Burkholderia pseudomallei suppresses Caenorhabditis elegans ...

2 downloads 0 Views 949KB Size Report
Sep 10, 2013 - Burkholderia pseudomallei is a Gram-negative soil bacterium that ... system. Degradation of ELT-2 requires the B. pseudomallei type.

Burkholderia pseudomallei suppresses Caenorhabditis elegans immunity by specific degradation of a GATA transcription factor Song-Hua Leea,1,2, Rui-Rui Wonga,1, Chui-Yoke China,3, Tian-Yeh Lima, Su-Anne Enga, Cin Konga, Nur Afifah Ijapa, Ming-Seong Laua, Mei-Perng Lima, Yunn-Hwen Ganb, Fang-Lian Hec,d,4, Man-Wah Tanc,d,5,6, and Sheila Nathana,e,6 a

School of Biosciences and Biotechnology, Faculty of Science and Technology, Universiti Kebangsaan Malaysia, 43600 UKM Bangi, Selangor, Malaysia; Department of Biochemistry, Yong Loo Lin School of Medicine, National University of Singapore, Singapore 117597; Departments of cGenetics and Microbiology and Immunology, Stanford University School of Medicine, Stanford, CA 94305-5120; and eMalaysia Genome Institute, Jalan Bangi, 43000 Kajang, Selangor, Malaysia b d

Edited by Frederick M. Ausubel, Harvard Medical School and Massachusetts General Hospital, Boston, MA, and approved August 7, 2013 (received for review June 24, 2013)

innate immunity

Taking advantage of the tractability and simplicity of C. elegans as a model, we investigated host responses to B. pseudomallei infection in the context of a whole organism. Wholegenome gene expression profiling in C. elegans identified a set of ELT-2–regulated transcripts that are progressively down-regulated over a time course of B. pseudomallei infection that is concomitant with a progressive and specific loss of nuclear ELT-2 in the worm intestine. Loss of ELT-2 requires the host ubiquitin–proteasome system (UPS) and is dependent on the B. pseudomallei type III secretion system (T3SS). Given the conserved role of GATA factors in epithelial immunity, this finding may contribute new insights into B. pseudomallei infection of humans. Significance Bacterial pathogens use multiple mechanisms to survive and proliferate within an infected host, including blunting the host’s ability to defend itself from pathogenic assaults. We identified a new immune suppression mechanism by Burkholderia pseudomallei, the causative agent of melioidosis, which a lifethreatening disease in humans. Analyses of whole-genome transcriptional responses of Caenorhabditis elegans to B. pseudomallei infection revealed that B. pseudomallei, through its type III secretion system, recruits the host ubiquitin–proteasome system to specifically degrade a GATA transcription factor. This GATA factor is critical for host immune defense; thus, its degradation leads to suppression of the host’s ability to mount an effective antimicrobial defense.

| ubiquitin–proteosomal system

B

acterial pathogens attack host cells to gain access to a privileged niche within a host, whereas hosts respond by activating their immune system aimed at restricting and eliminating infecting pathogens. Several conserved innate immune signaling pathways have been revealed from studies of host–pathogen interactions using vertebrate and invertebrate models. They include the Toll-like receptor signaling and a GATA transcription factor with a conserved role in epithelial immunity in Caenorhabditis elegans, Drosophila, and mammals (1, 2). The transcription factor FOXO, which is regulated by the conserved insulin signaling pathway, is another factor regulating the expression of antimicrobial peptides in worms, fly, and mammals (3–6). In the coevolutionary arms race between host and pathogens, selective pressures imposed by the host drive pathogens to evolve mechanisms to subvert or suppress host immune responses. For example, Pseudomonas aeruginosa suppresses the expression of a subset of FOXO-regulated immune defense genes in C. elegans by activating the DAF-2/DAF-16 insulin-like signaling pathway (7). Burkholderia pseudomallei is the causative agent of melioidosis, a severe infectious disease endemic to Southeast Asia, northern Australia, and other tropical areas (8). B. pseudomallei infections are responsible for up to 40% of sepsis-related mortality. No licensed vaccine is currently available for immunoprophylaxis of melioidosis. Despite much knowledge gained on the pathogenic factors and immunogenic agents of B. pseudomallei (8), molecular mechanisms underlying host susceptibility and response to infection remain poorly understood. www.pnas.org/cgi/doi/10.1073/pnas.1311725110

Author contributions: S.-H.L., C.-Y.C., M.-W.T., and S.N. designed research; S.-H.L., R.-R.W., C.-Y.C., T.-Y.L., S.-A.E., C.K., N.A.I., M.-S.L., M.-P.L., and F.-L.H. performed research; Y.-H.G. and M.-W.T. contributed new reagents/analytic tools; S.-H.L., R.-R.W., C.-Y.C., T.-Y.L., S.-A.E., C.K., N.A.I., M.-W.T., and S.N. analyzed data; and S.-H.L., R.-R.W., M.-W.T., and S.N. wrote the paper. The authors declare no conflict of interest. This article is a PNAS Direct Submission. Data deposition: The microarray data reported in this paper have been deposited in the Stanford Microarray Database, http://smd.stanford.edu. 1

S.-H.L. and R.-R.W. contributed equally to this work.

2

Present address: Department of Integrative Biology, University of California, Berkeley, CA 94720-3102.

3

Present address: Emory Vaccine Center, Emory University, Atlanta, GA 30329.

4

Present address: Department of Plant Biology, Carnegie Institution for Science, Stanford, CA 94305-4101.

