|
|
PLoS Pathog. 2008 July; 4(7): e1000105. Published online 2008 July 18. doi: 10.1371/journal.ppat.1000105. | PMCID: PMC2453101 |
Copyright Pujol et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Anti-Fungal Innate Immunity in C. elegans Is Enhanced by Evolutionary Diversification of Antimicrobial Peptides Nathalie Pujol,#1,2,3 Olivier Zugasti,#1,2,3 Daniel Wong,#1,2,3 Carole Couillault,1,2,3 C. Léopold Kurz,1,2,3 Hinrich Schulenburg,4 and Jonathan J. Ewbank1,2,3* 1Centre d'Immunologie de Marseille-Luminy, Université de la Méditerranée, Case 906, Marseille, France 2INSERM, U631, Marseille, France 3CNRS, UMR6102, Marseille, France 4Department of Animal Evolutionary Ecology, Zoological Institute, University of Tuebingen, Tuebingen, Germany Frederick M. Ausubel, Editor Massachusetts General Hospital, United States of America Received September 8, 2007; Accepted June 19, 2008. Encounters with pathogens provoke changes in gene transcription that are an integral part of host innate immune responses. In recent years, studies with invertebrate model organisms have given insights into the origin, function, and evolution of innate immunity. Here, we use genome-wide transcriptome analysis to characterize the consequence of natural fungal infection in Caenorhabditis elegans. We identify several families of genes encoding putative antimicrobial peptides (AMPs) and proteins that are transcriptionally up-regulated upon infection. Many are located in small genomic clusters. We focus on the nlp-29 cluster of six AMP genes and show that it enhances pathogen resistance in vivo. The same cluster has a different structure in two other Caenorhabditis species. A phylogenetic analysis indicates that the evolutionary diversification of this cluster, especially in cases of intra-genomic gene duplications, is driven by natural selection. We further show that upon osmotic stress, two genes of the nlp-29 cluster are strongly induced. In contrast to fungus-induced nlp expression, this response is independent of the p38 MAP kinase cascade. At the same time, both involve the epidermal GATA factor ELT-3. Our results suggest that selective pressure from pathogens influences intra-genomic diversification of AMPs and reveal an unexpected complexity in AMP regulation as part of the invertebrate innate immune response. We are interested in how exactly the nematode Caenorhabditi elegans, widely used in biological research, defends itself against fungal infection. Like most animals, this worm responds to infection by switching on defense genes. We used DNA chips to measure the levels of all the worm's 20,000 genes and discovered new inducible defense genes. Many of them encode small proteins or peptides that can probably kill microbes. By looking in other nematode species, we saw that these antimicrobial peptide genes are evolving rapidly. This means that they could be important for the worms' survival in their natural environment. We also looked at how some of these genes are regulated and uncovered a sophisticated control mechanism involving a series of proteins called kinases that relay signals within cells. The genes we looked at are active in the worm's skin. Some of the antimicrobial peptide genes that we looked at are also switched on in the skin by high salt, but in this case, the regulation doesn't involve the same cascade of kinases. The responses to both infection and high salt do, however, require the same transcription factor (the protein that actually switches genes on), in this case called a GATA factor. Two strategies exist for organisms that suffer from predation or infection in their natural environment. They can invest in constitutive defenses that will offer them permanent protection, and they can use inducible defenses that are activated only when they are in danger [1]. In C. elegans, the epidermis is at the interface with the environment and is expected to play a key role in defense. It is responsible for the production of the collagen-rich cuticle that surrounds the nematode and provides a permanent physical barrier to pathogens. Some bacteria, such as Microbacterium nematophilum, Xenorhabdus nematophila or Yersinia pestis adhere to the cuticle surface and while not physically penetrating the epidermis cause disease [2]– [4]. On the other hand, nematophagous fungi such as Drechmeria coniospora adhere to the cuticle and then infect nematodes directly via the epidermis [5]. In many animal species, infection of barrier epithelia results in the up-regulation of genes encoding antimicrobial peptides (AMPs) and proteins [6]– [9]. Much has been learnt about these inducible defenses through comparative transcriptional profiling. In the case of C. elegans, our previous study using microarrays with partial genome coverage showed that D. coniospora infection provokes increased AMP gene expression in the epidermis [10]. Many AMPs act by disrupting microbial cell membranes [11]. The efficiency of cell disruption by a single AMP may vary for different pathogens and depends on the exact structure of the microbial surface. Hence, hosts exposed to diverse pathogens may evolve a broader repertoire of AMPs that enhance their defensive potential, if this confers a selective advantage. Such AMP variation is found within the genome of Drosophila melanogaster. Of the 20 best-characterised AMPs, eight appear most efficient against fungi, eleven against Gram-negative and one against Gram-positive bacteria [7]. AMP diversity within a single genome may be achieved through gene duplication, a process considered to be one of the most important sources of evolutionary innovation [12]. In this study, we sought to characterize more completely the transcriptional response of C. elegans to natural fungal infection. We found that putative AMP genes constitute a major part of that response. We show that one group arose via gene duplication and that these duplicated genes are controlled by a complex regulatory mechanism. The transcriptional response of C. elegans to fungal infection To characterize the response of C. elegans to a natural fungal infection, we have analyzed changes in gene expression in worms infected with D. coniospora. In a previous study using cDNA nylon microarrays, one striking observation was the increased expression at 12 and 24 h post-infection of multiple genes potentially encoding glycine- and tyrosine-rich antimicrobial peptides (AMPs), members of the NLP (for neuropeptide-like protein) and CNC (Caenacin) families [10]. In our previous report, we only provided data for these nlp and cnc genes. While they do represent an important component of the response, a large number of other genes are up-regulated at both time points (, Tables S1A & S1B). The cDNA arrays correspond to fewer than 8,000 of the predicted 20,000 worm genes, so only give a partial coverage of the genome [13]. We therefore carried out an additional analysis using long oligonucleotide whole-genome microarrays [14], comparing the level of gene expression between uninfected controls and worms 24 h after infection with D. coniospora. In the top 20 up-regulated genes, ranked by fold-change, ( Table S1C), there were 8 nlp and cnc genes. In addition, there were two previously uncharacterized genes, one that we named grsp-2 (Glycine-Rich Secreted Protein 2), and the other fip-1 (Fungus-Induced Protein 1). Inspection of sequences of the next 50 genes led us to annotate 6 other FIPs and 29 FIP-related (FIPR) proteins (see Protocol S1; Figure S1). Based on comparisons to