Delivery of Epitopes by the Salmonella Type III Secretion System for Vaccine Development

See allHide authors and affiliations

Science  24 Jul 1998:
Vol. 281, Issue 5376, pp. 565-568
DOI: 10.1126/science.281.5376.565


Avirulent strains of Salmonella typhimurium are being considered as antigen delivery vectors. During its intracellular stage in the host, S. typhimurium resides within a membrane-bound compartment and is not an efficient inducer of class I–restricted immune responses. Viral epitopes were successfully delivered to the host-cell cytosol by using the type III protein secretion system ofS. typhimurium. This resulted in class I–restricted immune responses that protected vaccinated animals against lethal infection. This approach may allow the efficient use of S. typhimuriumas an antigen delivery system to control infections by pathogens that require this type of immune response for protection.

The success of global vaccination programs requires efficacious vaccines that are stable and easy to administer (1). Viable carrier systems offer the greatest potential for innovative approaches to develop polyvalent vaccines. Efficient protection against infectious agents often requires the action of both humoral and cellular immune mechanisms. Therefore, an ideal polyvalent antigen delivery system should be capable of stimulating all desired effector cell populations of the immune system. Live replicating bacteria and viruses that stimulate complex immune responses have been rendered avirulent and endowed with the ability to express foreign proteins derived from pathogenic microorganisms (2). Avirulent strains of Salmonella typhimuriumare being widely considered as delivery systems for heterologous antigens because of their ability to induce complex mucosal and systemic immune responses after oral administration (3). A characteristic feature of these bacteria is their ability to invade nonphagocytic cells such as those of the intestinal epithelium (4). After internalization, S. typhimuriumremains confined to a membrane-bound compartment insulated from the cytosolic environment of the host cell (5). Localization within the “internalization” vacuole prevents delivery of expressed foreign antigens to the major class I antigen presentation pathway, thereby hampering the use of Salmonella vaccine carriers when this type of response is crucial for protection (for example, viral infections) (6). An attempt to circumvent this problem has been the use of Salmonella to deliver plasmid DNA to express antigens within the host-cell cytosol (7).

Contact of S. typhimurium with host cells results in activation of a specialized protein secretion system (type III) that is encoded in a pathogenicity island at centisome 63 of its chromosome (4). This protein secretion system delivers a set of bacterial effector proteins into the host-cell cytosol, which leads to stimulation of signal transduction pathways that result in a variety of responses such as actin cytoskeleton reorganization and activation of transcription factors (4). In an effort to improve the ability of Salmonella to elicit class I–restricted immune responses to those epitopes, we investigated the potential of this system to deliver heterologous epitopes into the host-cell cytosol. To this end, we chose SptP, a S. typhimurium effector protein that is delivered into the host cell through the centisome 63 type III secretion system but is not required for efficient bacterial entry into nonphagocytic cells (8). We constructed a chimeric form of SptP that carries a class–I restricted epitope consisting of residues 366 to 374 from the influenza virus nucleoprotein (IVNP366–374) found to be immunodominant in mice of theH-2b haplotype (9). The epitope was introduced at a permissive site of SptP (10) located between the two predicted independent domains of this protein (Fig. 1) (8). The chimeric SptP-IVNP366–374 protein was secreted into the culture supernatant of both wild-type S. typhimurium and the isogenic avirulent aroA sptP mutant strain SB824 at concentrations indistinguishable from those of wild-type SptP (Fig. 1). Both strains efficiently delivered SptP-IVNP366–374 into the cytosol of infected cultured epithelial cells (Fig. 1). In contrast, and as expected, the isogenic S. typhimurium sipDmutant strain SB221 did not translocate the chimeric protein into the host-cell cytosol, although it could secrete SptP-IVNP366–374 to the infection medium (Fig. 1). SipD is essential for type III protein translocation into host cells, although it is completely dispensable for protein secretion (8, 11). Similar results were obtained when a different epitope derived from the lymphocytic choriomeningitis virus nucleoprotein (LCMVNP118–126) was introduced at the same site of SptP (Fig. 1). Thus, SptP can serve as a molecular courier to deliver foreign peptides of immunological interest to the cytoplasm of target cells.

