Modulation of Caenorhabditis elegans infection sensitivity by the LIN-7 cell junction protein.

In Caenorhabditis elegans, the LIN-2/7/10 protein complex regulates the activity of signalling proteins. We found that inhibiting lin-7 function, and also lin-2 and lin-10, resulted in enhanced C. elegans survival after infection by Burkholderia spp., implicating a novel role for these genes in modulating infection outcomes. Genetic experiments suggested that this infection phenotype is likely caused by modulation of the DAF-2 insulin/IGF-1 signalling pathway. Supporting these observations, yeast two-hybrid assays confirmed that the LIN-2 PDZ domain can physically bind to the DAF-2 C-terminus. Loss of lin-7 activity also altered DAF-16 nuclear localization kinetics, indicating an additional contribution by hsf-1. Unexpectedly, silencing lin-7 in the hypodermis, but not the intestine, was protective against infection, implicating the hypodermis as a key tissue in this phenomenon. Finally, consistent with lin-7 acting as a general host infection factor, lin-7 mutants also exhibited enhanced survival upon infection by two other Gram-negative pathogens, Pseudomonas and Salmonella spp.

Introduction

The soil nematode Caenorhabditis elegans has been used as a model system to dissect host–pathogen interactions for many microbes, including Burkholderia, Pseudomonas and Salmonella spp. (Aballay et al., 2000; Labrousse et al., 2000; Gan et al., 2002). In many cases, bacterial mutants exhibiting reduced virulence in mice also displayed attenuated virulence in the nematode, supporting the suitability of C. elegans as a model for studying selected aspects of mammalian host–pathogen interactions (Tan et al., 1999; Gan et al., 2002; Tenor et al., 2004). Use of the C. elegans model has also delineated several conserved host pathways modulating infection outcomes, including the p38 mitogen-activated protein kinase (MAPK), TGF-β and DAF-2 insulin/IGF-1 signalling pathways (Kim et al., 2002; Mallo et al., 2002; Garsin et al., 2003).

Upon pathogen exposure, activation of host defence pathways occur in a co-ordinated and integrated manner to elicit effective and appropriate host responses (Alper et al., 2007). Previous research has shown that many components of these host defence pathways are broadly expressed, making it likely that they need to be kept under strict regulatory control across different cells and tissues. Reflecting the importance of tissue specificity, different cellular compartments in C. elegans have also been shown to exhibit distinct pathogen responses under specific challenges. For example, in the intestine, p38 MAPK regulates the lys-2 lysozyme during bacterial infection (Ren et al., 2009); while in the hypodermis, p38 MAPK regulates the antimicrobial peptide nlp-29 during antifungal responses (Pujol et al., 2008; Ziegler et al., 2009).

DAF-2 insulin/IGF-1 signalling is one of the most extensively studied pathways regulating host infection outcomes in C. elegans. DAF-2 signalling is also involved in modulating lifespan, dauer formation and resistance to abiotic stresses (Kenyon et al., 1993; Murakami and Johnson, 1996; Kimura et al., 1997; Barsyte et al., 2001). Activation of the DAF-2 receptor tyrosine kinase (RTK) initiates a downstream phosphatidylinositol 3-kinase (PI3K) pathway, involving the AKT-1 kinase (Paradis and Ruvkun, 1998), which culminates in the phosphorylation of DAF-16, a forkhead-related FOXO transcription factor, and subsequent egression of phosphorylated DAF-16 from the nucleus into the cytoplasm (Lin et al., 1997; Ogg et al., 1997). DAF-16 regulates genes involved in stress responses, innate immunity and antimicrobial function (Murphy et al., 2003) and mediates the phenotypes exhibited by daf-2 loss-of-function mutants (Kenyon et al., 1993; Garsin et al., 2003). Besides DAF-16, the DAF-2 signalling pathway also affects heat shock factor 1 (HSF-1), a transcription factor regulating several heat shock proteins (HSPs) involved in the heat shock response. Upon induction, HSPs act as chaperones binding to unfolded or damaged proteins (Frydman, 2001). Similar to DAF-16, HSF-1 also contributes to the longevity of daf-2 mutants and their enhanced ability to survive pathogenic assaults (Hsu et al., 2003; Singh and Aballay, 2006).

In this study, we performed a targeted reverse genetic screen to identify new genes involved in modulating C. elegans infection outcomes. Among the genes screened, we identified lin-7, a cell junction gene (Simske et al., 1996), as a factor important for influencing C. elegans survival upon infection with bacterial pathogens. Genetic and biochemical experiments revealed that lin-7 modulates the DAF-2 insulin/IGF-1 signalling pathway, by binding directly to the DAF-2 RTK via LIN-2, and that both daf-16 and hsf-1 are required for the enhanced survival exhibited by lin-7 mutants upon infection. We also found that lin-7 functions predominantly in the C. elegans hypodermis to modulate infection outcomes. Our results thus reveal a regulatory connection between LIN-7 and the DAF-2 signalling pathway, and suggest an important role for the hypodermis in dictating the survival outcome of a nematode during bacterial infection.

Results

Burkholderia thailandensis accumulates in the C. elegans intestine during infection

Although Burkholderia spp. have been shown to kill C. elegans (O'Quinn et al., 2001; Gan et al., 2002; Köthe et al., 2003), the specific routes of infection used by Burkholderia spp. to infect nematodes have not been clearly determined. To address this, we infected nematodes with Burkholderia thailandensis E555, a strain expressing a capsular polysaccharide detectable by the monoclonal antibody 3015. Staining infected nematodes with the antibody 3015 allows visualization of each individual bacterium, defined by a clear circular band (Sim et al., 2010). Wild-type N2 nematodes infected with B. thailandensis E555, or with the reference strain B. thailandensis ATCC 700388, exhibited rapid death, with 100% of the nematodes demonstrating lethality after 4 days (P < 0.0001, Fig. 1A). Control nematodes fed with GFP-expressing Escherichia coli OP50 exhibited a broadly diffuse green fluorescence throughout the intestinal lumen, reflecting efficient bacterial destruction by the pharyngeal grinder (Fig. 1B, top panel). In contrast, immunofluorescence assays of B. thailandensis E555-infected nematodes revealed clearly defined intact B. thailandensis bacteria throughout the intestinal lumen from 8 h (Fig. 1B, bottom panel) until 32 h post infection (Fig. S1).

Fig. 1

During infection, B. thailandensis accumulates in the C. elegans intestine.

