A Protective Epitope in Type III Effector YopE Is a Major CD8 T Cell Antigen during Primary Infection with Yersinia pseudotuberculosis

Abstract

Virulence in human-pathogenic Yersinia species is associated with a plasmid-encoded type III secretion system that translocates a set of Yop effector proteins into host cells. One effector, YopE, functions as a Rho GTPase-activating protein (GAP). In addition to acting as a virulence factor, YopE can function as a protective antigen. C57BL/6 mice infected with attenuated Yersinia pestis generate a dominant H2-K^b-restricted CD8 T cell response to an epitope in the N-terminal domain of YopE (YopE_69-77), and intranasal vaccination with the YopE_69-77 peptide and the mucosal adjuvant cholera toxin (CT) elicits CD8 T cells that are protective against lethal pulmonary challenge with Y. pestis. Because YopE_69-77 is conserved in many Yersinia strains, we sought to determine if YopE is a protective antigen for Yersinia pseudotuberculosis and if primary infection with this enteric pathogen elicits a CD8 T cell response to this epitope. Intranasal immunization with the YopE_69-77 peptide and CT elicited a CD8 T cell response that was protective against lethal intragastric Y. pseudotuberculosis challenge. The YopE_69-77 epitope was a major antigen (∼30% of splenic CD8 T cells were specific for this peptide at the peak of the response) during primary infection with Y. pseudotuberculosis, as shown by flow cytometry tetramer staining. Results of infections with Y. pseudotuberculosis expressing catalytically inactive YopE demonstrated that GAP activity is dispensable for a CD8 T cell response to YopE_69-77. Determining the features of YopE that are important for this response will lead to a better understanding of how protective CD8 T cell immunity is generated against Yersinia and other pathogens with type III secretion systems.

INTRODUCTION

Virulence in many Gram-negative bacterial pathogens depends on specialized protein export pathways known as type III secretion systems (T3SSs) (21, 22, 33, 62). T3SSs are activated upon contact of bacteria with host cells and function to deliver effector proteins into or across the eukaryotic plasma membrane (13, 25, 34, 37, 46). T3SS protein effectors act to coopt or subvert numerous key eukaryotic cell biological processes to allow for pathogen adherence, invasion, and/or immune evasion (5, 14, 17, 24). In addition to promoting virulence, there is accumulating evidence that the mammalian host can detect the presence of T3SS effector proteins to generate protective innate and adaptive immune responses (15, 31, 64). Understanding how T3SS effector proteins contribute to both virulence and host protection can provide critical insights into bacterial pathogenesis as well as allow for the development of effective vaccines.

A plasmid-encoded T3SS has been shown to be essential for virulence in the following three Yersinia species that are pathogenic to humans: Y. enterocolitica, Y. pestis and Y. pseudotuberculosis (18). Y. pestis causes bubonic, septicemic, and pneumonic plague, while the other two Yersinia species are enteropathogens that commonly cause terminal ileitis and mesenteric adenitis in humans (39, 63). Y. pestis and Y. pseudotuberculosis are closely related at the genetic level, while Y. enterocolitica is genetically more distant (1, 63). The Yersinia effector proteins that are secreted by the plasmid-encoded T3SS are in general highly similar at the amino acid level among all pathogenic Yersinia spp. and are called Yops (19, 59). After translocation into the cytosol of eukaryotic cells, the seven known Yop effectors (YopH, YopM, YopO/YpkA, YopJ/YopP, YopK, YopT, and YopE), commandeer a number of different eukaryotic signaling pathways to counteract innate and adaptive immune responses of the host (16, 19, 59).

YopE is a 219-amino-acid effector that functions as a GTPase-activating protein (GAP) (2, 11, 60). The protein is comprised of an N-terminal secretion-translocation domain (residues 1 to 78) and a C-terminal GAP domain (residues 100 to 219). The eukaryotic host cell targets of YopE include the small GTPases RhoA, Rac1, and Rac2 (6, 11, 49, 60). YopE GAP activity inhibits phagocytosis and reactive oxygen species generation in host cells (11, 49). In addition, YopE GAP activity functions within the eukaryotic cytoplasm to regulate the amount of Yops that are expressed in Yersinia and translocated into host cells (3, 36). In mice infected with Y. enterocolitica expressing a YopE–beta-lactamase fusion protein, the reporter is translocated preferentially into B cells, macrophages, dendritic cells, and neutrophils (28), suggesting that these cell types are the major targets of YopE in vivo. YopE is an important Yersinia virulence factor in mice as measured by 50% lethal dose (LD₅₀) analysis of yopE mutants (4, 51). Yersinia yopE mutants are less virulent in part because they are defective for colonization of murine lymphatic tissues, including mesenteric lymph nodes (MLN) and spleen (11, 32, 49, 52, 57).

