|Home | About | Journals | Submit | Contact Us | Français|
Human polymorphonuclear leukocytes (PMNs, or neutrophils) are the most abundant innate immune cell and kill most invading bacteria through combined activities of reactive oxygen species (ROS) and antimicrobial granule constituents. Pathogens such as Yersinia pestis resist destruction by the innate immune system and are able to survive in macrophages and neutrophils. The specific molecular mechanisms used by Y. pestis to survive following phagocytosis by human PMNs are incompletely defined. To gain insight into factors that govern Y. pestis intracellular survival in neutrophils, we inactivated 25 two-component gene regulatory systems (TCSs) with known or inferred function and assessed susceptibility of these mutant strains to human PMN granule extracts. Y. pestis strains deficient for PhoPQ, KdpED, CheY, CvgSY, and CpxRA TCSs were selected for further analysis, and all five strains were altered for survival following interaction with PMNs. Of these five strains, only Y. pestis ΔphoPQ demonstrated global sensitivity to a panel of seven individual neutrophil antimicrobial peptides and serine proteases. Notably, Y. pestis ΔphoPQ was deficient for intracellular survival in PMNs. Iterative analysis with Y. pestis strains lacking the PhoP-regulated genes ugd and pmrK indicated that the mechanism most likely responsible for increased resistance to killing is 4-amino-4-deoxy-l-arabinose modification of lipid A. Together, the data provide new information about Y. pestis evasion of the innate immune system.
Human polymorphonuclear leukocytes (PMNs, or neutrophils) are essential components of the host innate immune system and are required for protection against invading bacterial pathogens. PMNs are recruited rapidly to sites of infection to remove bacteria through a process known as phagocytosis. Phagocytosis of microbial pathogens elicits production of reactive oxygen species (ROS) and release of antimicrobicidal granule constituents into pathogen-containing phagocytic vacuoles. Although the majority of microorganisms are readily killed by PMNs, some bacterial pathogens subvert the human innate immune system to cause disease in the immunocompetent host.
Yersinia pestis is a highly virulent human pathogen and the causative agent of plague. Plague is characterized by rapid systemic dissemination and progression to fulminant disease. The success of Y. pestis as a human pathogen is based, in part, on virulence factors that enable subversion of the innate immune response. Y. pestis effector Yop proteins target innate immune cells in vivo (31) and interfere with the complex signal transduction pathways that govern phagocyte function (50). The ability of Y. pestis to inhibit phagocytosis in macrophages and neutrophils (3, 7) is well characterized, and extracellular replication is a prominent feature of plague pathogenesis (5). Notwithstanding, several studies have demonstrated clearly that Y. pestis survives (4, 7) and replicates (40) within macrophage phagolysosomal compartments and that survival occurs independent of the type three secretion system (TTSS) (48). In contrast, intracellular survival of Y. pestis following PMN phagocytosis is suggested (20, 21, 47) but remains less well characterized. Further, specific mechanisms used by Y. pestis to survive following PMN phagocytosis are incompletely defined. Although PMNs and macrophages are important cells of the innate immune system and possess common functional features, they have distinct roles in the immune response. Whereas macrophages are important for mediating chronic inflammation, neutrophils are essential to initiation and execution of the acute inflammatory response and are the first-line cellular defense against bacterial pathogens (35).
The ability of bacterial pathogens to survive within host phagocytes is dependent on activity of complex transcriptional regulatory networks, including two-component signal transduction systems (TCSs). For example, the importance of PhoP for survival of Salmonella enterica serovar Typhimurium in macrophages has been characterized in detail (16, 32) and was followed by similar reports for Shigella flexneri (34), Y. pestis (36), and Yersinia pseudotuberculosis (14). The Y. pestis genome (strain CO92) contains elements predicted to encode approximately 29 putative TCSs, including four pseudogenes (30, 49). Although the function of many Y. pestis TCSs is undefined, several are homologous to stress response systems in Escherichia coli (57). To gain insight into mechanisms used by Y. pestis to avoid destruction by human neutrophils, we created 25 Y. pestis mutant strains with deletions of putative TCSs ([ΔTCS] excluding pseudogenes) and tested the strains for susceptibility to PMN primary granule extracts. We selected Y. pestis mutant strains susceptible to increased killing by granule constituents and assessed survival following PMN phagocytosis. The initial screening assay indicated that PhoPQ, KdpED, CheY, CvgSY, and CpxRA TCSs are involved in Y. pestis resistance to neutrophil bactericidal activity. Further, our results show that PhoP is important for Y. pestis survival in human neutrophils and that the mechanism most likely responsible for increased resistance to killing is 4-amino-4-deoxy-l-arabinose (4-aminoarabinose) modification of lipid A.
