The Conserved G-Protein Coupled Receptor FSHR-1 Regulates Protective Host Responses to Infection and Oxidative Stress

The innate immune system’s ability to sense an infection is critical so that it can rapidly respond if pathogenic microorganisms threaten the host, but otherwise maintain a quiescent baseline state to avoid causing damage to the host or to commensal microorganisms. One important mechanism for discriminating between pathogenic and non-pathogenic bacteria is the recognition of cellular damage caused by a pathogen during the course of infection. In Caenorhabditis elegans, the conserved G-protein coupled receptor FSHR-1 is an important constituent of the innate immune response. FSHR-1 activates the expression of antimicrobial infection response genes in infected worms and delays accumulation of the ingested pathogen Pseudomonas aeruginosa. FSHR-1 is central not only to the worm’s survival of infection by multiple pathogens, but also to the worm’s survival of xenobiotic cadmium and oxidative stresses. Infected worms produce reactive oxygen species to fight off the pathogens; FSHR-1 is required at the site of infection for the expression of detoxifying genes that protect the host from collateral damage caused by this defense response. Finally, the FSHR-1 pathway is important for the ability of worms to discriminate pathogenic from benign bacteria and subsequently initiate an aversive learning program that promotes selective pathogen avoidance.


Introduction
To survive an attack by a microbial pathogen, a host must first recognize that the pathogen is attempting to cause an infection so that it can then mount a response. Recognition can occur via direct binding of host pattern-recognition receptors (PRRs), including Toll-like receptors (TLRs) and NOD-like receptors (NLRs), to microbe-associated molecular patterns (MAMPs)

FSHR-1 delays pathogen accumulation in the intestine
Intestinal infection by Pseudomonas aeruginosa strain PA14 progresses more quickly in fshr-1 mutants than in wild-type C. elegans. Wild-type worms exposed to PA14 initially appear healthy, but by 8 hours post-infection they begin to show early symptoms of the infection, including a distension of the lumen of their intestine. By 24 hours post-infection, the worms may appear sickly and slightly small and the intestine has begun to accumulate live actively dividing bacteria [10]. To determine the effect of a null fshr-1(ok778) mutation on this progression of infection, we exposed worms to GFP-labeled PA14. At the light microscopic level, fshr-1(ok778) mutant worms follow the same sequence of infection as wild-type worms, but develop symptoms of infection more quickly. For example, by 18 hours of exposure to GFP-labeled PA14, most wild-type worms have very little fluorescence in the lumen of their intestines, suggesting they have not yet amassed a substantial amount of PA14 (Fig 1A and 1C). In contrast, after 18 hours of exposure to GFP-PA14, fshr-1(ok778) mutants have already accumulated a significantly greater amount of the fluorescent pathogen in their intestinal lumens (P<0.0001) (Fig 1B and 1C).
Both wild-type and fshr-1(ok778) mutant worms eventually succumb to a PA14 infection. However, infected fshr-1(ok778) worms have a mean survival time of 39 hours, significantly less than the wild-type mean survival time of 73 hours (P<0.0001) (Fig 1D). We previously showed that the slope of the survival curves is approximately the same but the fshr-1 survival curve is shifted to the left [9]. This observation, combined with the more rapid accumulation of pathogenic PA14, suggests that FSHR-1 protects C. elegans from bacterial infection by delaying the establishment or progression of the infection.
Could FSHR-1 be a pattern recognition receptor? Its extracellular domain contains seven Leucine-Rich Repeat domains, protein-protein interaction motifs that are also present in the ligand-binding domains of Toll-like receptors (TLRs) and Nod-like receptors (NLRs) [11]. However, fshr-1(ok778) mutants are more sensitive than wild-type worms to infection by diverse pathogens, including not only the Gram negative pathogen PA14, but also the Gram positive bacterial pathogens Staphylococcus aureus (P>0.01) and Enterococcus faecalis (P>0.05) (Fig 1D) [9], as well as the eukaryotic fungal pathogen Candida albicans (R. Pukkila-Worley and J. Powell, personal communication). Because these microbes are members of structurally dissimilar classes, it is unlikely that FSHR-1 detects infection by binding directly to a common MAMP possessed by these distinct phyla. Rather than serving as a direct PRR like fellow LRR-containing TLRs and NLRs, we propose that FSHR-1 is an indirect sensor of infection.