peptides with known antimicrobial activity [15], the fip, fipr and grsp genes could all potentially encode AMPs. Fold-change measurements are useful when performing exploratory analyses. They can be complemented by methods that evaluate the statistical significance of any observed differences [16]. We used two established statistical-tools, MAANOVA and BRB-ArrayTools (see Materials and Methods) to analyze our data. With the first method, 14 up-regulated and 26 down-regulated genes were found; with the second, 11 and 33, respectively (, Table S1D). Of the 11 up-regulated genes identified with both methods, 7 were nlp or cnc genes (). There is data for the expression pattern in uninfected worms for 9 of the 14 genes found to be up-regulated using MAANOVA. As judged by in situ hybridization (from the Kohara laboratory) or reporter gene expression, cnc-2 and ZK228.3 are not expressed at detectable levels, while far-3 and cnc-6 are expressed in the intestine. far-3 is also expressed in epidermis and around the vulva. Most of the genes for which there is data are, however, expressed specifically in the epidermis ( Table S1F). Together, these data reinforce the notion that inducible AMP genes expressed in the epidermis are key components of the innate immune response of C. elegans to natural fungal infection. Evolutionary diversification of AMP genes We analyzed the genomic distribution and evolutionary history of nlp genes from C. elegans and from C. briggsae and C. remanei, nematodes that belong to a different evolutionary lineage within the Elegans group of the genus Caenorhabditis [17]. In C. elegans, most of the infection-inducible nlp genes are found in a 12 kb region on the left arm of chromosome V, that we refer to as “the nlp-29 cluster”. Through close analysis and re-annotation of the available genomic sequences, we identified syntenic but highly divergent clusters in C. briggsae and C. remanei (). In C. elegans, these genes form an exclusive monophyletic group ( Figure S2), separate from the majority of other nlp genes that are likely to encode bona fide neuropeptides acting in the nervous system [18]. Within this monophylum, the genes nlp-27 to nlp-31 that are immediate neighbours, form a distinct clade (, S2 & S3). These 5 genes, and the adjacent gene nlp-34, which together make up the nlp-29 cluster are all induced by D. coniospora infection. We conducted a specific evolutionary analysis on these genes, as well as their respective orthologues from C. briggsae ( Cbr) and C. remanei ( Cre). The inferred phylogeny consistently identified three distinct clades for both the DNA and the protein datasets ( & data not shown). In the case of the inferred clades consisting of nlp-28 to nlp-31 and also with Cbr-nlp-34.2 and Cbr-nlp-34.3, the constituent genes are adjacent to one another within the respective genomes. This suggests that recent gene duplications occurred within the lineages leading to the different species. When we tested for adaptive sequence evolution across branches ( Protocol S1), we found that non-silent changes are more frequent than silent changes in four branches of the phylogenetic tree (; Table S2). Interestingly, all but one of the identified branches associate with clades of nlp genes found adjacent to each other within a single genome (). These results suggest that adaptive sequence evolution shaped the diversification of the nlp-29 cluster. Differential expression of AMP genes in the nlp-29 cluster in response to infection, injury and osmotic stress Our microarray analyses indicated that all the genes of the nlp-29 cluster were induced by fungal infection (). We confirmed this using qRT-PCR, and for all the genes except nlp-27, observed a statistically significant induction of expression post-infection ranging from 5 to 120-fold (p<0.05; Table S3). On the other hand, and in contrast to the other genes, the level of nlp-27 in non-infected worms was high (). We have recently shown that C. elegans responds to wounding by the up-regulation of nlp-29 and nlp-31 expression [19]. Using qRT-PCR, we found that following wounding all the genes in the cluster showed statistically significant increases of expression (p<0.05). It was least pronounced for nlp-27 (1.6-fold) whereas for the other genes there was at least a 3.5-fold induction. The relative magnitude of these inductions for the different genes was similar to that seen upon infection (; Table S3). Disrupting the cuticle or epidermal cell integrity is therefore sufficient to cause increased expression of all of the genes of the nlp-29 cluster. Both infection and injury cause cellular stress. To address the question of whether other stressors provoke AMP expression, we used a reporter strain with an integrated p nlp-29::GFP transgene that is strongly induced by D. coniospora infection and wounding [19]. We observed no increase in GFP expression when the worms were exposed to a number of stressful situations, including heat shock (1 hour at 37°C, or 10 minutes at 70°C), starvation (for up to 8 hours), paraquat or the heavy metals, cadmium and copper (results not shown). On the other hand, p nlp-29::GFP was highly induced by osmotic stress. Thus, exposure of these worms to high concentrations of NaCl (or 100 mM CaCl 2, MgCl 2, or MgSO 4) resulted in an increased level of p nlp-29::GFP expression that was dependent on the ionic strength ( and results not shown). We therefore investigated the effect of exposure to high salt on the expression of all the genes of the nlp-29 locus using qRT-PCR (). We observed a marked increase in the expression of nlp-28 and nlp-29 (3.0 and 3.6-fold, respectively; p<0.05), and a smaller, but significant induction of nlp-31 (2.0-fold; p<0.05). For the other genes of the cluster, nlp-27, nlp-30 and nlp-34, no significant changes were observed (p>0.05). Thus, for some of the genes of the nlp-29 cluster (e.g nlp-34), the fold-change in their expression was 2 orders of magnitude higher after infection than it was after osmotic stress, whereas for others (e.g. nlp-28) there was an equivalent induction after infection, wounding or exposure to high salt ( Table S3). We obtained results consistent with these qRT-PCR analyses using transgenic strains carrying different reporter constructs. For example, pnlp-27::GFP showed a strong constitutive level of fluorescence in the epidermis, while pnlp-30::GFP showed an increased level of GFP expression upon infection and wounding, but not osmotic stress (results not shown). These results clearly show that although all genes of the nlp-29 cluster are induced by natural fungal infection and by wounding, only some are induced by osmotic stress. Thus, the individual AMP genes of the nlp-29 cluster are subject to differential regulation and respond to distinct combinations of stimuli. Differential requirement for the p38 MAPK pathway in the response to infection, injury and osmotic stress Currently, one of the best characterized innate immunity signaling pathway in C. elegans is the p38 MAPK cascade. It is required for resistance to Pseudomonas aeruginosa infection [20]– [22]. It involves the MAP3K NSY-1, the MAP2K SEK-1 and the p38 MAPK PMK-1, acting downstream of the conserved adapter protein TIR-1 [23]. This pathway can also regulate the expression of nlp-29 in the C. elegans epidermis. Thus in nsy-1, sek-1, pmk-1 or tir-1( tm3036) mutants, there is essentially no increase in p nlp-29::GFP expression after fungal infection or wounding [19]. In clear contrast, nlp-29 up-regulation