|Home | About | Journals | Submit | Contact Us | Français|
Extracellular Yersinia pseudotuberculosis employs a type III secretion system (T3SS) for translocating virulence factors (Yersinia outer proteins [Yops]) directly into the cytosol of eukaryotic cells. Recently, we used YopE as a carrier molecule for T3SS-dependent secretion and translocation of listeriolysin O (LLO) from Listeria monocytogenes. We demonstrated that translocation of chimeric YopE/LLO into the cytosol of macrophages by Yersinia results in the induction of a codominant antigen-specific CD4 and CD8 T-cell response in orally immunized mice. In this study, we addressed the requirements for processing and major histocompatibility complex (MHC) class II presentation of chimeric YopE proteins translocated into the cytosol of macrophages by the Yersinia T3SS. Our data demonstrate the ability of Yersinia to counteract exogenous MHC class II antigen presentation of secreted hybrid YopE by the action of wild-type YopE and YopH. In the absence of exogenous MHC class II antigen presentation, an alternative pathway was identified for YopE fusion proteins originating in the cytosol. This endogenous antigen-processing pathway was sensitive to inhibitors of phagolysosomal acidification and macroautophagy, but it did not require the function either of the proteasome or of transporters associated with antigen processing. Thus, by an autophagy-dependent mechanism, macrophages are able to compensate for the YopE/YopH-mediated inhibition of the endosomal MHC class II antigen presentation pathway for exogenous antigens. This is the first report demonstrating that autophagy might enable the host to mount an MHC class II-restricted CD4 T-cell response against translocated bacterial virulence factors. We provide critical new insights into the interaction between the mammalian immune system and a human pathogen.
Protein antigens are recognized by T cells as short peptide fragments bound either to major histocompatibility class (MHC) I or to MHC class II molecules on the surface of antigen-presenting cells (APCs). The location of antigens in distinct intracellular compartments of APCs influences their proteolytic processing as well as access to MHC molecules (16, 39). Classically, peptides generated in the cytosol (e.g., derived from viral proteins) by proteasomal degradation are bound to MHC class I molecules after transport across the endoplasmic reticulum membrane by the transporters associated with antigen processing (TAP) (90). Subsequently, MHC class I molecules present these endogenous antigenic peptides to CD8 T cells. In contrast, exogenous antigens (e.g., antigens derived from engulfed bacteria and soluble antigens) are directed into the endosomal/lysosomal pathway for degradation (63). In late endosomal compartments, degraded protein fragments interact with MHC class II molecules and are further trimmed into peptides for presentation to CD4 T cells.
However, the synchrony of this system has been challenged by biochemical and functional studies of professional and nonprofessional APCs. Epitopes derived from a diverse pool of cytosolic antigens such as metabolic enzymes, cytoskeletal proteins, and viral and tumor antigens have been identified to be presented in the context of both murine and human MHC class II molecules (20, 51, 52, 53, 71, 88). Several different alternative pathways for delivering antigens into the MHC class II pathway have been described. Lich et al. reported that cytoplasmic processing by the proteasome and calpain is required for efficient processing of the autoantigen glutamate decarboxylase to CD4 T cells (51). Also, a proteasome/TAP-dependent pathway was shown to be important for the presentation of MHC class II peptides from the influenza virus (83). In contrast, other studies using human B lymphoblastoid cells revealed that TAP is not involved in the transport of cytosolic peptides to MHC class II molecules (52, 53). Alternatively, autophagy in the form of either macroautophagy (27, 30, 56, 60) or chaperone-mediated autophagy (19, 23, 24, 60, 91) results in the transport of cytosolic peptides and proteins directly into endosomes and lysosomes. Two essential components of the chaperone-mediated autophagy pathway are Lamp-2a and the heat shock cognate protein hsc70. The latter molecule is an accessory chaperone that intersects target proteins in the cytoplasm and facilitates their delivery to Lamp-2a, a lysosomal membrane protein.
Pathogenic yersiniae are Gram-negative bacteria that cause a wide range of diseases in humans, ranging from bubonic plague, caused by Yersinia pestis, to self-limiting gastroenteritis and lymphadenitis, caused by the enteric pathogens Yersinia pseudotuberculosis and Yersinia enterocolitica. Results from in vitro experiments and the mouse infection model revealed that the last two bacterial species have acquired a complex arsenal of effector proteins to overcome host defense mechanisms. These major pathogenicity factors are located on a 70-kb virulence plasmid, which encodes a protein export apparatus called the type III secretion system (T3SS) (21, 35). T3SS is a complex bacterial organelle that provides Gram-negative pathogens with a unique virulence mechanism enabling them to translocate bacterial effector proteins directly into the host cell cytosol. At least six effector proteins, Yersinia outer proteins (Yops), namely, YopE, YopH, YopM, YopO/YpkA, YopP/YopJ, and YopT, are injected into the cytosol of eukaryotic cells in a T3SS-dependent manner (22). The main function of these Yops is to inhibit the immune response of the host. Four Yops (YopE, YopH, YopO/YpkA, and YopT) are involved in inhibiting phagocytosis of yersiniae by disrupting the cytoskeleton of polymorphonuclear leukocytes and macrophages (12, 33, 40, 68). Thus, the consequence of this translocation process is that pathogenic yersiniae survive and proliferate at extracellular sites in the infected host (77).