A. Survival of wild-type nematodes was compared when fed E. coli OP50, B. thailandensis ATCC 700388 (P < 0.0001) or B. thailandensis E555 (P < 0.0001; all P values as compared with E. coli OP50). Each survival curve is representative of three independent experiments, each with three plates per strain and 40 nematodes per plate.

B. Wild-type nematodes were fed E. coli OP50::GFP for 8 h (top). Live nematodes were subsequently mounted for microscopy in PBS with NaN3. In these merged images, E. coli OP50 is shown in green and intestinal autofluorescence in blue. Wild-type nematodes were exposed to B. thailandensis E555 for 8 h (bottom). Infected nematodes were then fixed, permeabilized and labelled with monoclonal antibody 3015 and DAPI. In these merged images, B. thailandensis E555 is shown in red and DAPI in blue. Images are shown at 40× (left) or 100× magnification (right). The pharynx (P) and individual bacteria (white arrowheads) are indicated. Scale bar represents 0.01 mm.

A reverse genetic screen identifies lin-7 as a host factor to B. thailandensis infection

Since B. thailandensis was detected in the nematode intestinal lumen, the intestine may represent an important site for host–pathogen interactions. We hypothesized that screening genes expressed in the intestine might identify factors important in modulating host infection outcomes. Querying a publicly accessible database (WormBase, release WS180), we selected 81 genes reported to be expressed in nematode intestinal cells (Table S2). Using RNA interference (RNAi) (Timmons and Fire, 1998), we systematically tested each of the 81 genes for its effect in modulating nematode survival during infection. Briefly, wild-type nematodes were treated with E. coli HT115 clones producing gene-targeting double-stranded RNA (dsRNA) (Fire et al., 1998), followed by exposure to B. thailandensis (see Experimental procedures). We specifically identified genes that, when silenced, resulted in prolonged C. elegans survival during infection. This strategy eliminates potential confounding factors due to general sickness caused by specific RNAi treatment, which might result in nematodes exhibiting decreased survival during infection.

Of the 81 genes screened, RNAi knock-down of 4 genes (lin-7, myo-3, rfp-1 and rrt-2) conferred significant enhancements of survival (data not shown). Among these, we selected lin-7, encoding a cell junction protein (Simske et al., 1996), for further analysis. Wild-type nematodes treated with lin-7 dsRNA survived significantly longer than nematodes treated with parental E. coli HT115 upon B. thailandensis infection (P < 0.0001, Fig. 2A). To verify the use of parental E. coli HT115 as a control, wild-type nematodes treated with either parental E. coli HT115 or HT115 with an empty RNAi vector did not differ in their susceptibilities to B. thailandensis infection (P = 0.2896, Fig. S2).

Fig. 2

Loss of lin-7 confers enhanced survival upon B. thailandensis infection.

A. Wild-type nematodes were first grown on parental E. coli HT115 or exposed to lin-7 dsRNA before transferring to plates containing B. thailandensis ATCC 700388 (P < 0.0001).

B. Wild-type or lin-7 (e1449) nematodes were exposed to B. thailandensis ATCC 700388 (P < 0.0001).

C. Wild-type, lin-7 (e1449) (P < 0.0001), lin-7 (e1413) (P < 0.0001) or lin-7 (n106) (P = 0.0013; all P values as compared with wild-type) nematodes were first exposed to cdc-25.1 dsRNA before transferring to plates containing B. thailandensis ATCC 700388.

D. Wild-type or lin-7 (e1449) nematodes were first exposed to cdc-25.1 dsRNA until day 1 adult stage and subsequently after every 2 days, transferred to fresh plates containing the same HT115 strain for lifespan studies (P = 0.5661).

E. Wild-type or lin-7 (e1449) nematodes were exposed to B. thailandensis E555 for 24 h. Infected nematodes were then fixed, permeabilized and labelled with monoclonal antibody 3015 and DAPI. In these merged images, B. thailandensis E555 is shown in red and DAPI in blue. Images are shown at 40× magnification (right) and the pharynx (P) is indicated. Scale bar represents 0.01 mm.

F. Wild-type (▪) or lin-7 (e1449) (▴) nematodes were exposed to B. thailandensis ATCC 700388 for 24 h (P = 0.6610). Infected nematodes were then lysed mechanically and chemically to release intestinal bacteria. Lysates were plated on LB agar with gentamicin and amount of live bacteria per nematode was determined by cfu counts. Each symbol represents the average of 25–35 nematodes and horizontal lines indicate the geometric mean of triplicates.

To validate the RNAi results, we repeated the infection assays using the lin-7 (e1449) loss-of-function mutant (Ferguson and Horvitz, 1985). The lin-7 genetic mutant also survived significantly longer than the wild-type nematode upon exposure to B. thailandensis (P < 0.0001, Fig. 2B).

Previous studies have focused on LIN-7's role in C. elegans vulval epithelial cells, where it modulates the localization and activity of LET-23 RTK, a regulator of vulval development (Simske et al., 1996; Kaech et al., 1998). lin-7 (e1449) mutants are vulvaless and produce a characteristic ‘bag of worms’ phenotype (Ferguson and Horvitz, 1985). To address whether the lin-7-mediated infection phenotype was confounded by lin-7 mutants being prone to matricidal hatching, we treated an allelic series of three independent lin-7 loss-of-function genetic mutants, lin-7 (e1449), lin-7 (e1443) and lin-7 (n106) (Ferguson and Horvitz, 1985), with cdc-25.1 dsRNA to render them sterile (Evans et al., 2008; Shapira and Tan, 2008). Similar to our observations for fertile nematodes, sterile lin-7 (e1449) (P < 0.0001), lin-7 (e1413) (P < 0.0001) and lin-7 (n106) (P = 0.0013, Fig. 2C) mutants also exhibited enhanced survival during infection compared with sterile wild-type nematodes.

Previous research has shown that C. elegans pathways regulating lifespan are also involved in the modulation of host infection outcomes (Garsin et al., 2003), leading to speculations that both biological processes are controlled by the same underlying genetic mechanisms (Lithgow, 2003). However, subsequent work reported that enhanced survival during infection is not just a given consequence of longevity (Evans et al., 2008). We confirmed that the lin-7-mediated infection phenotype is not likely a secondary consequence of lifespan extension because the lifespan of sterile lin-7 (e1449) mutants was comparable to that of sterile wild-type nematodes grown on non-pathogenic E. coli (P = 0.5661, Fig. 2D). These results suggest a hitherto undescribed role for lin-7 during bacterial infection.

