In the present study, we have investigated the possibility to engage the Yersinia outer protein E (YopE) as a carrier molecule for heterologous Ag delivery by the type III secretion system of Salmonella typhimurium. Defined secretion and translocation domains of YopE were fused to the immunodominant T cell Ags listeriolysin O and p60 of Listeria monocytogenes. In vitro experiments showed that S. typhimurium allows secretion and translocation of large hybrid YopE proteins in a type III-dependent fashion. Translocation and cytosolic delivery of these chimeric proteins into host cells, but not secretion into endosomal compartments, led to efficient MHC class I-restricted Ag presentation of listerial nonamer peptides. Mice orally vaccinated with a single dose of attenuated S. typhimurium expressing translocated hybrid YopE proteins revealed high numbers of IFN-γ-producing cells reactive with listeriolysin O 91–99 or p60 217–225, respectively. This CD8 T cell response protected mice against a challenge with L. monocytogenes. In conclusion, these findings suggest that YopE is a versatile carrier molecule for type III-mediated foreign Ag delivery by Salmonella vaccine strains.
Live bacterial vaccines expressing heterologous Ags are attractive candidates to develop novel vaccination strategies (1, 2). An important requirement of such vaccines is their ability to induce a strong T cell-mediated immune response because immunity to viral and intracellular bacterial pathogens is often mediated by CTL. One of the most promising viable carrier systems is represented by attenuated Salmonella typhimurium strains (3, 4). These bacteria invade eukaryotic host cells and persist in endosomes during their entire intracellular life cycle (5). It is a well investigated and generally accepted concept in immunobiology that secreted rather than somatic Ags presented by viable carrier vaccines induce a stronger T cell priming in the vaccinated host (6). Unfortunately, the confinement of Salmonella to membrane-bound compartments complicates delivery of secreted heterologous proteins to the cytosol of host cells and subsequent MHC class I Ag presentation (7).
In an attempt to circumvent this problem, we have reported the use of the type III protein secretion system of S. typhimurium to target viral T cell epitopes directly to the cytosol of eukaryotic cells (7). Type III secretion systems are currently discovered in an increasing number of taxonomically diverse Gram-negative animal and plant pathogens (8). These systems are specialized for the export of bacterial virulence factors delivered directly into the cytosol of target cells to modulate host cellular functions (9).
In our previous study (7), mice were orally vaccinated with live replicating Salmonellae expressing the translocated type III effector protein Salmonella protein tyrosine phosphatase (SptP)4 fused to an immunodominant CD8 epitope of the murine lymphocytic choriomeningitis virus. In a subsequent challenge experiment, all vaccinated mice were protected against a lethal lymphocytic choriomeningitis virus infection. However, the use of SptP as a carrier protein for heterologous Ags was limited to deliver small protein fragments of 45–55 aa inserted in frame between two functional domains of the Salmonella type III protein. Because a versatile Ag delivery system used by attenuated S. typhimurium strains should be also capable of targeting large protein fragments derived from diverse microorganisms, we were interested in identifying a type III effector protein that could be used in Salmonella for this purpose.
Among different bacterial species, many components of type III secretion systems reveal functional conservation probably due to the fact that shared type III genes were recruited by horizontal transfer during evolution (8). Probably one of the best-studied type III effector proteins is the Yersinia outer protein E (YopE). During the interaction of Yersinia enterocolitica with professional phagocytes, YopE translocation mediates the ability of the bacteria to resist phagocytosis and to survive at extracellular sites (10, 11). Cytosolic delivery of YopE into eukaryotic cells was investigated by techniques based on a variety of reporter proteins fused to the type III effector molecule (12, 13, 14, 15). Based on results of these studies, transport signals of YopE for type III-dependent secretion and translocation have been well defined. The 23-kDa YopE molecule contains an N-terminal secretion domain of ∼11–15 aa and a translocation domain of at least 50 aa. The latter domain provides the binding site for the YopE-specific chaperone (SycE) that is required for YopE translocation. Recently, the versatility of YopE for heterologous protein delivery into the cytosol of target cells was expanded. Chaux et al. (16) used Y. enterocolitica expressing hybrid YopE/MAGE-A1 proteins for Ag translocation into dendritic cells, resulting in the induction of a proliferative cytotoxic T cell response. Moreover, for vaccination purposes, our laboratory made use of Yersinia to inject YopE fused to a heterologous immunodominant bacterial protein into the cytosol of infected host cells and to deliver foreign antigenic peptides to the MHC class I Ag presentation pathway (17).
In a previous study, Rosqvist et al. (18) have shown that full-length wild-type YopE can be secreted and translocated by S. typhimurium in a type III-dependent manner. This observation on one hand, and the promiscuity of YopE secretion and translocation domains to deliver foreign proteins to the cytosol of target cells in contrast, prompted us to investigate the possible employment of YopE for the delivery of heterologous Ags by attenuated Salmonella. In this study, we demonstrate that S. typhimurium allows secretion and translocation of chimeric YopE fused to large antigenic protein fragments of Listeria monocytogenes, which results in the induction of Ag-specific CD8 T cell responses in orally vaccinated mice and animal protection against a virulent L. monocytogenes challenge.