5

Present address: Department of Infectious Diseases, Genentech, Inc., South San Francisco, CA 94080.

6

To whom correspondence may be addressed. E-mail: [email protected] or [email protected] ukm.my.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. 1073/pnas.1311725110/-/DCSupplemental.

PNAS | September 10, 2013 | vol. 110 | no. 37 | 15067–15072

MICROBIOLOGY

Burkholderia pseudomallei is a Gram-negative soil bacterium that infects both humans and animals. Although cell culture studies have revealed significant insights into factors contributing to virulence and host defense, the interactions between this pathogen and its intact host remain to be elucidated. To gain insights into the host defense responses to B. pseudomallei infection within an intact host, we analyzed the genome-wide transcriptome of infected Caenorhabditis elegans and identified ∼6% of the nematode genes that were significantly altered over a 12-h course of infection. An unexpected feature of the transcriptional response to B. pseudomallei was a progressive increase in the proportion of down-regulated genes, of which ELT-2 transcriptional targets were significantly enriched. ELT-2 is an intestinal GATA transcription factor with a conserved role in immune responses. We demonstrate that B. pseudomallei down-regulation of ELT-2 targets is associated with degradation of ELT-2 protein by the host ubiquitin–proteasome system. Degradation of ELT-2 requires the B. pseudomallei type III secretion system. Together, our studies using an intact host provide evidence for pathogen-mediated host immune suppression through the destruction of a host transcription factor.

Results Transcriptional Changes in Response to Pathogenic B. pseudomallei.

To assess temporal transcriptional changes in B. pseudomalleiinfected C. elegans, we performed a genome-wide transcriptome analysis on age-matched adult worms infected with B. pseudomallei strain R15 (henceforth referred to as BpR15), a clinical isolate that is highly pathogenic on C. elegans (9). We compared transcript levels in worms exposed to BpR15 for 2, 4, 8, and 12 h to those exposed to Escherichia coli strain OP50—the standard laboratory food source as uninfected control—over the same time course. The initial time points precede the first deaths, whereas at 12 h postinfection (hpi), a mortality of less than 5% of the population was observed (Fig. 1A). At each experimental time point, microarray analysis compared gene expression between B. pseudomallei-infected and -uninfected control animals relative to a reference control, which comprised a mixed-stage population of C. elegans. The application of a reference control to normalize C. elegans gene expression enables an accurate comparison of the average gene expression between any two time points (10). B. pseudomallei provoked progressively greater transcriptional changes in C. elegans over the course of infection. Between 2 and 8 hpi, an increasingly larger subset of genes was up-regulated in response to BpR15 compared with E. coli (significance analysis of microarray analysis, at a false-discovery rate of ≤1%) (11). However, at 12 hpi, fewer genes were up-regulated (454 genes) and more were down-regulated (493 genes) (Table S1). Notably, the proportion of down-regulated genes increased with time of infection. Whereas only 16% of modulated genes were downregulated at 2 hpi, the proportion progressively increased to ∼30% at 4 and 8 hpi and to 53% at 12 hpi (Fig. S1), hinting at the possibility that suppression of host defense genes expression may be an important virulence mechanism used by B. pseudomallei. Cluster analysis of time course data from 2 to 12 hpi revealed that ∼6% (>1,200) C. elegans genes were differentially expressed in response to B. pseudomallei, of which 694 genes were up-regulated and 560 genes were down-regulated (Fig. 1B and Dataset S1). Although a majority of induced genes after 8 hpi were involved in defense response (Dataset S1), some that encode immune effectors, such as lys-7 and spp-1, were significantly repressed; the significance of this observation will be addressed below. Quantitative real-time PCR (qRT-PCR) measurements confirmed microarray results for 30 genes identified as infection and stress response genes with altered expression at 12 hpi (Table S2) (7). Functional category analysis revealed that defense-specific genes, including those with putative roles in

Fig. 1. B. pseudomallei induces a rapid transcriptional response. (A) Kinetics of C. elegans killing by BpR15. The dotted arrows indicate time points chosen for gene expression analyses. (B) Hierarchically clustered expression profile (rows) for 1,254 genes differentially expressed upon exposure to BpR15 compared with OP50 at 2, 4, 8, and 12 h after exposure. Data from three independent experiments (columns) are shown for each time point. The vertical bars mark clusters of genes induced (Upper) or repressed (Lower).