Neutrophils were isolated from heparinized venous blood of healthy individuals by use of standard dextran sedimentation and Ficoll-Paque Plus (Amersham Biosciences, Piscataway, NJ) gradient centrifugation as described previously (25). All studies involving human PMNs were performed in accordance with a protocol approved by the Institutional Review Board for Human Subjects, University of Idaho. Purity of PMN preparations and cell viability were assessed by flow cytometry (FACSCalibur; BD Biosciences, San Jose, CA) and isolated fractions routinely contained ~98% granulocytes.
Bacterial strains and plasmids used in this study are listed in Table Table1.1. Y. pestis strains were routinely grown overnight in brain heart infusion (BHI) medium (Becton Dickinson and Co., Sparks, MD) at 21°C. For each assay, the overnight cultures were diluted 20-fold in BHI supplemented with 2.5 mM CaCl2, grown at either 21 or 37°C with shaking (225 rpm), and harvested at mid-exponential growth phase (optical density at 600 nm [OD600] of 0.6). Y. pestis strains were verified to contain pCD1 by both selective growth on Congo red-magnesium oxalate agar (44) and PCR. Prior to each assay, bacteria were opsonized with 20% autologous human serum at 37°C for 30 min with shaking.
Y. pestis mutant strains (Table (Table1)1) were created in strain KIM5 by use of a PCR-based deletion strategy (8) with modifications. Gene deletion constructs were created by PCR amplification of the pKD4 kanamycin resistance cassette using primers containing 36-nucleotide (nt) 5′ extensions homologous to the open reading frame (ORF) of the gene of interest (Table (Table2)2) and were electroporated into competent Y. pestis cells (55). Double homologous recombination between the kanamycin gene deletion cassette and the target gene of interest was facilitated by lambda Red recombinase. Although the recombinase-encoding plasmid pKD46 is a temperature-sensitive replicon (8), we modified this plasmid by insertion of the levansucrase gene (sacB) to increase efficiency of screening for plasmid loss by negative selection (13). Briefly, the sacB cassette was PCR amplified from pMS20 (26) with primers containing 5′ BstXI restriction sites (Table (Table3),3), restriction enzyme digested, and ligated into the unique BstXI of pKD46 to create pJSO6. Plasmid loss was promoted by overnight growth at 39°C in Luria-Bertani (LB) broth and was followed by selection on LB agar plates containing 5% sucrose and kanamycin (50 μg/ml). Y. pestis-targeted gene deletions were verified by PCR (Table (Table3),3), and the kanamycin cassette was excised by a pCP20-encoded FLP recombinase. Y. pestis deletion strains were cured of pCP20 by overnight growth at 39°C in LB broth. For plasmid expression of Y. pestis genes (phoPQ, kdpED, cheY, cvgSY, and cpxRA), the native promoter and ORF were cloned into vector pCR2.1 TOPO (Invitrogen) for use in complementation experiments (Table (Table4),4), and expression was confirmed by real-time reverse transcription-PCR (RT-PCR) as described previously (51) (Table (Table55).
Subcellular fractionation of neutrophils was performed as described previously (37). Briefly, neutrophils were suspended in relaxation buffer (100 mM KCl, 3 mM NaCl, 3.5 mM MgCl2, 10 mM PIPES [piperazine-N,N′-bis(2-ethanesulfonic acid], and 1 mM ATP) and disrupted by nitrogen cavitation (400 lb/in2 for 30 min at 4°C). PMN lysates were collected into 100 mM EGTA (1.25 mM final concentration) and centrifuged at 200 × g for 6 min to remove remaining unbroken cells and nuclei. Neutrophil lysates were overlaid on Percoll step gradients (24) and centrifuged at 48,300 × g for 15 min at 4°C. Neutrophil granule fractions were collected as described previously (24), and residual Percoll was removed by ultracentrifugation at 100,000 × g for 90 min. Granule proteins were acid extracted by treatment with glycine buffer (pH 2.0) for 30 min at 25°C (37). Final protein concentration was determined by a bicinchoninic acid (BCA) protein assay (Thermo Scientific, Rockford, IL), and single-use aliquots of granule fractions were stored at −80°C. For experiments, granule proteins were diluted in RPMI 1640 medium (Invitrogen) supplemented with 5 mM HEPES (Sigma) (RPMI 1640/H medium) to 500 μg/ml and then added to an equal volume of bacteria (108/ml) in RPMI 1640/H medium. The suspension was rotated end over end at 37°C for 60 min. Bacteria were diluted in sterile saline, plated on LB agar, and incubated at 28°C for 2 days, and the output CFU were enumerated. Bacteria percent survival was calculated using the following equation: (number of CFU+granule/number of CFUcontrol) × 100, where CFU+granule indicates the CFU of the bacteria and granule suspension.
