Hypoxia and the hypoxic response pathway protect against pore-forming toxins in C. elegans

Abstract

Pore-forming toxins (PFTs) are by far the most abundant bacterial protein toxins and are important for the virulence of many important pathogens. As such, cellular responses to PFTs critically modulate host-pathogen interactions. Although many cellular responses to PFTs have been recorded, little is understood about their relevance to pathological or defensive outcomes. To shed light on this important question, we have turned to the only genetic system for studying PFT-host interactions-Caenorhabditis elegans intoxication by Crystal (Cry) protein PFTs. We mutagenized and screened for C. elegans mutants resistant to a Cry PFT and recovered one mutant. Complementation, sequencing, transgenic rescue, and RNA interference data demonstrate that this mutant eliminates a gene normally involved in repression of the hypoxia (low oxygen response) pathway. We find that up-regulation of the C. elegans hypoxia pathway via the inactivation of three different genes that normally repress the pathway results in animals resistant to Cry PFTs. Conversely, mutation in the central activator of the hypoxia response, HIF-1, suppresses this resistance and can result in animals defective in PFT defenses. These results extend to a PFT that attacks mammals since up-regulation of the hypoxia pathway confers resistance to Vibrio cholerae cytolysin (VCC), whereas down-regulation confers hypersusceptibility. The hypoxia PFT defense pathway acts cell autonomously to protect the cells directly under attack and is different from other hypoxia pathway stress responses. Two of the downstream effectors of this pathway include the nuclear receptor nhr-57 and the unfolded protein response. In addition, the hypoxia pathway itself is induced by PFT, and low oxygen is protective against PFT intoxication. These results demonstrate that hypoxia and induction of the hypoxia response protect cells against PFTs, and that the cellular environment can be modulated via the hypoxia pathway to protect against the most prevalent class of weapons used by pathogenic bacteria.

Figure 1. Cry21A PFT-resistant ye49 mutates egl-9.

Hermaphrodites from N2 wild-type and ye49 animals after feeding on (A) Bt spore crystal lysates for 72 h without or with Cry21A crystal protein or (B) E. coli expressing no toxin or Cry21A. ye49 animals on Cry21A from either Bt or E. coli are clearly healthier than wild-type animals on Cry21A, as they are bigger, darker, and more motile. (C) Complementation tests. Images of representative animals 48 h after feeding on E. coli expressing Cry21A. Top row: controls showing relative sickness of wild-type N2 animals on Cry21A compared to ye49 and egl-9(sa307) animals on Cry21A. Bottom rows: heterozygous over wild type controls showing egl-9(sa307)/+ and ye49/+ animals are sensitive to Cry21A; ye49/egl-9(sa307) animals showing resistance to Cry21A. These animals are also all heterozygous for dpy-17(e164), used as a marker to distinguish self from cross progeny. (D) Rescue experiments. Images of representative animals 48 h after feeding on E. coli expressing Cry21A. Top: controls showing relative health of wild-type N2 and ye49 animals on Cry21A. Bottom: animals of the genotype ye49 transformed with genomic wild-type egl-9 DNA showing that expression of egl-9 in ye49 animals rescues Cry21A resistance. (E) Location of amber mutation in egl-9 gene associated with ye49 (*) as well as locations of the egl-9(sa307) deletion (bar) and egl-9(RNAi) clone (dotted line) from the Ahringer library . Boxes represent exons. (F) RNAi of egl-9 results in animals resistant to Cry21A. L4440 is empty vector RNAi control (no gene knock down). RNAi animals were fed Cry21A for 24 hours. Scale bar is 0.5 mm in this and all figures unless otherwise specified.

Figure 2. Quantitative response of egl-9 and HIF-1 pathway mutants to Cry PFTs.

(A) Schematic illustrating O₂-dependent regulation of HIF-1 activity. The O₂-dependent prolyl hydroxylation of HIF-1 by EGL-9/PHD increases its affinity to VHL-1, leading to ubiquitylation and destruction. (B, C) Dose-dependent mortality assays were performed using (B) Cry21A spore crystal lysates or (C) purified Cry5B to quantitatively compare sensitivities of wild-type N2, egl-9 mutants, and HIF-1 pathway mutants to PFTs. Each data point shows the mean and standard errors of the mean of results from three independent experiments (three wells per experiment; on average, 180 animals per data point). Statistical differences between mutant strains and N2 are given for each concentration using P values represented by asterisks as follows: * P<0.05; ** P<0.01; *** P<0.001. Percent alive at specific doses and LC₅₀ values are reported in Table 1.