15068 | www.pnas.org/cgi/doi/10.1073/pnas.1311725110

pathogen detection, activation of immune-related signaling pathways, and candidate immune effector genes, were enriched upon B. pseudomallei infection, which is in concordance with responses to other bacterial pathogens (1, 12, 13). Numerous cellular processes were transcriptionally altered over the course of B. pseudomallei infection: genes involved in metabolic processes, longevity, and stress responses were enriched and host lipid metabolism genes were among the most significantly repressed. Comparing our infection-response gene list with published regulated gene lists for the p38 MAPK (12, 14), DAF-2 insulin-like (3, 15), and SMA/TGF-β pathways (16, 17) revealed significant enrichment of genes regulated by these factors (Fig. S2A) (hypergeometric probability, P < 1 × 10−6). Specifically, expression of 42 of 92 PMK-1–regulated genes was significantly altered during infection [representation factor (Rf) = 6.51]. Of the 440 (class 1 and 2) DAF-16–regulated genes (3), 149 genes were differentially expressed in infected worms (Rf = 4.83). Among the 1,122 genes identified as SMA-6 regulated (17), 144 of these overlapped with the B. pseudomallei infection response (Rf = 1.83). The GATA transcription factor ELT-2 is a positive regulator of intestinal immune responses (1, 18). A significant enrichment (Rf = 8.43, hypergeometric probability P < 1 × 10−6) was observed between ELT-2–regulated genes (19) and B. pseudomallei infection response genes, suggesting that ELT-2 might also contribute to the infection response of B. pseudomallei (Fig. S2A). By contrast, the noncanonical unfolding protein response (ncUPR), an important defense mechanism against bacterial pathogens (20) and represented by the pqn/abu genes, was not significantly induced (Fig. S2B), suggesting that B. pseudomallei infection did not elicit the ncUPR response. B. pseudomallei Down-Regulates ELT-2 Transcriptional Targets. ELT2 transcriptional targets are induced by P. aeruginosa PA14 (1). Unexpectedly, of the 42 ELT-2–regulated genes that were differentially expressed following B. pseudomallei infection, 83% were repressed over the course of infection (Fig. S3). They include genes predicted to encode for proteins involved in bacterial lysis, luminal degradation of macromolecules, detoxification, and stress response (Table S3) (19). When we compared the expression of a subset of ELT-2 transcriptional targets (elo-6, F55G11.2, F57F5.1, K10C2.3, K12H4.7, and spp-8) in worms infected by BpR15 and PA14 using qRT-PCR, an opposite pattern of gene expression was observed (Fig. 2A). ELT-2–regulated genes that were induced by PA14 were consistently suppressed by BpR15. Suppression of ELT-2 transcriptional targets by BpR15 does not appear to be a consequence of a general shutdown in transcription because ELT-2–independent transcripts (pgp-5, ugt-29, thn-1, pqm-1, and skr-3) were similarly induced by PA14 and BpR15 (Fig. 2B). To confirm that the repression of ELT-2 transcriptional targets by B. pseudomallei is associated with inactivation of elt-2, we measured the transcript levels of ELT-2 targets in adult worms following elt-2 knockdown by RNAi and upon BpR15 infection. We observed a more pronounced reduction of ELT-2 transcriptional targets (asp-3, mtl-1, mtl-2, elo-6, F55G11.2) expression in elt-2 RNAi-treated adult animals after BpR15 infection (Fig. 2C). To determine whether ELT-2 is required to protect against B. pseudomallei infection, we compared survival of elt-2 RNAitreated and vector control adult animals following BpR15 infection. Exposure to elt-2 RNAi resulted in a significant reduction in worm survival on BpR15 (Fig. S4A) but did not significantly alter normal life span (Fig. S4B). Thus, elt-2 is required to protect C. elegans from B. pseudomallei infection. B. pseudomallei Affects ELT-2 at the Posttranscriptional Level. The juxtaposition between the requirement of ELT-2 for defense against B. pseudomallei and down-regulation of ELT-2 transcriptional targets in BpR15-infected worms led us to hypothesize that this pathogen could target ELT-2 for immune suppression either at the transcriptional or posttranscriptional level. We ruled out the former by showing that elt-2 transcripts in BpR15-infected worms Lee et al.

Fig. 2. Down-regulation of ELT-2 transcriptional targets by B. pseudomallei. (A and B) Gene expression changes during infection by BpR15 (open columns) and PA14 (filled columns). Shown are fold relative to uninfected animals. Columns represent mean ± SD; n = 2. (A) ELT-2 and (B) non–ELT-2 transcriptional targets. (C) Effects of elt-2 RNAi knockdown on BpR15-infected worms measured by qRT-PCR. Shown are normalized fractions of specific RNA levels in BpR15-infected animals relative to uninfected animals in elt-2 RNAi animals (filled column) and animals fed on empty vector (open columns). The columns represent mean ± SEM; n = 2.

Lee et al.

Utan 320, K96243, Argus4533, and Goat 2124) were capable of causing loss of ELT-2::GFP (Fig. 4 E and F). Additionally, the extent of ELT-2::GFP loss correlated with levels of virulence, with the most marked ELT-2 loss seen in worms infected by the most virulent B. pseudomallei isolate. These results suggest that repression of ELT-2 transcriptional targets is associated with reduction of ELT-2 protein and that targeting ELT-2 is a unique pathogenic strategy of B. pseudomallei.