Individual virulence factors or live attenuated strains have been used as immunogens in mice to determine the basis of adaptive immune resistance to pathogenic Yersinia spp. (9, 10, 20, 48, 55). These studies showed that humoral and cellular immunity contribute to the adaptive immune response against Yersinia (9, 10, 20, 48, 55). For example, CD8 T cells have been shown to be important for resistance of immunized mice to Yersinia infection (8, 10, 31, 38, 40, 53). Evidence suggests that CD8 T cells mediate protection against Yersinia by production of cytokines (e.g., tumor necrosis factor alpha [TNF-α] and gamma interferon [IFN-γ]) (7, 53) and by killing bacteria-associated host cells to promote internalization by neighboring phagocytes (10).

Yersinia spp. have been utilized as experimental live vaccines in mouse infection models to prime protective immunity against virulent yersiniae and other pathogens (9, 42, 48, 58). For example, it has been shown that mice that survive infection with wild-type (wt) or attenuated Y. pseudotuberculosis are well protected against challenge with virulent Y. pestis (12, 47, 54, 66). In addition, the plasmid-encoded T3SS has been utilized for delivery of heterologous CD4 or CD8 antigens by live attenuated Y. pseudotuberculosis and Y. enterocolitica carrier strains (30, 43–45, 56, 58, 61). These carrier vaccine strains expressed YopE fused to model antigens such as listeriolysin O (LLO) or ovalbumin (OVA); after translocation into the host antigen-presenting cell (APC) cytosol, the antigens could be processed by the proteasome after polyubiquitination, then transported to the endoplasmic reticulum (ER) to be further trimmed, loaded onto major histocompatibility complex class I (MHC-I) molecules, and finally presented on the plasma membrane to CD8 T cells. The role of Yop effectors in modulating the CD8 T cell response to these model antigens in mice infected with Yersinia carrier vaccines has been examined (43, 44, 56, 58). YopP inhibited the CD8 T cell response to YopE-LLO in mice orally infected with Y. enterocolitica, an activity that could be correlated with the killing of APCs by this effector (56). In contrast, YopE was important for a maximal CD8 T cell response to Y. enterocolitica expressing YopE-LLO or YopE-OVA, most likely because this effector was required for bacterial colonization of deep murine lymphatic tissues such as spleen (30, 58, 61).

In addition to functioning as an efficient delivery vehicle for model antigens, YopE itself has been identified as a target of CD8 T cells during infection with pathogenic Yersinia spp. (23, 31). Results of studies in which Lewis rats were infected with Y. pseudotuberculosis suggested that CD8 T cells recognize an epitope in the N-terminal 128 amino acids of YopE (23). Recently, Lin et al. demonstrated that C57BL/6 mice vaccinated with attenuated Y. pestis and challenged with virulent Y. pestis generate a dominant H2-K^b-restricted CD8 T cell response against the epitope YopE_69-77 (within the N-terminal domain of YopE) (31). Experiments in which C57BL/6 mice were vaccinated with a YopE_69-77 peptide and challenged with virulent Y. pestis demonstrated that the CD8 T cell response to this epitope is protective against pulmonary infection (31). The amino acid sequences of YopE in most strains of Y. pestis and Y. pseudotuberculosis are identical. In addition, the YopE_69-77 epitope is present in several Y. enterocolitica strains. Thus, it is predicted that YopE_69-77 will function as a protective CD8 antigen in C57BL/6 mice infected with many different pathogenic Yersinia strains (31). In this study, we sought to determine if YopE_69-77 is a protective antigen for Y. pseudotuberculosis and if C57BL/6 mice generate a CD8 T cell response to this epitope during primary infection with this pathogen. In addition, the role of YopE GAP activity in the murine CD8 T cell response to YopE_69-77 was investigated.