Materials and Methods
Plasmids, bacterial strains, and growth conditions
Escherichia coli χ6060 was used as an intermediate host for cloning procedures. The M45 epitope tag (MDRSRDRLPPFETETRIL) is derived from the E4-6/7 protein of adenovirus (19). Plasmid pSB1187 bearing the genetic information for M45 (20) was used as template DNA to amplify the epitope tag by PCR (forward primer, M45 XhoI 5′-ATACTCGAGAACATTGCTAATAAA-3′; reverse primer, M45 KpnI 5′-TATGGTACCGCTTCTGCGTTCTGA-3′). The resulting DNA fragment was cloned into the XhoI and KpnI sites of the low-copy-number vector pWSK29 (21) to obtain pHR225.
Translational fusions between different lengths of YopE and listeriolysin O (LLO) or p60 were constructed by PCR cloning procedures. All resulting protein fusions were checked by DNA sequencing. Site-specific mutagenesis of the iap gene (cysteine 396→alanine) coding for p60 was performed according to the protocol of Nelson and Long (22). Briefly, plasmid pSK50 carrying the wild-type iap gene of L. monocytogenes strain sv1/2a EGD (23, 24) was used as template DNA for the first round of PCR amplification (forward primer, SSM1-iap 5′-CCAGATGCATCAAATGTAGTTGG-3′; reverse primer, SSM2-iap 5′-GACTGACTGACTGACTGACTTTGGAATTCTTATCTCATCATTT-3′). In primer SSM1-iap, the mutated triplet (TGC→GCA) was introduced. This led to the generation of an NsiI restriction site, which was used for restriction analysis. In primer SSM2-iap, an EcoRI site and the sequence (GACT)5 were inserted to allow cloning into pSK50 and amplification of the mutated strand, respectively. The resulting 1.8-kb PCR fragment was annealed with the wild-type iap gene from pSK50 in a second round of amplification to allow elongation of the mutated strand via two cycles. In a third PCR step, forward primer SSM3-iap (5′-GACTGACTGACTGACTGACT-3′) and reverse primer SSM4-iap (5′-TTACTGAATTCTTCATCATAC-3′) were added to the reaction mix of the second PCR, and 35 cycles were run. Subsequently, the resulting PCR product was separated by agarose gel electrophoresis, excised, and purified. The DNA fragment was then cleaved with PstI/EcoRI and inserted into plasmid pTZ18R for multiplication. The 1.09-kb mutated fragment was finally excised from pTZ18R by PstI/EcoRI digestion and used to replace the corresponding wild-type PstI/EcoRI fragment within the iap gene of pSK50. The resulting plasmid coding for p60 (Cys→Ala) was used as template DNA to amplify p60 (Cys→Ala) 130–484 (forward primer, p60–130 BamHI 5′-GGAAAAACTGGATCCGTTAACGGT-3′; reverse primer, p60–484 SalI 5′-CGAACTGCTTGGTCGACAGGTTAC-3′). Plasmids pHR220 and pHR222, expressing YopE 1–18/p60 130–484 and YopE 1–138/p60 130–484, respectively (17), were digested with BamHI and SalI to remove the iap gene fragment coding for p60 130–484, and subsequently religated by insertion of the iap gene fragment coding for p60 (Cys→Ala) 130–484 to obtain plasmids pHR220-Cys and pHR222-Cys. These latter two vectors were used as templates to amplify DNA fragments encoding YopE 1–18/p60 (Cys→Ala) 130–477 and YopE 1–138/p60 (Cys→Ala) 130–477, respectively (forward primer, YopE SacII 5′-CACCCGCGGCAGACCATCAATTTG-3′; reverse primer, p60 XhoI 5′-GCCCTCGAGATATTTACCCCAGCC-3′). Amplicons were cloned into SacII and XhoI sites of pHR225 to receive in-frame fusions of YopE/p60 (Cys→Ala) with M45 epitope tag. The resulting plasmid pHR226 encodes YopE 1–18/p60 (Cys→Ala) 130–477/M45, whereas pHR242 expresses YopE 1–138/p60 (Cys→Ala) 130–477/M45. In a next cloning step, the sycE gene was amplified using pHR220 (17) as template DNA. The PCR product was cloned into SacI and SacII sites of either pHR226 to obtain pHR240 or of pHR242 to construct pHR241. Genomic DNA from L. monocytogenes strain sv1/2a EGD was used to amplify a lisA gene fragment encoding LLO 51–363 (forward primer, LLO BamHI 5′-CCTAAGACGCGGATCCCAAAG-3′; reverse primer, LLO XhoI 5′-GTCCTCGAGGTTACCGTCGATGAT-3′). Plasmids pHR240, pHR241, and pHR242 were digested with BamHI and XhoI to remove the gene fragment encoding p60 130–477, and subsequently re-ligated by insertion of the lisA-PCR fragment. Thus, plasmids pHR230, pHR231, and pHR232 bear the genetic information for chimeric YopE/LLO/M45 fusion proteins. Plasmids constructed in this study were transformed into S. typhimurium SB824 (7) by electroporation. Strain SB824 was engineered by introducing sptP::kan mutant allele from strain SB237 (25) into the ΔaroA strain SL3261 (26) by P22HTint transduction. In some experiments, wild-type S. typhimurium strain SL1344 (26) and S. typhimurium strain SB161 were used. The latter strain carries a nonpolar insertion mutation in invG (27). S. typhimurium strains were grown in Luria Bertani (LB) medium supplemented with 0.3 M NaCl, pH 7, to allow expression of components and targets of the type III secretion system encoded by Salmonella Pathogenicity Island 1 (SPI1) (28). When required, the antibiotics ampicillin (100 μg/ml) and kanamycin (30 μg/ml) were added. L. monocytogenes strain sv1/2a EGD was used for challenge experiments in Salmonella-vacci- nated mice.
