally with 40 μg of recombinant full length MUL_3720 (aa 1–207) emulsified in Immune Easy adjuvant (Qiagen). Two weeks after the second immunization serum antibody titres against MUL_3720 (aa 1–207) as well as against the truncated MUL_3720 (aa 115–207) were determined by ELISA. Based on these results one BALB/c mouse was selected to receive a final intraperitoneal injection of 40 μg of recombinant MUL_3720 (aa 1–207) without adjuvant. Three days after this last booster dose, hybridoma cell lines were generated as described previously [23]. Briefly, the spleen of the selected mouse was removed and the spleen cells were fused with mouse myeloma cells (PAI cells). After a few days, cell culture supernatants were tested for the presence of anti-MUL_3720 (aa 1–207) as well as anti-MUL_3720 (aa 115–207) antibodies. Positive cell lines were cloned by limiting dilution and expanded. MAbs were purified using HiTrap rProtein A column (Amersham Biosciences). Two individual fusion experiments resulted in 24 and 17 MUL_3720-ELISA positive B-cell hybridoma cell lines, respectively. Of these, a total of 5 B-cell hybridoma cell lines (JD3.1, JD3.2, JD3.3, JD3.4, JD3.6, JD3.7) were successfully cloned and expanded for antibody production.
Rabbit polyclonal antibodies were generated and affinity purified by Eurogentech. New Zealand white Rabbits were injected intramuscularly with 20 μg recombinant MUL_3720 (aa 1–27) with Sigma Adjuvant System (SZ3398) or Imject Alum (SZ3403) on day 0, 14, 28 and 56. Total IgG was purified from antiserum collected on day 66 by protein A affinity chromatography.
96-well Immulon microtiter plates (Thermo Scientific) were coated with 1 μg recombinant MUL_3720 (aa 1–207) or MUL_3720 (aa 115–207) per well in 100 μl PBS and incubated ON at 4°C. Plates were washed three times with washing buffer (2.5% Tween 20 in dH2O) and blocked with 5% non-fat dry milk in PBS containing 0.1% Tween for 1 h at 37°C. After washing as described above, 100 μl of the primary antibody (mAbs or hybridoma supernatant) was added and incubated for 2 h at 37°C. Following an additional washing step, 100 μl of 1:30′000 diluted goat anti-mouse IgG (γ-chain specific) antibodies coupled to alkaline phosphatase (SouthernBiotech) was added to each well and incubated for 1 h at 37°C. Plates were washed and 100 μl/well of phosphatase substrate solution (1 mg/ml p-Nitrophenyl phosphate in substrate buffer) was added and incubated for 1 h at 37°C. Absorbance at 405 nm was measured with a microplate reader (Tecan Sunrise).
2 μg of mycobacterial protein lysates per lane were separated on NuPAGE Novex 4–12% Bis-Tris ZOOM Gels (Invitrogen) using NuPAGE MES SDS Running Buffer (Invitrogen) under reducing conditions. After electrophoresis proteins were transferred onto nitrocellulose membranes using an iBlot gel transfer device (Invitrogen). Membranes were blocked with blocking buffer (5% non-fat dry milk in PBS) ON at 4°C. Membranes were then incubated in blocking buffer containing anti-MUL_3720 IgG (mouse mAbs JD3.2, JD3.4 or rabbit polyclonal IgG SZ3398) or mouse mAb DD3.7 (specific for a conserved mycobacterial protein) serving as loading control for 1 h at RT. After washing, membranes were incubated with secondary goat anti-mouse IgG (γ-chain specific) (HRP, SouthernBiotech) or goat anti-rabbit IgG (Fc fragment specific) (HRP, Milan) for 45 min at RT. After washing, bands were visualized by chemiluminescence using the ECL Western Blotting substrate (Pierce).
Immunohistochemical analysis was performed on tissue or punch biopsies from different IS2404 qPCR reconfirmed patients. Tissue or punch biopsies of BU patients were removed aseptically and immediately fixed in 10% neutral buffered formalin for 24 hours. Afterwards the tissue was embedded into paraffin, cut into 5 μm thin sections and transferred onto microscopy glass slides. Immunohistochemical staining of the sections was performed after deparaffinisation, rehydration and antigen retrieval with citrate-pretreatment according to standard protocols [24]. Inactivation of endogenous peroxidase as well as prevention of unspecific binding was achieved by incubation in PBS containing 0.3% hydrogen peroxide and 1.5% horse serum for 20 min. Primary anti-MUL_3720 IgG was diluted in PBS containing 0.1% Tween-20 and added to the slides for 1 h at RT or ON at 4°C. After incubation with biotin-conjugated horse anti-mouse IgG, slides were stained using the Vector ABC and NovaRED system. Sections were counterstained with haematoxylin.