Neutrophil phagocytosis and bactericidal activity assays were performed as described previously (47) with modifications. To measure neutrophil bactericidal activity, PMNs (106) and Y. pestis strains (107 CFU) were combined in RPMI 1640/H medium in 96-well plates and incubated at 37°C for the times indicated in Fig. Fig.2,2, ,4,4, and and6.6. Alternatively, intracellular survival was assessed by addition of 250 μg/ml gentamicin (Sigma) at 15 min following phagocytosis to remove extracellular bacteria. Viable bacteria were recovered by use of a previously published protocol (40). Briefly, PMNs were lysed with 0.1% Triton X-100 (EMD Chemicals, Gibbstown, NJ) in phosphate-buffered saline (PBS) for 10 min on ice, and bacteria were dispersed by sonication with a microprobe for 3 pulses of 1 s each (150D microtip; Branson, Danbury, CT). Bacteria were diluted in sterile saline and plated on LB agar, and CFU were enumerated following 2 to 3 days of growth at 28°C. The percentage of surviving bacteria was calculated using the following equation: (number of CFU+PMN/number of CFUt=0 control) × 100, where CFU+PMN indicates the CFU of PMN-treated bacteria and CFUt=0 control indicates the CFU of the control at time zero.
Y. pestis strains were cultured to mid-exponential growth phase and suspended in RPMI/H medium to 108 CFU/ml. Peptides and proteins were reconstituted in 0.01% acetic acid unless indicated otherwise and diluted in RPMI 1640/H medium for use in assays: human neutrophil defensin 1 and 2 (Sigma), 100 μg/ml; neutrophil elastase (EMD Biosciences, San Diego, CA), 100 μg/ml; LL-37 (Phoenix Pharmaceuticals, Burlingame, CA), 25 μg/ml; neutrophil cathepsin G (EMD Biosciences), 500 mU/ml; human proteinase 3 (EMD Biosciences; 150 mM NaCl, 50 mM NaO acetate, and 0.1% Triton X-100 [pH 6.0]), 50 μg/ml; and human neutrophil azurocidin (Sigma), 50 μg/ml. Peptide or enzyme dilution buffer (control) was added to each Y. pestis strain, and samples were incubated at 37°C for 60 min. Samples were plated on LB agar and enumerated following 2 to 3 days of incubation at 28°C. Bacteria percent survival was calculated using the following equation: (CFU+peptide/CFUcontrol) × 100, where CFU+peptide indicates the CFU of peptide-treated bacteria.
Y. pestis strains (1.0 × 107 CFU) were cultured to early exponential phase of growth, incubated with 250 μg/ml granule proteins in RPMI 1640/H medium, and rotated end over end for 30 or 60 min at 37°C. RNA was extracted as described previously (42), and ugd and pmrK transcript levels were determined with TaqMan real-time RT-PCR (ABI 7000 thermocycler; Applied Biosystems, Foster City, CA) with primers and probes listed in Table Table5.5. Relative quantities of mRNA were normalized to the expression of the housekeeping reference gene proS, as previously described (42). Samples were assayed in triplicate from two biological replicates, and results are expressed as the relative transcript level of ugd and pmrK in Y. pestis KIM5 compared to expression in Y. pestis KIM5 ΔphoPQ.
Statistics were performed with a paired Student's t test or one-way analysis of variance (ANOVA) with Dunnett's or Tukey's posttests using GraphPad Prism, version 5.00 for Windows (GraphPad Software, San Diego, CA).