FSHR-1 regulates the expression of immune and stress-associated genes
To understand the effect of the FSHR-1 pathway on the innate immune response in C. elegans, we examined gene expression in infected worms. We previously showed that FSHR-1 is required in C. elegans for the transcriptional induction of a set of known infection response genes upon exposure to PA14 [9]. Many of these genes are thought to have antimicrobial activity that antagonizes the infecting pathogen [12]. To gain a more complete picture of the downstream transcriptional impact of the FSHR-1 pathway, we performed a full-genome microarray on wild-type or fshr-1(ok778) null mutant worms that were either infected with PA14 or grown on relatively non-pathogenic E. coli OP50 as a control [13]. We used quantitative RT-PCR (S1 four genes were expressed at greater than 2-fold higher levels (P<0.05) in wild-type worms relative to fshr-1(ok778) mutants grown on E. coli (S1 Table). Fifty-six genes were expressed at greater than 2-fold higher levels (P<0.05) in wild-type worms infected with PA14 relative to infected fshr-1(ok778) mutants (S2 Table). When we examined the reported functions of the genes whose expression in infected worms depends on fshr-1, we observed that many were known antimicrobial infection response genes; in addition, 18% of fshr-1-regulated genes were associated with some type of cellular stress (Fig 1E).

FSHR-1 acts in the intestine to mediate survival of heavy metal and oxidative stress
One way that host cells sense an infection indirectly is by assessing infection-induced cellular damage or stress. Because the loss of fshr-1 is associated with sensitivity to diverse pathogens and results in the misregulation of a set of stress-response genes, we hypothesized that the FSHR-1 pathway may contribute to the worm's immune defenses by responding to general -type and fshr-1(ok778) mutant worms exposed to different bacterial food sources. * P<0.05, **P<0.01, *** P<0.0001 in unpaired t-test. Error bars are SEM. (E) Functional annotation of genes whose mean expression from three microarray replicates in wild-type worms infected with PA14 for 4 hours was at least 2-fold higher than their mean expression in fshr-1(ok778) mutants infected with PA14 for 4 hours, with P<0.05. Genes known to be involved both in infection response and stress response were grouped with stress response. damage or cellular stress caused by infection. As a corollary, we predicted that the FSHR-1 pathway might also sense some types of cellular stress in the absence of infection. To test this prediction, we measured the survival of worms subjected to several types of stress.
To determine the effects of heavy metal stress on fshr-1(ok778) mutants, we exposed L4 worms to 100 μM cadmium and measured their survival. fshr-1(ok778) mutants died significantly more quickly in the presence of cadmium than wild-type worms (One-way ANOVA, F = 110.68; Tukey HSD test, P<0.05) (Fig 2A and 2B). Expression of fshr-1 in the intestine is necessary and sufficient for its role in the response to infection by PA14 [9], so we also tested whether fshr-1(+) expressed in the intestine could rescue the cadmium sensitivity phenotype of fshr-1(ok778) mutants. Interestingly, expression of fshr-1(+) from an intestinal promoter rescued the cadmium sensitivity phenotype and conferred resistance to cadmium exposure (Oneway ANOVA, F = 110.68; Tukey HSD test, P<0.01) (Fig 2A and 2B). These data suggest that the fshr-1 pathway is not only required for surviving cadmium exposure, but also could be sufficient in the intestine to protect worms against cadmium-induced damage independent of infection.