triggered by osmotic stress was largely independent of nsy-1, sek-1, pmk-1 or tir-1, especially when the reduced constitutive expression of the mutants is taken into consideration (). These results were confirmed by qRT-PCR: the induction by high salt of nlp-28 and nlp-29, the only two genes to be strongly up-regulated by osmotic stress was largely pmk-1-independent (; p>0.1 for the difference for either gene between wild-type and mutant). qRT-PCR was also used to demonstrate that the normal response to infection and injury of 5 of the 6 genes in the nlp-29 cluster required pmk-1, the exception being nlp-27 (, Table S3). These results indicate that p38 acts to control the response to fungal infection and wounding not only of nlp-29 but also of 3 other paralogous genes, and nlp-34. They also demonstrate that these responses can be genetically separated from the response to high salt. Osmotic stress resistant mutants have elevated nlp-29 expression We wished to explore further the link between osmotic stress and nlp-29 expression, to try to understand the physiological role of this regulation. To counter water loss in hypertonic environments, C. elegans increases its expression of gdph-1. This gene encodes the enzyme glycerol 3-phosphate dehydrogenase that catalyzes the rate-limiting step of glycerol biosynthesis. As a result, intracellular glycerol concentration increases and this has an osmoprotectant effect [24], [25]. Transgenic worms carrying a p gdph-1::GFP construct show an enhanced fluorescence upon exposure to increasing concentrations of salt. This reporter gene can thus be used as an in vivo sensor of the osmotic stress response [24]. There was no change in p gdph-1::GFP expression after D. coniospora infection or wounding (results not shown). Thus, these stimuli appear not to trigger an osmotic stress response. Certain mutants, including dpy-9(e12) and osm-11(n1604) have an elevated level of intracellular glycerol and a higher capacity to resist osmotic stress [24], [25]. We found that these mutants exhibited a high level of p nlp-29::GFP expression under normal culture conditions (). Dpy mutants are short and fat and manifest alterations of the epidermis and/or cuticle. We tested two other Dpy mutants, dpy-13(e184) and dpy-17(e164) that have a morphology similar to dpy-9(e12) but that are no more resistant to high salt than wild-type worms. They showed normal levels of p nlp-29::GFP expression, indicating that an altered morphology is not necessarily associated with elevated AMP expression (). The two mutants dpy-9(e12) and osm-11(n1604), which resist osmotic stress and have a high level of p nlp-29::GFP expression, also had a markedly increased resistance to D. coniospora infection (, p<0.001), while dpy-17 mutants were as susceptible as wild type worms. On the other hand, the lifespans of dpy-9(e12) and osm-11(n1604) mutants are no greater than that of wild type worms in the absence of infection (, p>0.05), suggesting that these mutants exhibit a specific pathogen resistance. To understand further the link between resistance to osmotic stress and increased nlp expression, we first generated a dpy-9;pmk-1 double mutant. Loss of pmk-1 function reduced the high constitutive level of p nlp-29::GFP expression seen in dpy-9 mutants ( Figure S4). But due to a synthetic interaction, the double mutants retained eggs (Egl phenotype) and were fragile, so that we were unable to carry out more analyses. The osm-11;pmk-1 double mutant, however, allowed us to assay the contribution of increased nlp expression to osmotic resistance. Compared to osm-11 mutants, these worms had a drastically diminished level of p nlp-29::GFP expression, like that of wild-type worms (). On the other hand, the osm-11;pmk-1 double mutant were still far more resistant to osmotic stress than wild-type worms () in standard tests of acute osmotic stress resistance (500 mM NaCl) [25]. These results suggest that nlp-29 is not involved in osmotic protection. Overexpression of AMP genes increases resistance to infection but not to osmotic stress When we assayed the survival of the null mutant strain nlp-29(tm1931), which cannot make any NLP-29, we saw no marked change in its resistance to D. coniospora infection nor its lifespan in the absence of infection (see below). Abrogation of the function of nlp-31 does not have a significant effect on survival either [10]. In both cases, this could reflect a redundancy in the function of single nlp genes in the nlp-29 cluster, especially given their high level of expression after infection. Therefore, to test whether the genes of the nlp-29 cluster could contribute directly in vivo to the capacity of C. elegans to resist infection, we generated transgenic strains carrying supernumerary copies of the entire nlp-29 cluster. By qRT-PCR we determined that there was an increase in the constitutive and inducible level of gene expression for all 6 genes in the cluster in the transgenic worms (a 3 to 8-fold increase for the different genes, ). We then assayed their resistance to D. coniospora infection and found a significantly greater survival when compared either to non-transgenic siblings or worms carrying an unrelated transgene (, S5A & S5B; p<0.001). The transgenic worms were, however, as susceptible as wild-type worms to 500 mM NaCl (results not shown). These results indicate that the nlp genes can contribute in vivo to increased resistance to fungal infection, but probably not to osmotic stress. The GATA transcription factor ELT-3 fulfils a generic requirement for nlp-29 expression Inspection of the upstream sequences of genes of the nlp-29 cluster revealed the presence of a conserved putative GATA site in the promoter regions of nlp-28 to nlp-31 ( Figure S6). The GATA factor ELT-2 has been shown to be important for the control of infection-inducible gene expression in the intestine [26]. There are 14 GATA factors encoded in the C. elegans genome [27]. We focused on those known to be expressed in the epidermis or seam cells, namely elt-1, 3 and 6 and egl-18 (previously known as elt-5) [28]– [30]. RNAi of egl-18, elt-1 and 6 did not have a significant effect (results not shown). We observed, however, that the constitutive expression of p nlp-29::GFP and its induction by infection or high salt was reduced upon elt-3 RNAi. We confirmed this effect using an elt-3 null mutant allele and found that GFP expression was knocked down by half following either of these treatments, as well as in untreated worms. The level of red fluorescence, from the p col-12::DsRed transgene was, on the other hand, essentially the same (+/−15%) in all cases (). To assay for a role of elt-3 in fungal resistance, we compared the survival of wild-type and mutant worms after D. coniospora infection. Unlike the nlp-29(tm1931) mutant, which behaved essentially like the wild type, there was a marked reduction in the resistance of the elt-3 mutants. These mutants, however, had a substantially reduced lifespan in the absence of infection. The same phenotypes were observed for tir-1(tm3036) mutants (). Thus, while being suggestive, we cannot definitively assign a specific role in fungal resistance to elt-3. Exposure to high salt up-regulates expression of the pgdph-1::GFP reporter. Unlike pnlp-29::GFP that is expressed specifically in the epidermis, it is expressed in both the epidermis and the intestine (). Interestingly, in the elt-3 mutant background, an abrogation of the epidermal expression of pgdph-1::GFP was seen, while