We reported recently that ~30% of Y. pestis survive at times up to 240 min following neutrophil phagocytosis and that intracellular survival was independent of TTSS inhibition of ROS production (47). Y. pestis is also reported to survive in guinea pig neutrophils (20, 21) and survive at low levels following exposure to rabbit neutrophil exudates (7). Based on these findings, we hypothesized that Y. pestis intracellular survival in human neutrophils requires resistance to granule-derived antimicrobial agents. Granule proteins are estimated to occupy ~40% of the PMN vacuolar volume (17) and achieve concentrations of up to 500 mg/ml (43). As a first step toward identifying factors that promote Y. pestis evasion of neutrophil oxygen-independent killing, we tested susceptibility of complete open reading frame deletion mutants for 25 known or putative TCSs identified in the annotated genome of Y. pestis strain KIM (9) to PMN primary granule extracts (Fig. (Fig.1A).1A). The neutrophil primary granule fraction contains a diversity of antimicrobial peptides and proteolytic enzymes such as α-defensins, elastase, cathepsin G, bactericidal permeability increasing protein, and lysozyme (10). Compared with Y. pestis strain KIM5, nine ΔTCS mutant strains showed increased susceptibility to 250 μg/ml solubilized α-granule constituents following a 60-min incubation at 37°C (P < 0.05 versus KIM5, by ANOVA with Dunnett's posttest). Y. pestis ΔTCS ΔphoPQ, ΔkdpED, ΔcheY, ΔcvgSY, and ΔcpxRA mutant strains were the most susceptible to bactericidal activity of the PMN α-granule fraction (e.g., survival at 60 min was 42.0% ± 4.4%, 50.1% ± 4.5%, 50.1% ± 4.4%, 53.0% ± 5.0%, and 56.5% ± 6.7%, respectively) and were chosen for further analysis. Complementation of Y. pestis ΔphoPQ, ΔkdpED, ΔcheY, ΔcvgSY, and ΔcpxRA strains with the respective Y. pestis KIM5 wild-type genes restored resistance to PMN α-granules (Fig. (Fig.1B).1B). These results suggest that multiple Y. pestis TCSs facilitate survival within human neutrophils, and this hypothesis is tested below.
To determine if increased sensitivity of Y. pestis ΔphoPQ, ΔkdpED, ΔcheY, ΔcvgSY, and ΔcpxRA strains to neutrophil α-granule extracts extends to increased susceptibility to neutrophil killing, PMN bactericidal activity was assessed following interaction with Y. pestis ΔTCS mutant strains (Fig. (Fig.2).2). Following a 105-min incubation with human PMNs, Y. pestis ΔphoPQ, ΔkdpED, ΔcheY, ΔcvgSY, and ΔcpxRA strains were killed efficiently compared to the Y. pestis KIM5 parent strain. In vitro growth characteristics were similar between Y. pestis ΔTCS mutant strains and KIM5 (data not shown), and there were no detectable differences in PMN phagocytosis (Fig. (Fig.2,2, inset). Taken together, these findings suggest that Y. pestis PhoP, KdpE, CheY, CvgS, and CpxR gene regulatory systems promote survival in human neutrophils.
To determine if Y. pestis PhoP, KdpE, CheY, CvgS, and CpxR gene regulatory systems provide resistance to specific neutrophil granule constituents, we tested susceptibility of the Y. pestis ΔTCS mutant strains to individual CAMPs and serine proteases that comprise, in part, PMN α-granules (Fig. (Fig.3).3). Survival of Y. pestis KIM5 and ΔTCS mutant strains was assessed following a 60-min incubation with human α-defensin 1 (Fig. (Fig.3A,3A, HNP-1) and α-defensin 2 (Fig. (Fig.3B,3B, HNP-2), neutrophil elastase (Fig. (Fig.3C),3C), azurocidin (Fig. (Fig.3D),3D), cathepsin G (Fig. (Fig.3E),3E), neutrophil proteinase 3 (Fig. (Fig.3F),3F), and the PMN secondary granule peptide cathelicidin LL-37 (Fig. (Fig.3G).3G). Consistent with the increased sensitivity of Y. pestis ΔphoPQ, ΔkdpED, ΔcheY, ΔcvgSY, and ΔcpxRA strains to human PMN α-granule extracts (Fig. (Fig.1B)1B) and PMN killing (Fig. (Fig.2),2), ΔTCS strains tested showed increased susceptibility to azurocidin, cathepsin G, and LL-37 (P < 0.0001 versus KIM5). However, only ΔphoPQ demonstrated increased susceptibility to all neutrophil granule constituents tested (Fig. 3A to G) (P < 0.0001 versus KIM5). Further, Y. pestis ΔphoPQ showed increased sensitivity to HNP-1, HNP-2, elastase, cathepsin G, proteinase 2, and LL-37 compared to all ΔTCS strains tested (Fig. 3A to C and E to G) (P < 0.0001 versus KIM5, ΔkdpED, ΔcheY, ΔcvgSY, and ΔcpxRA, by ANOVA with Tukey's posttest). These observations are consistent with previous reports demonstrating that Y. pestis PhoP regulates, in part, resistance to the CAMP polymyxin B (19, 41, 54) and is important for survival in macrophages (15, 36).