To determine whether the fshr-1 pathway plays a role in the survival of oxidative stress, we placed worms on plates containing 5 mM paraquat, which causes the production of reactive oxygen species (ROS). fshr-1(ok778) mutants died significantly more quickly than wild-type worms (One-way ANOVA, F = 40.96; Tukey HSD test, P<0.01) (Fig 2C and 2D). Additionally, intestinal expression of fshr-1(+) rescued the paraquat sensitivity phenotype (One-way ANOVA, F = 40.96; Tukey HSD test, P<0.01), suggesting that the fshr-1 pathway is important for the ability of worms to survive infection-independent oxidative stress in the same tissue in which it acts to promote survival of infection.
Wild-type worms subjected to thermal stress survive at 37°C for an average of 320 minutes (Fig 2E and 2F). The mean survival of fshr-1(ok778) mutants at 37°C is not significantly different from the survival of wild-type worms (One-way ANOVA, F = 0.42, P = 0.669), indicating that the fshr-1 pathway does not respond to the damage caused by high heat.
Because fshr-1(ok778) mutants are sensitive to oxidative and cadmium stress but not to thermal stress, the FSHR-1 pathway responds specifically to certain types of stress rather than more broadly to any type of cellular damage. Many studies have shown that cadmium exposure causes oxidative stress through the production of ROS (recently reviewed in [14]). Microarray experiments have also shown that cadmium exposure induces many of the same changes in gene expression as infection by Pseudomonas aeruginosa [12]. It is possible this common response is due to the fact that ROS play important roles in the interaction between host and pathogen during an infection [15]. Thus, host cells are subjected to oxidative stress both when infected and when exposed to toxic doses of cadmium. We propose a model in which the FSHR-1 pathway senses and responds to oxidative damage from a variety of sources, including exogenously added chemicals and the worm's own immune system. FSHR-1 controls the induction of the oxidative-stress response gene gcs-1 in response to infection C. elegans intestinal cells generate an oxidative burst upon infection with Enterococcus faecalis strain OG1RF [15]. The dual oxidase Ce-Duox1/BLI-3 is responsible for the formation of these ROS, and is required for the worm to fight the E. faecalis infection [16,17]. Because ROS react with and damage a variety of biological molecules, and do not discriminate against host or pathogen targets, C. elegans intestinal cells and other host cells that produce ROS as a form of pathogen defense must also induce an oxidative stress response to detoxify the ROS and protect itself from collateral damage. In worms infected with E. faecalis or PA14, this detoxification response is mediated by the Nrf transcription factor SKN-1 [18].
GCS-1 is a phase II detoxification enzyme and catalyzes the rate-limiting step of glutathione biosynthesis. It is an important target of SKN-1 and is a participant in the oxidative stress response in C. elegans. Wild-type worms induce the intestinal expression of a gcs-1::gfp transcriptional reporter within 5 hours of infection with PA14 (One-way ANOVA, F = 13.33; Tukey HSD test, P<0.01) (Fig 3A, 3B and 3E). Note that this change in gene expression occurs rapidly, before visible physical symptoms appear. In contrast, fshr-1(ok778) mutant worms fail to induce expression of gcs-1::gfp upon PA14 infection (Fig 3C-3E). These data, in combination with the microarray (Fig 1E), indicate that fshr-1 is required for the early activation of the host's protective oxidative stress response upon infection.