Detection of secreted proteins and Western blot analysis
Secreted proteins from Salmonella culture supernatants and bacterial cell lysates were prepared and detected, as described previously (29). Briefly, bacterial supernatants were passed through a 0.45-μm-pore-size syringe filter to remove bacteria. Protein in the bacteria-free medium was precipitated by the addition of cold TCA to 10% (v/v) and incubated for 2 h on ice. The protein was collected by centrifugation at 4°C, 10,000 × g for 20 min. Pellets were washed in 0.8 ml cold acetone, dried, and resuspended in PBS buffered with 80 mM Tris-HCl, pH 8. Samples corresponding to 100 μl whole bacterial culture and 200 μl culture supernatant were separated in a 10% discontinuous SDS-PAGE, and transferred to nitrocellulose membranes, as described previously (28). Hybrid YopE/p60/M45 and YopE/LLO/M45 proteins as well as Salmonella invasion proteins B and C (SipB and SipC) were detected by immunoblot analysis. Western blots were treated with a mAb against M45 (kind gift of P. Hearing, State University of New York, Stony Brook, NY) or mAbs directed against SipB and SipC (kind gift of J. E. Galán, Yale School of Medicine, New Haven, CT), followed by incubation with an AP-labeled anti-mouse Ab. Blots were developed by using a chemiluminescent detection kit.
Tissue culture cell invasion assay
The invasion assay was conducted as previously described (30). Briefly, P388D1 cells were grown for 2 days in DMEM supplemented with 5% FBS in 24-well dishes to reach 80% confluence. Next, 1 h before the addition of bacteria, the culture medium was replaced by 500 μl HBSS. Bacteria were grown overnight for 12 h in LB supplemented with 0.3 M NaCl, diluted 1/20 in fresh medium, and grown for another 4 h under mild aeration to reach an OD600 of 0.9. P388D1 cells were infected with S. typhimurium for 2 h with a multiplicity of infection (MOI) of 10 bacteria per cell. To determine the number of intracellular bacteria, cells were washed three times with HBSS and further incubated for 3 h with DMEM containing gentamicin (50 μg/ml) to kill extracellular bacteria before lysis with 0.1% sodium deoxycholate in PBS.
Immunofluorescence analysis of hybrid YopE protein translocation
P388D1 cells were grown on glass coverslips to 60% confluence. One hour before the addition of bacteria, DMEM was replaced by 500 μl HBSS. Bacteria were grown overnight for 12 h in LB supplemented with 0.3 M NaCl, diluted 1/20 in fresh medium, and grown for another 4 h under mild aeration to reach an OD600 of 0.9. P388D1 cells were infected with S. typhimurium at an MOI of 10 for 30 min at 37°C, 5% CO2. Cells were washed three times with HBSS and further incubated for 4.5 h with DMEM containing gentamicin (50 μg/ml) to kill extracellular bacteria, thus reducing extracellular growth of Salmonella. Afterward, cells were washed again three times with HBSS and fixed in PBS, 3.7% formaldehyde. Remaining extracellular bacteria were stained with an anti-Salmonella O-1,4,5,12 polyclonal rabbit antiserum (1:400 in PBS, 3% BSA; Difco, Detroit, MI) and a secondary anti-rabbit tetramethylrhodamine isothiocyanate (TRITC) conjugate (1:100 in PBS, 3% BSA; Sigma, St. Louis, MO). After permeabilization of P388D1 cells (3 min in PBS, 0.1% Triton X-100), extra- and intracellular bacteria were stained with a polyclonal anti-Salmonella antiserum and a secondary anti-rabbit 7-amino-4-methylcoumarin-3-acetic acid (AMCA) conjugate (1:100 in PBS, 3% BSA; Jackson ImmunoResearch, West Grove, PA). Translocated hybrid YopE/LLO/M45 or YopE/p60/M45 proteins were detected using a anti-M45 mAb and an anti-mouse FITC conjugate (1:100 in PBS, 3% BSA; Sigma). Coverslips were mounted on glass slides and analyzed by fluorescence microscopy. Experiments were repeated at least three times.
Female BALB/c mice, 6–8 wk old, were purchased from Charles River WIGA (Sulzfeld, Germany). All mice were kept under specific pathogen-free conditions (positive pressure cabinet) and were provided with food and water ad libitum.
Oral immunization of mice with recombinant Salmonella and in vivo protection assay
Groups of 10 mice were orally immunized with a single dose of 5 × 108 S. typhimurium SB824 expressing various hybrid YopE proteins or a sublethal i.p. dose of 5 × 103 CFU Listeria 8 wk before the challenge infection. Eight weeks after inoculation, four mice per group were sacrificed, and spleens were used for further ELISPOT analysis. The remaining six mice per group were challenged i.v. with 1 × 103 CFU of log phase L. monocytogenes strain sv1/2a EGD in 0.2 ml PBS. Three days after the challenge, CFU were determined by plating serial dilutions of spleen homogenates on PALCAM Listeria selective agar (Merck, Darmstadt, Germany). Colonies were enumerated after 48 h of incubation. Colony counts were corrected for dilution and averaged to yield CFU/organ. The level of protection was calculated as the log10 difference of the bacterial count from immunized mice and naive control mice. Each experiment was performed at least twice with similar results.