To further define the infection phenotype exhibited by lin-7 mutants, we performed immunofluorescence assays as described above. At 24 h post infection, we could not observe any significant difference in the intestinal colonization profiles between infected wild-type nematodes and lin-7 (e1449) mutants (Fig. 2E). Similar results were obtained when we directly measured intestinal B. thailandensis loads by quantifying colony-forming units (cfu) from infected nematodes – there was no significant difference between the abilities of the wild-type nematode and the lin-7 mutant to limit intestinal pathogen growth (P = 0.6610, Fig. 2F). The observation that the pathogen load was not substantially different between the wild-type nematode and the lin-7 mutant suggests that the enhanced survival exhibited by the lin-7 mutant is unlikely due to a more efficient restriction on intestinal bacterial growth.

LIN-7-associated LIN-2 binds to DAF-2 RTK

LIN-7 physically associates with two other proteins, LIN-2 and LIN-10, to form a multi-protein complex regulating the function of signalling receptors (Kaech et al., 1998; Kim and Sheng, 2004; Alewine et al., 2007). To determine if the LIN-2/7/10 complex modulates infection outcomes, we subjected lin-2 and lin-10 loss-of-function mutants to B. thailandensis infection. Similar to lin-7 mutants, lin-2 (e1309) (P < 0.0001) and lin-10 (n1402) (P < 0.0001) mutants (Horvitz and Sulston, 1980; Ferguson and Horvitz, 1985; Hoskins et al., 1996) also survived significantly longer than wild-type nematodes upon B. thailandensis infection (Fig. 3A ). These results suggest that LIN-7, LIN-2 and LIN-10 may function together as a tripartite complex to modulate infection outcomes.

Fig. 3

lin-7, lin-2 and lin-10 mutants exhibit similar infection phenotypes.

A. Wild-type, lin-7 (e1449) (P < 0.0001), lin-2 (e1309) (P < 0.0001), lin-10 (n1402) (P < 0.0001) or let-23 (sy1) (P = 0.1864; all P values as compared with wild-type) nematodes were exposed to B. thailandensis ATCC 700388.

B. Diagram highlighting the PDZ domains (shaded in black) of LIN-7, LIN-2 and LIN-10 proteins along with (1) known or (2) potential interacting receptors. The PDZ binding consensus motifs at the C-terminals of interacting partners are indicated in parentheses.

C. The PDZ domain of LIN-2 (amino acids 288–647) and the C-terminus of DAF-2 (amino acids 1445–1843) were expressed as fusion proteins with the GAL4 AD and DNA-BD respectively. SV40 large T antigen (T-Ag) and murine p53 (p53), as fusion proteins with the same GAL4 AD and DNA-BD, respectively, served as positive controls; human lamin C (Lam), fused to the GAL4 DNA-BD, was used as a negative control. Individually transformed haploid yeast cells were mated and diploids with positive protein–protein interactions were selected on synthetically defined media lacking leucine (-Leu), tryptophan (-Trp), adenine (-Ade) and histidine (-His).

In the C. elegans vulva, the LIN-2/7/10 complex has been shown to regulate the localization and activity of the LET-23 RTK in epithelial cells (Kaech et al., 1998). However, a let-23 (sy1) loss-of-function mutant did not exhibit enhanced survival when compared with wild-type nematodes upon B. thailandensis infection (P = 0.1864, Fig. 3A). This result indicates that the LET-23 RTK is unlikely to mediate the lin-7-associated infection phenotype during B. thailandensis infection.

LIN-7, LIN-2 and LIN-10 contain postsynaptic density-95, disc large, zona occludens (PDZ) domains that can interact with the C-terminals of receptors (Kaech et al., 1998). LIN-7 and LIN-10 contain type I PDZ domains that bind consensus motifs (S/T)X(V/I/L), while LIN-2 has a type II PDZ domain which binds Φ-X-Φ motifs, where Φ is a hydrophobic residue (Songyang et al., 1997). We therefore hypothesized that during infection, the LIN-2/7/10 complex might interact with other signalling receptors to regulate their function.

Based on the C-terminus binding motif information for the three proteins of the LIN-2/7/10 complex, we identified and tested five candidate receptors that have been shown or have the potential to interact with the LIN-2/7/10 complex (Fig. 3B). These included: INA-1, an alpha integrin subunit (Baum and Garriga, 1997); GPC-1 and GPC-2, both heterotrimeric guanine nucleotide-binding protein gamma subunits (Yamada et al., 2009); GLR-1, a glutamate receptor known to interact with LIN-10 at postsynaptic elements (Rongo et al., 1998); DAF-2, the insulin/IGF-1 RTK (Kenyon et al., 1993). We exposed nematodes carrying either loss-of-function mutations (glr-1, gpc-1 and daf-2) or wild-type nematodes pre-treated with gene-specific dsRNA (gpc-2 and ina-1) to B. thailandensis (Figs S3 and 4A). Of the five receptors tested, only daf-2 (e1370) mutants survived significantly longer than wild-type nematodes upon B. thailandensis infection (P < 0.0001, Fig. 4A).

Fig. 4

lin-7 acts in the DAF-2 signalling pathway during infection.

A. Wild-type or daf-2 (e1370) nematodes were first grown on parental E. coli HT115 before transferring to plates containing B. thailandensis ATCC (P < 0.0001). Wild-type (P < 0.0001) or daf-2 (e1370) (P = 0.1738; all P values as compared with its respective parental HT115) nematodes were also exposed to lin-7 dsRNA before infection by B. thailandensis ATCC 700388. Due to the longer assay time, P0 nematodes were transferred to fresh plates containing B. thailandensis ATCC 700388 after every 2 days.

B. Wild-type or akt-1 (mg144)gf nematodes were first grown on parental E. coli HT115 or exposed to lin-7 dsRNA before transferring to plates containing B. thailandensis ATCC 700388.

C. Wild-type (P = 0.8795) or lin-7 (e1449) (P < 0.0001; all P values as compared with its respective parental HT115) nematodes were first grown on parental E. coli HT115 or exposed to daf-16 dsRNA before transferring to plates containing B. thailandensis ATCC 700388.

D. Wild-type (P < 0.0001) or lin-7 (e1449) (P < 0.0001; all P values as compared with its respective parental HT115) nematodes were first grown on parental E. coli HT115 or exposed to hsf-1 dsRNA before transferring to plates containing B. thailandensis ATCC 700388.