JD3.4 and JD3.2 showed a comparable staining in intensity, specificity and sensitivity. JD3.2 gave a slightly lower unspecific background staining of the surrounding tissue and was used for IHC analysis.
IFA was carried out as described previously [25]. Briefly, a pellet of M. ulcerans bacteria (OD600~0.6) was resuspended in 1.5% low-melting agarose (BioWhittaker Lonza, Basel Switzerland) and transferred to cryomodules (Applied BioSystems). Agarose blocks were embedded into paraffin, cut in 3 μm sections and transferred onto microscopy glass slides (Thermo Scientific). Bacteria were stained with mAb JD3.2 and Alexa fluor488 (Invitrogen) conjugated goat anti-mouse IgG and mounted in ProLong Gold anti-fade reagent containing 4′,6-Diamidino-2-phenylindole (DAPI; Invitrogen).
The system described for investigating protein interactions by the functional reconstitution of a murine dehydrofolate reductase domain in M. tuberculosis [26] was modified here for use in M. ulcerans. N-terminal and C-terminal fusions of the bait domains to full length MUL_3720 were constructed using pUAB400 or pUAB200. Cloning was facilitated by the MfeI and ClaI restriction sites in the pUAB multiple cloning sites. All cloning was performed using E. coli DH10B and confirmed by Sanger sequencing. Plasmids were extracted from E. coli using mini-prep columns (Qiagen) and plasmid DNA was used to transform M. smegmatis MC2155 by electroporation as previously described [27].
M. ulcerans Agy99 genomic DNA libraries were prepared by partial AciI digestion. Digested DNA between 500 bp and 3 Kbp was purified using a gel purification kit (Qiagen), ligated into ClaI digested pUAB300 and used to transform E. coli DH10B. A number of colonies were randomly selected for PCR using primers F102 (5′-agaaccaccacgaggagctcat-3′) and R102 (5′-tgatgcctggcagtcgatcgta-3′) that flank the multiple cloning site on the vector to check for insertions containing inserts within the desired size range [26]. Approximately 2 × 105 clones were subsequently collected and cultured in LB ON. Plasmid DNA maxi-preps were performed on ON cultures according to the manufacturer’s instructions (Sigma-Aldrich).
Bacteria co-transformed with plasmids containing interacting, complementary mDHFR fragments were selected on 7H11 kanamycin (25 μg/ml) and hygromycin (50 μg/ml) plates. Colonies were patched onto 7H11 kanamycin-hygromycin-trimethoprim plates and colonies resistant to trimethoprim were selected for PCR. Using primers F102 and R102, PCR products were sequenced. The sequences were used to perform BLAST against the M. ulcerans Agy99 genome. Inserts containing open reading frames in the incorrect orientation were discarded. Also removed were inserts that matched non-coding genomic DNA or the dehydrofolate reductase from M. ulcerans.
M. ulcerans strain S1013 used for experimental infection of mice was isolated in 2010 from the ulcerative lesion of a Cameroonian BU patient [28]. Bacteria were cultivated in Bac/T medium for 6 weeks, recovered by centrifugation and a stock suspension in sterile PBS of 125 mg/ml wet weight was prepared. 30 μl of a 1:1000 dilution of the stock solution was injected subcutaneously into the left hind foot pad of 14 week old female BALB/c mice. On day 87 after infection, mice were euthanized and foot pads were aseptically removed. Foot pads were dipped into 70% ethanol, dried under the laminar flow, cut into four pieces with a scalpel and transferred to reinforced hard tissue grinding tubes (MK28-R, Precellys) containing 750 μl of Bac/T medium (bioMérieux). Tissue homogenization was performed with a Precellys 24-dual tissue homogenizer (3 × 20 s at 5000 rpm with 30 s break). After transferring the supernatant to a fresh tube, the residual tissue remains were homogenized a second time in 750 μl of Bac/T medium. Tissue lysates were pooled and stored at -80°C until further use. 500 μl of thawed tissue lysate was transferred into tough microorganism lysis tubes (VK05–2ml, Precellys), inactivated for 1 h at 85°C and centrifuged at 17′000 × g for 5 min. The pellet was resuspended in 250 μl PBS containing protease inhibitors (Roche, EDTA—free) and cells were disrupted with Precellys 24-dual tissue homogenizer (2 × 30s at 6800rpm with 1 min break in between). Lysates were cleared by centrifugation and tested by ELISA.