expression was maintained in the intestine (). This suggests that although elt-3 is necessary for expression of AMP genes, it acts as a generic transcription factor for inducible genes specifically in the epidermis. Transcriptional response of C. elegans to fungal infection In this study, after an unbiased microarray analysis of genes affected by natural fungal infection in the epidermis of C. elegans, we focused on putative AMP genes, as they are the most prominent class of up-regulated genes. Synthetic NLP-31 has demonstrated antimicrobial activity in vitro against D. coniospora [10]. The other infection-induced NLPs and the structurally-related CNCs are therefore candidate AMPs. Our sequence analysis showed that these proteins can be differentiated from most of the predicted NLP proteins. Indeed, it is important to emphasize that NLP-27 to NLP-34 (but not NLP-32) carry the name Neuropeptide-Like Protein only for historical reasons. With regards the many GRSPs, FIPs and FIPRs, while proteins in other species with similar sequences possess antimicrobial activities [15], expression and biochemical analyses are needed to test if the C. elegans proteins have such a function. A very recent study reported changes in host gene expression induced by the nematode-trapping fungus Monacrosporium haptotylum [31]. A comprehensive comparison cannot be made as the study with M. haptotylum used microarrays with probes to only a few hundred C. elegans genes, and of these only 20 are among the list of 800 genes potentially up-regulated by D. coniospora ( Tables S1B & S1C). Nevertheless, several nlp genes, including nlp-29, as well as cnc-4, were found to be induced by M. haptotylum [31]. While these genes are not induced by a number of bacterial pathogens that colonize the nematode intestine [14], [22], [26], another recent report indicates that infection of C. elegans by Leucobacter chromiireducens may provoke an upregulation of nlp-29 [32]. This pathogen infects worms via the uterus. A second Gram-positive bacterium, M. nematophilium, adheres to the nematode cuticle and causes disease, but does not induce the expression of nlp-29, or indeed any of the nlp or cnc genes [33]. On the other hand, wounding the epidermis also provokes an up-regulation of the expression of genes of the nlp-29 cluster, albeit via a genetically-distinct signalling pathway [19]. So both the nature of the pathogen and the route of infection likely play roles in determining the host's transcriptional response. Several of the robustly induced genes are nematode-specific and of unknown function. The predicted ß-lactamase LACT-1, on the other hand, is homologous to prokaryotic proteins that break down antibiotic ß-lactams produced by fungi. One might be tempted to speculate that this protein could act as an intra-cellular sensor for the presence of fungi. Intriguingly, a similar ß-lactamase (LACTB) is encoded in the human genome, but its function is currently unknown [34]. Other induced genes include far-3 that is induced by P. aeruginosa [22], and encodes one of eight related fatty acid- and retinol-binding proteins in C. elegans [35]. A class of structurally unrelated fatty acid-binding proteins (lipophorins) plays a role in clotting in arthropods [36], so FAR-3 might contribute directly to tissue repair. Finally, the gene T19B10.2/ phi-59 is also robustly up-regulated upon fungal infection. Abrogation of its function in worms defective for insulin signaling inhibits osmotic stress resistance [37]. So despite the clear dichotomy between the responses to osmotic stress and infection (e.g. the lack of p38-dependence for the former), as discussed below, it is likely that some genes that are induced upon D. coniospora infection affect pathogen resistance indirectly, not via antimicrobial effects, but by influencing other aspects of organismal physiology. Infection also provoked the specific down-regulation of many genes ( Tables S1A & S1C). A substantial proportion of these genes (including 14 of the 20 most repressed) encode cuticle collagens. This reduction could reflect a general decrease in gene transcription in the epidermis. On the other hand, the expression of many epidermal genes either does not change (e.g. col-12) or is increased (e.g. the genes of the nlp-29 cluster), leaving open the possibility that transcriptional repression could play a specific role in this innate immune response. This must be the subject of future studies. Adaptive evolution of innate immunity genes Parasites and pathogens can represent extremely powerful selective forces because of their ability to evolve rapidly. The resulting diversity of infectious agents favors hosts with a large repertoire of defense responses, including effector molecules with antimicrobial activity. A broad repertoire of AMP genes could evolve via gene duplication [12], [38]. Strong selection on evolution by gene duplication should result in clustered gene families, since gene duplications are usually more frequent across short genomic distances. This has been observed for immune-responsive genes in Drosophila [39]. The large majority of clustered gene families in the C. elegans genome appears to be associated with a function in the organism's interaction with the environment [40]. Consistent with this hypothesis, several clusters of duplicated genes are induced strongly by M. nematophilum infection [33]. Our study identified multiple small clusters of induced immune defense genes. A detailed analysis of the nlp-29 cluster indicated that it is undergoing rapid evolution. Although a sequence in the nlp-30 5′ UTR is found conserved in the 3′ UTR of Cbr-nlp-27, the most parsimonious explanation for the difference observed between the 3 Caenorhabditis species analyzed is that at the time of divergence of C. elegans from the common ancestor of C. briggsae and C. remanei (estimated at 3.1–12.2 MYR [41]) there were 2 genes at the nlp locus. One of them, the ancestral nlp-27, gave rise to 5 genes in C. elegans, while the other, the ancestral nlp-34, gave rise to 3 genes in C. briggsae. This is consistent with the presence of single nlp-27 and nlp-34 orthologues in the syntenic region of the C. brenneri genome (unpublished results). For C. elegans, C. briggsae and C. remanei, the diversification of the nlp genes is associated with adaptive sequence evolution, especially in the case of gene family expansions within species lineages (i.e. the four C. elegans genes nlp-28 up to nlp-31 and the three C. briggsae nlp-34 genes). The lineage-specific nlp expansions could reflect the stochastic nature of gene duplication and subsequent distinct selective pressures on C. elegans and C. briggsae. In the future, it will be interesting to test whether these differences in nlp genes translate into differences in resistance to the pathogens that the two species encounter in the wild. Since we have shown that introducing supernumerary copies of the nlp-29 cluster increases survival, the marginal extra cost of carrying an additional nlp gene is presumably outweighed by the advantage gained in a hostile environment. Further, we show, rather remarkably, that the C. elegans-specific genes nlp-28 and nlp-29 are up-regulated not only by infection and wounding but also by osmotic stress. This does not depend on the p38 MAPK pathway, suggesting that successive gene duplications were followed by divergence of regulatory regions, resulting in the co-option of an nlp gene by a pre-existing osmotic stress pathway, and the acquisition of