The DAF-2 RTK has a P-L-V motif at its extreme C-terminus (amino acids 1841–1843), raising the possibility that the DAF-2 C-terminus may potentially bind to the LIN-2 PDZ domain. To experimentally test this in silico prediction, we used the yeast two-hybrid system (Fields and Song, 1989). In this assay, the PDZ-containing portion of LIN-2 (amino acids 288–647 fused to the GAL4 transcriptional activation domain) interacted with the C-terminus of DAF-2 (amino acids 1445–1843 fused to the GAL4 DNA-binding domain), as detected by expression of both ade2 and his3 reporter genes downstream of two different GAL4-responsive promoters (Fig. 3C). Confirming the specificity of this interaction, the LIN-2 PDZ domain did not interact with human lamin C nor murine p53, and the DAF-2 C-terminus did not interact with SV40 large T antigen (Fig. 3C). These data suggest that LIN-7 binds to DAF-2 RTK via LIN-2, potentially regulating DAF-2 signalling activities.

LIN-7 interacts with the DAF-2 insulin/IGF-1 signalling pathway

The findings that daf-2 (e1370) mutants exhibited enhanced survival when exposed to B. thailandensis and that the LIN-2 PDZ domain bound to the DAF-2 C-terminus suggested that the LIN-2/7/10 complex could potentially modulate the DAF-2 signalling pathway. We thus tested the extent to which perturbations in DAF-2 signalling might affect the lin-7-mediated infection phenotype. Upon B. thailandensis infection, daf-2 (e1370) mutants died at highly similar rates irrespective of lin-7 status (P = 0.1738, Fig. 4A). This is consistent with the idea that lin-7 and daf-2 are acting in the same pathway and that lin-7 may be acting upstream of daf-2 as no additional survival advantage is imparted by lin-7 dsRNA when daf-2 is mutated.

If the lin-7-mediated infection phenotype involves DAF-2 signalling, then the enhanced survival exhibited by lin-7 mutants should also be abolished by constitutively activating AKT-1, the PI3K kinase downstream of DAF-2 (Paradis and Ruvkun, 1998). Indeed, akt-1 (mg144) mutants carrying gain-of-function mutations in the PI3K kinase suppressed the infection phenotype conferred by lin-7 RNAi treatment (P < 0.0001, Fig. 4B). This provides additional evidence that lin-7 acts in the daf-2 signalling pathway.

DAF-16 plays a pivotal role in phenotypes mediated by the DAF-2 signalling pathway; if LIN-7 acts in the same pathway as DAF-2 during infection, DAF-16 should also affect the survival ability of lin-7 mutants. To assess the contribution of DAF-16, we compared the infection outcomes of wild-type nematodes and lin-7 mutants in the presence or absence of daf-16 dsRNA. Previous studies have shown that in wild-type nematodes, inhibiting daf-16 alone is not sufficient to modulate survival during bacterial infection (Garsin et al., 2003; Evans et al., 2008). Our experimental results were consistent with these previous findings as we also found that daf-16 dsRNA did not alter the survival of infected wild-type nematodes (P = 0.8795, Fig. 4C). In contrast, daf-16 RNAi treatment significantly suppressed the enhanced survival exhibited by lin-7 mutants (P < 0.0001, Fig. 4C), similar to that observed with daf-2 mutants (Garsin et al., 2003). Collectively, these results suggest that the infection phenotype mediated by loss of lin-7 activity is likely to be, at least in part, dependent on daf-16 activity, supporting a role for lin-7 in positively regulating daf-2 signalling.

hsf-1 also contributes to the infection phenotype of lin-7 mutants

In addition to DAF-16, HSF-1 is required for the enhanced survival of daf-2 mutants during bacterial infection (Singh and Aballay, 2006). Together, HSF-1 and DAF-16 co-regulate certain subsets of genes, including the small HSPs (Hsu et al., 2003); these small HSPs were also found to be upregulated in daf-2 mutants (Hsu et al., 2003; McElwee et al., 2003). Recent studies have furthermore revealed that, analogous to DAF-2's role of inhibiting DAF-16, the insulin signalling pathway also compromises HSF-1 activity by directly regulating a protein complex which sequesters and suppresses HSF-1, further strengthening a functional interplay between the insulin signalling pathway and the heat shock response (Chiang et al., 2012).

To investigate the possible contribution of HSF-1 towards the enhanced survival exhibited by lin-7 mutants during infection, we compared the infection outcomes of wild-type nematodes and lin-7 mutants in the presence or absence of hsf-1 dsRNA. Consistent with previous findings (Singh and Aballay, 2006), the survival of infected wild-type nematodes was significantly reduced by hsf-1 dsRNA (P < 0.0001, Fig. 4D). Importantly, hsf-1 RNAi treatment also significantly suppressed the enhanced survival of lin-7 mutants (P < 0.0001, Fig. 4D). These results suggest that the infection phenotype exhibited by lin-7 mutants could also be, at least in part, dependent on hsf-1 activity.

lin-7 affects DAF-16 nuclear localization upon heat shock

HSF-1 and one of its downstream HSPs, HSP-1, have been shown to affect the kinetics of DAF-16 nuclear localization in C. elegans; nematodes treated with hsf-1 or hsp-1 dsRNA exhibited delayed DAF-16::GFP nuclear egression upon heat shock (Singh and Aballay, 2009). These prior observations and the finding that loss of hsf-1 activity suppressed the infection phenotype of lin-7 mutants led us to investigate whether lin-7 deficiency would also affect the kinetics of DAF-16 nuclear localization.

To study this, we used nematodes containing a DAF-16::GFP translational fusion protein to compare DAF-16 nuclear localization patterns under three different genetic backgrounds: no RNAi (parental HT115), lin-7 RNAi and hsf-1 RNAi. Consistent with previous reports (Henderson and Johnson, 2001), under non-stressed conditions (Brenner, 1974), DAF-16::GFP in all the three genetic backgrounds remained diffusely present throughout the nucleus and the cytoplasm in all tissues (‘unlocalized’; Fig. 5A–C, first and second panels). In our study, B. thailandensis infection alone was insufficient to effect any change in DAF-16::GFP nuclear localization in these nematodes (data not shown). Thus to assess the influence of lin-7 on DAF-16 nuclear localization, we implemented a heat shock regimen at 35°C for 30 min, previously shown to induce DAF-16::GFP nuclear localization (Henderson and Johnson, 2001). Following the acute heat stress, nematodes treated with parental HT115 exhibited DAF-16::GFP nuclear localization predominantly restricted to the 28–32 nuclei of the intestinal epithelial cells (‘intestinal nuclear’; Fig. 5A, third and fourth panels); while the majority of hsf-1 dsRNA-treated nematodes displayed DAF-16::GFP nuclear localization across all examined cell types, including the intestine and the hypodermis (‘all nuclear’; Fig. 5C, third and fourth panels). The number of hsf-1 dsRNA-treated nematodes exhibiting the ‘all nuclear’ DAF-16::GFP localization pattern was significantly higher than those without RNAi treatment (P = 0.0038, Fig. 5D), consistent with previous findings that hsf-1 dsRNA delayed DAF-16::GFP nuclear export (Singh and Aballay, 2009).