MATERIALS AND METHODS

Mice and immunizations with the YopE_69-77 peptide.

Female C57BL/6J mice (Jackson Laboratory) were used for all experiments. Animals that were 9 to 10 weeks old were anesthetized with a mixture of ketamine (60 mg/kg of body weight) and xylazine (10 mg/kg) delivered by intraperitoneal (i.p.) injection. A mixture of 1 μg of cholera toxin (CT; List Biological Laboratories Inc., Campbell, CA) and 10 μg of YopE_69-77 peptide (New England Peptide, Gardner, MA) in 15 μl of phosphate-buffered saline (PBS) was applied into the nares of each mouse. Control mice received CT in 15 μl of PBS only. The mice were then monitored for 1 h to ensure recovery from anesthetization before being returned to the animal facility. The immunization procedure was performed on days 0, 7, and 21.

Bacterial strains and infection conditions.

The Y. pseudotuberculosis strains used in this study are the serogroup O:1 strain 32777 (66) and its derivative, mE, which carries the catalytic inactive YopE(R144A) protein. The strain mE was created through allelic exchange with suicide plasmid pSB890-YopE^R144A, as described previously (67). Y. pestis strain KIM D27, used for protection studies, was described previously (66). The KIM D27 strain lacks the pgm locus and is exempt from select-agent guidelines.

Intragastric (i.g.) infection was carried out as described previously (66). To prepare bacteria, overnight cultures of Y. pseudotuberculosis grown in Luria-Bertani (LB) medium at 28°C were washed once and resuspended in phosphate-buffered saline (PBS) to achieve the desired number of CFU/ml. Naïve mice aged 9 to 10 weeks old or mice immunized as described above were fasted for 16 h and then inoculated with 200 μl of Y. pseudotuberculosis suspended in PBS through a 20-gauge feeding needle. Mice were provided with food and water thereafter. For primary intravenous (i.v.) infection with the mE mutant, bacterial samples prepared as described above were used to inoculate mice using 100-μl volumes delivered via the lateral tail vein. For protection studies with KIM D27, the bacteria were grown overnight at 28°C in heart infusion broth, washed, and resuspended in PBS to the desired number of CFU/ml as described previously (66). Mice that survived infection with mE 4 weeks prior were injected with 1,000 CFU (∼100 LD₅₀s) or 1 × 10⁵ CFU of KIM D27 in 100 μl of PBS via the lateral tail vein. At the indicated times postinfection, or when death was imminent, mice were euthanized by CO₂ asphyxiation. Mouse spleens and MLN were dissected aseptically, weighed, and dispersed with the plunger of a 5-ml syringe into single-cell suspensions in 5 ml of Dulbecco's modified Eagle medium (DMEM). Serial dilutions of 100-μl volumes were plated on LB agar to determine bacterial colonization by CFU assay; therefore, the limit of detection is 50 CFU or log₁₀ CFU of 1.7. All animal procedures were approved by the Stony Brook University Institutional Animal Care and Use Committee.

Flow cytometry.

Single-cell suspensions of spleens and MLN were prepared as described above. Single-cell suspensions from nonperfused lungs were processed with a gentleMACS dissociator obtained from Miltenyi Biotec by following the guidelines of the manufacturer. Suspended cells (1 × 10⁶) were blocked using anti-mouse CD16/CD32 (FcgIII/II receptor) clone 2.4G2 (BD Pharmingen) and labeled with fluorophore-conjugated antibodies or YopE_69-77-specific tetramer as described previously (31). Labeled cells were analyzed using BD FACSCalibur. Gating on side and forward scatter was used to focus on lymphocytes. Data were processed with WinList software. Isotype-matched antibodies were used to control for nonspecific binding. The antibodies used were anti-mouse CD3e-PerCP (clone 145-2C11; Pharmingen) and AlexaFluor488 anti-mouse CD8a (53-6.7; BD, BioLegend). The allophycocyanin-conjugated MHC-I tetramer K^bYopE_69-77 was provided by the NIH Tetramer Core Facility (Emory University, Atlanta, GA).

Statistical analysis.

Statistical analysis was performed with Prism 4.0 (GraphPad) software. The tests used are as indicated in the figure legends or main text. A P value of less than 0.05 was considered significant.