MICROBIOLOGY

were not significantly reduced at 4 and 12 hpi compared with uninfected (Fig. 3A). We next determined the effects of infection on ELT-2 protein levels in intact worms using a nuclear ELT-2:: GFP quantitation assay. First, we infected JM90 transgenic worms that express an ELT-2::GFP fusion protein under the control of its native promoter (21). Over the course of infection, we compared the intensity and distribution of ELT-2::GFP fluorescence (Fig. 3 B–D) and enumerated the number of fluorescent nuclei in the worm intestine (Fig. 3E). ELT-2::GFP fluorescence remain localized to intestinal nuclei of worms exposed to E. coli OP50 (Fig. 3B) and PA14 (Fig. 3C). In stark contrast, only faint ELT-2::GFP fluorescence could be detected in the intestinal nuclei of BpR15infected animals (Fig. 3D). Whereas the worms exposed to E. coli and PA14 retained the ELT-2::GFP fusion protein in the nucleus throughout infection (Fig. 3E), worms exposed to BpR15 showed a significant reduction in the number of nuclear GFP from 12 to 24 hpi. To demonstrate that the effect does not simply represent changes due to intestinal damage or the dying process due to the more rapid killing by BpR15, the number of nuclear localized GFP was enumerated in worms exposed to PA14 up to the point where 50% of the worm population was killed (Fig. S5A). These worms retained a similar number of nuclear GFP as worms exposed to E. coli over the course of infection (Fig. S5B). Thus, the loss of ELT-2 is specific to B. pseudomallei infection and not a consequence of intestinal damage. To further confirm that the level of ELT-2 protein had diminished over the course of BpR15 infection, we immunoblotted lysates obtained from age-matched JM90 adults exposed to E. coli, PA14, or BpR15, using a monoclonal antibody specific to either GFP or ELT-2 (21). Consistent with the nuclear enumeration assay, worms infected with PA14 for 12 and 18 h had comparable levels of ELT-2::GFP and ELT-2 as that observed in worms exposed to E. coli (Fig. 3F). In contrast, much lower levels of ELT-2::GFP and ELT-2 protein were observed in worms infected with BpR15 after 12 and 18 h (Fig. 3F). A similar observation was obtained in worms infected with BpR15 up to 24 h (Fig. S5C). We also enumerated nuclear ELT-2::GFP in worms infected with other Gram-negative (e.g., Salmonella typhimurium) and Gram-positive (e.g., Enterococcus faecalis and Staphylococcus aureus) bacteria known to cause a comparable killing rate (Fig. 4A). None of these pathogens induced significant loss of ELT-2::GFP over the course of infection (Fig. 4B). Infections by other Burkholderia species—B. thailandensis, B. cepacia, and B. vietnamiensis— also did not trigger the loss of ELT-2::GFP protein (Fig. 4 C and D). By contrast, all other B. pseudomallei isolates tested (Orang

Fig. 3. B. pseudomallei causes the loss of nuclear ELT-2 in the intestine of C. elegans. (A) elt-2 transcripts at 4 and 12 hpi with BpR15 measured by qRTPCR. Shown are normalized fractions of elt-2 RNA levels in infected relative to OP50-exposed animals. Shown are mean ± SD; n = 3. (B–D) Representative fluorescence micrographs of ELT-2::GFP-expressing adult worms with nuclear localized ELT-2::GFP fusion protein after 24-h exposure to (B) OP50, (C) PA14, and (D) BpR15. (E) Mean number of nuclei with detectable ELT-2::GFP signal after 4-, 12-, 18-, and 24-h exposure to OP50 (black columns), PA14 (white columns), and BpR15 (gray columns). *t test, P < 0.001 relative to OP50 at the corresponding time point. The columns represent mean ± SEM; n = 3. (F) Representative immunoblots of ELT-2::GFP protein in OP50-, PA14-, and BpR15-infected worms at 12 and 18 hpi using an ELT-2–specific mAb and anti-GFP antibody. Anti-actin antibody was used as loading control.

PNAS | September 10, 2013 | vol. 110 | no. 37 | 15069

The Loss of Nuclear Protein in the Intestine Is Specific to ELT-2. To test whether the loss of ELT-2 is mediated by B. pseudomallei infection rather than an indirect consequence of general stress, we exposed JM90 to cadmium, a toxic heavy metal known to induce stress responses in C. elegans (22). No loss of nuclear GFP was observed even with a high concentration of cadmium (50 mM) (Fig. S6A), indicating that the loss of nuclear ELT-2:: GFP is a specific response to B. pseudomallei infection and not to general stress. To test whether B. pseudomallei specifically targets nuclear-localized ELT-2, we compared the effect of BpR15 infection on ELT-2::GFP to other nuclear-localized transcription factors: (i) PHA-4, using a transgenic strain AD84 that expresses functional PHA-4::GFP fusion protein (23); and (ii) DAF-16, using a transgenic strain TJ356 that expresses functional DAF-16:: GFP fusion protein (24). The number and intensity of nuclear localized PHA-4::GFP (Fig. S6B) remained constant throughout the BpR15 infection. For DAF-16, which is distributed predominantly in the cytoplasm under normal growth conditions, we first localized DAF-16::GFP fusion protein to the nuclei by heat shock before exposing the worms to either BpR15 or E. coli. In contrast to ELT-2::GFP fluorescence, DAF-16::GFP fluorescence remained nuclear localized and at a constant intensity over the course of a 24-h infection (Fig. S6C). Heat treatment did not significantly affect the quantitation of ELT-2::GFP (Fig. S6D). We conclude that ELT-2::GFP loss was not due to a generalized translational shutdown nor a general loss of GFPfused proteins. Instead, it was specific to ELT-2 and specific to B. pseudomallei infection.

Fig. 4. Loss of ELT-2 in the intestine is B. pseudomallei specific. (A, C, and E) Kinetics of C. elegans killing by (A) other known pathogens, (C) different Burkholderia species, and (E) different B. pseudomallei isolates. (B, D, and F) Mean number of nuclei with detectable ELT-2::GFP signal on worms exposed to (B) other known pathogens for up to 6 d, (D) different Burkholderia species for up to 72 hpi, and (F) different B. pseudomallei isolates at 24 hpi. Shown are mean ± SD; n = 3.

