|Home | About | Journals | Submit | Contact Us | Français|
The ability of Yersinia pestis to forestall the mammalian innate immune response is a fundamental aspect of plague pathogenesis. In this study, we examined the effect of Ail, a 17-kDa outer membrane protein that protects Y. pestis against complement-mediated lysis, on bubonic plague pathogenesis in mice and rats. The Y. pestis ail mutant was attenuated for virulence in both rodent models. The attenuation was greater in rats than in mice, which correlates with the ability of normal rat serum, but not mouse serum, to kill ail-negative Y. pestis in vitro. Intradermal infection with the ail mutant resulted in an atypical, subacute form of bubonic plague associated with extensive recruitment of polymorphonuclear leukocytes (PMN or neutrophils) to the site of infection in the draining lymph node and the formation of large purulent abscesses that contained the bacteria. Systemic spread and mortality were greatly attenuated, however, and a productive adaptive immune response was generated after high-dose challenge, as evidenced by high serum antibody levels against Y. pestis F1 antigen. The Y. pestis Ail protein is an important bubonic plague virulence factor that inhibits the innate immune response, in particular the recruitment of a protective PMN response to the infected lymph node.
Three Yersinia species are pathogenic for humans and other animals. Yersinia enterocolitica and Yersinia pseudotuberculosis are not closely related phylogenetically but are both food-borne and waterborne enteric pathogens. Yersinia pestis and Y. pseudotuberculosis are very closely related but have different pathogenic lifestyles. Y. pestis is the agent of plague and is transmitted by fleas. Critical to the pathogenicity of all three species is the Yersinia virulence plasmid, which encodes a type III secretion system (T3SS) and secreted Yop effector proteins that act to prevent phagocytosis and other protective innate immune responses. Other virulence factors of Y. pestis are distinct from those of the enteropathogenic Yersinia species (and vice versa), in keeping with their different modes of transmission and routes of infection in the mammal.
One conserved factor that may contribute in different ways to virulence of the three pathogenic Yersinia species is Ail, an outer membrane protein of ~17 kDa. Ail was first discovered and named (attachment-invasion locus) in Y. enterocolitica (23). In that species, Ail has been shown to mediate adhesion to and invasion of epithelial cells in vitro (23) and to confer resistance to killing by serum complement (4, 5, 27). The fact that the ail locus is present in enteropathogenic strains and absent in nonpathogenic strains of Y. enterocolitica suggests that Ail plays a role in Y. enterocolitica pathogenesis, although Ail was not required for normal virulence in a mouse model (24, 39). Ail is also required for serum resistance in Y. pseudotuberculosis but shows less adherence and invasion properties than its counterpart in Y. enterocolitica (43). The role of Ail in Y. pseudotuberculosis virulence has not been directly assessed.
Y. pestis Ail and its role in plague pathogenesis have been the subject of recent investigations. As in the enteropathogenic Yersinia species, Y. pestis Ail mediates serum resistance and adherence to mammalian cells in vitro (1, 18), and Ail-mediated adhesion facilitates T3SS-dependent injection of Yop effector proteins (12, 13, 36). Deletion of ail resulted in greatly attenuated Y. pestis virulence in a rat model of pneumonic plague but in only a modestly increased time to death in a mouse model of pneumonic plague (17), which correlates with the sensitivity of Y. pestis ail mutants to killing by normal rat serum but not mouse serum (1, 17). The importance of Ail to bubonic plague pathogenesis has so far been tested only by direct intravenous injection of mice with attenuated Y. pestis strains, and conflicting results have been obtained (1, 12, 13). In this study, we report the effect of the mutational loss of ail on bubonic plague pathogenesis in mice and rats, using a fully virulent, wild-type Y. pestis strain background.