RESULTS

YopE_69-77 is a protective antigen for Y. pseudotuberculosis.

Recently YopE_69-77 (SVIGFIQRM) was identified as a protective CD8 T cell antigen for Y. pestis, as immunization of C57BL/6 mice with this peptide mediated significant protection against lethal pulmonary challenge (31). Analysis of genome information available through NCBI showed that the sequence of YopE_69-77 is well conserved in different Y. pestis and Y. pseudotuberculosis strains (31) (data not shown). Sequencing of the yopE gene in the strain of Y. pseudotuberculosis used here (32777) showed that YopE_69-77 is SVIGFIQRM (data not shown). To test whether the CD8 T cell response against this peptide is also protective against lethal Y. pseudotuberculosis challenge, a cohort of C57BL/6 mice were immunized intranasally with 10 μg of YopE_69-77 peptide combined with CT as an adjuvant (31). CT has been used as an efficient mucosal adjuvant to prime CD8 T cell responses (41). Control mice received CT only. After the last of three immunizations, the production of YopE-specific CD8 T cell responses in lung, spleen, and MLN was measured by tetramer (H2-K^b–YopE_69-77) staining and flow cytometry. Sixteen days after the last immunization, a significantly higher percentage of YopE_69-77-specific CD8 T cells was detected in both lungs and spleens of mice immunized with peptide plus CT compared to CT alone (Fig. 1A and B). In contrast, the percentages of YopE_69-77-specific CD8 T cells in MLN of mice immunized with peptide plus CT were similar to those seen in control mice receiving CT only (Fig. 1A and B).

Fig 1.

Immunization with the YopE_69-77 peptide induces YopE_69-77-specific CD8 T cell responses. C57BL/6 mice were immunized intranasally with 10 μg YopE_69-77 peptide mixed with cholera toxin (CT) as an adjuvant on days 0, 7, and 21. Control mice received CT alone at the same times. The presence of total and YopE_69-77-specific CD8 T cells in the indicated organs was determined by flow cytometry and tetramer staining 16 days after the last immunization. (A) Representative images of the analysis of the single-cell suspension prepared from mouse lung, spleen (SPL), and mesenteric lymph nodes (MLN) obtained from either control mice (CT; left) or peptide-immunized mice (YopE_69-77 plus CT; right). The percentage of total cells contained in each quadrant is shown. (B) Summary of the percentage of YopE_69-77-specific CD8 T cells within total CD8 T cell populations in the indicated organs from either CT-immunized mice (triangle) or YopE_69-77 peptide-plus-CT-immunized mice (square). Each symbol represents the value obtained from one mouse. The data shown are the summary of two independent experiments with cohorts of 2 to 3 mice. By Mann-Whitney test, “**” indicates P values of less than 0.01 compared to treatment with CT alone.

Additional cohorts of immunized mice were challenged i.g. with 5 × 10⁹ CFU/mouse (∼10 LD₅₀s) of Y. pseudotuberculosis strain 32777 17 days after the last immunization. Five days after i.g. challenge, the percentages of YopE_69-77-specific CD8 T cells remained high in lungs and were even higher in the spleens of the mice immunized with peptide plus CT than in those of the controls given CT alone (Fig. 2A). More importantly, the levels of such cells were now significantly higher in the MLNs of the mice immunized with peptide plus CT than in those of the controls given CT alone (Fig. 2A). Five days after i.g. challenge, the numbers of bacteria in spleen and MLN were similar between the peptide-plus-CT and CT alone groups (Fig. 2B), suggesting that the CD8 T cell response to YopE_69-77 was not important in controlling bacterial colonization in these organs at this stage of infection (Fig. 2B). However, when survival of the mice was monitored, all mice given CT alone died within a week postchallenge, while 64% of the mice immunized with YopE_69-77 peptide plus CT survived for over 28 days (Fig. 2C). The difference in survival between the two groups, determined by log rank test, was significant (P < 0.0001). Lin et al. (31) have shown that depletion of CD8 T cells significantly reduced protection against Y. pestis in mice vaccinated with CT plus YopE_69-77 under identical conditions to those used here. Therefore, the results obtained here suggest that YopE_69-77 is a protective CD8 T cell antigen for Y. pseudotuberculosis in C57BL/6 mice.