15070 | www.pnas.org/cgi/doi/10.1073/pnas.1311725110

Host UPS Mediates ELT-2 Degradation During B. pseudomallei Infection. The UPS plays a critical role in protein degradation

and is exploited by numerous pathogens to facilitate infection (25). Protein degradation via UPS involves two processes: (i) covalent attachment of ubiquitin (Ub) to target protein and (ii) degradation of ubiquitin-tagged protein by the 26S proteasome. Covalent attachment is a multistep process that involves Ub-activating (E1), Ub-conjugating (E2), and Ub-ligating (E3) enzymes, which selectively interacts with Ub-loaded E2, recruits and binds specific substrates, thus transferring Ub to specific target proteins. Ubtagged proteins are then degraded in the proteasome (Fig. S7). To investigate whether loss of ELT-2 during a B. pseudomallei infection is mediated by the host UPS, we enumerated nuclear ELT-2::GFP in RNAi-mediated knockdown of C. elegans ubq-1, ubq-2, or rpt-2 genes. ubq-1 and ubq-2 encode for C. elegans ubiquitin. RPT-2 is a regulatory proteasome particle and RNAi of rpt-2 has been shown to inhibit proteasome function in live C. elegans (26). As proteasome inactivation is lethal to developing worms, we initiated RNAi only in young adults. As expected, in adult animals exposed to E. coli OP50 and PA14, the number of nuclear ELT-2::GFP retained was indistinguishable between animals treated with ubq-1 or ubq-2 RNAi or control vector (Fig. 5A). By contrast, in BpR15-infected worms, although loss of ELT-2:: GFP was observed in worms treated with control vector, the number of ELT-2::GFP in ubq-1 or ubq-2 RNAi-treated worms was indistinguishable from those exposed to E. coli. The relative importance of ubq-1 vs. ubq-2 is not known as RNAi is expected to inactivate both genes due to their extensive homology. Similar results were obtained with rpt-2 RNAi-treated worms (Fig. 5B). Together, they indicate that UPS is required for the loss of ELT-2 and further supports the idea that ELT-2 is actively degraded rather than translationally inhibited. Specific E3 ligases coordinate ubiquitylation and subsequent degradation of targeted proteins. Given that only ELT-2 is specifically degraded, we hypothesized that this process may involve specific E3 ligase(s) that transfer Ub to ELT-2 and subsequently signal the ELT-2 for degradation by proteasomes. Alternatively, but not mutually exclusive, degradation of ELT-2 could be mediated by B. pseudomallei effectors that mimic host E3 ligase and subvert the normal ubiquitylation events as a pathogenic strategy (27). The C. elegans genome contains several hundred genes predicted to encode for E3 ligase. We explored the possible involvement of host E3 ligases in ubiquitylation and degradation of ELT-2 by looking for E3 ligase encoding genes in the microarray dataset that were induced by BpR15. Only two genes, F54B11.5 and ZK637.14, were significantly up-regulated by BpR15 but not by a nonpathogenic B. pseudomallei isolate (Table S4). We knocked down the expression of these genes individually by RNAi and tested for the loss of ELT-2 upon BpR15 infection. Knockdown of each of these genes was sufficient to rescue the loss of ELT-2::GFP in BpR15-infected animals with the effect being more pronounced for F54B11.5 (Fig. 5C). Knocking down another E3 ligase R1010A.2, of a similar class but whose expression was not significantly altered during BpR15 infection did not rescue the loss of ELT-2::GFP in BpR15-infected animals (Fig. 5C). The induction of F54B11.5 and ZK637.14 E3 ligases by BpR15 suggests a strategy used by B. pseudomallei to promote active degradation of ELT-2 as a means to counter the worm immune response. These experiments do not, however, rule out the possibility that bacterial effectors could also contribute to ELT-2 degradation by mimicking an E3 ligase. Loss of ELT-2 Requires B. pseudomallei T3SS. T3SS is important for virulence in many Gram-negative bacteria (28) and T3SS effector proteins have been implicated in hijacking the eukaryotic UPS (29). To determine whether T3SS is required for the loss of ELT-2 during B. pseudomallei infection, we enumerated nuclear ELT-2::GFP in JM90 worms infected with wild-type B. pseudomallei strains KHW or K96243 and their respective isogenic T3SS mutants in a KHW (bsaM and bspR) (30) or K96243 (bipB) (31) background that completely lack the T3SS. B. pseudomallei Lee et al.

BsaM is homologous to Salmonella PrgH, which together with InvG and PrgK/EscJ forms the multiring base of the needle complex. BspR is the TetR family transcription regulator required for the expression of structural and secretion components of T3SS, whereas BipB is a translocon in the T3SS apparatus. Similar to BpR15, the number of nuclear GFP was significantly reduced in worms exposed to B. pseudomallei strain KHW (Fig. 6A) and K96243 (Fig. 6B) at 24 hpi compared with worms fed with E. coli OP50. In contrast, the number of nuclear GFP retained in worms exposed to mutants bsaM, bspR (Fig. 6A), and bipB was similar to the control (P < 0.0001, Student t test) (Fig. 6B). This indicated a role for the T3SS in the loss of ELT-2 protein during B. pseudomallei infection. Discussion Whole-genome transcriptional analysis revealed that B. pseudomallei-infected C. elegans mount a complex transcriptional response that includes expression of a wide array of defensespecific as well as non–immune-related genes. A unique feature revealed from these studies is the repression of ELT-2–regulated transcripts and the concomitant loss of ELT-2 protein in infected worms. The loss of nuclear protein appears to be specific to ELT-2, requires the host UPS and an intact T3SS of B. pseudomallei. Together, they suggest that B. pseudomallei suppresses host immunity by degrading the ELT-2 protein. Several conserved transcription factors including GATA/ELT-2 (1), FOXO/DAF-16 (15), and bZIP/ZIP-2 (32) regulate the expression of immunity-related genes and protect C. elegans from lethal infection by P. aeruginosa (1, 32). We found that disruption of ELT-2 rendered worms hypersusceptible to B. pseudomallei infection, indicating that ELT-2 also protects C. elegans from a lethal B. pseudomallei infection. Down-regulation of ELT-2– regulated immunity genes, such as spp-8 and F55G11.2, during B. pseudomallei infection suggests that this pathogen is able to repress host defenses by subverting ELT-2. B. pseudomallei suppresses ELT-2 transcriptional targets by affecting ELT-2 protein stability, as evidenced by loss of nuclear-localized ELT2::GFP in the intestine and decrease in ELT-2 protein in B. pseudomallei-infected worm extracts. The progressive loss of ELT-2 protein from the intestinal nuclei is specific to ELT-2 as the PHA-4::GFP and DAF-16::GFP remained stably retained in the intestinal nuclei. The progressive loss of ELT-2 protein also appears to be specific to B. pseudomallei infection because ELT-2 Lee et al.