Fig. 5

During acute heat stress, lin-7 modulates the kinetics of DAF-16::GFP nuclear localization.

A–C. DAF-16::GFP mutants were first grown on parental E. coli HT115 or exposed to lin-7 or hsf-1 dsRNA before subjecting to heat shock at 35°C in a water bath. Nematodes were harvested every 15 min and assayed before and immediately after heat shock (+ HS). Representative images after 30 min of heat shock are shown at 125× (first and third panels) or 250× magnification (second and fourth panels). The pharynx (P) and non-intestinal nuclei (white arrowheads) are indicated. Scale bar represents 0.1 mm. Images are representative of at least 50 nematodes per RNAi treatment from three independent assays.

D and E. Nematodes were classified as exhibiting either uniformly distributed DAF-16::GFP (unlocalized), nuclear DAF-16::GFP in intestinal cells only (intestinal nuclear) or nuclear DAF-16::GFP in all cell types (all nuclear). The number of nematodes in each category was counted and shown as a percentage of total nematodes assayed immediately after acute heat stress of 30 min (D) or 45 min (E). Percentages that were significantly different from the parental HT115 treatment for the ‘all nuclear’ phenotype are indicated.

In contrast, lin-7 dsRNA-treated nematodes exhibited slower DAF-16::GFP nuclear localization kinetics when compared with those without RNAi; upon 30 min of heat shock, lin-7 dsRNA-treated nematodes predominantly exhibited ‘unlocalized’ DAF-16::GFP localization patterns (Fig. 5B, third and fourth panels; P = 0.0086, Fig. 5D). When the heat shock treatment was extended for another 15 min, nematodes treated with lin-7 dsRNA predominantly exhibited the ‘intestinal nuclear’ DAF-16::GFP localization pattern when compared with control (P = 0.0289, Fig. 5E). It is worth noting that at this time point, nuclear localization of DAF-16::GFP appeared to be delayed in the hypodermis of lin-7 dsRNA-treated nematodes. Further on, when heat shock was prolonged to 1 h, 100% of lin-7 dsRNA-treated nematodes eventually revealed an ‘all nuclear’ DAF-16::GFP localization pattern, identical to nematodes treated with parental HT115 or hsf-1 dsRNA (data not shown). These data show that lin-7 can significantly affect the kinetics of DAF-16 nuclear localization during heat shock, further corroborating a connection between LIN-7 and the DAF-2 signalling pathway. In addition, these results raised the possibility that perhaps the increased nuclear export of DAF-16::GFP in lin-7 dsRNA-treated nematodes could be attributed to hsf-1, given that reducing hsf-1 activity suppressed this phenomenon. Henceforth, our observations also suggest a functional relationship between LIN-7, DAF-2 and HSF-1.

Tissue-specific RNAi assays indicate hypodermal lin-7 regulates infection outcomes

Although our initial RNAi screen focused on genes expressed in the nematode intestine, lin-7 is known to function in non-intestinal tissues as well (Simske et al., 1996; Kaech et al., 1998). Various tissues in the nematode can also exhibit distinct aspects of pathogen defence (Pujol et al., 2008; Ren et al., 2009; Ziegler et al., 2009). These previous observations, coupled with the finding that lin-7 seems to affect nuclear localization of DAF-16::GFP in the hypodermis upon 45 min of heat stress (Fig. 5E), led us to investigate the roles of intestinal and hypodermal lin-7 in mediating the observed infection phenotype.

To study this, we performed tissue-specific RNAi experiments, specifically silencing lin-7 either in the intestine or in the hypodermis. Tissue-specific RNAi was achieved by feeding lin-7 dsRNA to rde-1 (ne219) mutants carrying the wild-type rde-1 transgene expressed either under the intestine-specific promoter pnhx-2 (Espelt et al., 2005) or under the hypodermis-specific promoter plin-26 (Qadota et al., 2007). rde-1 encodes a member of the Argonaute protein family, whose expression is necessary to initiate RNAi in a cell-autonomous manner (Tabara et al., 1999). Tissue specificity of RNAi in these strains was confirmed by feeding them with unc-22 dsRNA. As unc-22 expression is restricted to the muscles (Moerman et al., 1988), neither strain showed the characteristic unc-22 twitching phenotype as seen in wild-type nematodes (data not shown). As a further control for tissue specificity, we also treated these rde-1 (ne219) mutants with elt-2 dsRNA. Consistent with reports that elt-2 is specifically expressed in the intestine and protects against several bacterial pathogens (Fukushige et al., 1998; Kerry et al., 2006), we found that elt-2 RNAi rendered rde-1 (ne219) mutants carrying the intestinal pnhx-2::rde-1 transgene hypersensitive to B. thailandensis (P = 0.0140, Fig. 6A) but did not have an effect on rde-1 (ne219) mutants carrying the hypodermal plin-26::rde-1 transgene (P = 0.2679, Fig. 6B).

Fig. 6

lin-7 is a general host factor and functions specifically in the hypodermis during infection.

A. rde-1 (ne219) mutants carrying the intestinal pnhx-2::rde-1 transgene were first grown on parental E. coli HT115 or exposed to lin-7 (P = 0.7298) or elt-2 (P = 0.0140; all P values as compared with parental HT115) dsRNA before transferring to plates containing B. thailandensis ATCC 700388.

B. rde-1 (ne219) mutants carrying the hypodermal plin-26::rde-1 transgene were first grown on parental E. coli HT115 or exposed to lin-7 (P < 0.0001), daf-2 (P < 0.0001) or elt-2 (P = 0.2679; all P values as compared with parental HT115) dsRNA before transferring to plates containing B. thailandensis ATCC 700388.

C. Wild-type or lin-7 (e1449) nematodes were first exposed to cdc-25.1 dsRNA before transferring to NGM plates containing S. Typhimurium ATCC 14028 (P < 0.0001).

D. Wild-type or lin-7 (e1449) nematodes were first exposed to cdc-25.1 dsRNA before transferring to peptone-glucose-sorbitol plates containing P. aeruginosa PA14 (P < 0.0001).