Inasmuch as Y. pestis ΔphoPQ showed global sensitivity to PMN α-granules, CAMPs, and serine proteases and decreased survival following interactions with human neutrophils, we assessed next the role of PhoP in intracellular survival. PMN bactericidal activity was measured following synchronized phagocytosis of Y. pestis KIM5 and ΔphoPQ strains (Fig. (Fig.4A).4A). Although Y. pestis KIM5 and ΔphoPQ were rapidly killed within 45 min following phagocytosis, ΔphoPQ was killed to a higher extent at all times tested (Fig. (Fig.4A).4A). For example, survival of ΔphoPQ at 45 min was 17.6% ± 8.3% compared to 32.1% ± 3.3% for KIM5 (P < 0.001 for the ΔphoPQ strain versus KIM5). However, the PMN bactericidal activity assay measures total viable (ingested and uningested) Y. pestis (Fig. (Fig.2,2, inset) (e.g., phagocytosis was 70.0% ± 7.2% and 69.7% ± 13.6% for ΔphoPQ and KIM5, respectively) and does not measure directly intracellular survival. Thus, to determine Y. pestis survival in human PMNs, extracellular bacteria were removed by treatment with gentamicin following 15 min of phagocytosis (Fig. (Fig.4B).4B). Consistent with our previous report (47), survival of Y. pestis KIM5 was reduced in human PMNs at early times following phagocytosis (Fig. (Fig.4B)4B) (e.g., survival was 54.5% ± 13.2% at 45 min) and remained relatively constant at later time points (e.g., survival at 165 and 285 min was 32.7% ± 10.0% and 26.8% ± 7.1%, respectively). By comparison, ΔphoPQ was eliminated more effectively by human neutrophils, and survival was reduced to 5.7% ± 5.0% at 285 min following ingestion (P < 0.005 for the ΔphoPQ strain versus KIM5 at all times tested, by a paired Student's t test). Taken together, these observations indicate that Y. pestis survival following phagocytosis by human PMNs is dependent, at least in part, on resistance to antimicrobial peptides and proteases.
Modification of the lipid A portion of lipopolysaccharide is involved in resistance to antimicrobial peptides in several pathogenic Gram-negative bacteria (39). In both S. enterica serovar Typhimurium and Y. pestis, this property is conferred in part through covalent modification of lipid A phosphate residues with 4-aminoarabinose. Although important divergent regulatory mechanisms exist between genera, 4-aminoarabinose metabolism is controlled by enzymes encoded by ugd and within the pbgP operon (including pmrK) (38). PhoP is a global regulator of gene expression that governs a diversity of transcripts in Y. pestis, including those required for 4-aminoarabinose biosynthesis (15, 41, 54, 56). As a first step toward determining if 4-aminoarabinose biosynthesis is involved in Y. pestis resistance to PMN killing, we confirmed that transcript levels of ugd and pmrK were decreased in the Y. pestis KIM5 ΔphoPQ mutant strain (Fig. (Fig.5A).5A). Second, we tested the susceptibility of ugd (Δugd) and pmrK (ΔpmrK) deletion mutant strains to PMN α-granule extracts, HNP-1, elastase, azurocidin, cathepsin G, and LL-37 (Fig. 5B to G). Y. pestis Δugd and ΔpmrK strains showed increased susceptibility to PMN α-granule extracts following a 60-min incubation compared to Y. pestis KIM5 (Fig. (Fig.5B)5B) (e.g., survival of Δugd and ΔpmrK was 54.7% ± 4.6% and 66.5% ± 4.6%, respectively, versus 94.3% ± 3.9% for KIM5; P < 0.0001, by ANOVA with Dunnett's posttest). Consistent with the finding that Y. pestis ΔphoPQ showed increased susceptibility to PMN α-granule extracts (Fig. (Fig.1B)1B) and antimicrobial peptides and proteases (Fig. 3A to G), Δugd and ΔpmrK strains were susceptible to all microbicidal constituents tested (Fig. 5B to G) (P < 0.001 for Δugd and ΔpmrK versus KIM5 for all comparisons, by ANOVA with Dunnett's posttest). To determine if PhoP regulation of ugd and pmrK was the primary mechanism of Y. pestis resistance to PMN α-granules, we performed epistasis tests on ΔphoPQ, Δugd ΔpmrK, and ΔphoPQ Δugd ΔpmrK strains (Fig. (Fig.5H).5H). We found that there were minor differences in susceptibility of ΔphoPQ Δugd ΔpmrK and Δugd ΔpmrK strains to α-granule killing (Fig. (Fig.5H)5H) (e.g., 45.08% ± 6.72% versus 51.94% ± 5.41%, respectively), suggesting that there may be additional PhoP-regulated targets that facilitate Y. pestis survival.
