The nematode strains used in this study are listed in Table S1 . The E. coli O157:H7 clinical isolates strain EDL933 were from the Bioresource Collection and Research Center (BCRC), Taiwan. Caenorhabditis elegans strains TU3401, VP303, WM118, and SAL144 used in this work were kindly provided by the Caenorhabditis Genetics Center (CGC), which is funded by the NIH National Center for Research Resources (NCRR). iglr-2 deletion strains were generated by the CRISPR-Cas9 system as described (20, 21). iglr-2 overexpressed strains, transcriptional and translational reporter strains were created by microinjection as described (22, 23). The genotype of the above C. elegans strains were confirmed by single worm PCR, restriction fragment length polymorphism, and sequencing analysis. C. elegans strains were maintained on nematode growth medium (NGM) agar plates using the standard laboratory E. coli strain OP50 as described (24).

Several categories of genes are reported to contain specific motifs or domains considered as PRRs or potential PRRs (2) and hence were selected into our RNAi library e.g., C-type lectin-like domains (CTLDs), galectin, lipid binding protein (LBP), leucine-rich repeat domain (LRR), lysin motif domain (LysM), and peptidoglycan recognition protein (PGRP). We utilized WormMart on WormBase to identify the genes that contained domains of LRR, LBP, LysM, galectin, or PGRP. WormMart was terminated in 2016; however, a similar function can be found on BioMart. The study by Schulenburg et al. was used as a reference for genes containing C-type lectin-domains (25). RNAi clones are collected from the Ahringer RNAi or constructed from the Worm ORFeome library via Gateway cloning system. The constructed PRR library was frozen in a 96-well format with 8% glycerol (Sigma-Aldrich) at −80°C before use. The detailed screening method is as described and illustrated in the Supplemental Information ( Figure S1 ). The RNAi clones and the results of screening are listed in Table S2 .

All survival assays were performed as previously described (19, 26, 27). Briefly, E. coli O157:H7 EDL933 bacteria were cultured in Luria–Bertani (LB) broth for 16 to 18 h at 37°C. A 30 µL aliquot of bacterial culture (O.D.600 nm = 2) was plated on 5.0 cm nematode growth media (NGM) agar plates and incubated overnight at room temperature (25°C). Approximately fifty synchronized late larval stage 4 (L4) to young adult animals were transferred to each plate and maintained at 20°C. Worms were scored daily and considered as dead when animals did not respond to touch by platinum wire. Animals that crawled off the plate were scored as censored. Worms were transferred to fresh plates every day throughout the reproductive period. The assay was conducted independently at least three times with approximately 50 to 100 worms each time at 20°C. Data represent the sum of animals in multiple experiments in the figures. The number of independent survival assays and sample sizes for each of these experiments is given in Table S3 .

The assay was conducted as previously described with slight modifications (28, 29). Bacterial (E. coli OP50, E. coli O157:H7 EDL933 or P. aeruginosa PA14) lawns were prepared by inoculating individual bacterial colonies into 3 mL of Luria–Bertani (LB) broth and grown for 16 to 18 h in a 37°C incubator. A 30 µL aliquot of bacterial culture was plated onto the center of a 10 cm NGM plate and kept at room temperature (25°C) for 24 h. Approximately thirty synchronized late larval stage 4 (L4) animal worms were transferred onto plates’ center containing the bacterial lawn and scored after 16 h incubation at 20°C. Each animal was scored as inside or outside the lawn (Occupancy Index = N in/N total). Experiments were performed for at least three times independently.