Importantly, upon exposure to B. thailandensis, rde-1 (ne219) mutants carrying the hypodermal plin-26::rde-1 transgene exhibited enhanced survival when treated with lin-7 dsRNA (P < 0.0001, Fig. 6B). In contrast, treating rde-1 (ne219) mutants carrying the intestinal pnhx-2::rde-1 transgene with lin-7 dsRNA did not confer protection during infection (P = 0.7298, Fig. 6A), indicating that the lin-7-mediated infection phenotype probably involves mainly hypodermal and not intestinal lin-7. Interestingly, when treated with daf-2 dsRNA, rde-1 (ne219) mutants carrying the hypodermal plin-26::rde-1 transgene also exhibited enhanced survival upon B. thailandensis infection (P < 0.0001, Fig. 6B). This result provides further evidence for lin-7 's role in regulating daf-2 signalling during infection and highlights the spatial specificity of this functional relationship in the hypodermis.

Loss of lin-7 function is protective against multiple bacterial pathogens

It has been previously shown that the DAF-2 insulin signalling pathway, as well as HSF-1, protects the nematode against multiple bacterial pathogens such as Enterococcus faecalis, Pseudomonas aeruginosa (Garsin et al., 2003; Singh and Aballay, 2006), and Salmonella enterica serovar Typhimurium (Singh and Aballay, 2006; Jia et al., 2009). If lin-7 acts in the same pathway as daf-2, then lin-7 mutants should also possess the ability to survive infections by other bacterial pathogens and not just specifically by B. thailandensis. To test this possibility, we exposed wild-type nematodes and lin-7 mutants to two other bacterial pathogens, S. Typhimurium and P. aeruginosa, which are known to kill C. elegans by different mechanisms. S. Typhimurium kills C. elegans by slow intestinal colonization on NGM agar (Aballay et al., 2000; Labrousse et al., 2000), whereas P. aeruginosa produces a fast-killing nematocidal toxin on high osmolarity peptone-glucose-sorbitol agar (Mahajan-Miklos et al., 1999). We found that lin-7 mutants consistently survived longer upon infection with either pathogen (P < 0.0001, Fig. 6C and D). These results establish lin-7 as a general infection factor, consistent with the hypothesis that lin-7 regulates daf-2 signalling in C. elegans.

Discussion

In this study, we identified a novel role for LIN-7 during bacterial infection. Loss of lin-7 function imparted C. elegans the enhanced ability to survive infection by three distinct Gram-negative bacterial pathogens (B. thailandensis, P. aeruginosa and S. Typhimurium). Subsequent genetic and biochemical assays linked LIN-7 to DAF-2, a major regulator of the insulin/IGF-1 signalling pathway. Tissue-specific RNAi experiments revealed that lin-7 likely functions in the hypodermal cells of C. elegans to affect infection outcomes. Collectively, our results suggest that during infection, LIN-7 regulates the DAF-2 signalling pathway in nematode hypodermal tissues and this modulation of DAF-2 signalling becomes detrimental to the infected host.

Infection sensitivity in C. elegans can be modulated by the removal of germline signals (Miyata et al., 2008) and by mutations that increase overall nematode lifespan (Garsin et al., 2003). Our results suggest that the infection phenotype of lin-7 mutants is unlikely to be due to these two confounding factors. Specifically, lin-7 loss-of-function mutants are not known to be germline-deficient and are well characterized for their matricidal phenotype (Horvitz and Sulston, 1980; Ferguson and Horvitz, 1985). In addition, when exposed to B. thailandensis, fertile lin-7 (RNAi) nematodes, fertile lin-7 (e1449) mutants as well as sterile lin-7 (e1449) mutants [via cdc-25.1 RNAi (Evans et al., 2008; Shapira and Tan, 2008)] consistently exhibited enhanced survival over wild-type nematodes, indicating that the lin-7-mediated infection phenotype is not related to germline signalling. Furthermore, the lifespans of lin-7 mutants and wild-type nematodes when grown on non-pathogenic E. coli were indistinguishable, suggesting that the infection phenotype exhibited by lin-7 mutants is not simply a consequence of aberrant organismal development. These results thus support a more direct role for lin-7 in influencing infection outcomes.

LIN-7 is known to associate in a tripartite complex with LIN-2 and LIN-10 to positively regulate the subcellular localization and activity of various signalling proteins (Kaech et al., 1998; Kim and Sheng, 2004; Alewine et al., 2007). Similar to lin-7 mutants, lin-2 and lin-10 mutants also exhibited enhanced survival upon B. thailandensis infection. Our subsequent in silico analysis to identify potential interacting partners of the LIN-2/7/10 complex highlighted the DAF-2 RTK as a potential LIN-2 binding partner. Using the yeast two-hybrid system, we were able to confirm that the LIN-2 PDZ domain can indeed bind to the DAF-2 C-terminus. Along with genetic studies perturbing the insulin signalling pathway, our results suggest that LIN-7 binds, via LIN-2, to the DAF-2 C-terminus and may positively regulate DAF-2 signalling. Still, from our study we cannot exclude the possibility that during infection, LIN-7 can also bind and regulate hitherto unidentified proteins.

Our data suggest that the modulation of DAF-2 signalling by LIN-7 occurs predominantly in the C. elegans hypodermis. Interestingly, preceding research has implicated the hypodermis as an important tissue for influencing infection outcomes. For example, infection by the fungus Drechmeria coniospora activates signalling pathways in the hypodermis that in turn initiates an intestinal immune response (Pujol et al., 2008). In addition, intestinal E. faecalis infection has been shown to activate host NADPH oxidases to generate ROS in both the hypodermis and the intestine as a protective immune measure (Chávez et al., 2009). Thus, even though the host–pathogen interface is primarily localized to the intestine, tissues outside the intestine (such as the hypodermis) clearly can also contribute to the overall survival outcome of an infection.

DAF-16 is well established to positively regulate immune genes including lysozymes, catalases, saposins and superoxide dismutases (Murphy et al., 2003). The finding that inhibition of hsf-1 suppressed the infection phenotype exhibited by lin-7 mutants suggested that in addition to DAF-16, HSF-1 also plays a significant role in the ability of LIN-7 to modulate host infection outcomes. This is consistent with the findings that (i) LIN-7 and DAF-2 act in the same pathway (this study), and (ii) both DAF-16 and HSF-1 suppressed the enhanced survival of daf-2 mutants (Garsin et al., 2003; Singh and Aballay, 2006).

Although the exact interaction between the insulin signalling pathway and HSF-1 during infection is not yet well characterized, recently, this interplay has been dissected in the context of ageing. DAF-2 insulin signalling inhibits DDL-1 phosphorylation, thus allowing DDL-1 and DDL-2 to form a complex sequestering and inhibiting HSF-1 (Chiang et al., 2012). Hence, in a daf-2 mutant, HSF-1 is not sequestered by DDL-1/2. Considering these observations, our data suggest that LIN-7 may positively regulate DAF-2 signalling and hence inhibit HSF-1 activity. This view is supported by the findings that (i) the infection phenotype of lin-7 mutants was suppressed when treated with hsf-1 dsRNA, and (ii) loss of lin-7 resulted in a significant delay in DAF-16::GFP nuclear localization.