Nunc-Immuno Maxisorp 96-well plates (Thermo Scientific) were coated with 10 μg/ml JD3.4 mAb (50 μl per well) in PBS and incubated ON at 4°C. Plates were washed three times with washing buffer (2.5% Tween 20 in dH2O) prior to incubation with blocking buffer (5% non-fat dry milk in PBS) for 2 h at RT. After washing as described above, 50 μl of different dilutions of the purified recombinant full length MUL_3720, M. ulcerans lysate (NM20/02), or lysates from M. ulcerans infected tissue samples in PBS were added and incubated for 2 h at RT. Following an additional washing step, 50 μl anti-MUL_3720 rabbit IgG (5 μg/ml) in blocking buffer with detergent (0.5% non-fat dry milk in PBS containing 0.05% Tween 20) was added and incubated for 2 h at RT. After washing as described above, 50 μl goat anti-rabbit IgG coupled to horseradish-peroxidase (Milan) diluted 1:10′000 in blocking buffer with detergent was added and incubated for 1 h at RT. Plates were washed and TMB peroxidase substrate solution was added. After 10 min the reaction was stopped with 2 M sulfuric acid and absorbance was measured at 450 nm with a microplate reader (Tecan Sunrise).
For identification of suitable proteins that could be used as targets in diagnostic test formats, an M. ulcerans whole protein lysate was analysed by 2D gel electrophoresis (S1 Fig.). In total, 384 protein spots were detected, processed and subsequently subjected to MALDI-TOF-MS. Among the 384 spots, 118 peptide fragments were identified and attributed to 36 different genes. In order to select for proteins without orthologs in M. tuberculosis, M. bovis or M. leprae, a BLAST search against the Uniprot database was performed for all 36 proteins, resulting in the identification of three potential targets (MUL_3720, MUL_0343 and MUL_4023) suitable for a selective antigen capture assay. However, MUL_0343 and MUL_4023 presented very weak protein spots in the 2D gel, while MUL_3720 showed a high expression level and was therefore selected for further analysis.
The 624 bp MUL_3720 gene encodes a protein of 207 amino acids, with a molecular mass of 22 kDa. MUL_3720 is predicted to possess an N-terminal bulb-type mannose-specific lectin domain and a C-terminal peptidoglycan-binding Lysin Motif (LysM) linked by a proline-rich sequence (Fig. 1). Database comparisons revealed the presence of orthologs with a similar domain organisation in M. abscessus (MAB_2373), M. avium, M. colombiense, M. fortuitum, M. kansasii (MKAN_05370), M. marinum (MMAR_3773), M. smegmatis (MSMEG_3662) and M. xenopi. The M. marinum ortholog displayed a sequence identity of 99% (S2 Fig.).
For the generation of antibodies against MUL_3720, required for the detection of the protein in diagnostic assays, we immunized mice and rabbits with the full length protein, recombinantly expressed as a His-tagged fusion protein (predicted molecular mass 24 kDa) in E. coli BL21. Hybridoma cell lines producing antibodies against different epitopes of the protein were identified by analyzing their reactivity against MUL_3720 (aa 1–207) as well as the truncated version of MUL_3720 (aa 115–207), lacking the lectin domain and consisting only of the LysM motif and the proline-rich sequence. Five mAbs (JD3.2, JD3.3, JD3.4, JD3.6 and JD3.7), all of them mouse IgG1(κ) isotype, were generated, purified and further characterized. All the mAbs recognized recombinant full length MUL_3720 (aa 1–207) in ELISA. While all antibodies except for JD3.6 recognized recombinant full length MUL_3720 (aa 1–207) in Western Blot analysis (Fig. 2A), only JD3.2 and JD3.4 also reacted with recombinant truncated MUL_3720 (aa 115–207) (Fig. 2B) and the endogenous protein in M. ulcerans lysates (Fig. 2C) (S1 Table).