For CRISPR disruption of iglr-2, we used the technique previously described (20, 21) using the plasmids, pDD162 (21) (Peft-3::Cas9 + Empty sgRNA) (plasmid number 47549 on Addgene) and PU6::unc-119_sgRNA (21) (plasmid number 46169 on Addgene). Briefly, sgRNA targeting the 6^th and 10^th exons of iglr-2 was created by using overlap extension PCR and TA cloning system. We used PU6::unc-119_sgRNA as a template, U6prom EcoRI F (5′-CGG GAA TTC CTC CAA GAA CTC GTA CAA AAA TGC TCT-3′), iglr-2 (a + b) gRNA R (5′-TTA GAT ATC GAG CAG AGA ACA AAC ATT TAG ATT TGC AAT TCA ATT ATA TAG), iglr-2 (a + b) gRNA F (5′-GTT CTC TGC TCG ATA TCT AAG TTT TAG AGC TAG AAA TAG CAA GTT A-3′) and U6prom HindIII R (5′-CGG AAG CTT CAC AGC CGA CTA TGT TTG GCG T-3′) to construct sgRNA targeting the 6^th exon. As above, we used U6prom EcoRI F (5′-CGG GAA TTC CTC CAA GAA CTC GTA CAA AAA TGC TCT-3′), iglr-2 cytosol sgRNA R (5′-TTG AAT CCA CCA AAG AGC ACA AAC ATT TAG ATT TGC AAT TCAA TTA TAT AG-3′), iglr-2 cytosol sgRNA F (5′-GTG CTC TTT GGT GGA TTC AAG TTT TAG AGC TAG AAA TAG CAA GTT A-3′) and U6prom HindIII R (5′-CGG AAG CTT CAC AGC CGA CTA TGT TTG GCG T-3′) to construct sgRNA targeting the 10^th exon. N2 animals were injected with about 150 ng/μL of total DNA, and the ratio of pDD162 to sgRNA to marker pBCN40-R4R3 (plasmid number 34915 on Addgene) was 10:9:1, respectively. Transgenic F1 progeny was screened for mCherry expression and plated individually. After F2 progeny production, F1 animals were screened by single worm PCR and restriction enzyme digestion to identify iglr-2 homozygote mutants.

To analyze transcription of iglr-2, we PCR amplified the 218 bp region between the start codon of iglr-2 and the 3′ end of the adjacent upstream gene from genomic DNA of wild-type animals. The primers 5′-GGGGACAACTTTGTATAGAAAAGTTGAATTGTTAGATTTTATTTAGATATTATCCTT-3′ and 5′-GGGGACTGCTTTTTTGTACAAACTTGTTCTTCTTTTCTTTGTATCAAGACA-3′ were used for amplifying the region. This promoter region was constructed to plasmid pCG144 (plasmid number 21376 on Addgene) and pCM5.37 (plasmid number 17253 on Addgene) to destination plasmid pCG150 (plasmid number 17247 on Addgene) generating pWF274 [iglr-2p::mCherry::unc-54 3′UTR] through the Gateway cloning system. Plasmid, pWF237 [iglr-2p::mCherry::histone H2B::unc-54 3′UTR], was constructed using the same method except that mCherry::histone H2B was from pCM1.151 (plasmid number 21386 on Addgene). To analyze the expression of IGLR-2 protein, we PCR-amplified full-length, wild-type iglr-2 coding sequence isoform a (2,322 bp) and isoform b (1,875 bp) from cDNA, respectively and ligated it into the destination plasmid pCG150 via a Gateway cloning method to generate pWF294 [iglr-2p::iglr-2a::mCherry::unc-54 3′UTR] and pWF304 [iglr-2p::iglr-2b::mCherry::unc-54 3′UTR]. The primers (5′-GGGGACAAGTTTGTACAAAAAAGCAGGCTATGCGAAAATTTGTATTTTTCGTCGTAG-3′) (5′-ACCCTTTGAGACAGCAGCAGCAGCAGCAGCTCTCTTTTCTGGTGGAGAATCTG-3′) and the primers (5′-GGGGACAAGTTTGTACAAAAAAGCAGGCTATGCGAAAATTTGTATTTTTCGTCGTAG-3′) (5′-ACCCTTTGAGACAGCAGCAGCAGCAGCAGCTTCCTGAAGATGGTAAGTAAGTG-3′) were used for amplifying isoform a and isoform b, respectively.

Assays were carried out as previously described (19). Synchronized populations of wild-type N2 and iglr-2 overexpressed strain animals were grown to the late L4 larval stage and fed with E. coli OP50 or EDL933 for 12 and 24 h at 20°C. Total RNA was then extracted using TRI Reagent (Invitrogen) and reverse transcribed with M-MLV reverse transcriptase (Promega) using random hexamer primers. This cDNA was then subjected to qRT-PCR analysis using SYBR green detection on a 7,500 fast real-time PCR system (Applied Biosystems). Relative expression between samples was normalized to the nhr-23 as the reference gene. The primers (5′-GCCGAAGATGATGCCGAGAT-3′; 5′-GTCGCA GTGTCAAGAATCCC-3′) were used for nhr-23 gene detection, and the primers (5′-TTCTCC GAGATGGTGATTTGG-3′; 5′-TTCTCCGAGATGGTGATTTGG-3′) were used for iglr-2 gene detection. All experiments were performed a minimum of three times independently.