Our laboratory has described and analyzed in detail the potential of YopE to be a carrier molecule for heterologous antigen delivery by Yersinia (74). YopE is a translocated GTPase-activating molecule that can downregulate Rho activity, leading to actin filament disruption and inhibition of phagocytosis by macrophages (10, 69, 87). The N-terminal 18 amino acids (aa) of YopE fused to a large protein fragment of the p60 antigen from the intracellular pathogen Listeria monocytogenes were sufficient for T3SS-dependent secretion to the extracellular environment of Yersinia-infected target cells (74). In contrast, fusion of p60 to the N-terminal 138 aa of YopE resulted in translocation of the chimeric protein into the cytosol of host cells. As expected, T-cell activation assays revealed that the cytosolic delivery of this hybrid protein was a prerequisite to induce a p60-specific MHC class I-restricted CD8 T-cell response (74). Because translocated p60 easily enters the endogenous cytosolic antigen presentation pathway, it is processed like an endogenously synthesized antigen. In a more recent publication, we used Y. pseudotuberculosis expressing YopE fused to listeriolysin O (LLO) from Listeria to assess the influence of secreted versus translocated antigen display on in vitro antigen presentation and in vivo T-cell priming in the oral mouse infection model (72). We constructed two different plasmid-encoded hybrid YopE/LLO proteins. By engaging the above-mentioned well-defined secretion and translocation domains of YopE (79) fused to aa 51 to 363 of LLO, chimeric YopE was expressed either in secreted or in translocated form (72). Biochemical fractionation of Yersinia-infected macrophage-like P338D1 cells clearly revealed that YopE from aa 1 to 38 (YopE1-138)/LLO was translocated to the cytosol of host cells, whereas YopE1-18/LLO lacking the translocation domain was efficiently secreted to the culture medium but was not detected in the P338D1 cell lysate containing cytosolic proteins. Thus, it was theoretically expected that the well-secreted version of chimeric YopE/LLO could have the potential to enter the exogenous MHC class II antigen presentation pathway for proper CD4 T-cell priming. Strikingly, results from in vitro antigen presentation assays and also the enumeration of LLO-specific CD4 T cells from infected mice indicated a superior efficacy of translocated over secreted LLO for MHC class II antigen presentation and CD4 T-cell induction, respectively (72).
This study addresses the requirements for processing and MHC class II presentation of chimeric YopE proteins translocated into the cytosol of macrophages by the T3SS of Yersinia pseudotuberculosis. Our data demonstrate the ability of Yersinia to counteract exogenous MHC class II antigen presentation of secreted hybrid YopE by the action of wild-type YopE and YopH. However, a distinct MHC class II antigen presentation pathway was identified for YopE fusion proteins originating from the cytosol. Presentation of cytoplasm-derived chimeric YopE requires acidification of the phagolysosome and is sensitive to inhibitors of autophagy.
The plasmids (pACYC184 derivatives) and Y. pseudotuberculosis strains used in this study are listed in Table Table1.1. Translational fusions between different lengths of YopE or YopH with LLO or p60 were constructed by standard PCR cloning procedures, as described previously (72, 73). All resulting protein fusions were checked by DNA sequencing. Escherichia coli χ6060 was used as an intermediate host for cloning experiments. The chimeric proteins used in this study are tagged at their C termini with an M45 epitope tag (MDRSRDRLPPFETETRIL) that is derived from the E4-6/7 protein of adenovirus (58). Overnight cultures of Y. pseudotuberculosis strains were grown in Luria-Bertani (LB) medium at 27°C. On the next day, cultures were diluted and incubated at 37°C for 4 h to allow expression of components and targets of the T3SS encoded by the Yersinia virulence plasmid (74). When required, the antibiotics kanamycin (30 μg/ml) and chloramphenicol (20 μg/ml) were added.
Secreted proteins from Yersinia culture supernatants were prepared as follows. Briefly, bacterial supernatants were passed through a 0.45-μm-pore-size syringe filter to remove bacteria. Protein in the bacterium-free medium was precipitated by the addition of cold trichloroacetic acid to 10% (vol/vol) and incubated for 2 h on ice. The protein was collected by centrifugation at 10,000 × g and 4°C for 20 min. The pellets were washed in 0.8 ml cold acetone, dried, and resuspended in phosphate-buffered saline (PBS) buffered with 80 mM Tris-HCl, pH 8. Samples corresponding to 500 μl culture supernatant were separated in a 10% discontinuous sodium dodecyl sulfate (SDS)-polyacrylamide gel and transferred to nitrocellulose membranes. Hybrid proteins were detected by immunoblot analysis. Western blots were treated with a monoclonal antibody against M45, followed by incubation with a horseradish peroxidase-labeled antimouse antibody. Blots were developed by using a chemiluminescence kit.