Fig 2.

Immunization with the YopE_69-77 peptide mediates protection against lethal Y. pseudotuberculosis challenge. Groups of C57BL/6 mice were immunized with YopE_69-77 peptide plus CT or with CT alone, as described in the legend of Fig. 1. (A, B) Seventeen days after the last immunization, the mice received 5 × 10⁹ CFU of 32777 (10 LD₅₀s) i.g. (A) At 5 days postchallenge, cohorts of mice were euthanized, and the percentages of YopE_69-77-specific CD8 T cells among total CD8 T cell populations in the indicated organs from either CT- or peptide-plus-CT-immunized mice were determined by flow cytometry and tetramer staining. (B) Bacterial colonization levels in the indicated organs were determined by CFU assay (number of log₁₀ CFU/organ) at 5 days postchallenge. Each symbol represents the value obtained from one mouse. The limit of detection was log₁₀ CFU of 1.7. The data shown are the summary of two independent experiments with cohorts of 2 to 3 mice. By Mann-Whitney test, “*” indicated P values of less than 0.05, while “**” indicates P values of less than 0.01 compared to CT-immunized mice. (C) Survival of immunized mice that were challenged was followed for 28 days. Numbers in parentheses indicate the number of mice in the treatment group. The data shown are from one experiment. The P value obtained by log rank test is indicated.

Primary infection with wild-type Y. pseudotuberculosis elicits a CD8 T cell response to YopE_69-77.

YopE_69-77 was shown to be a dominant antigen recognized by CD8 T cells in C57BL/6 mice vaccinated with live attenuated Y. pestis and then challenged with virulent Y. pestis (31). This finding prompted us to investigate if a YopE_69-77-specific CD8 T cell response is produced during primary Y. pseudotuberculosis infection. C57BL/6 mice were left uninfected or infected i.g. with 5 × 10⁷ CFU of Y. pseudotuberculosis 32777, a dose shown previously to allow ∼80% survival (66). At 7, 14, and 21 days after infection, surviving mice were euthanized, and the numbers of YopE_69-77-specific CD8 T cells were determined in spleens and MLNs. In comparison to uninfected mice, there was a significant increase in YopE_69-77-specific CD8 T cells in spleen, as measured by percentage (Fig. 3A) or number of cells (data not shown). Seven days postinfection, an average of 6.3% of CD8 T cells from spleen were labeled positive with the YopE tetramer, while only 0.78% of such cells were found in mice left uninfected (Fig. 3A). A large increase in the percentage of these cells was seen 14 days after infection—an average of 30.9% of CD8 T cells from infected spleens labeled positive with the tetramer (Fig. 3A). The percentage of YopE_69-77-specific CD8 T cells decreased to an average of 22.5% at 21 days postinfection (Fig. 3A). In MLN, YopE_69-77-specific CD8 T cells were detected at 7 days postinfection, with an average of 8.8% of CD8 T cells labeled positive with the tetramer (Fig. 3B). In comparison, only an average of 0.51% of YopE_69-77-specific CD8 T cells were detected in MLN from mice left uninfected (Fig. 3B). The percentage of YopE_69-77-specific CD8 T cells in MLN remained relatively steady at 9.1% and 10.6% on 14 and 21 days postinfection, respectively (Fig. 3B). Collectively, these results indicated that YopE_69-77 is a major CD8 T cell antigen during primary Y. pseudotuberculosis infection of C57BL/6 mice.

Fig 3.

Generation of a major CD8 T cell response to YopE_69-77 in response to primary infection with Y. pseudotuberculosis. Groups of C57BL/6 mice were infected with wild-type 32777 (wt) or the yopE catalytic inactive mutant (mE) or left uninfected (UI). At the indicated days postinfection, mice were euthanized, and the percentages of YopE_69-77-specific CD8 T cells among total CD8 T cell populations in spleen (A) or MLN (B) were determined by flow cytometry and tetramer staining. Each symbol represents the value obtained from one mouse. The data are pooled from 3 independent experiments with cohorts of 1 to 5 mice. When the values obtained from the spleens of mice infected with the wt were compared to those from mice infected with mE at the day 14, the difference was significant (*, P < 0.05, Mann-Whitney test).