Fig. 6. B. pseudomallei T3SS is required for the loss of ELT-2. (A and B) Mean number of nuclei with detectable ELT-2::GFP signal after 24-h exposure to B. pseudomallei. (A) KHW and mutants bsaM and bspR; (B) K96243 and mutant bipB. ***t test, P < 0.0001. The columns represent mean ± SEM; n = 3. Each experiment included 17–20 animals per group.

PNAS | September 10, 2013 | vol. 110 | no. 37 | 15071

MICROBIOLOGY

Fig. 5. B. pseudomallei infection triggers ELT-2 degradation by the host ubiquitin–proteosome system. (A–C) Mean number of nuclei with detectable ELT-2::GFP signal of empty vector-treated worms and (A) ubq-1 and ubq-2, (B) rpt-2, and (C) E3 ligases knockdown worms after 24-h exposure to OP50 and BpR15. **t test, P < 0.005; ***t test, P < 0.0001. n = 3.

protein levels remained stable over the course of infection by several other bacteria as well as other Burkholderia species. How do pathogens target host proteins? Pathogenic bacteria could suppress host immune response by inhibiting host translation (33, 34), interfering with ubiquitination of signaling intermediates (35), or disrupting signaling complexes (36). For example, P. aeruginosa infection inhibits mRNA translation in the intestine via the endocytosed translation inhibitor Exotoxin A (33, 34). Other bacterial pathogens interfere with the NF-κB pathway by deubiquitination (37). In this study, we demonstrate that B. pseudomallei subverts the host UPS to specifically degrade the intestinal transcription factor ELT-2, which has a significant role in protecting worms from infection. We showed that disruption of important components in the UPS in worms such as ubq-1, ubq-2, and rpt-2 limited the loss of ELT-2 seen in B. pseudomallei-infected worms. This is in concordance with numerous findings demonstrating that subversion of the UPS is one of the mechanisms used by bacterial pathogens to manipulate host cells (25). Similarly, a recent report also demonstrated that B. pseudomallei actively inhibits NF-қB and type I IFN-related pathway activation by interfering with the ubiquitination of critical signaling intermediates including TNFR-associated factor-3 and factor-6, and IқBα through TssM (a broad-base deubiquitinase) (38). Many bacterial effector proteins act as E3 Ub ligase mimics and can interact with a host Ub-bound E2 enzyme and facilitate ubiquitination of host target. Our study, however, suggests that host E3 ligases are responsible for the observed ubiquitination event. We identified two C. elegans E3 ligases that were up-regulated upon B. pseudomallei infection and knockdown of these E3 ligases limited ELT-2 loss in infected worms. Loss of ELT-2 required an intact B. pseudomallei T3SS. Previous findings indicate that T3SS and T4SS effector proteins are important in facilitating ubiquitination of target host proteins (29). We used a transgenic strain JM90 that overexpresses ELT2::GFP to directly observe degradation of ELT-2 within an intact host. Although this implies that the conclusions are based on overexpression studies, other data, including the down-regulation of ELT-2–regulated transcripts in wild-type worms are consistent with B. pseudomallei using the T3SS to induce specific host E3 ligases to target ELT-2 for destruction by UPS, resulting ultimately in the down-regulation of ELT-2–regulated transcripts and increased host susceptibility. Further work is needed to unravel the identity of B. pseudomallei type 3 effectors and the mechanism by which T3SS functions to subvert the host UPS to degrade ELT-2. Our study also identified genes with putative roles in activation of immunity-related signal transduction pathways, such as the p38 MAPK, DAF-2/DAF-16, and Sma/TGF-β pathways as well as their respective immune effectors. For example, up-regulation of tir-1 and vhp-1, which encode defense proteins TIR-1 (39) and VHP-1, a regulator of the c-Jun N-terminal kinase (JNK) and

epithelial immunity may lead to insights that are generalizable to other hosts of B. pseudomallei, including humans.