a supplementary function that was then retained in nlp-28 and nlp-29. Currently, the genetic control of the osmotic stress response is incompletely characterized. Studies underway in other laboratories to delineate the molecular cascades involved should allow such a hypothesis to be tested in the future. Wounding, infection and the osmotic stress response If wounding and infection were associated with an alteration of the worm's osmotic balance, activating AMP gene expression under conditions of osmotic stress could then be a way to protect C. elegans from pathogens without the need to detect the exact nature of the threat. But infection does not affect pgdph-1::GFP expression, nor does wounding (unpublished results). Thus, the tissue damage associated with fungal infection or a needle prick does not trigger an osmotic stress response under laboratory culture conditions. At the same time, there is clearly a link between the response to infection and osmotic stress, exemplified by the up-regulation of certain genes, including nlp-29, under both conditions. Further, the genes dpy-9 and osm-11, which affect the osmotic stress response, act as regulators of nlp-29 expression. While DPY-9 is a cuticular collagen [42], OSM-11 has been proposed to be secreted by the epidermis and to play a sentinel role in monitoring external conditions and mediation of stress responses [25]. If this is the case, under normal conditions, osm-11 would repress nlp-29 expression, and upon osmotic stress or infection, osm-11 activity would decrease, leading to AMP expression. Importantly, however, the increased expression of p nlp-29::GFP seen in either osm-11 or dpy-9 mutant is suppressed in osm-11;pmk-1 or dpy-9; pmk-1 double mutants. By contrast, the pmk-1 p38 pathway is not involved in the regulation of glycerol levels or acute osmotic stress resistance [25], [43]. Nor is it required for the induction of nlp-29 under high salt conditions. Thus the induction of nlp-29 seen in dpy-9 and osm-11 worms might arise from a problem of structural integrity and consequent triggering of the p38 pathway in the mutants, independently of the pathway controlling osmotic stress resistance. Therefore, the level of expression of certain AMP genes could be controlled by the balance of negative OSM-11-dependant and positive PMK-1-dependant regulation. Since the nlp genes appear not to contribute in vivo to increased resistance to osmotic stress, the physiological reason for their induction by salt remains unclear. And finally, although an increased expression of AMP genes may contribute to the resistance of dpy-9 and osm-11 mutants to fungal infection, other factors, most notably high levels of intracellular glycerol, could affect the growth and virulence of D. coniospora. One common facet of the upregulation of nlp-29 following infection or osmotic stress was the partial dependence upon the GATA factor ELT-3. This transcription factor therefore appears to play a necessary role in the epidermis in the regulation of genes that respond to environmental stimuli. As such, it is different from ELT-2 that has been proposed to have a specific role in the regulation of innate immune genes in the intestine [26]. The response to infection, injury, and osmotic stress is not, however, part of a general stress-response mechanism, since no induction of p nlp-29::GFP expression was seen when the worms were exposed to a number of other stressful situations, such as heat shock or starvation. Thus AMP genes appear to contribute directly in vivo to the capacity of C. elegans specifically to resist infection after epidermal damage. Concluding remarks In the current work we have analyzed the host transcriptional changes associated with natural fungal infection in C. elegans. Our findings reinforce the importance of AMP genes in invertebrate innate immunity. The observation that there has been a recent expansion of the AMP-encoding nlp genes, together with the evidence for their in vivo role and the positive selection of the nlp-29 cluster suggest that these genes are important for the survival of C. elegans. In conclusion, this study advances significantly our knowledge of host defenses in the nematode and illustrates how new function may arise during evolution through gene duplication and co-option into existing regulatory mechanisms. Nematode strains The mutant strain sek-1(ag1) [20] was kindly provided by F. Ausubel, nsy-1(ky397) [44] by C. Bargmann, tir-1(tm3036) and nlp-29(tm1931) by S. Mitani (Japanese National Bioresource Project) and fer-15(b26ts), sek-1(km4), pmk-1(km25), dpy-9(e12), dpy-13(e184), dpy-17(e164), osm-11(n1604) and elt-3(gk121) by the Caenorhabditis Genetics Center (CGC). All strains were maintained on nematode growth media (NGM) and fed with E. coli strain OP50, as described [45]. RNA preparation Synchronized populations of either fer-15(b26ts) or N2 worms were cultivated at 25°C until the mid-L4 stage for cDNA or oligo microarrays, respectively. Worms were then transferred to plates spread with fresh D. coniospora spores and harvested after 12 or 24 h, in M9 buffer. In each case, total RNA from 4 independent samples were extracted with Trizol (Invitrogen). Identification of differentially regulated genes We used cDNA microarrays that partially cover the genome (6424 non-redundant cDNA probes). For both 12 h and 24 h fungal infection datasets (4 biological replicates for each time point), genes with background-normalized, photostimulated luminescence (PSL) ratios (infected/control) >1.01 or <0.99 in at least three out of four arrays were initially considered. Normalized data for cDNA arrays can be found in Table S1A. Differentially regulated genes, corresponding to the uppermost 18.75 th percentile of each dataset can be found in Table S1B. We also used oligo microarrays with full genome coverage, containing 23232 features against 20334 unique transcripts. “Per Spot and Per Chip: Intensity Dependent (Lowess) Normalization” in GeneSpring GX version 7.3 (Agilent Technologies) was used to normalize all data (4 biological replicates). Differentially regulated genes based on fold change, corresponding to the uppermost 18.75 th percentile of datasets formed using genes with normalized, expression ratios (infected/control) >1.01 or <0.99 in at least ten out of fourteen arrays are shown in Table S1C. The specificity of the probes corresponding to the nlp and cnc gene is listed in Table S1E. Primary data is deposited at ArrayExpress (E-MEXP-767, E-MEXP-768 and E-MEXP-479). Microarray statistical analyses MAANOVA Various tools as implemented in the software package, J/MAANOVA version 1.0a ( http://www.jax.org/staff/churchill/labsite/) were used. Briefly, raw data was normalized using “Joint Lowess intensity-spatial Lowess” transformation. Normalized data was then analyzed with a variant of the “Mixed Effects ANOVA Model”, in which three components of variance were assumed. Two “fixed” components were array specific effects, condition (pathogen or control) and a “random” component was attributed to the different biological replicates used. Within J/MAANOVA, a Fs-test [46] based on the James-Stein estimator [47] was used to identify genes differentially expressed between our two conditions of interest. Robustness of ANOVA data was tested using a permutation test; means were randomly permuted 500 times and test statistics were recalculated for differences between the two conditions. Agreement between ANOVA and permutation test results would indicate the