The fully virulent, wild-type Y. pestis CO92 strain was used in this study (26). We also used an isogenic mutant in which the ail gene (YPO2905; annotated as ompX [y1324] in the Y. pestis KIM strain) was deleted and replaced with a kanamycin resistance gene generated by the pKOBEG-sacB lambda red-based mutagenesis method (9). A PCR fragment composed of the Tn903-derived aminoglycoside phosphotransferase (aph) kanamycin resistance gene flanked by 46- and 54-bp sequences present in the upstream and downstream regions of ail was generated using oligonucloetide primers ail1F (5′-CTGTACGAATATCCATATTTTTCATGTGTCAGATATTTGTTAATATTTGGCTGGGGGGAAAGCCACGTTGTGTC-3′) and ail1R (5′-GTTAAAAAATCGTCTATGAGCCAGAAGCAGCCCGGTATTCATTGGTCTGAGGTCTGCCTCGTGAAGAA-3′) and the aph gene from pUC4K as the template. This DNA fragment was purified and introduced into Y. pestis CO92(pKOBEG-sacB) by electroporation. Kanamycin-resistant transformants were isolated, and pKOBEG-sacB-cured derivatives of them were obtained by selection on LB agar plates containing 7.5% sucrose. DNA sequencing confirmed that the entire ail gene, 57 bp of the upstream sequence, and 212 bp of the downstream sequence had been deleted and replaced by aph.
The ail mutant was complemented by chromosomal insertion of the wild-type ail gene into the glmS-pstS intergenic region using a Tn7-based method (6). The wild-type ail gene and upstream (440-bp) and downstream (248-bp) flanking regions were PCR amplified from CO92 DNA using the ail2 primer set (5′-GATCCTCGAGGGCTGTCACCGTCCTGG-3′ and 5′-CTCACTAGTCGTCTATGAGCCAGAAGCAGC-3′) and ligated into pUC18R6K-mini-Tn7T-Gm (kindly provided by Herbert Schweizer, Colorado State University) that had been linearized by restriction enzymes XhoI and SpeI. The resulting recombinant plasmid was used to transform Escherichia coli DH5α λpir by electroporation. The purified recombinant pUC18R6K-mini-Tn7T-Gm plasmid and the pTNS2 helper plasmid (also provided by H. Schweizer) were introduced into the Y. pestis CO92 Δail::aph mutant by electroporation. A complemented mutant clone with the wild-type ail gene integrated at the glmS locus was selected and its gentamicin resistance cassette was excised as described previously (6). DNA sequencing confirmed that a wild-type copy of ail had been inserted.
Bacterial resistance to the complement in fresh serum that we collected from Brown Norway rats and RML Swiss Webster mice and in human serum from PAA Laboratories was determined as described previously (1). Heart infusion broth containing 2.5 mM CaCl2 was inoculated with frozen stocks of the Y. pestis strains and incubated at 28°C for 18 h with aeration at 200 rpm. Fresh media were inoculated with a 1:100 dilution of the primary culture and incubated under the same conditions. Bacteria in 1 ml of the second culture were collected by centrifugation, washed twice, and resuspended in 1 ml of phosphate-buffered saline (PBS), pH 7.4. Bacterial aggregates were disrupted by treating the suspensions for 15 s in a FastPrep FP120 homogenizer using lysing matrix H (Qbiogene, Carlsbad, CA); the suspensions were then diluted in PBS to A620 = 0.2. Fifty microliters of diluted bacterial suspension was added to paired 1.5-ml tubes containing 200 μl of serum, one of which had been previously incubated at 56°C for 20 min to heat inactivate the complement. Tubes were incubated at 37°C for 2 h in an Eppendorf Thermomixer R heat block rotating at 1,000 rpm; appropriate dilutions were plated onto tryptose agar containing 5% defibrinated sheep blood and 1 μg/ml irgasan (blood agar), and the plates were incubated for 48 h at 28°C. Percent survival was calculated by dividing the average number of CFU from the normal serum samples by the average number of CFU from the heat-inactivated serum samples.