Figure 1

Translocation of SptP chimeric proteins into cultured cells. (A) Henle-407 cells were infected with differentS. typhimurium strains carrying the indicated plasmids and the presence of SptP fusion proteins in the different fractions examined as described in (15). Lane 1, whole cell lysate of non-cell–associated bacteria; lane 2, bacteria-free infection medium; lane 3, Triton X-100 insoluble fraction containing internalized bacteria; lane 4, Triton X-100 soluble Henle-407 cell lysate containing translocated proteins. SptP and fusion derivatives were detected by protein immunoblotting with a monoclonal antibody to SptP. The relevant phenotype of the infecting strains is indicated at the bottom of each panel. (B) Henle-407 cells were infected with wild-type S. typhimurium or the isogenic translocation defective, hypersecreting sipD mutant strain carrying the plasmid pSB762, which encodes SptP-IVNP366–374. Infected cells were then stained with a monoclonal antibody to SptP as described in (15). Micrographs of the SptP stained cells and the corresponding DIC images are shown. Similar results were obtained with the same strains carrying pSB775.

We then examined the ability of wild-type, aroA, andaroA sptP S. typhimurium expressing the chimeric protein SptP-IVNP366–374 to deliver the influenza nucleoprotein (NP) epitope to a class I–restricted antigen presenting pathway. Murine RMA thymoma cells (H-2b ) were infected with the different S. typhimurium strains and the ability of the infected cells to present the influenza NP epitope to the class I–restricted T cell hybridoma 12.164 was assessed by measuring its interleukin-2 (IL-2) secretory response (12, 13) (Fig. 2). The RMA thymoma cells infected withS. typhimurium strains expressing SptP-IVNP366–374 were efficiently recognized by the epitope-specific T cell hybridoma. Antigen presentation was strictly dependent on the cytosolic delivery of the epitope by the S. typhimurium type III system, as RMA cells infected with asipD mutant strain, which can secrete SptP-IVNP366–374 but cannot deliver it into the cell cytosol (Fig. 1), did not stimulate the T cell hybridoma (Fig. 2). Salmonella typhimurium strains expressing a deletion mutant of SptP-IVNP366–374(SptPΔ35–62-IVNP366–374) lacking the binding site for its specific chaperone, and therefore efficiently secreted but not translocated into the host cell (14), did not present antigen in infected RMA cells (Fig. 2). Similarly, the COOH-terminal half of the influenza virus NP fused to InvJ (InvJ-IVNP335–498), a protein substrate of the type III system that is secreted but it is not translocated into host cells (15, 16), also did not present antigen in infected RMA cells (Fig. 2). Salmonella typhimurium is therefore capable of delivering foreign epitopes to the class I antigen presenting pathway with proteins translocated by the type III secretion system.

Figure 2

Antigen presentation by RMA (A) or TAP-deficient RMA-S (B) cells infected with S. typhimurium. RMA or RMA-S cells were infected with wild-type, aroA, aroA sptP, or sipD S. typhimurium strains bearing a plasmid expressing SptP-IVNP366–374 or SptP-LCMV NP118–126before exposure to 12.164 hybridoma cells. The concentration of IL-2 secreted into the medium was determined by ELISA (13). When indicated, control RMA and RMA-S cultures were treated with 1 μM IVNP366–374 peptide or LCMV NP118–126 peptide or were infected with influenza virus A/PR/8/34. Error bars indicate SEM. Results are representative of three independent experiments.

Peptides delivered to the cytosol by S. typhimurium required a functional peptide transporter system (TAP) for their transfer to the endoplasmic reticulum and loading by class I molecules, because TAP2-defective RMA-S mutant cells (17) infected withS. typhimurium SptP-IVNP366–374 were markedly impaired in their capacity to stimulate the T cell hybridoma compared with wild-type RMA cells (Fig. 2). This result provides independent evidence that peptides displayed at the RMA cell surface by class I molecules occupied a cytosolic compartment before they were transferred into the endoplasmic reticulum.

We then examined the potential of avirulent S. typhimuriumstrains expressing SptP-IVNP366–374 to induce cytotoxic T lymphocytes (CTLs) in vivo. C57BL/6J mice orally inoculated with an avirulent S. typhimurium sptP aroA mutant strain expressing SptP-IVNP366–374 developed CTL precursors that lysed target cells infected with influenza virus or loaded with the synthetic peptide NP366–374 (Fig. 3A) (18). In contrast, and consistent with the in vitro antigen presentation results, mice inoculated intragastrically withSalmonella expressing either SptP-LCMVNP118–126 or InvJ-IVNP335–498failed to stimulate CTLs directed to influenza virus epitopes (Fig. 3A). Thus, cytoplasmic targeting of the type III system hybrid substrates dictates the outcome of the class I–restricted response to a foreign epitope: pronounced CTL responses were induced by theS. typhimurium mutant strain expressing SptP-IVNP366–374 but not by the strain expressing the nontranslocated InvJ-IVNP335–498 chimeric protein.