p38 MAPKs (40), respectively, implicates the involvement of the p38 MAPK signaling pathway during B. pseudomallei infection. Interestingly, several known DAF-2/DAF-16–regulated immune effectors such as lys-7, spp-1, and thn-2 are significantly repressed upon B. pseudomallei infection similar to that previously seen in a P. aeruginosa infection (7). P. aeruginosa accomplished this by inducing INS-7, a ligand of the DAF-2/DAF-16 pathway that through a series of phosphorylation events ultimately leads to ejection of DAF-16 from the intestinal nuclei and down-regulation of immune effectors (4, 7). By contrast, B. pseudomallei retains DAF-16 in the nuclei of intestinal cells over the course of infection, indicating an immune suppression mechanism that is distinct from P. aeruginosa (Fig. S6C). Thus, gene expression and protein localization studies showed that B. pseudomallei and P. aeruginosa use distinct immune suppression mechanisms. Further work is in progress to decipher the underlying mechanism that contributes to the down-regulation of DAF-16 transcriptional targets while retaining nuclear localized DAF-16. In summary, we have identified a mechanism by which B. pseudomallei suppresses host immunity. B. pseudomallei targets ELT-2, a component of a multipathogen defense pathway (1), by actively degrading the ELT-2 protein leading to down-regulation of ELT-2 transcriptional targets and suppression of host immunity. This process is mediated by host UPS and requires the bacterial T3SS. The conserved role for GATA factors in

ACKNOWLEDGMENTS. We thank Yee-Chin Wong for her technical assistance and the Institute for Medical Research, Kuala Lumpur, for provision of the B. pseudomallei isolates. We also thank Prof. James McGhee (University of Calgary) for providing anti–ELT-2 antibodies, Prof. Andrew Dillin (University of California, Berkeley) for providing PHA-4::GFP strain, and Prof. Sunee Korbsrisate (Mahidol University) for providing the bipB mutant. This study was supported by the Malaysia Genome Institute–Stanford University International Research Grant awarded to S.N. and M.-W.T. by the Government of Malaysia. S.-H.L. was supported by a National Science Fellowship provided by the Ministry of Science, Technology and Innovation, Malaysia.

1. Shapira M, et al. (2006) A conserved role for a GATA transcription factor in regulating epithelial innate immune responses. Proc Natl Acad Sci USA 103(38):14086–14091. 2. Senger K, Harris K, Levine M (2006) GATA factors participate in tissue-specific immune responses in Drosophila larvae. Proc Natl Acad Sci USA 103(43):15957–15962. 3. Murphy CT, et al. (2003) Genes that act downstream of DAF-16 to influence the lifespan of Caenorhabditis elegans. Nature 424(6946):277–283. 4. Kawli T, Tan MW (2008) Neuroendocrine signals modulate the innate immunity of Caenorhabditis elegans through insulin signaling. Nat Immunol 9(12):1415–1424. 5. Becker T, et al. (2010) FOXO-dependent regulation of innate immune homeostasis. Nature 463(7279):369–373. 6. Jensen VL, Simonsen KT, Lee YH, Park D, Riddle DL (2010) RNAi screen of DAF-16/ FOXO target genes in C. elegans links pathogenesis and dauer formation. PLoS One 5(12):e15902. 7. Evans EA, Kawli T, Tan MW (2008) Pseudomonas aeruginosa suppresses host immunity by activating the DAF-2 insulin-like signaling pathway in Caenorhabditis elegans. PLoS Pathog 4(10):e1000175. 8. Wiersinga WJ, van der Poll T, White NJ, Day NP, Peacock SJ (2006) Melioidosis: Insights into the pathogenicity of Burkholderia pseudomallei. Nat Rev Microbiol 4(4):272–282. 9. Lee SH, Ooi SK, Mahadi NM, Tan MW, Nathan S (2011) Complete killing of Caenorhabditis elegans by Burkholderia pseudomallei is dependent on prolonged direct association with the viable pathogen. PLoS One 6(3):e16707. 10. Reinke V, et al. (2000) A global profile of germline gene expression in C. elegans. Mol Cell 6(3):605–616. 11. Tusher VG, Tibshirani R, Chu G (2001) Significance analysis of microarrays applied to the ionizing radiation response. Proc Natl Acad Sci USA 98(9):5116–5121. 12. Troemel ER, et al. (2006) p38 MAPK regulates expression of immune response genes and contributes to longevity in C. elegans. PLoS Genet 2(11):e183. 13. Wong D, Bazopoulou D, Pujol N, Tavernarakis N, Ewbank JJ (2007) Genome-wide investigation reveals pathogen-specific and shared signatures in the response of Caenorhabditis elegans to infection. Genome Biol 8(9):R194. 14. Kim DH, et al. (2002) A conserved p38 MAP kinase pathway in Caenorhabditis elegans innate immunity. Science 297(5581):623–626. 15. Garsin DA, et al. (2003) Long-lived C. elegans daf-2 mutants are resistant to bacterial pathogens. Science 300(5627):1921. 16. Mallo GV, et al. (2002) Inducible antibacterial defense system in C. elegans. Curr Biol 12(14):1209–1214. 17. Roberts AF, Gumienny TL, Gleason RJ, Wang H, Padgett RW (2010) Regulation of genes affecting body size and innate immunity by the DBL-1/BMP-like pathway in Caenorhabditis elegans. BMC Dev Biol 10:61. 18. McGhee JD, et al. (2009) ELT-2 is the predominant transcription factor controlling differentiation and function of the C. elegans intestine, from embryo to adult. Dev Biol 327(2):551–565. 19. McGhee JD, et al. (2007) The ELT-2 GATA-factor and the global regulation of transcription in the C. elegans intestine. Dev Biol 302(2):627–645. 20. Sun J, Singh V, Kajino-Sakamoto R, Aballay A (2011) Neuronal GPCR controls innate immunity by regulating noncanonical unfolded protein response genes. Science 332(6030):729–732. 21. Fukushige T, Hawkins MG, McGhee JD (1998) The GATA-factor elt-2 is essential for formation of the Caenorhabditis elegans intestine. Dev Biol 198(2):286–302. 22. Cui Y, McBride SJ, Boyd WA, Alper S, Freedman JH (2007) Toxicogenomic analysis of Caenorhabditis elegans reveals novel genes and pathways involved in the resistance to cadmium toxicity. Genome Biol 8(6):R122.