HSF-1 and its downstream HSPs have been shown to protect the infected host by reducing protein aggregation detrimental to host cellular tissues. This undesired protein aggregation has been proposed to be caused by reactive oxygen species generated in the nematode in response to infection (Mohri-Shiomi and Garsin, 2008). The inhibitory function of DAF-2 on HSF-1 activity may thus provide a possible explanation for the reduction in protein aggregation observed in daf-2 mutants (Chávez et al., 2009). Similarly, in lin-7 mutants, this protective pathway could also have been enhanced to counter infection-induced protein aggregation. This further hints at the possibility that lin-7 mutants, instead of being resistant (which we have shown earlier on not to be the case), could be tolerant of infection, restricting the damage inflicted on host tissues (Schneider and Ayres, 2008).

Another possible explanation for the infection phenotype exhibited by lin-7 mutants revolves around the fact that throughout evolution from nematodes to mammals, it is absolutely crucial to have a tight surveillance over host defence pathways as both insufficient and excessive activity can prove detrimental. Although DAF-16 is known to positively regulate immune-related genes (Murphy et al., 2003), excessive DAF-16 transcriptional activity can also result in enhanced susceptibility to bacterial infections (Singh and Aballay, 2009). In that study, it was shown that nematodes with additional gene copies of daf-16, coupled with either acute heat stress or daf-2 loss-of-function mutations, actually exhibited an increased rather than reduced susceptibility to bacterial infections, indicating that while DAF-16 is essentially protective during infection, excessive DAF-16 nuclear activity could also be detrimental rather than beneficial.

At present it remains unclear exactly why excessive DAF-16 nuclear accumulation is detrimental. It is possible that DAF-16 hyperactivation upregulates AQP-1, a membrane water channel protein, and this results in dysregulated water homeostasis, host cellular damage and reduced immune responses (Singh and Aballay, 2009). It is thus not far-fetched to speculate why DAF-2 signalling also has an inhibitory effect on HSF-1. When daf-2 mutants are exposed to bacterial pathogens, excessive DAF-16 nuclear accumulation may occur due to lack of DAF-16 phosphorylation by upstream PI3K kinases. However, along with an enhanced HSF-1 activity which promotes DAF-16 nuclear export (Singh and Aballay, 2009), daf-2 mutants would be able to maintain a homeostatic check on DAF-16 nuclear activity. In the same way, lin-7 mutants may have the ability to maintain advantageous levels of DAF-16 in the hypodermal tissues and this, in part, confers protection to the whole nematode.

In summary, our findings revealed a role for the C. elegans cell junction protein, LIN-7, in the modulation of hypodermal DAF-2 signalling and hence host infection outcomes. Interestingly, several other C. elegans genes functioning in early developmental or physiological processes also appear to have been reused in the adult nematode for pathogen defence (Lin et al., 1997; Honda and Honda, 1999; Irazoqui et al., 2010). In vertebrates, LIN-7 homologues are expressed ubiquitously in various tissues, including the junctional complex regions of kidney cells (Irie et al., 1999), synaptic junctions in neurones (Butz et al., 1998) and basolateral membranes of epithelial cells (Yan et al., 2009). The LIN-2/7/10 complex is also conserved from C. elegans to mammals: mammalian LIN-7 interacts with the LIN-2 homologue CASK (Cohen et al., 1998) and the LIN-10 homologue X11/Mint (Rongo et al., 1998) to establish cell junctions in epithelial cells and synapses. Given that some pathogen defence pathways are evolutionarily conserved from invertebrates to mammals, it is intriguing to speculate that perhaps this is one of the earlier functions of LIN-7 and that during evolution, LIN-7 homologues have acquired additional roles to cope with more complex systems and processes in higher organisms.

Experimental procedures

Nematode, bacteria and yeast strains

Nematode, bacteria and yeast strains used in this study are listed in Table S1. All nematode strains were cultured and maintained at 20°C on modified nematode growth media (NGM, 0.35% peptone) agar and fed with E. coli strain OP50, as described (Brenner, 1974). Except for daf-2 (e1370) mutants, this strain was cultured and maintained at 15°C to suppress dauer formation. Bacteria strains were grown in Luria–Bertani (LB) broth at 37°C. Yeast strains were grown in yeast extract peptone dextrose (YPD) broth at 30°C.

Survival assays

Burkholderia thailandensis ATCC 700388 and E555, P. aeruginosa PA14 and S. Typhimurium ATCC 14028 were grown overnight in LB broth at 37°C. B. thailandensis and S. Typhimurium lawns were prepared by spreading 100 μl of overnight culture on modified NGM agar and grown for 24 h at 37°C. P. aeruginosa lawns were prepared similarly by spreading on peptone-glucose-sorbitol agar (Mahajan-Miklos et al., 1999). Unless specified otherwise, 40 L4-staged nematodes were added to each lawn and infected as per described (Powell and Ausubel, 2008). Nematodes were set down on bare agar before transferring to pathogen-containing lawns to minimize the transfer of E. coli. No visible E. coli growth on pathogen-containing lawns was observed at locations where nematodes were added nor was there any crowding of nematodes at such locations. To further test for E. coli contamination, nematodes were removed 24 h post infection; pathogen-containing lawns were harvested, diluted appropriately in M9 buffer (Brenner, 1974) and tested for E. coli and B. thailandensis by plating on neat LB agar and LB agar supplemented with gentamicin 25 μg ml−1 (Thermo Fisher Scientific Waltham, MA). No E. coli contamination on pathogen-containing lawns was observed (Fig. S4). Nematode survival was scored at 24°C and nematodes were considered dead upon failure to respond to gentle touch by a platinum wire. Results are representative of three independent experiments.