T cell lines and Ag presentation assay
CD8 T cell lines specific for p60 217–225 (31) and LLO 91–99 (32) were derived from L. monocytogenes-infected BALB/c mice and propagated by repeated restimulation in the presence of the appropriate synthetic peptide, as described previously (33). The detection limit of the CD8 T cell lines used was between 10−11 and 10−12 M peptide. CD8 T cell activation by infected P388D1 macrophage-like APC was measured by the detection of IFN-γ in culture supernatants (34). Briefly, P388D1 cells were infected with S. typhimurium in 96-well flat-bottom microwell plates for 10 min by 200 × g centrifugation. After 2 h at 37°C, infected APC were washed and culture medium supplemented with 25 μg/ml gentamicin was added. CD8 T cells were added to each well, and after 12–18 h at 37°C, supernatants were harvested and the IFN-γ concentration was measured by means of an IFN-γ-specific ELISA that binds and detects IFN-γ with a pair of specific mAb. Results were corrected for dilution of the sample to yield the sample concentration in ng/ml.
The frequency of T lymphocytes in mice immunized with attenuated S. typhimurium was determined with an IFN-γ-specific ELISPOT assay (34, 35). Assays were performed in nitrocellulose-backed 96-well microtiter plates (Nunc, Wiesbaden, Germany) coated with rat anti-mouse IFN-γ mAb (RMMG-1; BioSource International, Camarillo, CA). Unseparated splenocytes (6 × 105/well) were stimulated for 6 h in round-bottom microtiter plates in the presence of 10−8 M peptide. Subsequently, activated cells (4 × 105/well or 4 × 104/well) were transferred to ELISPOT plates and incubated overnight. ELISPOT plates were developed with biotin-labeled rat anti-mouse IFN-γ mAb (clone XMG1.2; PharMingen, San Diego, CA), HRP streptavidin conjugate (Dianova, Hamburg, Germany), and aminoethylcarbazole dye solution. The frequency of Ag-specific cells was calculated as the number of spots per splenocytes seeded. The specificity and sensitivity of the ELISPOT assay were controlled with IFN-γ-secreting CD8 T cell lines specific for p60 217–225 and LLO 91–99. Recovery of seeded CD8 T cells was higher than 90% for both cell lines. CD4 T cells did not contribute to the fraction of IFN-γ-secreting cells. Recently, we have demonstrated that both MHC class I-restricted nonamer peptides do not stimulate LLO- or p60-specific CD4 T cells (36).
The statistical analysis of the results of in vitro experiments was performed with the Newman-Keuls multiple comparison test at the 0.05 significance level. The statistical significance of the results of in vivo experiments was checked with the nonparametric Tukey multiple comparison test at the 0.05 significance level. All tests were performed using WINKS statistical analysis software (Texasoft, Cedar Hill, TX).
Construction of hybrid YopE proteins
After invasion of target cells, it is a characteristic feature of L. monocytogenes to escape from the host phagosome into the cytosol (37). Destruction of the phagosomal membrane is mediated by the secretion of the pore-forming virulence factor LLO (38, 39). Another protein that is constitutively secreted by Listeria is a murein hydrolase designated as p60 (40). Bacterial secretion of LLO and p60 into the cytosol of infected cells directs these listerial Ags to the MHC class I Ag-processing pathway, leading to presentation of Ag-derived peptides to MHC class I-restricted CD8 T cells (41). Analysis of CD8 T cell clones from Listeria-infected mice revealed that one such clone recognizes residues 91–99 of LLO in the context of the H2-Kd MHC class I molecule (32). Further investigations indicated that p60 217–225 is presented by H2-Kd as well (31, 42). CD8 T cells specific for these epitopes were shown to transfer protective immunity against L. monocytogenes to naive mice (35).
To assess the possible employment of YopE for the delivery of two independent heterologous Ags by attenuated Salmonella, large portions of LLO and p60 consisting of >300 aa were fused to various amino-terminal parts of YopE (Fig. 1⇓). It is important to emphasize that both Ag fragments do not elicit any membrane- or bacterial cell-lysing activity (data not shown). This was achieved by deleting the N-terminal 50 and the C-terminal 165 aa of LLO (43) and by exchanging cysteine 396 of p60, which has been shown to be directly involved in cell division to alanine (44). Plasmids pHR230 and pHR240 encode the N-terminal 18 aa of YopE fused to aa 51–363 of LLO or aa 130–477 of p60 (Cys→Ala), respectively. The resulting chimeric proteins are both tagged at their C terminus with an M45 epitope and contain the secretion, but lack the translocation domain of YopE. In contrast, both plasmids pHR231 and pHR 241 bear the genetic information for hybrid YopE proteins containing the N-terminal secretion and translocation domains of YopE (Fig. 1⇓). Plasmid pHR231 encodes the N-terminal 138 aa of YopE fused to LLO 51–363/M45, whereas pHR241 encodes YopE 1–138 fused to p60 (Cys→Ala) 130–477/M45. Plasmids pHR230, pHR231, pHR240, and pHR241 also bear the sycE gene coding for SycE. To assess the influence of rSycE produced in direct vicinity to hybrid YopE proteins on the translocation and Ag presentation efficiency of LLO and p60, plasmids pHR232 and pHR242 were constructed lacking the sycE gene (Fig. 1⇓).