Groups of 10 Brown Norway rats (Rattus norvegicus) and RML Swiss Webster mice (all 9- to 12-week-old females) were injected intradermally in the right thigh with 25 μl of PBS containing 30 to 30,000 Y. pestis CFU. Inocula were prepared from overnight Luria broth cultures incubated at 21°C without aeration. Bacteria were quantified by using a Petroff-Hausser bacterial counting chamber and diluted in PBS. The number of Y. pestis CFU injected was verified by CFU counts of each inoculum preparation. Animals were observed at least three times daily for 3 weeks and euthanized upon signs of terminal plague (moribund state as evidenced by lethargy, hunched posture, and reluctance to respond to external stimuli). Spleen and blood were collected after euthanization; spleens were weighed, placed in a lysing matrix H tube with 1 ml PBS, and homogenized briefly using a Mini-Beadbeater (BioSpec, Bartlesville, OK). Bacterial load in the spleen and blood was determined by CFU counts on blood agar plates. Fifty-percent lethal dose (LD50) values were determined by logistic regression (JMP statistical discovery software version 9.0; SAS Institute, Inc., Cary, NC) and by probit analysis (41). All animal experiments were done in compliance with the Animal Care and Use Committee of the NIAID/NIH.
Dissected right inguinal lymph nodes were fixed in 10% neutral buffered formalin, embedded in paraffin, processed, and 4-μm-thick sections were stained with hematoxylin and eosin. Polymorphonuclear leukocytes (PMNs) and Y. pestis in other lymph node sections were differentially stained by immunohistochemistry (IHC) as described previously (7), using anti-myeloperoxidase antibody AF3667 (R&D Systems, Minneapolis, MN) and the ChromoMap DAB detection kit (Ventana Medical Systems) followed by anti-Y. pestis antibody (30) and the RedMap detection kit (Ventana Medical Systems). Sections were counterstained with hematoxylin. Species-specific nonimmune IgG was used for negative staining controls.
Wells of Immulon 2HB (Thermo Scientific) 96-well enzyme-linked immunosorbent assay (ELISA) plates were coated with 0.1 μg of Y. pestis F1 (Caf1) protein in 200 mM NaCl-20 mM Tris (pH 7.5), washed with PBS-0.05% Tween (PBS-T), blocked with PBS-T containing 5% dry nonfat milk, and rewashed. Serial dilutions of serum samples (100 μl) were added, and the plates were incubated at room temperature for 3 h. Wells were then washed six times, and a 1:10,000 dilution of horseradish peroxidase-conjugated goat anti-rat secondary antibody (Pierce) was added. Following 3 h of incubation, wells were washed six times and developing reagent (TMB-ELISA substrate; Pierce) was added. After 15 min, 100 μl of 2 M sulfuric acid was added to stop the reaction, and the A450 was measured. Results from two independent experiments done in triplicate were analyzed by one-way analysis of variance (ANOVA) with Dunnet's posttest to compare naïve control and infected mouse serum samples (GraphPad Prism 5 software). Triplicate wells containing no antigen, no primary antibody, or no secondary antibody were included as negative controls; anti-Y. pestis polyclonal serum (30) was used for a positive control.
The Y. pestis CO92 ail mutant and complemented mutant strains were first tested for serum sensitivity as an additional means to verify their genotype. As expected based on previous studies (1, 17, 18), the mutant was sensitive to normal human and rat serum but not to mouse serum; and normal resistance to human serum was restored in the complemented mutant (Fig. 1).
Virulence of the Y. pestis CO92 ail mutant was assessed first in a Brown Norway rat model of bubonic plague (30). Intradermal injection of challenge doses of up to 3 × 104 CFU of the ail mutant produced no mortality and no signs of systemic plague morbidity (Fig. 2A). Plate counts of triturated spleens collected following euthanasia of all surviving rats at 17 to 20 days after infection with the ail mutant were all negative except for one rat infected with 3 × 104 CFU, whose spleen contained 3 × 103 CFU/g. In contrast, spleens of rats infected with 50 CFU of wild-type Y. pestis CO92 contained 107 to 1010 CFU/g when the rats were euthanized upon signs of terminal disease at 2 to 5 days after infection. These results indicate that Ail is an essential virulence factor in the rat model of bubonic plague.