We next tested susceptibility of the Y. pestis strains to neutrophil killing (Fig. (Fig.6).6). PMN bactericidal activity was measured following synchronized phagocytosis of Y. pestis KIM5 strains (Fig. (Fig.6)6) by use of assays identical for determining survival of ΔphoPQ (Fig. (Fig.4).4). Consistent with increased susceptibility to antimicrobial peptides and proteases (Fig. 5A to F), Y. pestis Δugd and ΔpmrK strains were killed more efficiently by human PMNs than the KIM5 parent strain. For example, intracellular survival of Y. pestis Δugd and ΔpmrK strains at 285 min following PMN phagocytosis was 11.0% ± 4.4% and 12.0% ± 3.3%, respectively, compared to 27.5% ± 8.8% for KIM5 (P < 0.0001, by ANOVA with Dunnett's posttest). These findings are consistent with a previous report which identified both ugd and pmrK as PhoP-regulated genes that promote Y. pestis survival in macrophages (15). Taken together, these findings indicate that PhoP-dependent regulation of genes required for 4-aminoarabinose modification of Y. pestis lipopolysaccharide (LPS) is important for survival in human PMNs.
Neutrophils are the first-line cellular defense against invading bacterial pathogens and are essential to maintenance of host health. The importance of PMNs in maintenance of human health is highlighted by increased susceptibility to severe bacterial infections in patients with either qualitative or quantitative neutrophil defects (27). Notwithstanding, bacterial pathogens such as Y. pestis subvert neutrophil killing to cause disease in otherwise healthy individuals. Y. pestis inhibits PMN phagocytosis through activity of TTSS effector Yop proteins (47), thus facilitating extracellular replication during plague pathogenesis (29). However, the fate of intracellular Y. pestis following phagocytosis by human PMNs is less clear. As a step toward understanding the ability of Y. pestis to survive in human PMNs, we screened 25 Y. pestis mutant strains with deletions of putative TCSs (excluding pseudogenes) for increased susceptibility to PMN granule constituents and killing (Fig. (Fig.1).1). Herein, we demonstrate that Y. pestis resistance to PMN primary granule extracts and individual neutrophil microbicides (Fig. (Fig.3)3) is influenced by key two-component gene regulatory systems that contribute to intracellular survival following phagocytosis by human neutrophils (Fig. (Fig.22).