RNAi experiments were performed as described with some modifications (30). In brief, RNAi E. coli strains HT115 transformed with RNAi plasmids were spread on NG-IC plates which are NGM plates with 50 µg/mL carbenicillin and 1 mM isopropyl-β-D-1-thiogalactopyranoside (IPTG). Plates with RNAi bacteria were incubated at 25°C overnight to induce the double-stranded RNA (dsRNA) expression. Synchronized wild-type N2, RNAi-sensitized strain rrf-3(pk1426), and tissue specific RNAi knockdown strains L1 stage larvae were obtained using standard alkaline hypochlorite protocols (24). L1 stage larvae cultured on L4440 or plates expressing RNAi of interest at 20°C until F1 progenies grown to L4 stage. Survival assays were conducted by mixing control OP50-GFP or EDL933-GFP (19) (which were carrying an ampicillin resistant plasmid) with RNAi bacteria of 1:1 ratio. Mixed bacteria were spread on NGIC plates (30 µL per plate). Approximately fifty worms were picked to the prepared OP50/EHEC-RNAi plates and scored their survival as mentioned above. E. coli HT115 with L4440, an empty vector, was used as a negative control of RNAi. All experiments were performed a minimum of three times independently. Data represent the sum of animals in multiple experiments in the figures.

Late L4 to young adult stage of iglr-2 reporter strains were plated on NGM plates with E. coli OP50 at 20°C for 12 h. For K08D8.5 GFP expression analysis, late L4 to young adult stage of C. elegans strains with K08D8.5::GFP transgene in wild-type and iglr-2(wf275) background were infected with EHEC at 25°C for 8 h, respectively. Animals were then mounted on glass slides with 2% agarose pads and anesthetized with 25 nM sodium azide (NaN₃). The fluorescent signal in the nematode was observed by differential interference contrast (DIC) imaging with Nomarski optics and epifluorescence imaging with corresponding filters using a Nikon Eclipse Ti inverted microscope system with DP72 CCD camera. The quantification of K08D8.5::GFP signal was measured by imageJ.

To generate nsy-1(ums8);iglr-2(wf295) double mutant, wild-type N2 male animals were crossed to iglr-2(wf295) hermaphroditic gravid adults. Then, male animals of crossed iglr-2 heterozygotes were picked to mate nsy-1(ums8) hermaphrodites. Their progenies of F2 generation were confirmed for iglr-2(wf295), nsy-1(ums8) or nsy-1(ums8);iglr-2(wf295) by single worm PCR and restriction fragment length polymorphism. The animal exhibited wild-type genotype of iglr-2(wf295), and nsy-1(ums8) was selected as a control. iglr-2(wf295) was confirmed via using the primers (5′-TCTGAATGGAACTGTAATCC-3′; 5′-ATTATCGCGAAGAAT TATGA-3′), and the amplicon was then digested by EcoRV enzyme. nsy-1(ums8) was confirmed via using the primers (5′-CGATTTTGACAAATTGGACT-3′; 5′-AGACAAACAACG TCTGGACT-3′), and the amplicon was then digested by HpyCH4V enzyme. nsy-1(ums8) gain-of-function allele was a kind gift from Dr. Pukkila-Worley (31). To generate iglr-2(wf275);K08D8.5::GFP, male animals of SAL144 (32) were mated to iglr-2(wf275) hermaphroditic gravid adults. Their progenies of F2 generation showed GFP expression were isolated and confirmed for iglr-2(wf275) by the method as described above. The animal exhibited wild-type genotype of iglr-2(wf275) was selected as a control. The worm strain, SAL144 pha-1(e2123);denEx22 [K08D8.5::GFP + pha-1(+)] was a kind gift from Dr. Schwartz (32).

The assay was performed as described (33) with slight modifications. Briefly, the RPW-24 agar plates were prepared by adding 5 µM of RPW-24 (MCE, catalog number: HY-W035409) compounds to 3.5 cm plates containing of 4 mL NGM agar. 10 µL overnight cultured EHEC bacteria were seeded to these RPW-24 agar plates and incubated for 24 h at 25°C. Then, thirty to thirty-five synchronized L4 to young adult staged iglr-2 mutants were placed to RPW-24 agar plates and maintained at 20°C. NGM agar plates supplemented with 0.5% DMSO were used as negative controls. Animals were scored daily and considered as dead when nematodes did not respond to touch by platinum wire. Worms that crawled off the plate were scored as censored and were transferred to plates every day throughout the reproductive period. The assay was performed independently for four times.