Figure 3

(A) Lysis of IVNP366–374 peptide-sensitized EL-4 target cells by restimulated splenocytes from mice orally immunized with the aroA sptP S. typhimurium strain SB824 expressing SptP-IVNP366–374, an irrelevant peptide (SptP-LCMVNP118–126), or InvJ-IVNP335–498(18, 20). Control mice were vaccinated intraperitoneally with live influenza virus A/PR/8/34 or recombinant vaccinia virus expressing influenza nucleoprotein (vaccinia IVNP)(25). Effector-to-target cell ratios (E:T) are indicated. Error bars indicate SEM. Results are representative of three independent experiments. Open squares, SptP-IVNP366–374; open circles, SptP-LCMVNP118–126; closed squares, vaccinia IVNP; open triangles: influenza virus strain A/PR/8/34; closed triangles, InvJ-IVNP335–498. (B and C) Immune response of mice immunized with avirulent S. typhimurium expressing SptP-LCMVNP118–126. (B) Survival of vaccinated mice after LCMV intracerebral challenge inoculation (20). Groups of BALB/c mice were alternatively orally vaccinated with the avirulent aroA sptP S. typhimurium strain SB824 expressing SptP-LCMVNP118–126 or SptP fused to an irrelevant epitope (SptP-IVNP366–374), an avirulent S. typhimurium sipD strain, defective in translocation but not in secretion of SptP-LCMVNP118–126, or mock vaccinated with broth. Open squares, aroA(SptP-LCMVNP118–126) (n = 8); closed squares, aroA (SptP-IVNP366–374) (n= 7); closed triangles, sipD(SptP-LCMVNP118–126) (n = 4); closed circles, mock control (n = 6). (C) Percent specific lysis by in vitro restimulated splenocytes from mice inoculated with aroA sptP S. typhimurium expressing SptP-LCMVNP118–126 or SptP fused to an irrelevant peptide (SptP-IVNP366–374). Control mice were vaccinated by intraperitoneal inoculation with live LCMV. Error bars indicate SEM. Results are representative of three independent experiments. Open squares, aroA (SptP-LCMVNP118–126); closed triangles, aroA (SptP-IVNP366–374); closed squares, LCMV.

To assess the ability of the type III–mediated S. typhimurium antigen delivery system to induce protective class I–restricted CTLs, we chose murine lymphocytic choriomeningitis virus (LCMV) infection because CTLs play a dominant role in protection against this viral disease (19). In this model, intracerebral inoculation with LCMV results in lethal choriomeningitis, which can be prevented by a single clonal population of LCMV-specific CTLs (19). BALB/c mice immunized intragastrically with thearoA sptP S. typhimurium mutant strain expressing SptP-LCMVNP118–126 were completely protected against lethal intracerebral challenge with a virulent strain of LCMV (20). In contrast, mice immunized intragastrically withS. typhimurium aroA mutant strain expressing an irrelevant epitope (SptP-IVNP366–374) succumbed to the same challenge (Fig. 3B). Consistent with the hypothesis that protection was mediated by an H-2–restricted immune response, the same S. typhimurium aroA mutant strain expressing SptP-LCMVNP118–126 did not protect C57BL/6J mice against an identical infection with LCMV (21). Furthermore, protection required cytosolic delivery of the epitope by the type III secretion system, as vaccination with aS. typhimurium sipD strain defective in translocation but not in secretion of SptP-LCMVNP118–126did not protect mice against LCMV infection (Fig. 3B). Induction of protective immunity in mice by Salmonella vaccination was correlated with the presence of LCMV-specific CTLs although quantities of splenic CTL precursor were lower in mice vaccinated withSalmonella than in mice infected with a sublethal dose of LCMV (Fig. 3C).

We show here that delivery of epitopes through the S. typhimurium type III secretion system results in efficient stimulation of class I–restricted protective antiviral immune responses. Use of this system will expand the efficient use of S. typhimurium as a carrier of heterologous antigens to vaccinate against infections in which this type of response is crucial for protection (22). In addition, avirulent S. typhimurium expressing tumor-specific antigens may allow use of this system for treatment of neoplastic diseases by induction of cancer cell–specific class I–restricted CTLs (23).

  • * Present address: Max von Pettenkofer-Institut, Pettenkofer Strasse 9a, 80336 Munich, Germany.

  • To whom correspondence should be addressed. E-mail: galan{at}


View Abstract

Stay Connected to Science

Navigate This Article