Macrophage-like P388D1 cells were grown on glass coverslips to 60% confluence. One hour before the addition of bacteria, Dulbecco's modified Eagle's medium (DMEM) was replaced by 500 μl of Hank's balanced salt solution (HBSS). P388D1 cells were infected with Y. pseudotuberculosis strain pIB102 or pIB251 at a multiplicity of infection (MOI) of 5 to 10 for 15 min. Then, macrophage-like cells were incubated with 50 μg/ml dye-quenched (DQ) ovalbumin (OVA)-fluorescein isothiocyanate (FITC) (Molecular Probes, Eugene, OR) for 30 min at 37°C, washed three times with HBSS to remove non-cell-associated bacteria and DQ-OVA-FITC, and fixed in PBS-3.7% formaldehyde. Extracellular yersiniae were stained with an anti-Y. pseudotuberculosis lipopolysaccharide (LPS) polyclonal rabbit antiserum (a kind gift of R. Rosqvist, Umeå University, Umeå, Sweden) and a secondary anti-rabbit tetramethylrhodamine isothiocyanate (TRITC) conjugate (1:100 in PBS-3% bovine serum albumin; Sigma, St. Louis, MO). After permeabilization of P388D1 cells (3 min in PBS-0.1% Triton X-100), extra- and intracellular yersiniae were stained with a polyclonal anti-Y. pseudotuberculosis LPS antiserum and a secondary anti-rabbit 7-amino-4-methylcoumarin-3-acetic acid (AMCA) conjugate (1:100 in PBS-3% bovine serum albumin; Jackson Immuno Research, Newmarket, United Kingdom). Coverslips were mounted on glass slides and analyzed by fluorescence microscopy. Experiments were repeated at least three times.
CD8 T-cell lines specific for p60217-225 and LLO91-99 and CD4 T cells against the CD4 T-cell epitope p60301-312 were derived from spleens of L. monocytogenes-infected BALB/c mice. CD4 T-cell lines specific for the H-2b-restricted CD4 T-cell epitopes p60177-188 and LLO190-201 were established from spleens 14 days after intravenous infection of C57BL/6 mice with 1 × 103 CFU L. monocytogenes. All CD8 T-cell lines were propagated by repeated restimulation with P815 cells transfected with the human B7.1 gene (P815/B7) in the presence of the appropriate synthetic peptide in medium supplemented with interleukin-2 (IL-2) as described previously (38). All CD4 T-cell lines were repeatedly restimulated with syngeneic mitomycin C-inactivated splenocytes as APCs in the presence of 10−6 M peptide. The T-cell culture medium was an alpha modification of Eagle's medium supplemented with glutamine, penicillin, streptomycin, 10% fetal calf serum (FCS), 100 U/ml recombinant murine IL-2 (R&D Systems, Minneapolis, MN), and 2 × 10−5 M 2-mercaptoethanol.
All antigen presentation studies were performed with bone marrow-derived macrophages (BMMs). Female C57BL/6 JO1aHsd (H-2b) and CB6F1 (H-2b × H-2d) mice were purchased (Centre d'Elevage R. Janvier, Le Genest-St. Isle, France), kept under conventional conditions, and used as bone marrow donors at 8 to 16 weeks of age. Macrophages were seeded at a density of 1 × 105 cells per well in 96-well flat-bottom tissue culture plates in DMEM supplemented with 10% FCS and 10 ng/ml granulocyte-macrophage colony-stimulating factor (GM-CSF; R&D Systems). BMM cultures were fed with 100 μl GM-CSF-supplemented medium every 4 days and were used after 12 to 16 days of culture. Nonadherent cells were removed by thorough washing before use.
Direct antigen presentation assays were performed as described previously (78) with some modifications. BMMs to be used in antigen presentation assays were cultured for 24 h in medium supplemented with 10 ng/ml gamma interferon (IFN-γ; R&D Systems). Macrophages were infected with yersiniae for 4 h at the indicated MOI. Cells were fixed with 1% paraformaldehyde before addition of T cells. After overnight incubation, T-cell activation was assessed by measuring the IFN-γ concentration in culture supernatants with an IFN-γ specific enzyme-linked immunosorbent assay (ELISA) that binds and detects IFN-γ with a pair of specific monoclonal antibodies. Results were corrected for dilution of the sample to yield the sample concentration in ng/ml.
Inhibition experiments were performed in the presence of proteasome inhibitors (epoxomycin, lactacystin), an inhibitor of tripeptidyl peptidase II (TPPII; alanyl-alanyl-phenylalanine chloromethyl ketone [AAF-CMK]), an inhibitor of endosomal proteases (leupeptin), an inhibitor of lysosomal acidification (chloroquin), or inhibitors of the autophagy pathway (3-methyladenine [3-MA] and wortmannin) at the indicated concentrations. Brefeldin A was used to inhibit the export of newly folded MHC class II molecules. Inhibitors were generally added 30 to 60 min before infection of cells.
For exogenous loading experiments either purified p60, whole concentrated L. monocytogenes culture supernatant (SN), or nonconcentrated supernatant from Yersinia pseudotuberculosis was used. The bacterial murein hydrolase p60 was purified from supernatants of L. monocytogenes MR1(pGB363-1p60) cultures by preparative SDS-PAGE and subsequent gel elution, as described previously (36). Concentrated (100 ×) L. monocytogenes sv1/2a EGD supernatant was prepared as described previously (61). Yersinia pseudotuberculosis supernatant was prepared after 6 h infection (MOI, 10) of CBA/J-derived BMMs (H-2k haplotype) with pIB102(pHR429) or pIB102(pHR430). Supernatants were filtrated through a 0.22-μm-pore-size filter before use.