To identify potential pattern recognition receptors (PRRs) in C. elegans, we constructed and screened a PRR focused RNAi library. These RNAi clones target 315 genes that encode various PRRs specific motifs or domains ( Figure 1A and Table S2 ). We identified that knockdown of seven genes, including lron-9 (P < 0.05), iglr-2 (P < 0.001), mog-2 (P < 0.05), sds-22 (P < 0.05), soc-2 (P < 0.05), pxn-2 (P < 0.05), zmp-2 (P < 0.001), increased susceptibility of animals to EHEC infection when compared to the empty vector, L4440 by one-way ANOVA test. Among them, only iglr-2 RNAi exhibited a significantly decreased survival of EHEC to C. elegans and did not cause developmental defect to compromise our analysis in the adult animals; we therefore turned our focus to analyze the role of iglr-2 during EHEC infection in C. elegans. IGLR-2 is a transmembrane protein with an extracellular leucine-rich repeat domain at the N-terminus, an immunoglobulin-like domain in the center, and a transmembrane domain and an intracellular domain at the C-terminus ( Figure 2A ). To reconfirm our result from RNAi liquid-toxicity assay, we performed EHEC plate-toxicity assay with RNAi of iglr-2. Indeed, RNAi-mediated knockdown of iglr-2 that targets the leucine-rich repeat (LRR) and immunoglobulin (IG)-like domains in wild-type N2 strain exhibited hypersusceptibility to EHEC ( Figure 1B , P < 0.001 by Log-rank test).

To strengthen our hypothesis that iglr-2 is required for C. elegans to defend against EHEC O157:H7 strain EDL933 infection, we generated iglr-2 deletion mutants by the clustered regularly interspaced short palindromic repeats and CRISPR-associated protein 9 (CRISPR-Cas9) genome-editing technology (20, 21). We designed single guide RNA (sgRNA) targeting at the LRR domain of iglr-2 and generated two independent mutants ( Figure 2A ). iglr-2(wf275) mutant had four nucleotides deleted causing a tyrosine to change to a stop codon in codon 284 of the LRR domain by prediction of amino acid sequences. iglr-2(wf295) mutant not only had four nucleotides deleted; it also had one nucleotide changed (Adenosine to Guanosine) and a leading tyrosine changed to a stop codon in codon 284 of the LRR domain as well ( Figure 2A ). Furthermore, we also designed sgRNA targeting at the transmembrane domain of iglr-2 and generated two independent mutants iglr-2(wf297) had one nucleotide alteration (Thymidine to Adenosine) and seven nucleotide insertion causing a start codon (Methionine) to change to a stop codon in codon 543 of the transmembrane domain by prediction of amino acid sequences. iglr-2(wf298) had one nucleotide deleted also causing a start codon (Methionine) to change to a stop codon in codon 605 of the transmembrane domain ( Figure 2B ). Svensk et al. revealed that iglr-2 mutant exhibited a withered tail tip defect when grown at 15°C for 144 h (34). Consistent with their findings, we also observed that four of our iglr-2 alleles shared the same withered tail tip defect phenotype at 15°C ( Figure S2 ), implying the function of IGLR-2 is disrupted in these mutants. We next examined the susceptibility of these iglr-2 null mutants to EHEC. iglr-2 mutants disrupted either in LRR domain or transmembrane domain were more susceptible to EHEC compared to the wild-type N2 ( Figures 2C, D , All P < 0.001 by Log-rank test). Moreover, the colonization of EHEC in iglr-2(wf275) was significantly increased ( Figure S3A ), but the defecation cycle and pumping rate were compatible to N2 ( Figures S3B, C ), which implied that the increased bacterial accumulation was not contributed by rapid pumping and/or slow defecation in iglr-2(wf275) animals. Taken together, our data indicated that iglr-2 loss-of-function in C. elegans confers animals hypersusceptibility to EHEC infection.