Establishment of a mouse infection model with a Y. pseudotuberculosis yopE catalytic mutant.

Previous results obtained using Y. enterocolitica expressing YopE-LLO or YopE-OVA to infect mice i.g. suggest that YopE function is important for inducing efficient CD8 T cell responses against these model antigens (30, 56, 61), most likely because GAP activity is necessary to allow efficient bacterial colonization of murine lymphoid tissues such as spleen (11, 32, 57). We assumed that GAP activity would be important for a CD8 T cell response to YopE_69-77 because it would promote efficient bacterial colonization of lymphoid tissues in C57BL/6 mice infected with 32777. Thus, we hypothesized that GAP activity would be dispensable for the immunogenicity of YopE per se. To determine if YopE GAP activity is dispensable for the CD8 T cell response to YopE_69-77, it was important to establish a mouse infection model that would allow adequate colonization of murine lymphoid tissues by a Y. pseudotuberculosis strain lacking YopE GAP catalytic activity. Toward this goal, a derivative of strain 32777 was constructed by mutating codon 144 in yopE. The resulting strain, designated mE, expresses a YopE protein in which the essential catalytic Arg at position 144 is changed to Ala (R144A) (11). Because the substitution is positioned 67 residues from the CD8 T cell epitope, we assumed it would not interfere with antigen processing of YopE. We first determined the maximal dose of mE that would allow the majority of infected C57BL/6 mice to survive (maximal sublethal dose). Groups of mice were infected i.g. with different doses of mE (ranging from 5 × 10⁷ to 5 × 10⁹ CFU), and survival was monitored for 28 days. At the infection dose of 5 × 10⁹ CFU/mouse, none of the infected mice survived, at 5 × 10⁸ CFU/mouse, 86% of the mice survived, and 100% of the mice survived when the infection dose was 5 × 10⁷ CFU/mouse (data not shown). A dose of 5 × 10⁸ CFU/mouse was used as the maximal sublethal dose for subsequent infections with mE, which is 10-fold higher than the infection dose used for wt strain 32777 (5 × 10⁷ CFU/mouse), which results in survival of ∼80% of the infected mice (66).

The degree of virulence attenuation of mE was also assessed through the i.v. route of infection. Doses of 1 × 10³ and 1 × 10⁴ CFU were sublethal for C57BL/6 mice, while 1 × 10⁵ CFU was lethal (data not shown). By comparison, an i.v. dose of 1 × 10³ CFU is lethal for the majority of mice infected with wt 32777 (35). These results confirmed that mE is substantially attenuated.

We previously demonstrated that C57BL/6 mice that survive primary i.g. infection with 5 × 10⁷ CFU of wt strain 32777 are well protected against secondary challenge with virulent Y. pestis (66). To test whether i.g. inoculation with 5 × 10⁸ CFU of mE is sufficient to protect against subsequent Y. pestis infection, the vaccinated mice were i.v. challenged with 1,000 CFU (∼100 LD₅₀s) of Y. pestis KIM D27 4 weeks after primary infection with mE. As expected, all unimmunized mice died between 4 and 5 days postchallenge, while 100% of mice vaccinated with mE survived the challenge until 28 days later (Fig. 4). Next, we increased the challenging dose to 10⁵ CFU of Y. pestis KIM D27. Again, none of the unimmunized mice survived, while all the mice immunized with mE survived until 28 days postchallenge (Fig. 4). This result showed that vaccination of mice with 5 × 10⁸ CFU of mE elicited protective immune responses against Y. pestis.

Fig 4.

Mice vaccinated by infection with Y. pseudotuberculosis mE are protected from lethal challenge by Y. pestis. Groups of C57BL/6 mice were immunized (I) by inoculation i.g. with 5 × 10⁸ CFU of mE or left uninfected (UI). Four weeks later, mice were i.v. challenged with either 10³ CFU or 10⁵ CFU of Y. pestis KIM D27, and survival was monitored for 28 days. The number of mice in each treatment group is indicated in parentheses. The data shown are the summary of two independent experiments, with one experiment using the 10³ CFU dose and the second using the 10⁵ CFU dose.

The GAP activity of YopE is dispensable for a CD8 T cell response to YopE_69-77 in mice infected with Y. pseudotuberculosis.