The detection of translocated chimeric YopE/LLO was carried out as described previously (72). Briefly, BMM monolayers were grown in 100-mm-diameter tissue culture plates in DMEM supplemented with 5% FCS. Chemical inhibitors were generally added 30 to 60 min before infection of cells. BMMs were infected with Y. pseudotuberculosis strain pIB102(pHR430) expressing translocated YopE/LLO proteins with an MOI of 10 in 2.5 ml of HBSS at 37°C. Prior to infection, bacterial overnight cultures (LB medium, 27°C) were diluted and incubated at 37°C for 4 h. After infection for 5 h, nonadherent bacteria were removed and the cells were washed with HBSS. The infection supernatant was combined with the material from the washes and centrifuged at 8,000 × g for 20 min. The pellet containing nonadherent bacteria was resuspended in 200 μl of phosphate-buffered saline (fraction of non-cell-associated bacteria). Infected BMMs were incubated for 30 min with DMEM containing 100 μg of gentamicin/ml to kill the cell-associated extracellular bacteria. The cells were then treated with 30 μg of proteinase K/ml in HBSS for 15 min at 37°C to eliminate cell surface-associated Yops. After proteinase K treatment, 3 ml of chilled HBSS containing 2 mM phenylmethylsulfonyl fluoride was added. Cells detached during the proteinase K treatment and were subsequently collected by low-speed centrifugation (600 × g for 10 min) and lysed in 1 ml of HBSS containing 0.1% Triton X-100 and 1 mM phenylmethylsulfonyl fluoride. Then, the cell lysate was centrifuged at 15,000 × g for 10 min. The supernatant was filtered through a 0.45-μm-pore-size syringe filter, and proteins were precipitated in the presence of 10% trichloroacetic acid (the fraction of Triton X-100-soluble P388D1 cell lysate containing translocated proteins).
The statistical significance of the results was analyzed using the t test at the 0.05 significance level and, if more than two experimental groups were compared, the Tukey multiple-comparison test at the 0.05 significance level (84). Calculations were performed using WINKS statistical analysis software (TexaSoft, Cedar Hill, TX).
In this study, two previously described plasmid vectors have been used (72). Plasmid pHR429 encodes the N-terminal 18 aa of YopE fused to aa 51 to 363 of LLO, resulting in a hybrid protein that contains the secretion domain but that lacks the translocation domain of YopE. Plasmid pHR430 bears the genetic information for YopE1-138/LLO51-363. This hybrid protein contains both the secretion and translocation domains of YopE. To facilitate recognition, both chimeric molecules were tagged at their C termini with an adenoviral M45 epitope (58, 72). Transcription of plasmid-borne gene fusions is mediated by the wild-type yopE promoter that is regulated by the Yersinia T3SS.
In a first set of experiments, we wanted to answer the question whether both types of hybrid YopE/LLO proteins are presented by MHC class II after exogenous loading of macrophages with these antigens. In Fig. Fig.1A1A it is demonstrated that wild-type Y. pseudotuberculosis strains pIB102(pHR429) and pIB102(pHR430) secrete similar amounts of YopE1-18/LLO or YopE1-138/LLO into bacterial culture supernatants. BMMs derived from CB6F1 mice were incubated with different concentrations of these supernatants (Fig. (Fig.1B),1B), and MHC class II antigen presentation was tested with an H-2Ab-restricted CD4 T-cell line specific for LLO190-201 (37). Loading of APCs with either chimeric YopE1-18/LLO or YopE1-138/LLO resulted in pronounced stimulation of LLO190-201-specific CD4 T cells. T-cell activation in the two experimental groups did not differ significantly.
In Y. pseudotuberculosis, three translocated T3SS proteins are known to disturb the cytoskeleton dynamics of macrophages and inhibit phagocytosis (40). YopE is a GTPase-activating protein that is active toward G proteins from the Rho family (10, 87). The tyrosine phosphatase YopH (42) disturbs focal adhesion sites by dephosphorylating the focal adhesion kinase (FAK), p130Cas, Fyn-binding protein, paxillin, and SKAP-HOM (Src Kinase-associated phosphoprotein 55 homologue) (9, 11, 45, 62). The threonine kinase YpkA is activated by actin and binds to GTP- and GDP-bound forms of Rho GTPases (31, 49).
In a next set of experiments, we investigated whether the antiphagocytic molecules YopE, YopH, and YpkA influence endosomal MHC class II antigen processing of secreted YopE/LLO. Therefore, BMMs derived from CB6F1 mice were incubated with various Yersinia deletion mutants and used as APCs. The aim of this experiment was to compare the effect of the respective yop deletion on the relative strength of antigen presentation of the secreted versus the translocated form of LLO that served as the internal control. Infection of BMMs with wild-type Y. pseudotuberculosis strain pIB102(pHR430) translocating YopE/LLO into the host cell cytosol resulted in a pronounced stimulation of LLO190-201-specific CD4 T cells (Fig. (Fig.2A).2A). In contrast, APCs infected with pIB102(pHR429) secreting YopE/LLO to the extracellular environment stimulated a significantly weaker antigen-specific CD4 T-cell response. Thus, results from Fig. Fig.2A2A confirm previous observations (72).