Although none of the rats infected with the Y. pestis CO92 ail mutant showed any of the typical signs of systemic plague (lethargy, hunched posture, reluctance to move) and remained alert and active, very large buboes (enlarged lymph nodes) developed during the second week after infection in several of the rats injected with 3,000 or 30,000 CFU (Fig. 3). Upon dissection, the inguinal buboes often contained large amounts of pus, which we have never observed in buboes produced by wild-type Y. pestis (7, 30). The range of lymph node histopathology due to infection with the Y. pestis ail mutant is shown in Fig. 4. At its most extreme, the lymph node parenchyma was destroyed and completely replaced by intact and degraded polymorphonuclear leukocytes (PMNs) that surrounded fields of bacteria (Fig. 4A to C). Other lymph nodes contained large abscesses encapsulating smaller numbers of PMNs and bacteria (Fig. 4D and E), and many of the bacteria were associated with intact PMNs (Fig. 4F). Evidence of a chronic inflammatory response was seen in other lymph nodes, which contained large, apparently sterile and resolving abscesses (Fig. 4G and H) associated with large numbers of fibroblasts and macrophages and rare multinucleated giant cells in addition to PMNs (Fig. 4I). These lymph nodes also contained many pronounced germinal centers filled with immunoglobulin-producing plasma cells (Fig. 4G). In contrast, buboes produced by wild-type Y. pestis showed massive accumulation of extracellular bacteria within 3 to 5 days that was not associated with a robust PMN response (Fig. 4J and K), as described previously (7, 30). Another marked difference was that hemorrhage, typical of bubonic plague lymphadenopathy in the rat (7, 30), was seen only in lymph nodes infected with wild-type Y. pestis and not in any of the lymph nodes infected with the ail mutant (Fig. 4).
We also assessed the role of Ail in virulence in a mouse model of bubonic plague. The Y. pestis CO92 ail mutant was significantly attenuated for virulence in RML Swiss Webster mice but less so than in rats (Fig. 2). The estimated LD50 of the ail mutant for mice was 8.3 × 103 CFU (95% confidence interval of 1.9 × 103 to 1.5 × 105 CFU) by logistic regression analysis and 7.8 × 103 CFU (95% confidence interval of 1.5 × 103 to 3.9 × 104 CFU) by probit analysis, approximately 1,000-fold higher than the LD50 of the wild-type CO92 strain. The LD50 of the ail mutant for rats was not determined, but no mortality was seen even with the highest challenge dose of 3 × 104 CFU (Fig. 2A). Complementation of the ail mutant by chromosomal integration of a single copy of the wild-type ail gene restored full virulence, indicating that loss of Ail was solely responsible for the attenuated virulence of the ail mutant. The lymph node histopathology in mice infected with the ail mutant was similar to that seen in rats, with enlargement and abscess formation that was sometimes grossly purulent (data not shown).
Anti-Y. pestis antibody was quantified in the serum of five rats 17 to 25 days after infection with 30, 300, 3,000, or 30,000 CFU of the Y. pestis ail mutant (Fig. 5). Nine of the 10 rats that received the two highest challenge doses had very high titers against Y. pestis F1 antigen, indicative of a protective immune response.
A primary underlying cause of the fulminant, extreme virulence of plague is the lack of the normal protective innate immune response to infection (7, 15, 20). The rapid recruitment of phagocytic cells, especially PMNs, which are able to ingest and kill invading microorganisms, is an important aspect of innate immunity. Our results indicate that a major in vivo function of the Y. pestis outer membrane protein Ail is to prevent PMN recruitment to the lymph node during the progression of bubonic plague. The purulent, resolving lymphadenopathy produced by the ail mutant in highly susceptible laboratory mice and rats (Fig. 4) is identical to clinical descriptions of devolving buboes of humans that spontaneously recovered from bubonic plague in the preantibiotic era, and suppuration of the bubo was recognized as a favorable sign (28, 32). Purulent buboes are also seen commonly in relatively resistant wild rodents with a form of disease that was termed subacute plague (22). Thus, the ability of Y. pestis to prevent PMN recruitment correlates with disease progression and mortality and appears to be an essential aspect of the anti-innate immunity pathogenesis strategy.