Salmonella type III-dependent secretion of hybrid YopE proteins
To assess the ability of S. typhimurium aroA sptP mutant strain SB824 to secrete recombinant hybrid YopE/LLO/M45 and YopE/p60/M45 proteins, strains carrying the respective plasmids were grown under conditions that allow expression of all components of the type III secretion system encoded by SPI1 (see Materials and Methods). Fig. 2⇓A reveals that all six constructed chimeric YopE proteins were stably expressed in the cytosol of SB824. Furthermore, similar amounts of these proteins could be detected in bacterial culture supernatants.
To demonstrate that the observed secretion of hybrid YopE fusion proteins is Salmonella type III dependent rather than unspecific leakage, the isogenic S. typhimurium invG mutant strain SB161 was used for in vitro secretion studies. The invG gene belongs to SPI1 and encodes an essential component of the type III secretion system (45). Thus, SB161 is deficient in secretion of Salmonella type III effector molecules. As shown in Fig. 2⇑B, no secretion of chimeric YopE by SB161 could be detected, even though all fusion proteins were found in the bacterial cytosol.
Taken together, it is demonstrated that the N-terminal 18 aa of YopE are engaged by the SPI1 type III secretion system, resulting in sufficient secretion of large antigenic protein fragments fused to YopE.
Type III-dependent secretion of hybrid YopE proteins does not alter the invasion phenotype of Salmonella
The invasion phenotype of Salmonella is crucial for colonization of the host and is mediated by components and effectors of the invasion-associated SPI1 type III secretion system. It is conceivable that concomitant secretion of hybrid YopE proteins and Salmonella type III effectors could lead to saturation of the export machinery, resulting in reduced secretion of essential Salmonella type III invasion molecules.
To investigate the possible influence of type III-dependent hybrid YopE protein engagement on secretion of SPI1 effector molecules, the secretion efficiencies of SipB and SipC by various Salmonella strains were compared. As shown in Fig. 3⇓, the nontransfected vaccine strain SB824 secreted both type III proteins into the culture supernatant. Under these in vitro conditions, it is a characteristic feature of SipC that a relatively small amount of the protein was detected in the bacterial whole cell lysate as compared with SipB (28). SB824 expressing recombinant YopE/LLO/M45 or YopE/p60/M45 produced SipB and SipC in comparable amounts as nontransfected SB824 (whole cell lysates). Moreover, coexpression of chimeric YopE did not lead to differences in secretion of SipB and SipC into culture supernatants. To further verify this observation, a comparison of the invasion competence of all SB824 used in this study was conducted. Fig. 4⇓ reveals that no statistically significant difference among the invasion competence of Salmonella wild-type SL1344, the attenuated carrier vaccine strain SB824, and SB824 expressing hybrid YopE proteins was detected. In contrast, negative control strain SB161 lacking the essential type III molecule InvG was found to be invasion defective.
These results indicate that Salmonella’s type III machinery has the ability to concomitantly engage its own effector molecules and heterologous YopE fused to antigenic protein fragments without any measurable effect on its functionality.
Cytosolic delivery of hybrid YopE proteins into APC
Infections of P388D1 cell monolayers with SB824 were conducted to study the potential of Salmonella to deliver chimeric YopE proteins into the cytosol of infected cells. Five hours after infection, cells were fixed and processed for differential immunofluorescence staining with Abs directed against M45 and Salmonella’s O-Ag, as indicated in Materials and Methods.
Fig. 5⇓, upper panel, shows typical images obtained by an overlay of the fluorescence signals from TRITC (extracellular bacteria) and AMCA (intra- and extracellular bacteria). Thus, internalized bacteria (AMCA+ and TRITC−) appear blue, whereas extracellular salmonellae (AMCA+ and TRITC+) exhibit a purple fluorescent color (mixture of blue and red). No fluorescent anti-M45 FITC signal (green color) was detected in P388D1 cells infected with Salmonella SB824 expressing YopE 1–18/LLO/M45 or YopE 1–18/p60/M45 encoded by plasmids pHR230 or pHR240, respectively (Fig. 5⇓, middle panel). In contrast, both YopE 1–138/LLO/M45 and YopE 1–138/p60/M45 gave strong fluorescent FITC signals in samples infected with SB824 (pHR231) or SB824 (pHR241). Optical sectioning and confocal microscopy revealed that these hybrid proteins were equally distributed throughout the cytoplasm of Salmonella-infected P388D1 cells (data not shown). Thus, chimeric YopE containing the secretion and translocation domain of YopE was efficiently delivered into the cytosol of host cells by internalized bacteria. Besides coding for hybrid YopE proteins, plasmids pHR231 and pHR241 also bear the genetic information for SycE (Fig. 1⇑). Translocation of hybrid YopE 1–138 was dependent on the presence of rSycE, because no visible fluorescent FITC signals were observed in cells infected with SB824 that were transformed with plasmids pHR232 or pHR242, which lack the sycE gene (Fig. 5⇓, middle panel).
The present observations suggest that, supported by the action of SycE, the N-terminal translocation domain of YopE is sufficient to facilitate delivery of large antigenic protein fragments into the cytosol of target host cells by S. typhimurium.