robustness of the ANOVA model. False discovery rate (FDR) control adapted from algorithms discussed by Y. Benjamin [48] and J. Storey [49] was applied to provide 95% confidence. BRB-ArrayTools A second analysis was performed using tools within BRB-ArrayTools version 3.4.1 developed by the Biometric Research Branch of the US National Cancer Institute ( http://linus.nci.nih.gov/BRB-ArrayTools.html). Lowess intensity dependent normalization was initially used to adjust for differences in labeling intensities of the Cy3 and Cy5 dyes. The adjusting factor varied over intensity levels [50]. Subsequently using “Class Comparison” with dye-swapped experiments being averaged, we identified genes that were differentially expressed among two classes, infected and control, by using a multivariate permutation test. We used this test with 90% confidence so that the false discovery rate was less than 10%. The false discovery rate is the proportion of the list of genes claimed to be differentially expressed that are false positives. The test statistics used were random variance t-statistics for each gene [51]. Although t-statistics were used, the multivariate permutation test is non-parametric and does not require an assumption of Gaussian distributions. Phylogenetic analyses The general phylogenetic position of the antimicrobial nlp and cnc genes was reconstructed in relation to the remaining nlp genes from C. elegans, using peptide sequences. For the nlp-29 clade including the C. elegans genes nlp-27 to nlp-31 and nlp-34, phylogenetic relationships with their 4 orthologues from C. briggsae and 2 from C. remanei were inferred from both peptide and DNA sequences. Sequences were obtained from Wormbase ( www.wormbase.org) Syntenic regions were identified for C. remanei and C. briggsae. They were re-annotated manually using Blast and the new gene predictions submitted to Wormbase. All alignments were generated with the help of CLUSTALW [52]. Phylogenetic analysis was based on the maximum likelihood (ML) approach. The optimal substitution model was identified with the Akaike information criterion and the program Prottest 1.3 for protein and the program Modeltest 3.7 for DNA sequences [53]- [55]. It was then employed for ML tree reconstruction using the program PHYML [56], [57] for protein and the program PAUP* 4.0b10 for DNA sequences [58]. The robustness of the inferred topology was assessed with the help of non-parametric bootstrapping [59], based on 500 replicate data sets. The presence of adaptive Darwinian evolution (i.e. mutations that lead to amino acid changes are selectively favored) was assessed for the nlp-29 clade by analysis of the non-synonymous versus synonymous substitution rate ratio ( dN/dS) across the different branches of the phylogenetic tree. We focused on the peptide regions without the signal sequence, where selection for diversifying functions is expected. Based on the program PAML 3.15 [60], two different approaches were employed [61]. On the one hand, we inferred dN/dS ratios for each individual branch using the free-ratio model. Ratios larger than one indicate adaptive sequence evolution. Their significance was tested by non-parametric bootstrapping using 100 replicate data sets. On the other hand, we compared the likelihood of test trees, in which one branch was allowed to vary in dN/dS ratio (2-ratio model), to the likelihood of the null model tree, in which all branches had identical dN/dS ratios (1-ratio model). If the varying dN/dS ratio was larger than one and if the difference between test and null model tree was significant according to a likelihood ratio test (LRT) [60], [62], then this was taken as an indication for the presence of adaptive sequence evolution. Reporter gene constructs and generation of transgenic lines The frIs7 transgenic strain containing the p nlp-29::GFP and p col-12::DsRed reporters is described elsewhere [19]. p nlp-27::GFP, p nlp-30::GFP and p nlp-31::GFP [10], [18] were injected at 20 µg/ml into N2 with the transformation marker p col-12::DsRed (80 µg/ml). At least three independent lines for each p nlp::GFP reporter were characterized. The IG615 ( frEx217) transgenic strain for overexpression of the nlp locus was produced by injecting the entire nlp locus into wild type N2 worms with the co-injection marker pB unc-53::GFP [63]. The entire locus was included in a 11 kb SpeI- ClaI fragment from the cosmid B0213 that was purified after migration on 0.4% agarose gel into a buffer-filled trough cut in the gel [64]. Infection, wounding and osmotic stress Infections with a freshly harvested solution of D. coniospora spores were done as described [65]; worms were analyzed after 24 h at 25°C. Worms were pricked in the tail region using a microinjection needle under a dissecting microscope and analyzed 6 h later for GFP induction or 2 hours later for qRT-PCR. Osmotic stress was done in liquid by incubating young adult worms in 300 mM NaCl for 6 h, or on NGM plates containing 300 mM NaCl after RNAi treatment with analysis 24 h later. Analyses with the Biosort worm sorter Upregulation of p nlp-29::GFP reporter gene levels were quantified with the COPAS Biosort (Union Biometrica). Generally 100 to 2000 synchronized worms were analyzed for size (TOF), extension (EXT), green (GFP) and red (RFP) fluorescence (see [19], for further details). The fluorescence ratio (Green/Red) was then calculated to normalize the GFP for variations in size and health of individual worms. Mean values for FR were calculated and the values for the different samples within a single experiment normalized so that the control ( wt;frIs7) worms had a fluorescence ratio of 1. As discussed more extensively elsewhere, [19] direct numerical comparisons can be made between age-matched populations in single experiments, and qualitative comparisons can be made between experiments performed on different days. The results shown are representative of at least 3 independent experiments. qRT PCR 2.5 µg of total mRNA from infected and non-infected worms were used for a reverse transcription using a standard protocol. Primers were designed to detect specific transcripts (see Protocol S1). Using 1/50 of cDNA in 12.5 µl of SYBRgreen mix (Applied Biosystem) and 0.3 µM of primers, qRT PCR were performed on a Gene Amp 5700 Sequence detector. Results were normalized to act-1, and then relative expression calculated using 2 ((A+10)−x), A being the normalized cycle number for nlp-27 in the non-infected sample and x the value of interest. Control and experimental conditions were tested in the same run. Means and standard deviations were calculated from a minimum of 3 independent experiments. Statistical analyses used the paired bilateral Student's test within Excel (Microsoft software) ( Table S3). Infection and osmotic assays 50–70 worms at the L4 stage were infected at 25°C with D. coniospora and the surviving worms were counted every day as described elsewhere [65] except that the NGM plates were seeded with heat killed OP50. Killing assays were conducted at 25°C. Statistical analyses used one-sided log rank test within Prism (Graphpad software). The resistance to acute osmotic stress was assayed after 10 min on NGM plate containing 500 mM NaCl with young adult worms as described [25]Protocol S1 Supporting Materials and Methods (0.03 MB DOC) Figure S1 New classes of putative anti-microbial peptides and proteins. Raw output from CLUSTALW multiple sequence alignments for GRSP, FIP and FIPR protein sequences. The genes identified as