In contrast to the results shown in Fig. Fig.2A,2A, it is shown in Fig. 2B and C that at MOIs of 0.1 and 1, no significant differences in MHC class II antigen presentation were observed using Y. pseudotuberculosis ΔyopE and ΔyopH mutant strains pIB522 and pIB29 expressing secreted and translocated YopE/LLO, respectively. Interestingly, at an MOI of 10, this effect of YopE or YopH on exogenous MHC class II antigen presentation was not observed. However, by employing the ΔyopE ΔyopH double mutant strain pIB251 (Fig. (Fig.2E),2E), the levels of CD4 T-cell activation were comparable after secretion or translocation of YopE/LLO at MOIs of 0.1, 1, and 10. As demonstrated in Fig. Fig.2D,2D, YpkA did not contribute to the inhibition of exogenous MHC class II antigen presentation by APCs infected with Y. pseudotuberculosis. Thus, with the exception of the ΔyopE ΔyopH double-deletion mutant, T-cell activation was always more efficient after infection with the strain translocating LLO. The difference between translocated and secreted LLO was also significantly reduced in the ΔyopE and ΔyopH single-deletion mutants, indicating that expression of these genes inhibited exogenous antigen presentation.
To verify these observations, BMMs were coincubated with purified p60 from L. monocytogenes and various Y. pseudotuberculosis strains. Similar to the results obtained after coinfection with yersiniae secreting hybrid YopE/LLO proteins, the presentation of exogenous p60 was inhibited by the Y. pseudotuberculosis wild-type strain pIB102 (Fig. (Fig.3A).3A). This inhibitory activity was dependent on the presence of YopE and YopH, as demonstrated by the significantly improved presentation of exogenous p60 after coincubation with the ΔyopE ΔyopH double mutant strain pIB251. In contrast to native p60, the presentation of the cognate epitope p60301-312 was inhibited to a much lesser extent (Fig. (Fig.3B3B).
In order to visualize the influence of YopE and YopH on exogenous antigen uptake, macrophage-like P388D1 cell monolayers were incubated with fluorescently labeled OVA and with wild-type Y. pseudotuberculosis strain pIB102 or ΔyopE ΔyopH double mutant strain pIB251. Cells were fixed and processed for differential immunofluorescence staining with an antibody directed against the Yersinia LPS, as indicated in Materials and Methods. Figure Figure4,4, upper panels, shows typical images obtained by an overlay of the fluorescence signals from TRITC (extracellular Yersinia) and AMCA (intra- and extracellular Yersinia). Thus, internalized bacteria (AMCA positive and TRITC negative) appear blue, whereas extracellular yersiniae (AMCA and TRITC positive) exhibit a purple fluorescent color (mixture of blue and red). Wild-type Y. pseudotuberculosis strain pIB102 inhibited its own uptake and almost completely inhibited the engulfment of OVA-FITC by macrophage-like cells, whereas in P388D1 cells infected with pIB251, the number of endosomes labeled with OVA-FITC was comparable to that in noninfected macrophages.
In summary, these results indicate that the concerted action of YopE and YopH has a significant inhibitory impact on exogenous antigen uptake and MHC class II presentation.
In further experiments, two important questions were addressed. Does the alternative MHC class II antigen presentation of translocated hybrid LLO depend on YopE as a carrier molecule, and is this type of antigen presentation restricted to LLO? As an alternative to YopE used for cytosolic foreign antigen delivery by Yersinia, the translocated type III effector protein YopH, a tyrosine phosphatase (42), was employed. In addition to LLO, p60 was engaged as a second antigen. Thus, four plasmid vectors were constructed (Table (Table1).1). Plasmid pHR514 encodes the N-terminal 32 aa of YopH fused to residues 51 to 363 of LLO, resulting in a hybrid protein that contains the secretion domain but lacks the translocation domain of YopH. Plasmid pHR507 bears the genetic information for YopH1-138/LLO51-363. This hybrid protein contains both the secretion and translocation domains of YopH. In contrast, plasmids pHR116 and pHR119 encode the translocated and secreted YopH/p60 fusion proteins, respectively (Table (Table11).
Antigen presentation assays using Y. pseudotuberculosis wild-type strain pIB102 expressing the newly constructed chimeric YopH proteins yielded results comparable to those obtained from assays employing hybrid YopE proteins. MHC class II antigen presentation of LLO (Fig. (Fig.5A)5A) as well as p60 (Fig. (Fig.5B)5B) was dependent on YopH-mediated translocation of the respective fusion protein into the cytosol of the APCs. In contrast, pIB102 carrying plasmid pHR514 or pHR119, coding for secreted hybrid YopH1-32/LLO51-363 and YopH1-32/p60130-477 proteins, respectively, was not recognized by LLO190-201-or p60301-312-specific CD4 T cells.
Taken together these data indicate that alternative MHC class II antigen presentation of proteins translocated by the Yersinia T3SS is not restricted to a specific heterologous antigen or a specific Yop carrier molecule.