FSHR-1 mediates the avoidance of pathogens
Another important component of immunity is a behavioral response to pathogen exposure. Worms increase their survival when they avoid a lawn of pathogenic bacteria, presumably because they ingest fewer pathogen cells [19][20][21][22]. To determine whether fshr-1 contributes to the behavioral avoidance of pathogens, we measured the occupancy of worms on lawns of pathogenic and non-pathogenic bacteria. Both wild-type and fshr-1(ok778) mutant worms prefer to forage on a lawn of relatively non-pathogenic E. coli strain OP50, with more than 90% of the worms found occupying the lawn (Fig 4A and 4B). Wild-type worms can distinguish between pathogenic and non-pathogenic bacteria and initiate an aversive learning program to actively avoid feeding on the pathogen [20,23,24]. Initially, wild-type worms are attracted to pathogenic PA14 and occupy a lawn at levels similar to a non-pathogenic OP50 lawn (Fig 4A  and 4C). After 9 hours of exposure to PA14, however, only 29% of wild-type worms occupy a lawn of the pathogen (Fig 4D). Although fshr-1(ok778) mutants are indistinguishable from wild-type in their occupancy of non-pathogenic OP50 lawns and of their initial occupancy of pathogenic PA14 lawns (Fig 4A, 4B and 4C), fshr-1(ok778) mutants are impaired in their ability to learn pathogen avoidance. 45% of fshr-1(ok778) mutants remain on a lawn of PA14 after 9 hours of exposure (One-way ANOVA, F = 8.89; Tukey HSD test, P<0.01) (Fig 4D). Additionally, fshr-1(RNAi) worms show significantly lower lawn occupancy (51%) after 9 hours of exposure to PA14 than control RNAi worms (21%, two-sample T-test, P<0.001), confirming that FSHR-1 is required for the ability of worms to avoid pathogenic bacteria (S3 Fig).
Learned behavioral avoidance requires that the worm identify the microbe and associate that identity with the effects of the infection. Microbial identification is mediated by a chemosensory circuit that detects secondary metabolites and activates a neuroendocrine signaling pathway [23,24]. Although the detection of secondary bacterial metabolites associated with virulence can indicate not only the presence of a bacterium but also an infection, it is likely this neuronal signal is integrated with signals from the site of infection. To determine whether FSHR-1 acts in the intestine to influence pathogen avoidance, we measured the lawn occupancy of worms expressing FSHR-1 exclusively in the intestine. Intestine-specific expression of the wild-type fshr-1(+) from the integrated transgene agIs35[gut::fshr-1(+)] was sufficient to rescue the PA14 avoidance defect of fshr-1(ok778) null mutants (One-way ANOVA, F = 8.89; Tukey HSD test, P<0.01), suggesting that FSHR-1 does function in the intestine for this role (Fig 4D).
Interestingly, worms with the agIs35[gut::fshr-1(+)] transgene inappropriately avoid bacterial lawns in the absence of a pathogen. Only 60% of gut::fshr-1(+) transgenic worms occupy a lawn of OP50 after 9 hours of exposure, while 90% of wild-type and fshr-1(ok778) mutant worms are present in the lawn (Fig 4B) (One-way ANOVA, F = 54.84, P<0.0001). Even in the absence of aversive learning, naïve transgenic worms show a modest but statistically significant 10% reduction in their occupancy of a bacterial lawn on which they were recently placed ( Fig  4A and 4C) (OP50: One-way ANOVA, F = 8.38; Tukey HSD test, P<0.01. PA14: One-way ANOVA, F = 9.85; Tukey HSD test, P<0.01). It is possible that the heterologous intestinal promoter and multiple copies in which the gut::fshr-1(+) transgene is present cause a hyper-activation of the FSHR-1 pathway, which could incorrectly or constitutively signal the need to avoid a microbe.
To determine the extent to which an altered avoidance behavior contributes to the infection sensitivity phenotype of fshr-1(ok778) mutants, we performed killing assays using standard or big lawns. In a standard killing assay, the lawn of pathogen covers only part of the plate, giving the worms the option to avoid this lawn and reduce their consumption of the pathogen. A big lawn assay removes this choice by covering the entire surface of the agar with the pathogen. Wild-type worms die more quickly in a big-lawn killing assay than in a standard small-lawn killing assay (One-way ANOVA, F = 255.04; Tukey HSD test, P<0.01), confirming previous reports that avoiding the lawn is one successful strategy for fighting infection (Fig 5A and 5B) [19,25,26]. Infected fshr-1(ok778) mutants die along a nearly identical time course in standard small-and big-lawn killing assays, providing additional support for the conclusion that the FSHR-1 pathway regulates a behavioral pathogen avoidance component of the immune system. Importantly, fshr-1(ok778) mutants are still more sensitive to infection than wild-type worms in a big-lawn killing assay (One-way ANOVA, F = 255.04; Tukey HSD test, P<0.01) ( Fig 5A and 5B), indicating that regulation of pathogen avoidance is not the only contribution of the FSHR-1 pathway to innate immune defense.