To examine the relationship between iglr-2 expression and EHEC infection, we detected the change in the expression level of iglr-2 after EHEC infection by qRT-PCR analysis. The data demonstrated that iglr-2 transcript expression level was significantly upregulated after wild-type N2 animals fed on EHEC for either 12 h post infection (h p.i.) (P < 0.01 by unpaired t-test) or 24 h p.i. (P < 0.001 by unpaired t-test) compared to that of the control OP50 ( Figure 3A ). These results indicated that iglr-2 is induced upon EHEC infection. We were then interested in the effect of iglr-2 overexpression in C. elegans. We constructed the iglr-2 expression plasmid with whole iglr-2 sequence and injected the plasmid into wild-type N2 to generate two independent iglr-2 overexpressed strains iglr-2 o/e-1 (strain YQ322) and iglr-2 o/e-2 (strain YQ323). First, we examined whether the expression level of iglr-2 was elevated in these overexpressed strains by qRT-PCR analysis. Expression level of iglr-2 transcripts in iglr-2 o/e-1 and iglr-2 o/e-2 was significantly enhanced regardless of feeding on OP50 to young adult stage (both of P < 0.001 by unpaired t-test) or gravid adult stage (both of P < 0.001 by unpaired t-test) ( Figure 3B ). Next, we fed N2, iglr-2 o/e-1 and iglr-2 o/e-2 with EHEC to examine their sensitivity to EHEC. The results showed that iglr-2 overexpressed strains were significantly more resistant to EHEC compared to that of N2 ( Figure 3C , both P < 0.001 by Log-rank test). Taken together, our data suggested that iglr-2 plays an important role in protecting C. elegans against EHEC infection.

In order to determine the expression pattern and identify the site of action of iglr-2, we constructed a plasmid that specifically expresses red fluorescence under the control of the iglr-2 5′ promoter and injected the plasmid into wild-type animals to generate two independent transcriptional reporter animals iglr-2p::mCherry (strain YQ305) and iglr-2p::mCherry::histone H2B (strain YQ326). We fed these two transcriptional reporter strains with E. coli OP50 and examined the expression pattern of red fluorescence at their young adult stage. iglr-2p::mCherry transgenic animals displayed red fluorescence expression in the neurons of the head and anterior/posterior part of the intestinal cells, which suggested that iglr-2 is expressed in these tissues ( Figures 4A–I ). iglr-2p::mCherry::histone H2B transgenic animals also displayed red fluorescence expression pattern resembling iglr-2p::mCherry animals except for the red fluorescent signal only being detected in the nuclei of these tissues ( Figures 4J–R ). Furthermore, iglr-2p::iglr-2a::mCherry (strain YQ338) and iglr-2p::iglr-2b::mCherry (strain YQ362) animals showed two IGLR-2 protein isoforms, which were also expressed in the neurons of the head and anterior/posterior part of the intestine ( Figure S4 ). To reconfirm the site of action of iglr-2, we used tissue-specific RNA knockdown strains to silence iglr-2 in specific tissues and then examined their survival on EHEC infection. Figure 5A shows the survival curves of iglr-2 silencing in RNAi-sensitized strain rrf-3(pk1426) resulting in C. elegans exhibiting susceptible phenotype to EHEC compared to that of empty vector control, L4440 (P < 0.001 by Log-rank test). Similar results are shown in Figures 5B, C ; when knockdown of iglr-2 was specific to neurons or the intestine in worms, they were significantly more susceptible to EHEC (P < 0.001 by Log-rank test). However, specific knockdown of iglr-2 in the muscle did not alter the susceptibility of animals to EHEC compared to the empty vector control ( Figure 5D , P = 0.3725 by Log-rank test). These data are consistent with the images ( Figure 4 ) that show that iglr-2 is only expressed in the neurons and the intestines. Taken together, these data suggested that iglr-2 plays a role in the neural system and intestine to defend against EHEC.