We next infected groups of C7BL/6 mice i.g. with 5 × 10⁷ CFU of the wt or 5 × 10⁸ CFU of mE and assessed weights of spleens and MLN and levels of bacterial colonization of these organs on days 7, 14, and 21 postinfection. Infection with wt Y. pseudotuberculosis caused splenomegaly peaking on day 14 (Fig. 5A), which is largely due to infiltration of neutrophils and macrophages (66). Spleen weight increases were also observed in mice infected with mE, yet the splenomegaly was significantly reduced compared to that seen in wt-infected mice (Fig. 5A). In contrast, the increase in weights of MLNs in wt-infected and mE-infected animals was not significantly different on days 7, 14, and 21 after infection (Fig. 5B).

Fig 5.

Comparison of weights and bacterial colonization levels of organs in mice infected with the Y. pseudotuberculosis wt or mE strains. Groups of C57BL/6 mice were left uninfected (UI) or infected i.g. with 5 × 10⁷ CFU of the wt or 5 × 10⁸ CFU of mE. At the indicated days postinfection, the weights of spleen (A) and MLN (B) in grams were determined. Bacterial colonization levels in spleens (C) and MLNs (D) was determined by CFU assay (number of log₁₀ CFU/organ). Each symbol represents the value obtained from one mouse. The data shown are the summary of three independent experiments with cohorts of 1 to 5 mice. (C, D) The dashed line indicates the limit of detection. By Mann-Whitney test, “*” indicates P values of less than 0.05, and “**” indicates P values of less than 0.01 when comparing the values obtained from wt-infected mice to those obtained from mice infected with mE at the same time point.

Even at a 10-fold higher infection dose, the mE strain showed a defect in colonizing spleen compared to the wt (Fig. 5C). Seven days postinfection, significantly reduced colonization levels were observed in spleens of mE-infected mice compared to those in mice infected with wt (Fig. 5C). In MLN, the colonization levels of mE were comparable to those of the wt on days 7 and 10 postinfection; however, at day 21, significantly lower colonization levels were observed in mice infected with mE (Fig. 5D).

The percentage of YopE_69-77-specific CD8 T cells in spleens and MLN of mE-infected mice was determined and compared to the results obtained with wt-infected mice. Similar percentages of splenic CD8 T cells were labeled positive with the YopE_69-77-specific tetramer in mice infected with the wt or mE on 7 and 21 days postinfection (Fig. 3A). Only at 14 days postinfection was a higher percentage of YopE_69-77-specific CD8 T cells observed in the spleens of mice infected with wt bacteria than in those of mice infected with mE (Fig. 3A). In MLN of mice infected with the wt or mE, the levels of YopE_69-77-specific CD8 T cells were not significantly different through the time points analyzed (Fig. 3B). Overall, these results indicated that although the GAP activity of YopE is required for maximal splenic colonization and splenomegaly in mice infected with Y. pseudotuberculosis, it is dispensable for a CD8 T cell response to YopE_69-77.

DISCUSSION

Understanding how the adaptive immune response is produced is of fundamental importance for the control and prevention of infectious diseases. More than a decade ago, the potential of the translocated effectors of bacterial T3SSs to function as antigens of CD8 T cells was recognized (50). By vaccination of mice with purified effector protein, Starnbach and Bevan (50) raised murine CD8 T cells that recognize a peptide epitope from YopH in the context of the MHC class I molecule H-2 Db. This epitope was presented by cells that produce YopH endogenously and by epithelial cells infected with Y. enterocolitica (50). Several T3SS effector proteins, including YopE, have been used as delivery vehicles for model CD8 T cell antigens such as OVA and LLO expressed by live attenuated carrier vaccine strains (42). However, only recently has it been demonstrated that a CD8 T cell response can be generated against a native epitope in a type III effector protein during bacterial infection (31). The CD8 T cell response to the YopE_69-77 epitope in C57BL/6 mice infected with attenuated Y. pestis was shown to be dominant (31). In addition, mice immunized with the YopE_69-77 peptide were protected against lethal pulmonary challenge with virulent Y. pestis, and CD8 T cells from immunized mice could be induced to produce TNF-α and IFN-γ upon ex vivo stimulation with YopE_69-77, indicating that they have an effector T cell phenotype (31). Here, we extend these results by showing that C57BL/6 mice vaccinated with the YopE_69-77 peptide are protected from lethal i.g. challenge with wt Y. pseudotuberculosis and that the YopE_69-77 epitope is a major CD8 T cell antigen during primary i.g. infection of mice with this pathogen.