Conclusions
In summary, the conserved GPCR FSHR-1 mediates multiple defensive and protective facets of the C. elegans innate immune system. FSHR-1 is required for the survival both of infection by diverse pathogens and of xenobiotic insults that cause oxidative stress. Upon infection, FSHR-1 activates expression of a collection of known antimicrobial infection response genes that defend the worm by attacking the pathogen. FSHR-1 also activates expression of a collection of oxidative stress response genes that detoxify ROS produced as part of the worm's defense and thereby protect the worm against oxidative damage. FSHR-1 is additionally required for the protective ability of worms to sense an infection and modify their behavior to avoid pathogens.
We have previously shown that FSHR-1 is necessary and sufficient in the intestine for its role in promoting survival of infection by ingested pathogens [9]. Here we provide evidence that this receptor coordinates multiple facets of innate immunity from the intestine, the site of infection. An intestinal role for FSHR-1 in the response to infection-induced oxidative stress is not unexpected because the worms must sense and respond to oxidative damage at the site of ROS production, the intestine. FSHR-1 also acts in the intestine for its protective role against infection-independent oxidative stress, suggesting the same mechanism can be used to sense xenobiotic ROS production despite the fact that the worm's exposure to ROS-generating cadmium and paraquat is not restricted to the intestinal epithelium.
Learned pathogen avoidance behavior is a more complex phenomenon that requires the integration of sensory signals from multiple tissues. ASJ chemosensory neurons detect secondary metabolites produced by bacteria via a G-protein signaling pathway and communicate with interneurons via a TGF-β neuroendocrine signal [26]. The intestine and hypodermis sense perturbations in essential cellular processes caused by bacterial toxins at the site of an infection [27]. The bacterial identity signal from the neurons is integrated with an infection signal from the intestine and/or hypodermis to modulate the worm's behavior and promote learned avoidance of a pathogen via a suppression of aerotaxis and activation of serotonin signaling [23,24,27]. Although G-protein signaling is important for the neuronal identification of FSHR-1 Regulates Immune and Oxidative Stress Responses bacteria, and it is a formal possibility that FSHR-1 also acts in the neurons, our data indicate that the GPCR FSHR-1 acts in the intestine to promote the avoidance of pathogens, suggesting that FSHR-1 contributes to the intestinal detection of infection rather than the neuronal identification of bacteria.
How might the FSHR-1 pathway sense an infection? The production of ROS by host cells is a widely utilized mechanism of pathogen defense that also results in collateral damage to host tissues [15]. In addition, multiple pathogens have been shown to use ROS production as an important virulence determinant, suggesting that oxidative damage could be classified as a pattern of pathogenesis [28][29][30][31]. Regardless of the source of oxidative damage-bacterial attack, host defense, or xenobiotic assault-a critical response is the activation of detoxification enzymes that mitigate the damage. In addition, epithelial cells sensing oxidative stress would increase the survival of the host by initiating or amplifying local anti-pathogen defenses such as antimicrobial infection response gene expression, and endocrine signals that trigger an aversive learning program. Because FSHR-1 is required both for the survival of xenobiotic oxidative stress, and for the activation of oxidative stress response and infection response genes upon infection, we propose a model in which the FSHR-1 pathway could sense oxidative damage as a pattern of pathogenesis and then activate multiple facets of the innate immune response.
The ligand for FSHR-1 is unknown. FSHR-1 belongs to the LGR class of GPCRs; homologs bind the heterodimeric glycopeptide hormone FSHα/β [32][33][34]. The worm genome does not contain an identifiable FSHα subunit and no known role in immunity has been associated with the gene that most closely resembles FSHβ, suggesting that FSHR-1 may bind a noncanonical ligand. Because FSHR-1 is important for the defense against multiple dissimilar pathogens, it is not likely to be a PRR that binds a MAMP. We have shown that the FSHR-1 pathway functions in the absence of infection to promote survival of heavy metal and oxidative stress, suggesting that FSHR-1's ligand must be endogenously produced rather than bacterial in origin. A recently described example of a GPCR that recognizes infection in C. elegans by binding an endogenous ligand is the DAMP receptor DCAR-1 [4]. HPLA is a tyrosine metabolite that is produced by the worm upon physical disruption of the cuticle and underlying hypodermis by a sterile wound or by the piercing hyphae of the pathogenic fungus Drechmeria coniospora. Rather than detecting physical damage as a sign of infection, it is tempting to speculate that FSHR-1 could bind an endogenously-produced DAMP that signals chemical damage caused by ROS production. An important future step in understanding the recognition and response to infection in epithelial cells is the identification of the FSHR-1 ligand.