Brefeldin A has been widely used to differentiate two distinct pathways of MHC class II molecule access to antigenic peptides (47, 57, 64, 81). This inhibitor of anterograde movement from the endoplasmic reticulum to the Golgi apparatus blocks presentation of epitopes by newly synthesized class II proteins in the late endosome without influencing the process of peptide editing in which peptides are displayed by recycling class II molecules in early endosomal vesicles. Brefeldin A did not significantly alter the presentation of the synthetic LLO190-201 peptide in macrophages, as was expected on the basis of the binding of the peptide to preexisting cell surface class II complexes (Fig. (Fig.6).6). However, brefeldin A blocked LLO190-201 presentation in macrophages infected with pIB102(pHR430), indicating that processing of translocated cytosolic YopE/LLO was dependent on the binding of the LLO190-201 epitope to newly synthesized MHC class II α/β heterodimers en route to the cell surface.
We used various inhibitors of antigen processing to characterize the alternative endogenous antigen presentation pathway which allows MHC class II presentation of translocated hybrid YopE proteins in the presence of YopE/YopH-mediated inhibition of exogenous antigen presentation. For inhibition experiments, we employed the previously described Y. pseudotuberculosis yopK null mutant strain pIB155 (48) that was shown to hypertranslocate chimeric YopE/LLO (72). Enhanced presentation of hypertranslocated antigen allowed a more sensitive assessment of the activity of protease inhibitors. Processing and presentation of CD4 T-cell epitope p60177-188 or LLO190-201 after infection of macrophages with pIB155(pHR414) or pIB155(pHR430) hypertranslocating the hybrid YopE1-138/p60 and YopE1-138/LLO proteins, respectively, were very sensitive to the inhibition of endosomal acidification by chloroquin but independent of the inhibition from endosomal proteases by leupeptin (Fig. 7A and B). MHC class II antigen presentation was also not inhibited by the highly specific inhibitors of the proteasome lactacystin or epoxomycin, which selectively inhibited the presentation of the CD8 T-cell epitopes p60217-225 or LLO91-99 (Fig. 7A and B). Remarkably, MHC class II but not MHC class I antigen presentation was also significantly (~50 to 70%) inhibited in the presence of AAF-CMK, an inhibitor of TPPII. TPPII is a large cytosolic enzyme with tripeptidyl aminopeptidase and endopeptidase activity that has been implied in proteasome-independent antigen presentation (66, 89).
Efficient MHC class II presentation also occurred in macrophages derived from TAP−/− mice, indicating that endogenous antigen processing of translocated hybrid YopE proteins was TAP independent (Fig. (Fig.7C).7C). It is important to mention that similar results were obtained when the Y. pseudotuberculosis wild-type strain pIB102 was used for inhibition experiments (data not shown).
As shown in Fig. Fig.4,4, the ΔyopE ΔyopH double mutant strain pIB251 is internalized by macrophages. Therefore, it is expected that YopE1-18/LLO is secreted into the phagolysosomal vacuole and YopE1-138/LLO is translocated into the cytosol of the host cell. Similar to antigen translocated by wild-type strain pIB102 or hypertranslocating strain pIB155, MHC class II presentation of translocated YopE1-138/LLO expressed by pIB251 was not inhibited by the lysosomal protease inhibitor leupeptin (Fig. (Fig.7D),7D), indicating that the translocated antigen is the major source of MHC class II-restricted antigen presentation after infection with a strain that is capable of antigen translocation. In contrast, MHC class II presentation of YopE1-18/LLO secreted by pIB251 was inhibited by leupeptin, indicating that antigen secreted by Y. pseudotuberculosis principally can be presented via the endosomal MHC class II antigen presentation pathway. This result clearly indicates that translocated hybrid YopE proteins enter a distinctive alternative MHC class II antigen presentation pathway that is not affected by YopE/YopH-mediated inhibition of exogenous antigen uptake.
Endogenous MHC class II processing of cytosolic antigen is a long-known phenomenon and has been demonstrated for cytosolic cellular proteins as well as for various viral antigens (29, 51, 65, 70, 71). More recently, intracellular autophagy was defined as a pathway that delivers cytosolic antigens for further presentation in the context of MHC class II molecules (27, 30, 56, 60, 91). Two different pathways have been defined. While macroautophagy requires processing in the phagolysosomal compartment (56, 60), in chaperone-mediated autophagy, antigens are processed in the cytoplasm (91). The macroautophagy pathway has been characterized in detail in cells transfected with the cytosolic model antigen neomycin phosphotransferase II (NeoR) (56) as well as in cells latently infected with Epstein-Barr virus (EBV) (60). In contrast to chaperone-mediated autophagy, macroautophagy can be inhibited by the inhibitors wortmannin and 3-MA. Thus, the experimental characteristic of the macroautophagy pathway is that it is equally sensitive to inhibitors of lysosomal antigen presentation and the macroautophagy inhibitors (56, 60). Therefore, in order to investigate the possible involvement of macroautophagy in the presentation of translocated hybrid YopE/LLO, we tested the effects of two inhibitors of macroautophagy, 3-MA and wortmannin, on MHC class II antigen presentation. Each of the inhibitors was added 1 h before infection of BMMs with pIB102(pHR430) or before external loading with concentrated L. monocytogenes culture supernatant as a control antigen. As shown in Fig. Fig.8,8, both 3-MA and wortmannin significantly suppressed the recognition of the translocated hybrid YopE1-138/LLO protein but did not affect the presentation of externally loaded antigen.