MHC class I-restricted Ag presentation of cytosolic hybrid YopE proteins
The ability of S. typhimurium expressing chimeric YopE/LLO/M45 or YopE/p60/M45 to deliver listerial nonamer peptides to the MHC class I-restricted Ag-presenting pathway was examined. Murine P388D1 cells were infected with various Salmonella strains and used as APC. The ability of infected cells to present the immunodominant epitopes LLO 91–99 and p60 217–225 to CD8 T cells with corresponding specificities was assessed in an in vitro Ag presentation assay by measuring secreted IFN-γ in the culture supernatant (Fig. 6⇓). P388D1 cells infected with SB824 (pHR231) expressing and translocating YopE 1–138/LLO/M45 were efficiently recognized by LLO 91–99-specific T cells (Fig. 6⇓A), whereas APC infected with SB824 (pHR230) or SB824 (pHR232) expressing and secreting hybrid YopE proteins stimulated a much weaker LLO 91–99-specific T cell response. In contrast, the negative control strain SB824 did not induce any measurable level of IFN-γ secretion into the culture supernatant. Similar results were obtained when APC infected with p60-expressing Salmonella strains were tested with the p60 217–225-specific T cell line (Fig. 6⇓B). P388D1 cells infected with SB824 (pHR240) or SB824 (pHR242) expressing and secreting, but not translocating chimeric YopE/p60/M45 proteins were less efficiently recognized by p60 217–225-specific T cells than APC infected with SB824 (pHR241), which translocates p60 into the cytosol of the host cell. Thus, Ag presentation and CD8 T cell stimulation were significantly enhanced by cytosolic delivery of the respective listerial nonamer peptide by the Salmonella type III secretion/translocation system.
Interestingly, in all Ag presentation assays, a decline in IFN-γ production at higher MOI was observed. This phenomenon is probably due to the fact that the infection of macrophages with higher numbers of Salmonella can lead to the death of phagocytes by apoptosis (46, 47).
In vivo induction of LLO- and p60-specific T cell responses by Salmonella translocating hybrid YopE proteins
The potential of attenuated S. typhimurium expressing chimeric YopE proteins to induce cytotoxic CD8 T cells in vivo was investigated. For this purpose, BALB/c mice were orally inoculated with a single dose of 5 × 108 S. typhimurium SB824 harboring the indicated plasmids. Control groups received a sublethal i.p. dose of 5 × 103 Listeria (Fig. 7⇓). Eight weeks after inoculation, ELISPOT assays were performed to determine the frequency of LLO- and p60-specific T cells in vivo. The frequency of LLO 91–99- and p60 217–225-specific CD8 T cells was calculated as the number of IFN-γ spots generated per 1 × 105 spleen cells in the presence of the corresponding synthetic peptide. Mice immunized with SB824 (pHR241) translocating YopE 1–138/p60/M45 revealed similar numbers of IFN-γ-producing cells reactive with p60 217–225 as mice infected with L. monocytogenes (Fig. 7⇓). In contrast, immunization of animals with SB824 (pHR240) secreting YopE 1–18/p60/M45 resulted in induction of a significantly lower number of p60-reactive CD8 T cells. The difference between the potential of secreted Ags on one hand and translocated Ags in contrast to elicit peptide-specific CD8 T cell priming was even more pronounced, when mice were immunized with the LLO-expressing strains SB824 (pHR231) or SB824 (pHR230). The frequency of IFN-γ-producing cells reactive with LLO 91–99 in the former group of animals indicated a prominent Ag-specific T cell response, whereas the number of LLO-specific T cells in mice of the latter group was below the detection limit of the ELISPOT assay. In summary, these experiments demonstrate that in vivo type III-mediated cytosolic delivery of chimeric YopE proteins by Salmonella results in superior efficacy of MHC class I-restricted Ag presentation as compared with secreted Ag display.
Impact of translocated vs secreted hybrid YopE proteins on vaccine-induced protection
Eight weeks after oral immunization with Salmonella expressing hybrid YopE proteins, mice were i.v. challenged with 1 × 103 L. monocytogenes. To compare the contribution of translocated vs secreted LLO and p60 on vaccine-induced protection, CFU were determined in spleens 3 days after the challenge. Spleens of uninfected mice and of animals infected with the nontransfected vaccine strain SB824 were colonized with 1.6 ± 1.2 × 105 and 1 ± 0.6 × 105 CFU of Listeria, respectively (Fig. 8⇓). In contrast, no bacteria were detected in spleens of mice that had received a sublethal i.p. dose of 5 × 103 CFU of Listeria 8 wk before the challenge infection. Mice immunized with SB824 (pHR240) or SB824 (pHR230) secreting but not translocating p60 or LLO, respectively, showed a significant (p < 0.05) decrease of the bacterial load in spleens (1.9 ± 1.8 × 104 and 1.1 ± 1.5 × 104 CFU) as compared with nonimmunized mice. However, animals orally inoculated with SB824 (pHR241) or SB824 (pHR231) translocating p60 or LLO revealed a more pronounced reduction of bacterial colonization in their spleens (0.9 ± 1.1 × 102 and 0.6 ± 0.8 × 102 CFU). Thus, in comparison with secreted Ag display, MHC class I-restricted Ag presentation and CD8 T cell stimulation of translocated listerial Ags resulted in a significantly better protection against L. monocytogenes in vivo.