C. elegans strains and maintenance
C. elegans were grown at 20°C on NGM agar plates seeded with E. coli strain OP50 as previously described [35]. For experiments requiring developmentally synchronized worms, gravid hermaphrodites were bleached to obtain embryos, which hatched in sterile M9 buffer overnight at room temperature. Arrested L1s were pipetted to the appropriate seeded plates to resume development.

Killing assays
Infection-based "slow killing" pathogen assays were performed as described [36,37]. Briefly, 10 μl of a saturated overnight culture of P. aeruginosa PA14 was pipetted onto 3.5 cm SK plates (NGM, with 0.35% Bacto-peptone) and spread slightly (standard lawn) or to cover the plate (big lawn). The plates were incubated at 37°C for 24 hours, followed by 25°C for 24 hours [38]. FUDR (75 μg/ml) was added to the perimeter of the agar 30-60 minutes before assay set-up.
L4 larvae were transferred to the plates, incubated at 25°C, and monitored for survival over time [36,37].
For oxidative and cadmium stress killing assays, 3.5 cm NGM plates were poured with a supplement of paraquat or CdCl 2 . Concentrations of paraquat between 0.2 mM and 200 mM are commonly used to induce cytoplasmic oxidative stress so we used 5 mM paraquat in our assays [39,40]. We used 100 μM cadmium to induce heavy metal stress, as previously described [41]. Plates were seeded with 20x concentrated OP50 and FUDR (75 μg/ml) was added to the perimeter of the agar 30-60 minutes before assay set-up. L4 larvae were transferred to the plates and monitored for survival over time using a nose-touch response assay.
For heat stress killing assays, synchronized L1s were pipetted onto NGM plates seeded with OP50 and grown at 20°C until the young adult stage. Plates were transferred to 37°C and the numbers of live and dead worms were scored at multiple time points using a nose-touch response assay. To minimize temperature fluctuations throughout the assay, each plate was scored only once and then discarded.
All killing assays were performed in triplicate with approximately 40 worms per plate, and each experiment was repeated at least three times. Mean survival time was calculated for each plate as described [37]. Conditions were compared using a one-way ANOVA and Tukey HSD test.

Pathogen accumulation
To measure accumulation of pathogens, N2 wild-type or fshr-1(ok778) L4 worms were transferred to SK plates prepared as for a pathogen killing assay, but with GFP-labeled PA14. After 18 hours of incubation at 25°C, worms were anesthetized with 25 mM NaN 3 , then photographed using a Nikon 90i compound epiflourescence microscope. To quantify the pathogen accumulation, the sum of the fluorescence intensity in the worm was determined for at least 10 worms per condition.

Avoidance assays
To measure bacterial avoidance, 7 μl of PA14 or 10 μl of OP50 saturated overnight culture was pipetted onto a 3.5 cm SK plate. The plates were grown at 37°C for 24 hours, followed by 25°C for 24 hours. L4 worms were transferred to the center of the bacterial spot and incubated at 25°C. All plates were scored for avoidance of the lawn at least 30 minutes but no more than 1 hour after set-up, and again 9 hours after set-up. Worms were scored as "in" the lawn if at least 50% of their posterior body were in contact with the lawn or if their head were in contact with and foraging in the lawn. Worms were scored as "out" of the lawn if their head and >50% of their body were outside the lawn. All experiments were performed in triplicate and repeated at least three times. The mean lawn occupancies for different conditions were compared using a one-way ANOVA and Tukey HSD test.