The ability of mouse mAbs JD3.2 and JD3.4 to detect endogenous MUL_3720 in lysates of M. ulcerans strains from different geographical regions was examined by Western Blot analysis. While the protein was recognized by JD3.2 and JD3.4 in all M. ulcerans strains, isolates belonging to the classical M. ulcerans lineage (Ghana, Côte d’Ivoire, Togo and Australia) showed higher expression levels as compared to isolates belonging to the ancestral lineage (China and Japan) (Fig. 3).
Interspecies cross-reactivity of MUL_3720 was determined by Western Blot analysis with lysates of a range of different mycobacterial species (Fig. 4). In accordance with the BLAST search for MUL_3720 orthologs (S2 Fig.), rabbit polyclonal anti-MUL_3720 IgG reacted with proteins in lysates of M. fortuitum, M. marinum, M. smegmatis and M. xenopi. In agreement with a shorter linker between the lectin and the LysM domains (S2 Fig.), the M. xenopi ortholog was detected at a lower molecular weight. The predicted orthologous proteins in M. abscessus, M. avium and M. kansasii were not recognized by the rabbit polyclonal IgG. Furthermore, a protein band was observed in lysates of M. gordonae, M. malmoense and M. terrae for which no sequence information is available (Fig. 4A). The detected proteins in M. malmoense and M. terrae were slightly smaller than MUL_3720.
Analysis with JD3.4 led to a similar staining pattern among the mycobacterial lysates, except for the protein expressed by M. gordonae, which was not recognized (Fig. 4B). JD3.2 only reacted with protein in lysates of M. malmoense and M. marinum, suggesting that the two mAbs JD3.2 and JD3.4 recognize different epitopes of MUL_3720 (Fig. 4C).
In order to confirm the expression of MUL_3720 by M. ulcerans and the ability of the anti-MUL_3720 mAb JD3.2 to detect the protein in vivo, we performed immunohistochemistry and immunofluorescence stainings. M. ulcerans could be detected in punch biopsies of human BU patients with the mAbs JD3.2 (Fig. 5) and JD3.4. ZN staining (Fig. 5 A, C, E) and JD3.2 staining (Fig. 5 B, D, F) of serial sections showed identical localization at the same tissue region. AFBs detected by ZN staining in tissue sections revealed a homogeneous staining pattern, whereas immuno-staining with mAb JD3.2 exhibited a heterogeneous staining pattern of the bacteria with intensively stained poles (Fig. 5), indicating a higher expression of the protein in these areas in the natural environment of the bacteria. In contrast, IFA of in vitro cultivated bacteria showed a more homogeneous distribution of the protein on the bacterial cell surface. This localization was confirmed in a MUL_3720 overexpressing M. ulcerans strain (Fig. 6).
As a first step to begin to understand the role of MUL_3720 we used a bait and prey approach to identify other M. ulcerans proteins that interacted with this protein. We employed the mycobacterium-specific protein fragment complementation (M-PFC) system. An M-PFC bait clone using a N-terminal fusion of MUL_3720 and co-transformation with a random library of M. ulcerans genomic DNA fragments in pUAB300 (prey) resulted in approximately 150 trimethoprim resistant colonies. Subsequent clones were patched and screened using primers F102 and R102 to determine the identity of the DNA sequence present in pUAB300 (Table 1). Multiple independent clones were identified for sequences encoding DesA1 (MUL_0445) and a PE-PGRS protein (MUL_0572), together with 13 other single-hit CDS, including an interaction with MUL_3720 itself. Many of the putative interacting proteins had a predicted cell wall location or role in cell wall biosynthesis, in line with the localization data for MUL_3720 revealed by mAb staining (Table 1 and Fig. 6). No interacting proteins were identified using the C-terminal MUL_3720 bait fusion, consistent with the predicted cell wall location for this domain of MUL_3720. The M-PFC detects cytoplasmic protein-protein interactions only [26].
We analyzed different combinations of the generated mAbs and polyclonal IgG as MUL_3720 capturing and detecting antibodies in an antigen capture sandwich ELISA. The application of mAb JD3.4 as capturing and polyclonal rabbit IgG as detecting reagent enabled a highly sensitive detection of recombinant MUL_3720 (Fig. 7A) and the endogenous protein present in lysates of in vitro cultivated M. ulcerans (Fig. 7B).
In order to test if the antigen capture ELISA is able to detect MUL_3720 expressed by bacteria in infected tissue samples, we analyzed lysates of M. ulcerans infected mouse foot pads. MUL_3720 could be detected in lysates of all five infected tissue samples analyzed, while only background readouts were obtained for lysates of uninfected foot pads (Fig. 8).