Although our challenge model used i.g. inoculation to mimic the natural enteric route of Y. pseudotuberculosis infection, we chose to use intranasal immunization with YopE_69-77 peptide plus CT because the conditions for eliciting robust CD8 T cell responses had been previously established using this route (31). Intranasal immunization with YopE_69-77 peptide plus CT was shown to elicit a CD8 T cell response in lungs as well as spleen and peripheral blood leukocytes (31), and therefore, we expected that this vaccination protocol would protect against lethal systemic Y. pseudotuberculosis infection. Although YopE_69-77 peptide-vaccinated mice were protected against lethal disease following i.g. challenge with Y. pseudotuberculosis (Fig. 2C), a corresponding decrease in bacterial colonization in spleen and MLN was not seen at day 5 postinfection in immunized mice (Fig. 2B). It is possible that bacterial colonization in immunized mice was reduced at this time point in organs that were not studied (e.g., liver) or that the CD8 T cell response to YopE_69-77 was effective at clearing Y. pseudotuberculosis from tissues at later time points.

At the peak of the response during primary infection of C57BL/6 mice with wt Y. pseudotuberculosis, up to 40% of the splenic CD8 T cells were specific for YopE_69-77 (Fig. 3A, day 14). To our knowledge, this level of splenic CD8 T cell response is unprecedented for a primary infection with a bacterial pathogen. This finding is especially surprising because YopE_69-77 was not predicted to be a good CD8 T cell antigen. Several epitope prediction algorithms indicated that many peptides carried by other Yersinia type III-secreted proteins should bind H2-K^b far more effectively than YopE_69-77 (31). In fact, a screen of 90 high-scoring pCD1-encoded peptides predicted by the BIMAS and SYFPEITHI algorithms failed to identify a Yersinia CD8 T cell antigen (31).

The extent to which the biology of YopE accounts for its immunodominance in the C57BL/6 mouse infection model is presently unclear, although several features that could be involved include its specific delivery into APC subsets, levels of translocation, and/or effector functions. To initiate studies in this direction, we investigated whether the effector GAP activity of YopE is required for the production of YopE_69-77-specific CD8 T cells. By infecting mice with a higher-than-normal dose of mE expressing YopE(R144A), adequate colonization of murine lymphoid tissues by the attenuated strain was obtained, allowing us to show that GAP activity is not required for the production of a CD8 T cell response to YopE_69-77 (Fig. 3). Specifically, similar percentages of YopE_69-77-specific CD8 T cells were detected in spleens of mice infected with the wt and mE at day 21 (∼25%) (Fig. 3). This was true despite the fact that at days 7, 14, and 21, there were lower levels of splenomegaly (Fig. 5A) and bacterial colonization (Fig. 5C) in spleens infected by the mutant than in those infected by the parent. In vitro studies showed that more of the catalytically inactive YopE(R144A) protein is translocated into host cells than active YopE (4). It is therefore possible that the R144A substitution actually enhances the immunogenicity of YopE by increasing the amount of this effector that is translocated into APCs. An additional advantage of the R144A substitution is that it decreases the virulence of Y. pseudotuberculosis by both the i.g. and i.v. routes of inoculation without comprising the priming of immunity against Y. pestis (Fig. 4), and therefore, the mE strain can be considered an improved live vaccine for plague compared to the wt (66).

The amount of YopE that is translocated into APCs could regulate its presentation as an antigen to CD8 T cells, as discussed above. In addition, the ability of YopE to bind to host membranes (27, 29, 65) or to undergo ubiquitination (26) could influence antigen presentation. YopE is the first type III effector to be identified as a protective CD8 T cell antigen. Since antigen processing and presentation is the first step in generating an adaptive immune response, it will be important to understand the features of YopE that drive its efficient presentation and exceptional antigenicity during Yersinia infection. Such information could aid in the identification of T3SS effector proteins in other pathogens as protective CD8 T cell antigens.