Immunofluorescence assays

Nematodes infected by B. thailandensis E555 were prepared for immunohistochemical staining using a freeze-crack method (Duerr et al., 1999) and fixed using 50% methanol (2 min) and 50% acetone (4 min). After washing, slides were blocked for 1 h in 5% bovine serum albumin (BSA) in antibody buffer [0.5% Triton X-100, 1 mM EDTA, 0.1% BSA and 0.05% sodium azide (NaN3) in phosphate-buffered saline (PBS)], followed by 1 h primary antibody incubations using monoclonal antibody 3015 IgG1 (Sim et al., 2010). Secondary antibody incubations were performed using donkey anti-mouse Texas Red (1:500) (Jackson ImmunoResearch, West Grove, PA) for 4 h. All incubations were performed at 24°C. Slides were mounted in anti-photobleaching media with DAPI (Vector Laboratories, Burlingame, CA) and visualized on a LEICA DMRE microscope. Images were analysed by GNU Image Manipulation Program (version 2.6.3). For GFP experiments, nematodes fed on E. coli OP50::GFP lawns were harvested and mounted for microscopy in PBS with 25 mM NaN3. Nematodes were visualized on a Carl Zeiss Axiovert 200m inverted microscope and images were analysed by Metamorph software (version 6.3r7). Images are representative of at least 50 nematodes from three independent assays.

RNAi assays

Unless specified otherwise, RNAi assays were carried out at 20°C by feeding nematodes with parental E. coli HT115 (DE3) strain or E. coli HT115 clones expressing gene-specific dsRNA (Timmons and Fire, 1998). Each clone identity was verified by direct sequencing using specific oligonucleotides targeting the L4440 vector: pL4440-F (gTTTTCCCAgTCACgACgTT) and pL4440-R (TggATAACCgTATTACCgCC) (Rual et al., 2004). RNAi assays were performed by growing each clone for 8 h in LB broth supplemented with ampicillin 50 μg ml−1 (Sigma-Aldrich, St. Louis, MO) and seeding on isopropyl β-d-1-thiogalactopyranoside (IPTG)-containing modified NGM agar. Nematode embryos, generated by hypochlorite treatment, were propagated on these seeded plates until the L4 stage. Nematodes were subsequently transferred to pathogen-containing lawns.

For experiments involving sterile nematodes, embryos were exposed to cdc-25.1 dsRNA at late embryogenesis till day 1 adult stage, before transferring them to pathogen-containing lawns. Under such conditions, cdc-25.1 RNAi resulted in nematodes with an Emb phenotype (Evans et al., 2008; Shapira and Tan, 2008).

For experiments involving hsf-1 RNAi, nematode embryos were exposed to hsf-1 dsRNA at 15°C until the L4 stage, before transferring them to pathogen-containing lawns. Under such conditions, hsf-1 RNAi did not result in larva developmental arrest observed when the assay was carried out at higher temperatures (Walker et al., 2003; Singh and Aballay, 2006).

Nematode bacterial load analysis

Nematodes were infected with B. thailandensis ATCC 700388 as per described. At 24 h post infection, infected nematodes were harvested and set down on bare agar before transferring to M9 buffer to minimize the contamination of bacteria. Nematodes were washed thrice with M9 buffer, followed by 1 h incubation in M9 buffer containing trypsin-EDTA (Life Technologies, Carlsbad, CA) to remove bacteria present on the exterior of the nematode. Nematodes were then washed thrice with M9 buffer only to remove trypsin-EDTA, and subsequently lysed by vortexing with 400 g silicon-carbide sharp particles (Biospec, Bartlesville, OK) and 0.2% sodium dodecyl sulfate. Lysates were diluted appropriately in M9 buffer and plated on LB agar supplemented with gentamicin to select for B. thailandensis. After 2-day incubation at 37°C, amount of live bacteria per nematode was determined by cfu counts. At least 25 nematodes were harvested per nematode strain and experiments were performed in triplicates.

Yeast two-hybrid assays

Plasmids and primers used in this assay are listed in Table S3. The PDZ domain of LIN-2 (amino acids 288–647) and the C-terminus of DAF-2 (amino acids 1445–1843) were PCR-cloned using template cDNA from a mixed population of wild-type N2 nematodes. PCR fragments were cloned, at EcoRI and BamHI restriction sites for daf-2, or at NdeI and XmaI sites for lin-2, in-frame into the GAL4 DNA-binding domain (DNA-BD) vector pGBKT7 or GAL4 activation domain (AD) vector pGADT7 (Clontech Laboratories, Mountain View, CA) respectively. Reagents provided by the manufacturer included: (i) positive controls – murine p53 and SV40 large T antigen, as fusion proteins with the same GAL4 DNA-BD and GAL4 AD, respectively, and (ii) negative control – human lamin C, fused to the GAL4 DNA-BD.

Y2HGold reporter and Y187 mating haploid yeast strains were transformed individually with either the GAL4 DNA-BD or AD construct, according to the manufacturer's instructions. Successfully transformed haploids were selected on plates lacking tryptophan or leucine accordingly. Six pairwise combinations of transformed Y2HGold and Y187 haploid strains (see Fig. 3C) were allowed to mate overnight in YPD broth and diploids were selected on synthetically defined media lacking leucine, tryptophan, adenine and histidine. Growth within 3 days of incubation at 30°C indicated positive protein–protein interactions within mated pairs. Clone identities were verified by extracting plasmids from yeast diploids harbouring positive LIN-2/DAF-2 interactions and rescuing them in E. coli TG1 strain, followed by direct sequencing of plasmids isolated from TG1 using the following pairs of oligonucleotides: daf2-check-F (TgACgATTCAgAAgCACTgg) and daf2-check-R (CATCTTgTCCACCACgTgTC); lin2-check-F (AgTggCAggTTTgACgAgAC) and lin2-check-R (AgTTgTCggAgT TCCAATgC).

DAF-16 nuclear localization assays

Embryos from DAF-16::GFP nematodes, generated by hypochlorite treatment, were propagated on plates containing parental E. coli HT115 strain or E. coli HT115 clones expressing lin-7 or hsf-1 dsRNA. Nematodes were grown at 20°C for parental HT115 and lin-7 dsRNA, or at 15°C for hsf-1 dsRNA until the L4 stage. Acute heat shock was performed at 35°C for intervals of 15 min by placing sealed plates into a water bath. Nematodes before and immediately after heat shock were visualized using an Olympus MVX10 dissecting microscope under 125× or 250× total magnification and images were analysed by DP controller software (version 3.1.1.267). Nematodes were classified as exhibiting either uniformly distributed DAF-16::GFP (unlocalized), nuclear DAF-16::GFP restricted to the 28–32 nuclei of the intestinal epithelial cells only (intestinal nuclear) or nuclear DAF-16::GFP in all cell types (all nuclear). At least 50 nematodes were counted per RNAi treatment and experiments were performed in triplicates.

Statistical analysis

Survival curves were analysed using the PRISM (version 5.0) software. Kaplan-Meier survival curves with P values < 0.05 were considered significantly different from the control. Student's t-test was used to analyse nematode bacteria loads and DAF-16::GFP nuclear localization results.