To elicit suitable CD8 T cell responses, a live vaccine vector must be capable of delivering pathogen-derived antigenic peptides in the vaccinated individual in a way that they are made available for binding to MHC class I molecules on appropriate APC (48, 49). Cytosolic delivery of secreted heterologous proteins into APC by bacterial carrier vaccines and subsequent MHC class I Ag presentation can be achieved by several approaches (50). On one hand, the use of L. monocytogenes for vaccination purposes has the advantage that the vaccine carrier is already located in the cytoplasm of the infected host cell (51, 52), which facilitates MHC class I processing of secreted heterologous Ags. In contrast, endosomal-bound Salmonella or extracellular Yersinia translocating foreign proteins into the cytosol of APC via their type III secretion systems are attractive alternative bacterial species for the use of live vaccine vehicles (7, 17).
The present study describes the efficient protection against the intracellular pathogen L. monocytogenes by oral vaccination with attenuated S. typhimurium translocating two defined listerial Ags, LLO and p60, via the SPI1 type III secretion system. The first major finding of our experiments is that the N-terminal translocation domain of the Yersinia type III effector molecule YopE is engaged by Salmonella’s type III secretion system, resulting in excellent secretion and translocation of large heterologous antigenic proteins. In Yersinia, N-terminal YopE transport domains were shown to be sufficient for secretion and translocation of fused reporter proteins. Sory et al. (12) used recombinant Y. enterocolitica expressing truncated YopE proteins fused to a calmodulin-dependent adenylate cyclase of Bordetella pertussis, resulting in an increase in cAMP in the eukaryotic cytosol. In another approach, Lee et al. (14) constructed hybrid YopE-neomycin phosphotransferase proteins to study type III effector protein targeting. Furthermore, Jacobi et al. (15) used the green fluorescent and the firefly luciferase proteins fused to various residues of YopE to investigate translocation. However, this is the first report that describes the use of YopE for heterologous Ag delivery in Salmonella. Use of YopE by S. typhimurium’s type III secretion system allows secretion and translocation of large antigenic proteins, thus avoiding limitations to single epitopes, which might occur using Salmonella’s type III effector molecule SptP for vaccination purposes (7). Remarkably, the concomitant secretion of hybrid YopE proteins and Salmonella type III effectors did not lead to saturation of the export machinery that would result in reduced secretion of essential Salmonella type III invasion molecules. Impaired invasion and persistence of Salmonella is not a desirable feature of a carrier vaccine strain because these phenotype alterations are clearly associated with reduced immunogenic properties.
By using differential immunofluorescence staining of Salmonella-infected APC, we demonstrate that LLO and p60 fused to the N-terminal translocation domain of YopE comprised of 138 aa were efficiently delivered into the cytosol of P388D1 cells. Protein translocation resulted in efficient listerial Ag presentation and peptide-specific CD8 T cell stimulation in vitro. In contrast, no cytosolic delivery of these hybrid YopE proteins was observed in the absence of recombinant, plasmid-borne SycE. SycE is a small homodimeric protein that interacts with YopE in the Yersinia cytoplasm (53), and this interaction is absolutely required for YopE translocation (14), whereas YopE secretion is not affected. Obviously, SycE serves a similar essential purpose when expressed in Salmonella. Hybrid YopE proteins containing the N-terminal secretion domain (aa 1–18), but lacking the translocation domain, could not be detected in the cytosolic compartment of APC as well. The presence of secreted but nontranslocated chimeric YopE proteins in the Salmonella-containing endosomal compartment could not be visualized by immunofluorescence staining. It is tempting to speculate that after secretion, a rapid degradation of these proteins occurs. As compared with translocation, hybrid YopE protein secretion resulted in significantly less efficient, but measurable listerial Ag presentation and weak peptide-specific CD8 T cell stimulation in vitro. Obviously, a minimal portion of secreted but nontranslocated chimeric YopE was introduced from the endosome to the MHC class I processing pathway. Whether this is due to endosomal loading of empty MHC class I molecules (54) with listerial peptides after rapid degradation of hybrid YopE proteins, or release of these peptides into the cytosol of APC with subsequent classical MHC class I Ag processing and presentation (55) remains to be elucidated.
The second major finding of our study is that in accordance with the in vitro results, translocated rather than secreted Ags elicited a much more pronounced peptide-specific CD8 T cell priming in immunized mice, resulting in the induction of stronger protective immunity. As determined by ELISPOT assays, animals that were orally inoculated with SB824 (pHR241) expressing translocated hybrid YopE/p60/M45 protein revealed a significantly higher number of IFN-γ-producing splenocytes reactive with p60 217–225 than mice immunized with SB824 (pHR240) expressing secreted but nontranslocated hybrid YopE/p60/M45 protein. The effect that a nontranslocated listerial Ag induces a weak but measurable induction of peptide-specific CD8 T cells in vivo was observed for p60, but not for LLO. However, in vivo protection assays showed that oral vaccination of mice with either Salmonella strain SB824 (pHR230) or SB824 (pHR240) led to a slight but significant decrease of the bacterial load in spleens after challenge with Listeria as compared with control mice. It remains to be elucidated whether this low-level protection could be also due to the induction of LLO- or p60-specific CD4 T cells by secreted hybrid YopE comprised of >300 aa of the respective listerial Ag. In fact, adoptive transfer experiments revealed that p60-specific Th1 clones mediate significant protection against L. monocytogenes infection (56). In contrast, Salmonella species have been reported to inhibit activated macrophages in their ability to present peptides from homologous or heterologous Ags in the context of MHC class II molecules (57). In further studies, we will focus on aspects of MHC class II-restricted immune responses and CD4 T cell priming elicited by type III-mediated Ag display by recombinant Salmonella.