Two other Y. pestis virulence factors have previously been found to have a role in moderating the PMN response to infection. The best-studied antiphagocytic virulence factor of Y. pestis is the virulence plasmid-encoded type III secretion system (T3SS) and Yop cytotoxic effector proteins that are secreted intracellularly into host cells upon close contact. Normally, Y. pestis disseminates to the draining lymph node and multiplies extensively to produce the bubonic stage of plague without stimulating a strong PMN response (Fig. 4J and K) (7, 30). Because of the lack of a normal innate immune response in general, and a PMN response in particular, the infection is not contained in the lymph node but quickly spreads systemically to produce fatal septicemic plague. The Y. pestis virulence plasmid (pYV or pCD1) is required for sustained bacterial growth in the lymph node and for limiting PMN recruitment; infection with pYV-negative Y. pestis results in a large influx of PMNs and little or no bacterial replication in the draining lymph node and other tissues (7, 37, 38).
A second Y. pestis virulence factor that contributes to a limited PMN response is the plasminogen activator (Pla), a cell surface protease. Like our ail mutant, Y. pestis pla mutants are highly attenuated in bubonic plague models: they can disseminate from a subcutaneous injection site to the draining lymph node but are successfully contained and eliminated there (31, 33, 42). Increased PMN infiltration and abscess formation have been noted in lymph node, skin, spleen, and liver infected with Pla−, but not Pla+, Y. pestis strains (31, 33). How Pla dampens the normal PMN response has not been determined but appears to be mediated by plasminogen or fibrin, because an increased PMN response is seen following infection with either Pla+ or Pla− Y. pestis in plasminogen-deficient (Plg−/−) mice (8).
The ail mutant used in this study contained the virulence plasmid, pla, and all other virulence factors, differing from the wild-type CO92 parent strain only in the absence of the ail gene. Therefore, our results indicate that the Y. pestis T3SS and pla are necessary but not sufficient to prevent an acute inflammatory cell response. Ail is also essential for this. In fact, loss of ail resulted in a much greater infiltration of PMNs to the lymph node (Fig. 4) than did loss of pYV or pla (7, 31), although this may be due to the much longer persistence of the ail mutant in the lymph node. The Y. pestis ail mutant was greatly attenuated for virulence (mortality), even though it caused an atypical, nonfatal form of bubonic plague at high challenge doses. Notably, the lymphadenitis produced by the ail mutant was characterized by a very robust PMN response, purulent abscess formation, and containment in the lymph node without causing the septicemic form of the disease. It is also likely that any ail-negative bacteria that escape the lymph node are effectively eliminated as well, because Y. pestis KIM5 ail mutants are attenuated for virulence in mice when injected intravenously (12, 13).
We hypothesize that the ability of Ail to prevent a strong PMN response in vivo is related to its ability to protect against serum complement (Fig. 1) (1, 17, 18). The complement system is a key part of innate immunity and an important first line of defense against infection based on its three major effector mechanisms: opsonization, induction of inflammation, and bacteriolysis (10). In the absence of a specific antibody, opsonization occurs via C3b fragments that are generated and covalently bound to the surface of invading microbes by the alternative pathway and which are recognized by receptors CR1 and CR3 on PMNs and macrophages. An acute inflammatory response is induced via the anaphylotoxins C3a and C5a that are released during complement activation and which have potent vasodilation and PMN chemotactic effects. Complement-mediated lysis depends on the formation of the C5b-C9 membrane attack complex built up on the microbial surface.