Figure 3. egl-9 mutation confers resistance to V. cholerae VCC.

(A) The survival of wild-type N2 and egl-9(sa307) mutant animals on V. cholerae CVD109(VCC+) and CVD110(VCC−) are shown. (B) The survival of wild-type N2, hif-1(ia04), and egl-9(sa307) hif-1(ia04) mutant animals on V. cholerae CVD109(VCC+) and CVD110(VCC−) are shown. Quantitative data and statistical analyses for the representative experiment shown here are included in Table 2. Data for two other independent repetitions of the experiment are shown in Table S1.

Figure 4. egl-9 and hif-1 mutant phenotypes on other stressors and on aging.

(A) The survival of wild-type N2 and egl-9 mutant animals on P. aeruginosa PA14 are shown for one representative assay (n>50). (B) egl-9 mutant and N2 animals were scored for viability after 12 h exposure to 35°C. Data are averaged from three independent experiments with n>30 for each. Error bars represent standard error of the mean, * represents P values<0.05 and ** P values<0.01. (C) Comparison of egl-9 mutant animals to wild-type animals on hydrogen peroxide. Representative worms are shown for each strain 6 h after continual treatment. Wild-type N2 animals are still alive but paralyzed, whereas egl-9 mutant animals are active and healthy, indicative of resistance. (D) Lifespan of N2 and egl-9 mutants feeding on E. coli OP50 (n>30 animals per strain). One of three representative experiments shown. (E–G) hif-1(ia04) and egl-9(sa307) hif-1(ia04) mutant animals challenged with (E) P. aeruginosa PA14 (one of three representative experiments shown), (F) heat shock at 35°C for 14 h, and (G) oxidative stress (hydrogen peroxide; both mutant strains are still active, unlike N2 animals). (H) Lifespan of hif-1(ia04) and egl-9(sa307) hif-1(ia04) mutant animals feeding on E. coli OP50. Quantitative data and statistical analyses for the experiments shown here are included in Table 2. Independent repeats for the PA14 and lifespan assays are shown in Table S1.

Figure 5. Intestinal specific expression of egl-9 is sufficient to rescue Cry21A PFT resistance.

Resistance to Cry21A was compared among wild type N2, egl-9(sa307), egl-9(sa307) transformed with cpr-1::gfp, and egl-9(sa307) animals transformed with cpr1::egl-9. Two representative worms are shown for each strain 48 h after feeding on E. coli-expressed Cry21A. Wild-type animals and egl-9(sa307) animals transformed with cpr1::egl-9 are similarly sensitive to Cry21A whereas egl-9(sa307) animals and egl-9(sa307) animals transformed with cpr-1::gfp are relatively resistant.

Figure 6. nhr-57, a HIF-1-target, is required for loss of egl-9-mediated resistance to PFT.

Wild-type N2 and egl-9(sa307) mutant animals were treated to RNAi using either empty vector (L4440) or dsRNA for nhr-57 and exposed to E. coli-expressed Cry21A PFT for 48 hours. Shown are representative animals for each condition. All animals are healthy in the absence of Cry21A. On Cry21A, whereas egl-9(sa307) animals are resistant (third row), nhr-57(RNAi) egl-9(sa307) animals are not.

Figure 7. Cry5B PFT activate the hypoxia pathway and hypoxia confers protection against PFTs.

(A) Quantitative real-time PCR analysis of nhr-57 expression in glp-4(bn2) animals upon treatment on Cry5B for 1, 2, 4 and 8 hours relative to expression levels in no-toxin treatment animals. Data are averaged from three independent experiments. Error bars represent standard error of the mean, * represents P value<0.05 and *** P value<0.001. (B) Resistance to Cry5B PFT was compared among wild-type N2 worms in normoxia (top two rows) and in hypoxia (2% O₂) (rows 3 and 4) and among hif-1(ia04) mutant animals in normoxia (rows 5 and 6) and in hypoxia (bottom two rows). Wild-type worms co-treated with hypoxia and Cry5B are significantly healthier (larger, darker color, more embryos, more motile) than worms treated with Cry5B under normoxia. In contrast, hif-1(ia04) mutant animals on Cry5B PFT look similarly sick either in the normoxic or in the hypoxic environment. Scale bar is 0.2 mm. (C) Schematic illustrating our results and relationship between the HIF-1 pathway and PFT INCED. In response to PFT, the hypoxia response is activated by suppression of EGL-9, either by low oxygen and/or other means. HIF-1 activates expression of target genes that protect against PFT intoxication, such as nhr-57. Activation of HIF-1 is also able to activate the XBP-1 arm of the UPR defense pathway.