However, a much stronger impact on vaccine-induced protection was achieved by immunization of mice with SB824 (pHR231) or SB824 (pHR241), which deliver chimeric YopE/LLO or YopE/p60 proteins directly to the cytosol of infected host cells, thus mimicking the situation of a L. monocytogenes infection. In additional experiments, we were interested in investigating how long this protection against listeriosis is able to persist. Therefore, 100 days after a single immunization with Salmonella vaccine strains, ELISPOT assays and challenge infections with Listeria were conducted as described above (data not shown). Interestingly, the results of these experiments showed no significant differences than the data obtained from animals 8 wk after immunization. Thus, in contrast to secreted Ag display, translocation of chimeric YopE by recombinant Salmonella induced prolonged protective immunity against listeriosis based on prominent Ag-specific CD8 T cell priming.
Remarkably, in mice orally immunized with a single dose of Salmonella, type III-dependent translocation and cytosolic delivery of hybrid YopE proteins into host cells resulted in the induction of prominent LLO- and p60-specific T cell responses, which were comparable with the levels of peptide-specific CD8 T cell priming in animals immunized with a sublethal dose of wild-type Listeria. This observation emphasizes the efficacy of type III-mediated induction of MHC class I-restricted immune responses. Obviously, recombinant Salmonella translocating a truncated version of a single immunodominant Ag from Listeria did not elicit the same protective ability as an immunizing sublethal dose of wild-type L. monocytogenes, which naturally displays a variety of listerial peptides from different Ags to CD8 T cells of the vaccinated host. Despite that it has been shown that selected or cloned CD8 T cells from Listeria immune mice are able to adoptively transfer antilisterial protection (6), for a potent vaccine candidate against listeriosis it might be necessary to display several pathogen-derived Ags. In future studies, we will investigate whether type III-mediated dual translocation of LLO and p60 by a single Salmonella vaccine strain enhances the protective ability against listeriosis.
In the experimental setup described in this study, mice were immunized with attenuated Salmonella expressing a single immunodominant listerial Ag, either LLO or p60, which led to significant protection of animals against a Listeria challenge infection. Similar results were obtained by Hess et al. (6), who used the E. coli hemolysin export apparatus to secrete LLO or p60 by recombinant Salmonella. In this study, the authors compared the efficacy of secreted vs somatic listerial Ag display. Secretion of LLO or p60 resulted in prominent peptide-specific CD8 T cell priming. This discrepancy to our study is probably due to the fact that we took special care of constructing hybrid YopE/LLO and YopE/p60 proteins that lack any membrane-lysing activity (43, 44). By truncating N- and C-terminal domains of the pore-forming protein LLO and by exchanging cysteine 396 to alanine of the autolysin p60, we were able to reduce secondary leakage from the endosomal compartment to the cytosol of infected cells. In our hands, type III-mediated translocation of antigenic hybrid proteins is mandatory for efficient MHC class I-restricted Ag presentation elicited by Salmonella carrier vaccines.
Taken together, we demonstrate for the first time that a type III effector molecule from Yersinia can be used as a carrier protein for translocation of large Ags by Salmonella’s type III secretion system. Our study shows that efficient stimulation of peptide-specific CD8 T cells and protection against listeriosis require cytosolic delivery of the respective listerial Ag. By engaging well-defined secretion and translocation domains of YopE, targeting of Ags into different subcellular compartments of APC can be orchestrated, which might expand the use of attenuated S. typhimurium strains for future oral vaccination strategies.
We thank J. E. Galán (Yale School of Medicine, New Haven, CT) for helpful discussions and for providing Salmonella mutant strains as well as mAbs. Also, we thank P. Hearing (State University of New York, Stony Brook, NY) for providing M45 mAb.
↵1 H.R. was supported by the AIDS-Stipendienprogramm from the Bundesministerium für Bildung, Wissenschaft, Forschung, und Technologie of Germany, and by the Deutsche Forschungsgemeinschaft (Schwerpunktprogramm “Neue Vakzinierungsstrategien”). W.-D.H. received grants from the Deutsche Forschungsgemeinschaft and Volkswagen- Stiftung.
↵2 Address correspondence and reprint requests to Dr. Holger Rüssmann, Max von Pettenkofer-Institut für Hygiene und Medizinische Mikrobiologie, Ludwig Maximilians Universität München, Pettenkoferstrasse 9a, 80336 München, Germany. E-mail address:
↵3 Current address: Microbiological Analytics, Merck KGaA, 64271 Darmstadt, Germany.
4 Abbreviations used in this paper: SptP, Salmonella protein tyrosine phosphatase; AMCA, 7-amino-4-methylcoumarin-3-acetic acid; LB, Luria Bertani; LLO, listeriolysin O; MOI, multiplicity of infection; Sip, Salmonella invasion protein; SPI1, Salmonella pathogenicity island 1; SycE, YopE-specific chaperone; TRITC, tetramethylrhodamine isothiocyanate; YopE, Yersinia outer protein E.
- Received January 8, 2001.
- Accepted April 18, 2001.
- Copyright © 2001 by The American Association of Immunologists