To rule out any interference of the chemical inhibitors with the amount of proteins translocated into the cytosol of macrophages, BMMs were infected with Yersinia pIB102(pHR430) in the presence of all individual inhibitors used in this study. As shown in Fig. 9A and B, the inhibitors did not influence either the amount of YopE/LLO production by the bacteria (whole-cell lysates of non-cell-associated bacteria) or the amount of antigen translocated into BMMs (Triton X-100-soluble BMM lysates). It is conceivable that the treatment with 0.1% Triton X-100 results in a permeabilization of the bacteria, and therefore, the Triton X-100-soluble BMM lysate could include YopE/LLO in the bacteria internalized by these cells. Thus, YopE/LLO may not have been translocated into BMMs, as shown in Fig. Fig.9B.9B. To rule out this effect, a Y. pseudotuberculosis yopB mutant strain (pIB604) was used as a control. This strain is deficient in translocation of Yops but still synthesizes these effector proteins (44). BMMs were infected with Yersinia wild-type pIB102(pHR430) or pIB604(pHR430). As shown in Fig. Fig.9C,9C, both strains produced comparable amounts of chimeric YopE/LLO (whole-cell lysates of non-cell-associated bacteria). However, in contrast to BMMs infected with wild-type Yersinia, YopE/LLO was not detected in the Triton X-100-soluble BMM lysate from cells incubated with the yopB mutant strain (Fig. (Fig.9D9D).
In summary, the inhibition of the MHC class II-restricted presentation of translocated antigen by two inhibitors of macroautophagy and by the inhibition of lysosomal acidification together indicate that MHC class II presentation of heterologous proteins translocated by the Y. pseudotuberculosis T3SS into the cytosol of APCs depends on an alternative macroautophagy-dependent antigen presentation pathway.
In the mouse infection model, enteropathogenic yersiniae invade Peyer's patches by entering through specialized epithelial cells called M cells (41, 46). Subsequently, yersiniae disseminate to the lymph nodes, spleen, and liver, where they form microcolonies/microabscesses (5, 17). In contrast to other enteric pathogens such as Salmonella and Shigella, Yersinia is predominantly an extracellular pathogen (46, 77). Initially innate host defenses, such as polymorphonuclear leukocytes, macrophages, and natural killer cells, are involved in controlling Yersinia infection (3, 18, 50), but subsequently, a robust adaptive immune response is required to overcome Yersinia infections. Specific antibodies (86) as well as IFN-γ-producing CD4 and CD8 T cells (2, 4, 5, 13, 32) play an essential role in clearing Yersinia infections and have been shown to mediate protection in adoptive transfer experiments (2, 4, 5). It has been demonstrated that infection of rats with Y. pseudotuberculosis leads to the induction of an MHC class I-restricted cytotoxic CD8 T-cell response, but the antigens that are presented to CD8 T cells in an MHC class I context are unknown (32). However, cells infected with Yersinia have been shown to present an epitope of YopH to MHC class I-restricted CD8 T cells (80).
In the current report, we demonstrate for the first time that the concerted action of YopE and YopH from Y. pseudotuberculosis blocks the uptake and MHC class II presentation of secreted bacterial antigens. Remarkably, macrophages infected with Y. pseudotuberculosis still present epitopes in the context of MHC class II molecules after translocation of antigenic proteins into the cytoplasm of APCs by the Yersinia T3SS. This MHC class II antigen presentation of translocated proteins clearly occurs via an alternative processing pathway that involves macroautophagy and requires lysosmal acidification but is independent of the proteasome and TAP.
A number of previous studies have addressed the effect of Yops on the uptake (12, 33, 40, 68) and presentation (6) of exogenously loaded model antigens. In the present study, we extend these previous observations by directly demonstrating the effects of YopE and YopH from Y. pseudotuberculosis on the presentation of bacterially secreted model antigens. The synergistic suppressive effect of YopE and YopH on antigen uptake and presentation also confirms the findings of a previous study that suggest that Y. enterocolitica inhibits antigen degradation by dendritic cells through the interaction of multiple Yops (1). While YopE inhibits phagocytosis by its Rho GTPase-inactivating activity (10, 69, 87), the protein tyrosine phosphatase activity of YopH counteracts phagocytosis by the inhibition of early steps in the β1 integrin signaling pathway (85). It can be assumed that simultaneous inhibition of multiple RhoA activation pathways is required in order to obtain a significant inhibition of phagocytosis. However, the transport of de novo-synthesized MHC class II molecules from the endoplasmic reticulum to the phagosomal compartment and further peptide loading were not affected by YopE and YopH.