Microarray
To measure genome-wide gene expression, young adult N2 wild-type or fshr-1(ok778) worms were transferred to SK plates with PA14 or OP50 prepared as if for a pathogen killing assay. Samples were prepared as previously described [12]. Briefly, total RNA was harvested in TriReagent (Molecular Research Center) after 4 hours of incubation on SK plates. cDNA was synthesized following the recommended RetroScript (Life Technologies) protocol. After second strand synthesis, double-stranded cDNA was purified using an RNeasy (Qiagen) column. Samples were hybridized on Affymetrix full-genome GeneChips for C. elegans at the Harvard Medical School Biopolymer Facility. Each experiment was repeated three times. The data were analyzed using GenePattern (Broad Institute) and Microsoft Excel. The mean expression for the three replicates of each gene was compared between conditions using a two-tailed two-sample t-test. A gene was considered differentially expressed between conditions if there was a greater than 2-fold difference in mean expression with P<0.05.
To validate the microarray data, quantitative RT-PCR was performed as previously described [12] on four genes that were induced in infected worms according to microarray data. Briefly, wild-type N2 and fshr-1(ok778) mutant young adults were transferred to SK plates with PA14 or OP50 prepared as if for a pathogen killing assay. After 4 hours, worms were washed off the plates with M9 buffer and washed two more times to remove excess bacteria. Total RNA was extracted using TRI-Reagent. cDNA was synthesized using a Retro-script kit (Ambion) with oligo-dT primers, and then amplified on a MyiQ2 iCycler (Bio-Rad) with iQ supermix containing SYBR green (Bio-Rad). Primers are available upon request. Normalized expression values were compared among three independent biological replicates. For each experiment, the induction was calculated as the expression in worms exposed to PA14 divided by the expression in worms exposed in parallel to OP50. A two-sample T-test was used to compare the induction in wild-type worms with the induction in fshr-1(ok778) mutants.

Reporter induction
To measure induction of the ldIs3 [gcs-1::gfp] oxidative stress response reporter, L4 LD1171 ldIs3 or JRP1024 fshr-1(ok778); ldIs3 worms were transferred to SK plates with PA14 or OP50 prepared as if for a pathogen killing assay. After 5 hours, worms were anesthetized with 25 mM NaN 3 , then photographed using a Nikon 90i compound epiflourescence microscope. To quantify reporter expression, the intestine was manually outlined using the Nikon Elements Advanced Research software. The mean fluorescence intensity of this Region of Interest was determined by the software for at least 10 worms per condition.
To validate the microarray data, reporter strains BC11778 sEx11778 [gpx-1::gfp] and BC14266 sEx14266 [gst-38::gfp] [42] were transferred as L4 larvae to SK plates with PA14 or OP50 prepared as described above. After 5 hours, worms were anesthetized and photographed as described above. The exposure time was held constant for all conditions within an experiment.  Fig. FSHR-1 is required for the avoidance of pathogenic PA14. L4 fshr-1 (RNAi) and L4440 control (RNAi) worms were transferred to lawns of pathogenic P. aeruginosa PA14. The lawn occupancy was measured 1 hour and 9 hours after the transfer and the means of the two RNAi conditions were compared at each time point with a T-test. ÃÃÃ P<0.001. Error bars are SEM. (TIF) S1 Table. Ratio of basal gene expression in uninfected worms. The ratio of gene expression was calculated as the expression in wild-type N2 worms / fshr-1(ok778) mutant worms grown on E. coli OP50. The genes shown have at least 2-fold higher expression in wild-type worms than in mutants, with P<0.05 in a two-sample T-test. (XLSX) S2 Table. Ratio of gene expression in infected worms. The ratio of gene expression was calculated as the expression in wild-type N2 worms / fshr-1(ok778) mutant worms grown on E. coli OP50 until the young adult stage and then transferred to P. aeruginosa PA14 for 4 hours. The genes shown have at least 2-fold higher expression in wild-type worms than in mutants, with P<0.05 in a two-sample T-test. (XLSX)