23. Panowski SH, Wolff S, Aguilaniu H, Durieux J, Dillin A (2007) PHA-4/Foxa mediates diet-restriction-induced longevity of C. elegans. Nature 447(7144):550–555. 24. Henderson ST, Johnson TE (2001) daf-16 integrates developmental and environmental inputs to mediate aging in the nematode Caenorhabditis elegans. Curr Biol 11(24):1975–1980. 25. Collins CA, Brown EJ (2010) Cytosol as battleground: Ubiquitin as a weapon for both host and pathogen. Trends Cell Biol 20(4):205–213. 26. Hoppe T, et al. (2004) Regulation of the myosin-directed chaperone UNC-45 by a novel E3/E4-multiubiquitylation complex in C. elegans. Cell 118(3):337–349. 27. Hicks SW, Galán JE (2010) Hijacking the host ubiquitin pathway: Structural strategies of bacterial E3 ubiquitin ligases. Curr Opin Microbiol 13(1):41–46. 28. Coburn B, Sekirov I, Finlay BB (2007) Type III secretion systems and disease. Clin Microbiol Rev 20(4):535–549. 29. Janjusevic R, Abramovitch RB, Martin GB, Stebbins CE (2006) A bacterial inhibitor of host programmed cell death defenses is an E3 ubiquitin ligase. Science 311(5758): 222–226. 30. Sun GW, et al. (2010) Identification of a regulatory cascade controlling Type III Secretion System 3 gene expression in Burkholderia pseudomallei. Mol Microbiol 76(3): 677–689. 31. Suparak S, et al. (2005) Multinucleated giant cell formation and apoptosis in infected host cells is mediated by Burkholderia pseudomallei type III secretion protein BipB. J Bacteriol 187(18):6556–6560. 32. Estes KA, Dunbar TL, Powell JR, Ausubel FM, Troemel ER (2010) bZIP transcription factor zip-2 mediates an early response to Pseudomonas aeruginosa infection in Caenorhabditis elegans. Proc Natl Acad Sci USA 107(5):2153–2158. 33. Dunbar TL, Yan Z, Balla KM, Smelkinson MG, Troemel ER (2012) C. elegans detects pathogen-induced translational inhibition to activate immune signaling. Cell Host Microbe 11(4):375–386. 34. McEwan DL, Kirienko NV, Ausubel FM (2012) Host translational inhibition by Pseudomonas aeruginosa Exotoxin A triggers an immune response in Caenorhabditis elegans. Cell Host Microbe 11(4):364–374. 35. Angot A, Vergunst A, Genin S, Peeters N (2007) Exploitation of eukaryotic ubiquitin signaling pathways by effectors translocated by bacterial type III and type IV secretion systems. PLoS Pathog 3(1):e3. 36. Cirl C, et al. (2008) Subversion of Toll-like receptor signaling by a unique family of bacterial Toll/interleukin-1 receptor domain-containing proteins. Nat Med 14(4): 399–406. 37. Shames SR, Auweter SD, Finlay BB (2009) Co-evolution and exploitation of host cell signaling pathways by bacterial pathogens. Int J Biochem Cell Biol 41(2):380–389. 38. Tan KS, et al. (2010) Suppression of host innate immune response by Burkholderia pseudomallei through the virulence factor TssM. J Immunol 184(9):5160–5171. 39. Shivers RP, Kooistra T, Chu SW, Pagano DJ, Kim DH (2009) Tissue-specific activities of an immune signaling module regulate physiological responses to pathogenic and nutritional bacteria in C. elegans. Cell Host Microbe 6(4):321–330. 40. Kim DH, et al. (2004) Integration of Caenorhabditis elegans MAPK pathways mediating immunity and stress resistance by MEK-1 MAPK kinase and VHP-1 MAPK phosphatase. Proc Natl Acad Sci USA 101(30):10990–10994. 41. Shapira M, Tan MW (2008) Genetic analysis of Caenorhabditis elegans innate immunity. Methods Mol Biol 415:429–442.

15072 | www.pnas.org/cgi/doi/10.1073/pnas.1311725110

Materials and Methods All experimental procedures are detailed in SI Materials and Methods. RNA extraction, microarray experiments, and qRT-PCR were performed as described (1). C. elegans RNAi knockdown and survival assays were performed as previously described (41). Nuclear localized ELT-2 was assayed by enumerating total visible ELT-2::GFP nuclei from 1-d-old individual worm strain (JM90) exposed to various pathogenic bacteria, bacterial mutants (bsaM, bspR, and bipB), or heavy metal cadmium under 400× total magnification using an upright fluorescent microscope. For immunoblot analyses, ELT-2– specific mouse monoclonal antibody (a gift from James McGhee at the University of Calgary, Calgary, AB, Canada), anti-GFP antibody (Calbiochem, EMD Biosciences), and anti-actin antibody (Sigma-Aldrich) were used.

Lee et al.