To date, endogenous MHC class II processing has been reported for cytosolic cellular proteins as well as for various viral antigens (29, 51, 65, 70, 71). A novel finding of our current study is that this immune mechanism also applies to bacterial antigens translocated into the cytoplasm of infected macrophages. MHC class II antigen presentation of cytosolic bacterial proteins is reported for two independent listerial T-cell antigens (LLO and p60) and for two different Yersinia effector proteins (YopE and YopH) used as translocated carrier proteins. Furthermore, we show that the antigen presentation pathway involved in MHC class II presentation of translocated hybrid proteins was independent of the proteasome and TAP but was inhibited by alkalinization of the phagolysosome or by blockade of the macroautophagy pathway through the action of 3-MA and wortmannin. It is known that autophagy plays an important role for the innate immune defense against intracellular bacteria such as Mycobacteria, Salmonella, and Shigella (8, 26, 43, 55, 59). In a more recent study, Deuretzbacher and colleagues analyzed the effects of Y. enterocolitica on autophagy in macrophages (28). The authors showed that autophagy was mediated by the Yersinia adhesins invasin and YadA and particularly depended on the engagement of β1 integrin receptors. Several autophagy-related events followed β1 integrin-mediated engulfment of the bacteria, including the formation of autophagosomes, processing of the marker protein LC3, redistribution of green fluorescent protein-LC3 to bacteria-containing vacuoles, and the segregation of intracellular bacteria by autophagosomal compartments. However, owing to the importance of autophagy as a host defense response, wild-type Y. enterocolitica suppressed autophagy by mobilizing type III secretion (28). It was speculated that the subversion of autophagy may be part of the Y. enterocolitica virulence strategy that supports bacterial survival when β1 integrin-dependent internalization and autophagy activation by macrophages are deleterious for the pathogen (28). Our data now provide a direct link between macroautophagy and acquired immunity against obligate extracellular yersiniae which translocate Yops into the cytoplasm of infected cells by means of their T3SSs. Obviously, suppression of autophagy during the initial interaction of macrophages and Yersinia by the concerted activities of Yops (28) does not lead to a complete impairment of autophagy. Strikingly, translocated Yops promote subversion of phagocytosis and autophagy on the one hand but are also protein substrates of remaining levels of autophagy on the other hand.
The link between macroautophagy and endogenous MHC class II antigen presentation was demonstrated in a number of recent publications (27, 30, 56, 60), and it is now also known that macroautophagy represents a process that is constitutively active in various MHC class II-expressing cells, including dendritic cells and macrophages (75). Nimmerjahn et al. have shown that endogenous presentation of an epitope derived from the cytosolic model antigen NeoR on MHC class II is mediated by autophagy and requires lysosomal processing (56). Similarly, studies with cells latently infected with EBV have shown the requirement of an autophagy- and lysosome-dependent pathway for the MHC class II-restricted presentation of EBV nuclear antigen (60). Our data clearly characterize the alternative MHC class II antigen presentation pathway of translocated chimeric YopE as being dependent on macroautophagy and lysosomal acidification. However, as MHC class II-restricted presentation of translocated hybrid YopE was not inhibited by the inhibitor of lysosomal proteases, leupeptin, we cannot exclude the possibility that antigen processing occurred independently of lysosomal processing and that phagolysosomal acidification was required just for peptide loading on MHC class II molecules in the phagolysosome. Interestingly, we present evidence that TPPII (38), a cytoplasmic protease, might be involved in endogenous MHC class II presentation of translocated chimeric YopE fusion proteins. The involvement of TPPII in cytoplasmic MHC class I processing and presentation has been the matter of some debate (7, 34, 66, 76, 89). Despite the suppressive effect of TPPII inhibition on endogenous MHC class II presentation of cytosolic hybrid YopE, we were not able to demonstrate any influence of inhibitors of the proteasome on this processing pathway. In control experiments, however, proteasome inhibitors clearly inhibited the MHC class I-restricted presentation of epitopes derived from the same YopE fusion protein. The requirement of cytoplasmic antigen processing for endogenous MHC class II antigen presentation of chimeric YopE is a characteristic that also plays a central role in the chaperone-mediated autophagy-dependent MHC class II presentation pathway described by Zhou et al. (91) and in the endogenous MHC class II presentation pathway described by Lich et al. (51), both of which involve cytoplasmic antigen presentation by the proteasome. Interestingly, compared to macroautophagy, the chaperone-dependent pathway is active in APCs showing diminished levels of macroautophagy (25, 54, 82, 90).
In the current study, we have used LLO and p60 as bacterial model antigens fused to N-terminal secretion or translocation domains of YopE or YopH in order to study the processing and presentation of defined CD4 T-cell epitopes. The inhibition of exogenous antigen presentation of these hybrid proteins after infection of cells with Y. pseudotuberculosis and the presence of an alternative endogenous MHC class II antigen presentation pathway lead to the prediction that translocated wild-type Yops bear possible CD4 T-cell epitopes and are excellent CD4 T-cell antigens. The employment of the autophagy pathway enables the host to mount an MHC class II-restricted CD4 T-cell response even in the presence of microbial immunoevasive mechanisms which suppress the presentation of microbial antigens by the exogenous antigen presentation pathway. In addition, the presence of an alternative MHC class II antigen presentation pathway for translocated bacterial Yops explains why vaccination using the Yersinia T3SS as an antigen delivery system is an excellent tool for CD8 and CD4 T-cell priming (72, 74).
We thank S. Schenk for expert technical assistance.
This study was supported by Deutsche Forschungsgemeinschaft (DFG) grants RU 838/2-2 (to H. Rüssmann, B. Köhn, and S. Jellbauer) and GE 1081/2-2 (G. Geginat).
Editor: J. B. Bliska
Published ahead of print on 27 September 2010.