Gram-positive bacteria are not susceptible to lysis, and the chemotactic inflammatory and opsonophagocytic properties are considered to be the most important antibacterial effector mechanisms of the complement system (34). For example, although mice are not more vulnerable to Gram-negative infection than other animals, it has been known for decades that mouse complement has little or no bactericidal activity. This has been attributed to a nonfunctional C5 convertase, the enzyme complex that normally generates C5a and C5b, the first component of the membrane attack complex (11). In this regard, it is interesting that the Y. pestis ail mutant was attenuated more in the rat than in the mouse model of bubonic plague following intradermal injection (Fig. 2). Furthermore, in rat and mouse models of pneumonic plague, a Y. pestis CO92 ail mutant was greatly attenuated following intranasal infection of rats, but in mice only a delayed time to terminal symptoms was seen (17). A more pronounced exudative response to Y. pestis infection occurs in the lung than in the lymph node, suggesting that complement-mediated lytic activity may be more important in pneumonic (and septicemic) plague than in bubonic plague.
Our results suggest that Y. pestis Ail inhibits the proinflammatory as well as the bacteriolytic defense mechanisms of the complement. Given the very strong PMN response to infection with the ail mutant in both our rat and mouse bubonic plague models, we speculate that Ail interferes with the complement cascade before the pivotal step of C3 (and subsequently C5) activation, preventing the generation of the anaphylotoxins C3a and C5a. Such a mechanism remains to be demonstrated; however, complement evasion mechanisms are a common theme of microbial pathogenesis. Several bacteria have evolved resistance mechanisms that involve outer membrane proteins that directly bind and inhibit, or degrade, complement components or that bind to one of the host's circulating endogenous negative regulators of complement activation (19). Notably, both Y. pestis and Y. enterocolitica can bind the classical pathway regulator C4b-binding protein, and this is partly mediated by Ail in Y. enterocolitica (16, 25). Y. enterocolitica Ail also binds factor H, a regulator of the alternative pathway (2). Interestingly, mutational changes to Ail that decreased serum resistance of Y. enterocolitica did not affect factor H binding (3), suggesting that different domains of Ail may interact with different complement components or complement regulators to separately interfere with opsonic, proinflammatory, or lytic effector mechanisms.
An alternative mechanism is suggested by the fact that, in addition to its role in complement resistance, Ail is able to mediate the bacterial binding to host cells required for T3SS delivery of cytotoxic Yop proteins in vitro (13, 36). The Yersinia T3SS targets PMNs in vivo (21) and inhibits PMN phagocytosis and chemotaxis in vitro (14, 29, 35, 40). If loss of Ail-mediated binding to target cells sufficiently reduces T3 secretion in vivo, increased phagocytosis by PMNs may stimulate the production and release of chemotactic mediators of PMN recruitment, such as the chemokines interleukin 8 (IL-8) (CXCL8) and macrophage inflammatory protein 2 (MIP-2) (CXCL2). Thus, the adhesin function of Ail may also be important for suppression of the cellular inflammatory response and full virulence.
As a surface-exposed, essential virulence factor, Ail is an attractive target for medical countermeasures against plague, including use as a vaccine component. Human serum, like rat serum, is bactericidal to ail-negative Y. pestis, suggesting that Ail would be required for pathogenesis of human plague. In addition, because laboratory animals succumb so rapidly to plague, convalescent-phase serum samples are infrequently obtained. Convalescent human serum is also a rare commodity. The Y. pestis ail mutant, in contrast to pYV- and pla-negative attenuated strains which are cleared rapidly, is able to produce a persistent infection in the lymph node. The ability to reliably reproduce subacute plague and to generate high titers of convalescent-phase serum in rats following infection with ail-negative Y. pestis (Fig. 5) may be useful in studies to identify new correlates of immune protection against Y. pestis infection.
This work was supported by the Division of Intramural Research, NIAID, NIH, and by a Trans-NIH/FDA Intramural Biodefense Program grant (to S.K.B.).
We thank Dan Long and Rebecca Rosenke, RML Veterinary Pathology Section, for help with histology, Jeff Skinner, Bioinformatics and Computational Biosciences Branch, NIAID, for help with statistical analysis, and Herbert Schweizer, Colorado State University, and Elisabeth Carniel and Anne Derbise, Institut Pasteur, for providing plasmids used for mutagenesis and complementation. We also acknowledge Scott Kobayashi, Jeff Shannon, and Justin Spinner for their review of the manuscript.
Published ahead of print on 3 October 2011.