being strongly up-regulated by D. coniospora are highlighted in yellow. (0.05 MB DOC) Figure S2 Phylogenetic tree for the C. elegans NLP and CNC proteins, inferred with maximum likelihood from protein sequences. Branches are drawn in proportion to the estimated number of substitutions per site. The results of bootstrapping (200 replicates) are indicated next to the branches. Only values of at least 50 are shown. (1.44 MB EPS) Figure S3 Tree topology for the nlp genes with branch labels as tested for the presence of adaptive sequence evolution (see Table S2). (0.95 MB EPS) Figure S4 The high expression of nlp-29 in dpy-9 mutants is almost entirely dependant on pmk-1. Fluorescent pictures (A) and quantification with the Biosort of the normalized fluorescent ratio (green/red) (B) of worms carrying the integrated frIs7 transgene in different mutant backgrounds, dpy-9(e12), pmk-1(km25) and dpy-9(e12);pmk-1(km25). The number of worms used in each quantification is shown in parenthesis. (1.84 MB EPS) Figure S5 The presence of the cosmid containing the nlp-29 locus increases resistance to infection. (A) Worms were infected by D. coniospora, then transferred to OP50 seeded NGM plates containing the anti fungal agent nystatin (12.5 µg/ml) after 3 or 12 h. Transgenic worms carrying the cosmid B0213 (in black) or their non-transgenic sibling (in red) were scored as live or dead over 6 days. In both cases, the difference in survival is highly significant (p<0.001, one-side log rank test). The IG368 ( frEx75) transgenic strain contains the whole cosmid B0213 including the nlp locus and the co-injection marker sur-5::GFP [66]. (B) The resistance of strain CX6760 ( wt; kyEx749[ F13B10.1;ofm-1::GFP]) [67] is indistinguishable from that of wild-type worms under standard conditions [68]. (0.66 MB EPS) Figure S6 Alignment of the promoter sequences of the genes from the nlp-29 locus. Raw output from CLUSTALW multiple alignments for the proximal 500 bp 5′ sequences for the 6 genes of the nlp-29 cluster. Putative minimal GATA sites are highlighted in yellow; one is shared between 3 nlp genes. (0.06 MB DOC) Table S1 Expression patterns of 14 selected genes. (0.82 MB XLS) Table S2 Results of the analysis of adaptive sequence evolution for individual branches of the nlp tree. (0.06 MB DOC) Table S3 qRT-PCR. (0.04 MB XLS) We thank Y. Duverger and S. Scaglione for worm sorting, L. Sofer (UMR1112-INRA/University Nice Sophia Antipolis) for expert technical assistance, and A. Coulson for cosmids. Microarray experiments and worm sorting were carried out using the facilities of the Marseille-Nice Genopole. Some nematode strains used in this work were provided by the Caenorhabditis Genetics Center, which is funded by the NIH National Center for Research Resources (NCRR), or by the National Bioresource Project coordinated by S. Mitani. We thank C. Braendle, S. Cypowyj, K.-Z. Lee, and K. Ziegler for discussion and T. Lamitina and U. Theopold for comments on a previous version of the manuscript. 1. Schmid-Hempel P, Ebert D. On the evolutionary ecology of specific immune defence. Trends Ecol Evol. 2003;18:27–32. 2. Hodgkin J, Kuwabara PE, Corneliussen B. A novel bacterial pathogen, Microbacterium nematophilum, induces morphological change in the nematode C. elegans. Curr Biol. 2000;10:1615–1618. [PubMed] 3. Darby C, Hsu JW, Ghori N, Falkow S. Caenorhabditis elegans: plague bacteria biofilm blocks food intake. Nature. 2002;417:243–244. [PubMed] 4. Couillault C, Ewbank JJ. Diverse Bacteria Are Pathogens of Caenorhabditis elegans. Infect Immun. 2002;70:4705–4707. [PubMed] 5. Jansson HB, Jeyaprakash A, Zuckerman BM. Differential adhesion and infection of nematodes by the endoparasitic fungus Meria coniospora (Deuteromycetes). Appl Envir Microbiol. 1985;49:552–555. 6. Laube DM, Yim S, Ryan LK, Kisich KO, Diamond G. Antimicrobial peptides in the airway. Curr Top Microbiol Immunol. 2006;306:153–182. [PubMed] 7. Lemaitre B, Hoffmann J. The Host Defense of Drosophila melanogaster. Annu Rev Immunol. 2007 8. Lievin-Le Moal V, Servin AL. The front line of enteric host defense against unwelcome intrusion of harmful microorganisms: mucins, antimicrobial peptides, and microbiota. Clin Microbiol Rev. 2006;19:315–337. [PubMed] 9. Schroder JM, Harder J. Antimicrobial skin peptides and proteins. Cell Mol Life Sci. 2006;63:469–486. [PubMed] 10. Couillault C, Pujol N, Reboul J, Sabatier L, Guichou JF, et al. TLR-independent control of innate immunity in Caenorhabditis elegans by the TIR domain adaptor protein TIR-1, an ortholog of human SARM. Nat Immunol. 2004;5:488–494. [PubMed] 11. Shai Y. Mode of action of membrane active antimicrobial peptides. Biopolymers. 2002;66:236–248. [PubMed] 12. Ohno S. New York: Springer Verlag; 1970. Evolution by gene duplication. 13. Mochii M, Yoshida S, Morita K, Kohara Y, Ueno N. Identification of transforming growth factor-beta- regulated genes in Caenorhabditis elegans by differential hybridization of arrayed cDNAs. Proc Natl Acad Sci U S A. 1999;96:15020–15025. [PubMed] 14. Wong D, Bazopoulou D, Pujol N, Tavernarakis N, Ewbank JJ. Genome-wide investigation reveals pathogen-specific and shared signatures in the response of C. elegans to infection. Genome Biol. 2007;8:R194. [PubMed] 15. Fjell CD, Hancock RE, Cherkasov A. AMPer: a database and an automated discovery tool for antimicrobial peptides. Bioinformatics. 2007;23:1148–1155. [PubMed] 16. Allison DB, Cui X, Page GP, Sabripour M. Microarray data analysis: from disarray to consolidation and consensus. Nat Rev Genet. 2006;7:55–65. [PubMed] 17. Kiontke K, Sudhaus W. Ecology of Caenorhabditis species. The C. elegans Research Community, editor. WormBook. 2006. pp. 1–14. 18. Nathoo AN, Moeller RA, Westlund BA, Hart AC. Identification of neuropeptide-like protein gene families in Caenorhabditis elegans and other species. Proc Natl Acad Sci U S A. 2001;98:14000–14005. [PubMed] 19. Pujol N, Cypowyj S, Ziegler K, Millet A, Astrain A, et al. Distinct Innate Immune Responses to Infection and Wounding in the C. elegans Epidermis. Curr Biol. 2008;18:481–489. [PubMed] 20. Kim DH, Feinbaum R, Alloing G, Emerson FE, Garsin DA, et al. A conserved p38 MAP kinase pathway in Caenorhabditis elegans innate immunity. Science. 2002;297:623–626. [PubMed] 21. Kim DH, Liberati NT, Mizuno T, Inoue H, Hisamoto N, et al. Integration of Caenorhabditis elegans MAPK pathways mediating immunity and stress resistance by MEK-1 MAPK kinase and VHP-1 MAPK phosphatase. Proc Natl Acad Sci U S A. 2004;101:10990–10994. [PubMed] 22. Troemel ER, Chu SW, Reinke V, Lee SS, Ausubel FM, et al. p38 MAPK Regulates Expression of Immune Response Genes and Contributes to Longevity in C. elegans. PLoS Genet. 2006;2:e183. doi:10.1371/journal.pgen.0020183. [PubMed] 23. Liberati NT, Fitzgerald KA, Kim DH, Feinbaum R, Golenbock DT, et al. Requirement for a conserved Toll/interleukin-1 resistance domain protein in the Caenorhabditis elegans immune response. Proc Natl Acad Sci U S A. 2004;101:6593–6598. [PubMed] 24. Lamitina T, Huang CG, Strange K. Genome-wide RNAi screening identifies protein damage as a regulator of osmoprotective gene expression. Proc Natl Acad Sci U S A. 2006;103:12173–12178. [PubMed] 25. Wheeler JM, Thomas JH. Identification of a novel gene family involved in osmotic stress response in Caenorhabditis elegans. Genetics. 