Attempts to develop a diagnostic tool based on serological approaches have been equivocal [9–11], so we decided to focus on direct detection of M. ulcerans antigens in BU patient specimens. In the present study, we identified the MUL_3720 protein as a promising target in antigen capture-based diagnostic tests for M. ulcerans. Based on 2D gel electrophoretic analyses, MUL_3720 is one of the most highly expressed proteins in vitro. The high expression of MUL_3720 is considered an advantage with respect to developing a sensitive antigen detection test for the diagnosis of BU.
While the biological role of MUL_3720 is not known, clues to its function are suggested by its two-domain structure—a conserved bulb-type mannose-binding lectin domain and a Lysin Motif (LysM) domain—predicted to be involved in alpha-D-mannose recognition and in binding to peptidoglycan, respectively. Some bacterial species retain certain proteins attached to peptidoglycan by their LysM domains [29]. We used a mycobacteria-specific two-hybrid system to search for M. ulcerans proteins interacting with MUL_3720 and we had hits to a range of proteins known or predicted to be cell-wall associated or involved in cell wall synthesis (Table 1). Many of the interacting proteins—such as DesA1 (Table 1)—are involved in biosynthesis or modification of cell wall molecules. In other mycobacteria, the resulting double bonds from the DesA1-mediated catalysis of a desaturation reaction of saturated alkyl chains that arise during mycolic acid synthesis are required for subsequent position specific modifications such as epoxidation and cyclopropanation of this key cell wall metabolite [30, 31]. MUL_3720 appears to be arranged in an operon structure with two adjacent putative cell wall-associated protein coding genes (MUL_3721, MUL_3722). Immunofluorescence stainings of M. ulcerans bacilli confirmed the cell wall localization of MUL_3720.
With its cell wall location, the two-domain structure including a mannose-binding N-terminal cytoplasmic component and C-terminal peptidoglycan-binding component, its operon structure and a substantial list of potential interacting proteins, MUL_3720 may be an adaptor protein for multiple cell wall biosynthetic pathways. MUL_3720 might play a role in cell attachment and cell-cell interactions given its presence at the cell surface as revealed by immunofluorescence microscopy and immunohistochemical analyses.
The cell-surface localization of MUL_3720 is an additional advantage with respect to developing a sensitive diagnostic test, since the protein is expected to be easily accessible and detectable in tissue specimens of BU lesions. Potential shedding of the protein from the cell surface may facilitate ready detection in body fluids, which will be examined in future experiments.
Monoclonal and polyclonal antibodies against MUL_3720 were generated for the development of antigen capture assays. These antibodies recognized in vitro grown M. ulcerans bacilli as well as bacteria in biopsies of human BU patients, proving the expression of MUL_3720 in BU lesions.
Since these antibodies did not react with orthologs of MUL_3720 in other pathogenic mycobacterial species prevalent in the BU endemic regions, prospects for the development of a test with the desired specificity, excluding in particular cutaneous tuberculosis [5], are good. The monoclonal antibodies used for the antigen capture test bind to an epitope on the proline/rich linker and/or the LysM domain. Since the LysM domain is a widespread protein module present in more than 4000 proteins of both prokaryotes and eukaryotes [29], the potential cross-reactivity of the anti-MUL_3720 antibodies with those proteins remains to be analyzed.
Importantly, this capture assay specifically detected MUL_3720 protein in tissue lysates of M. ulcerans infected mouse footpads. Furthermore, initial results revealed that MUL_3720 could be detected in swab samples from human BU lesions with a high bacterial burden (manuscript in preparation). Ongoing optimization of the applied reagents as well as the assay format is aiming at the development of a simple test format appropriate for low-resource laboratory settings with suitable test sensitivity.
Antibiotic treatment of BU in its early stages leads in most of the cases to complete healing of the lesions with little or no trauma, whereas treatment at later stages often requires adjunct surgical treatment and is associated with prolonged hospitalization and long-term sequelae. The development of a simple and rapid diagnostic test, whose key elements are provided in the work presented here, will be of immediate benefit to BU patients in rural endemic communities. Clinical findings could directly be reconfirmed by this point-of-care test helping to avoid a false diagnosis and to facilitate a prompt onset of adequate treatment.
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.