In a previous study we showed that intracellular Y. pestis survives at times up to 4 h following phagocytosis and that survival is independent of TTSS inhibition of PMN ROS production (47). The finding that Y. pestis survives within PMNs is somewhat unexpected, given that mechanisms used by microbial pathogens to subvert macrophage killing often fail in neutrophils based on the enhanced microbicidal potential of PMNs (1). Consistent with that notion, Y. pestis is more susceptible to extracts from rabbit neutrophils than to similar homogenates from macrophages (7). Notwithstanding, the majority of studies on Y. pestis interaction with PMNs indicate that although Y. pestis survival is initially reduced, killing is incomplete (4, 7, 20, 21, 22, 28, 47), and Y. pestis is relatively resistant to CAMPs (2, 12). Taken together, these findings suggest that Y. pestis survival in PMNs requires resistance to neutrophil granule-derived microbicidal agents. In support of this hypothesis, our data provide evidence that the Y. pestis PhoPQ, KdpED, CheY, CvgSY, and CpxRA global gene regulatory systems are important for resistance to PMN primary granule extracts (Fig. (Fig.1).1). Y. pestis TCSs with increased sensitivity to PMN granule fractions were subsequently more susceptible to neutrophil killing (Fig. (Fig.2).2). Of the five Y. pestis TCSs identified in the screening assay, PhoPQ remains the only well-described Y. pestis TCS and is predicted to confer resistance to CAMPs (15, 36, 41). Of the TCSs conferring enhanced PMN survival, only PhoPQ provided universal resistance to the granule constituents tested (Fig. (Fig.3).3). This finding extends previous reports that PhoPQ is important for Y. pestis resistance to polymyxin B (18, 41, 54).
The outer membrane of Y. pestis contains the major surface determinant lipopolysaccharide that consists of a conserved oligosaccharide moiety linked to lipid A. The phosphate residues of lipid A have a net negative charge that functions as a target for cationic antimicrobial peptides and proteases (46). In Y. pestis, PhoPQ signaling regulates expression of both the ugd gene and the pbgP operon to facilitate 4-aminoarabinose biosynthesis and subsequent modification of lipid A phosphate residues. The 4-aminoarabinose modification decreases the net negative charge of the bacterial outer membrane and provides resistance to cationic antimicrobial agents (54, 56). Consistent with this mechanism of resistance, Y. pestis strains deficient for Ugd and PmrK showed increased sensitivity to PMN primary granule fractions, CAMPs, and serine proteinases (Fig. (Fig.5)5) and were important for survival in neutrophils (Fig. (Fig.6).6). These findings are consistent with a recent report identifying the importance of PhoP-regulated genes ugd and pmrK for Y. pestis survival in macrophages (15). PhoP plays an important role in survival in macrophages and pathogenesis of Y. pestis (15, 36), Y. pseudotuberculosis (14), and S. enterica serovar Typhimurium (11, 32). Taken together, these studies indicate that modification of lipid A may represent a common strategy employed by Gram-negative bacterial pathogens to evade PMNs in addition to macrophages to facilitate pathogenesis.
The role of the other four TCSs identified by our PMN survival screen is less clear. Both CvgS and CpxR have been identified as virulence factors in the related Yersinia enteropathogens. CvgSY was identified as a novel virulence factor in Y. pseudotuberculosis (23). CpxRA functions in maintenance of the bacterial envelope and promotes intracellular survival of Yersinia enterocolitica through htrA transcription regulation (18, 45). HtrA encodes a protease that degrades misfolded periplasmic proteins. Likewise, Carlsson et al. showed that CpxRA modulates type III secretion and host cell cytotoxicity (6). The KdpED system has not been studied in Yersinia but has been shown to regulate turgor pressure in other enteric bacteria (52). Finally, our results showing that CheY confers intracellular survival in PMNs is the most surprising. CheY is well studied in Salmonella and E. coli regarding its role in motility signal transduction via flagellar rotation (53). Given that Y. pestis is nonmotile and that loss of motility is thought to enhance innate immune evasion by loss of flagellin expression (33), the role of CheY is enigmatic. However, we were able to detect cheY transcripts in KIM5 using real-time RT-PCR (data not shown).
In summary, this study demonstrates that two-component regulatory systems in Y. pestis contribute to bacterial survival during human PMN interactions by increasing resistance to PMN granule-derived antimicrobial agents. Y. pestis intracellular fate in the human neutrophils is dependent, in part, on PhoP-regulated surface modification of lipid A phosphate residues with 4-aminoarabinose. It is likely that there are additional factors involved in Y. pestis resistance to PMN killing, as evidenced by the finding that four other TCSs contributed to increased survival. Thus, additional studies are required to elucidate resistance mechanisms conferred by these Y. pestis global regulatory systems. An enhanced understanding of the ability of Y. pestis to evade the innate immune system will be important for developing novel therapeutic strategies.
We thank B. Joseph Hinnebusch for kindly providing Y. pestis strain KIM5.
Funding for this research was provided by NIH awards 1K22AI6274 and P20RR015587 to S.D.K. and U54AI5714 and P20RR016454 to S.A.M.
Editor: J. B. Bliska
Published ahead of print on 23 November 2009.