We noticed that part of the iglr-2 signal appears in the neurons ( Figure 4 and Figure S4 ). In the C. elegans nervous system, the olfactory nerve has been reported to play an important role in detecting pathogens and to mediate host–pathogen interactions through chemosensation (29). C. elegans is attracted by most bacteria, but some pathogens like P. aeruginosa stimulate C. elegans to avoid them (35). We demonstrated that the iglr-2 expression was located with neurons in the head and tail ( Figure 4 and Figure S4 ). Therefore, we hypothesize that iglr-2 may be involved in avoidance behavior to EHEC. We transferred N2, iglr-2 mutant and iglr-2 overexpressed worms to the plates containing bacterial lawn and scored the occupancy of animals on the bacterial lawn after keeping them at 20°C for 16 h. The occupancy index was calculated as the number of animals on the bacterial lawn divided by the total number of animals on plates. Indeed, loss of iglr-2 resulted in more animals lingering on OP50 (P < 0.001 by unpaired t-test) or EHEC (P < 0.001 by unpaired t-test) bacterial lawn compared to that of wild-type N2 animals ( Figures 6A, B ). Overexpression of iglr-2 reversed the avoidance behavior, which was similar to wild-type N2 on either the OP50 or EHEC bacterial lawn ( Figures 6A, B ). Of note, iglr-2 deficiency animals exhibited similar avoidance behavior to P. aeruginosa PA14 compared to that of wild-type N2 ( Figure 6C , P = 0.14 by unpaired t-test), which suggested that iglr-2 showed selectivity to E. coli and was not involved in the avoidance behavior to P. aeruginosa in C. elegans. Taking these results together, we demonstrated that iglr-2 is required for behavioral avoidance to pathogenic E. coli, EHEC, and the relatively weak pathogen, E. coli OP50 (36) in C. elegans. P. aeruginosa PA14 is considered an acute pathogen to C. elegans which kills animals within one day (37). We noticed that animals on PA14 bacterial lawn showed a low percentage (~0%) of occupancy for exposing 16 h. To test whether the fast killing of PA14 may hinder the potential increased avoidance of animals, we reduced the assay time of avoidance behavior experiment. We first scored the occupancy index of N2 animals on PA14 in a time-dependent manner. The occupancy of N2 on PA14 was approximately 40% for 8 h exposure and was dropped to 10% for 12 h ( Figure S5 ). We then measured the occupancy of iglr-2 mutant on PA14 for 8-h treatment and found that iglr-2(wf275) exhibited a comparable aversion behavior compared to N2 ( Figure S6 , P = 0.15 by unpaired t-test), suggesting that the occupancy assay might not be affected by fast killing of PA14.