2006 26. Shapira M, Hamlin BJ, Rong J, Chen K, Ronen M, et al. A conserved role for a GATA transcription factor in regulating epithelial innate immune responses. Proc Natl Acad Sci U S A. 2006;103:14086–14091. [PubMed] 27. Reece-Hoyes JS, Deplancke B, Shingles J, Grove CA, Hope IA, et al. A compendium of Caenorhabditis elegans regulatory transcription factors: a resource for mapping transcription regulatory networks. Genome Biol. 2005;6:R110. [PubMed] 28. Koh K, Rothman JH. ELT-5 and ELT-6 are required continuously to regulate epidermal seam cell differentiation and cell fusion in C. elegans. Development. 2001;128:2867–2880. [PubMed] 29. Gilleard JS, McGhee JD. Activation of hypodermal differentiation in the Caenorhabditis elegans embryo by GATA transcription factors ELT-1 and ELT-3. Mol Cell Biol. 2001;21:2533–2544. [PubMed] 30. Gilleard JS, Shafi Y, Barry JD, McGhee JD. ELT-3: A Caenorhabditis elegans GATA factor expressed in the embryonic epidermis during morphogenesis. Dev Biol. 1999;208:265–280. [PubMed] 31. Fekete C, Tholander M, Rajashekar B, Ahren D, Friman E, et al. Paralysis of nematodes: shifts in the transcriptome of the nematode-trapping fungus Monacrosporium haptotylum during infection of Caenorhabditis elegans. Environ Microbiol. 2008;10:364–375. [PubMed] 32. Muir RE, Tan MW. Leucobacter chromiireducens subsp. solipictus Exerts Virulence on Caenorhabditis elegans, Characterization of a Novel Host-Pathogen Interaction. Appl Environ Microbiol. 2008 33. O'Rourke D, Baban D, Demidova M, Mott R, Hodgkin J. Genomic clusters, putative pathogen recognition molecules, and antimicrobial genes are induced by infection of C. elegans with M. nematophilum. Genome Res. 2006;16:1005–1016. [PubMed] 34. Liobikas J, Polianskyte Z, Speer O, Thompson J, Alakoskela JM, et al. Expression and purification of the mitochondrial serine protease LACTB as an N-terminal GST fusion protein in Escherichia coli. Protein Expr Purif. 2006;45:335–342. [PubMed] 35. Garofalo A, Rowlinson MC, Amambua NA, Hughes JM, Kelly SM, et al. The FAR protein family of the nematode Caenorhabditis elegans. Differential lipid binding properties, structural characteristics, and developmental regulation. J Biol Chem. 2003;278:8065–8074. [PubMed] 36. Karlsson C, Korayem AM, Scherfer C, Loseva O, Dushay MS, et al. Proteomic analysis of the Drosophila larval hemolymph clot. J Biol Chem. 2004;279:52033–52041. [PubMed] 37. Lamitina ST, Strange K. Transcriptional targets of the DAF-16 insulin signaling pathwayprotect C. elegans from extreme hypertonic stress. Am J Physiol Cell Physiol. 2004 38. Nei M, Rooney AP. Concerted and birth-and-death evolution of multigene families. Annu Rev Genet. 2005;39:121–152. [PubMed] 39. De Gregorio E, Spellman PT, Tzou P, Rubin GM, Lemaitre B. The Toll and Imd pathways are the major regulators of the immune response in Drosophila. EMBO J. 2002;21:2568–2579. [PubMed] 40. Thomas JH. Analysis of homologous gene clusters in Caenorhabditis elegans reveals striking regional cluster domains. Genetics. 2006;172:127–143. [PubMed] 41. Cutter AD, Felix MA, Barriere A, Charlesworth D. Patterns of nucleotide polymorphism distinguish temperate and tropical wild isolates of Caenorhabditis briggsae. Genetics. 2006;173:2021–2031. [PubMed] 42. Simmer F, Moorman C, Van Der Linden AM, Kuijk E, Van Den Berghe PV, et al. Genome-Wide RNAi of C. elegans Using the Hypersensitive rrf-3 Strain Reveals Novel Gene Functions. PLoS Biol. 2003;1:e12. doi: 10.1371/journal.pbio.0000012. [PubMed] 43. Solomon A, Bandhakavi S, Jabbar S, Shah R, Beitel GJ, et al. Caenorhabditis elegans OSR-1 regulates behavioral and physiological responses to hyperosmotic environments. Genetics. 2004;167:161–170. [PubMed] 44. Troemel ER, Sagasti A, Bargmann CI. Lateral signaling mediated by axon contact and calcium entry regulates asymmetric odorant receptor expression in C. elegans. Cell. 1999;99:387–398. [PubMed] 45. Stiernagle T. Maintenance of C. elegans. The C. elegans Research Community, editor. WormBook. 2006. pp. 1551–8507. http://www.wormbook.org: WormBook. 46. Cui X, Hwang JT, Qiu J, Blades NJ, Churchill GA. Improved statistical tests for differential gene expression by shrinking variance components estimates. Biostatistics. 2005;6:59–75. [PubMed] 47. Lindley DV. Discussion of Professor Stein's paper, ‘Confidence sets for the mean of a multivariate normal distribution’. J R Stat Soc [Ser B] 1962;24:265–296. 48. Benjamin Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc [Ser B] 1995;57:289. 49. Storey JD. A direct approach to false discovery rates. J R Stat Soc [Ser B] 2002;64:479–498. 50. Yang YH, Dudoit S, Luu P, Lin DM, Peng V, et al. Normalization for cDNA microarray data: a robust composite method addressing single and multiple slide systematic variation. Nucleic Acids Res. 2002;30:e15. [PubMed] 51. Wright GW, Simon RM. A random variance model for detection of differential gene expression in small microarray experiments. Bioinformatics. 2003;19:2448–2455. [PubMed] 52. Thompson JD, Higgins DG, Gibson TJ. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 1994;22:4673–4680. [PubMed] 53. Posada D, Crandall KA. MODELTEST: testing the model of DNA substitution. Bioinformatics. 1998;14:817–818. [PubMed] 54. Posada D, Buckley TR. Model selection and model averaging in phylogenetics: advantages of akaike information criterion and bayesian approaches over likelihood ratio tests. Syst Biol. 2004;53:793–808. [PubMed] 55. Abascal F, Zardoya R, Posada D. ProtTest: selection of best-fit models of protein evolution. Bioinformatics. 2005;21:2104–2105. [PubMed] 56. Guindon S, Gascuel O. A simple, fast, and accurate algorithm to estimate large phylogenies by maximum likelihood. Syst Biol. 2003;52:696–704. [PubMed] 57. Guindon S, Lethiec F, Duroux P, Gascuel O. PHYML Online–a web server for fast maximum likelihood-based phylogenetic inference. Nucleic Acids Res. 2005;33:W557–559. [PubMed] 58. Swofford D. PAUP*. Phylogenetic analysis using parsimony (*and other methods). Version 4 Sinauer Associates, Sunderland, Massachusetts. 2003 59. Felsenstein J. Confidence limits on phylogenies-an approach using the bootstrap. Evolution Int J Org Evolution. 1985;39:783–791. 60. Yang Z. PAML: a program package for phylogenetic analysis by maximum likelihood. Comput Appl Biosci. 1997;13:555–556. [PubMed] 61. Schulenburg H, Boehnisch C. Diversification and adaptive sequence evolution of Caenorhabditis lysozymes (Nematoda: Rhabditidae). BMC Evol Biol. 2008;8:114. [PubMed] 62. Yang Z, Wong WS, Nielsen R. Bayes empirical bayes inference of amino acid sites under positive selection. Mol Biol Evol. 2005;22:1107–1118. [PubMed] 63. Stringham E, Pujol N, Vandekerckhove J, Bogaert T. unc-53 controls longitudinal migration in C. elegans. Development. 2002;129:3367–3379. [PubMed] 64. Hansen H, Lemke H, Bodner U. Rapid and simple purification of PCR products by direct band elution during agarose gel electrophoresis. Biotechniques. 1993;14:28–30. [PubMed] 65. Powell JR, Ausubel FM. Models of Caenorhabditis elegans Infection by Bacterial and Fungal Pathogens. In: Ewbank J, Vivier E, editors. Methods Mol Biol: Humana Press; 2008. pp. 403–427. 66. Gu T, Orita S, Han M. Caenorhabditis elegans SUR-5, a novel but conserved protein, negatively regulates LET-60 Ras activity during vulval induction. Mol Cell Biol. 1998;18:4556–4564. [PubMed] 67. Chuang CF, Bargmann CI. A Toll-interleukin 1 repeat protein at the synapse specifies asymmetric odorant receptor expression via ASK1 MAPKKK signaling. Genes Dev. 2005;19:270–281. [PubMed] 68. Powell JR, Ausubel FM. Models of Caenorhabditis elegans infection by bacterial and fungal pathogens. Methods Mol Biol. 2008;415:403–427. [PubMed]
|
This article has been 