Next, we aimed to identify the downstream signal of iglr-2. TOL-1, MOM-4, and PMK-3 function together in a signaling pathway in the chemosensory BAG neurons (29). Therefore, we were interested in whether the TLR signaling pathway is also involved in iglr-2 protecting hosts from death by EHEC in C. elegans. We performed RNAi-mediated knockdown of mom-4 in iglr-2 overexpressed animals (iglr-2 o/e, YQ322) upon infecting EHEC. Survival of iglr-2 o/e with RNAi silencing of mom-4, a MAP kinase kinase kinase (MAP3Ks), in TLR signaling, was comparable to that of the empty vector control L4440 ( Figure S7 , P = 0.44 by Log-rank test), which suggested that the MAP3K- involved TLR signals are not the downstream signaling pathway of iglr-2 in C. elegans. Several evolutionarily conserved innate immune responses have been reported to be involved in defending pathogenic microbes in C. elegans, for example, the insulin-like signaling pathway (38), the unfolded protein response (UPR) signaling pathway (38), the autophagy pathway (39), and the p38 MAPK pathway (38). We therefore tested these known innate immune pathways in C. elegans to determine whether these pathways are involved in the iglr-2-mediated defense system against EHEC. When insulin-like signaling pathway DAF-16/FOXO transcriptional factor, daf-16, was knocked down by RNAi-mediated silencing in iglr-2 o/e, no significant difference was observed in the survival compared to empty vector, L4440 control ( Figure S8A , P= 0.14 by Log-rank test). Similar results were also found when xbp-1 was knocked down in iglr-2 o/e, which is spliced by IRE-1 during endoplasmic reticulum (ER) stress to activate UPR. The survival was compatible to the L4440 control ( Figure S8B , P = 0.24 by Log-rank test). Knockdown of a helix–loop–helix transcription factor regulating autophagy gene expression, TFEB/hlh-30, in iglr-2 o/e C. elegans resulted in slightly enhanced sensitivity compared to the control ( Figure S8C , P < 0.05 by Log-rank test). Our previous study demonstrated that the p38 MAPK pathway is activated in C. elegans for defending EHEC infection (19). To determine whether the p38 MAPK pathway is involved in iglr-2 regulating immune response, we examined the survival of iglr-2 o/e strain silencing of genes involved in p38 MAPK pathway upon EHEC infection. iglr-2 o/e subjected to pmk-1 RNAi was significantly susceptible to EHEC compared to the empty vector control, L4440 ( Figure 7A , P < 0.001 by Log-rank test). Consistently, RNAi knockdown of nsy-1, tir-1 and sek-1 suppressed the extended survival of iglr-2 o/e to EHEC, respectively ( Figures S9A–C ). Furthermore, survival of RNAi against pmk-1 to iglr-2(wf275) was compatible to that of control ( Figure 7B , P = 0.22 by Log-rank test), suggesting iglr-2 and pmk-1 act in the same pathway to defend EHEC in C. elegans. Together, these data indicated that the p38 MAPK signaling cascades act downstream of iglr-2. To further reconfirm the notion that the p38 MAPK signaling cascades act downstream of iglr-2, we performed genetic epistasis analysis using a nsy-1 gain-of-function allele, nsy-1(ums8), which constitutively activates the p38 MAPK cascades (31). We generated nsy-1(ums8);iglr-2(wf295) double mutant and examined its survival to EHEC. As shown in Figure 7C , hyperactivation of p38 MAPK pathway in iglr-2(wf295) significantly increased the survival of iglr-2 mutant animals to EHEC infection. Moreover, nsy-1(ums8);iglr-2(wf295) animals were more susceptible to EHEC infection compared to nsy-1(ums8) (P < 0.001 by Log-rank test), and the survival curve of nsy-1(ums8);iglr-2(wf295) was similar to that of wild-type (WT) animals (P = 0.835 by Log-rank test). These results reconfirm that the iglr-2, at least partly, regulated the p38 MAPK pathway. An alternative explanation of this result is that p38 MAPK pathway hyperactivation also induces aversion behavior of C. elegans to EHEC, and disruption of iglr-2 suppresses the behavior contributing nsy-1(ums8);iglr-2(wf295) susceptible to EHEC. However, we found that nsy-1(ums8) did not escape from EHEC infection, and iglr-2 loss of function in nsy-1(ums8) animal exhibited a compatible aversion behavior to iglr-2(wf295) and nsy-1(ums8) ( Figure S10 ), suggesting that suppression of the increased EHEC resistance in nsy-1(ums8);iglr-2(wf295) was not due to the absence of aversion behavior from iglr-2 loss of function. In addition to activate the p38 MAPK immune pathway in the iglr-2 mutant genetically, we also induced the p38 MAPK pathway pharmacologically to strengthen our hypothesis. RPW-24, a small molecule drug, had been reported to stimulate immune genes through p38 MAPK pathway in C. elegans (33). We therefore supplemented RPW-24 to iglr-2 mutant worms upon EHEC infection and examined the survival. Treatment of 5 µM of RPW-24 increased the survival of iglr-2 mutant animals against EHEC compared to the vehicle control ( Figure 7D , P < 0.001). An interpretation of this result is that RPW-24 may directly inhibit or kill EHEC bacteria. However, EHEC treated with 5 µM RPW-24 exhibited similar growth curves to untreated control ( Figure S11 ) suggesting that treatment of RPW-24 did not affect the growth of EHEC. Next, we asked whether the expression of p38 MAP kinase-dependent genes is changed in iglr-2 mutant. We examined the GFP expression of K08D8.5::GFP strain, a gene that encoded CUB-like domain regulated by the p38 MAPK signaling (32), under iglr-2(wf275) background. The expression pattern of K08D8.5::GFP was throughout the intestine of the worms extensively while subjected to EHEC ( Figure 7E ). In contrast, iglr-2(wf275);K08D8.5::GFP strain showed a significant reduction of GFP level to wild-type control ( Figures 7E, F , P < 0.01). These results indicated that iglr-2 is involved in the regulation of p38 MAPK-dependent gene. It did not escape our attention that iglr-2 had been reported to regulate NHR-49 (34) (ortholog of the mammalian nuclear-hormone receptor PPARα), which is correlated with C. elegans’ immune response (40, 41). We then examined the susceptibility of iglr-2 o/e C. elegans RNAi knockeddown of nhr-49 to EHEC infection. iglr-2 o/e treated with nhr-49 RNAi lived significantly shorter than the control ( Figure S12 , P < 0.001), suggesting that nhr-49 might be one of the downstream of iglr-2 in C. elegans defending EHEC. All together, our results demonstrated that iglr-2 is required for the host defense in Caenorhabditis elegans, at least in part, through the p38 MAPK immune pathway.