OUP user menu

Protection against murine tuberculosis by an attenuated recombinant Salmonella typhimurium vaccine strain that secretes the 30-kDa antigen of Mycobacterium bovis BCG

Jürgen Hess, Leander Grode, Jacqueline Hellwig, Peter Conradt, Ivo Gentschev, Werner Goebel, Christoph Ladel, Stefan H.E. Kaufmann
DOI: http://dx.doi.org/10.1111/j.1574-695X.2000.tb01441.x 283-289 First published online: 1 April 2000


A recombinant (r-) Salmonella typhimurium aroA vaccine that secretes the naturally secreted protein of Mycobacterium bovis strain BCG, Ag85B, by means of the HlyB/HlyD/TolC export machinery (termed p30 in the following) was constructed. In contrast to r-S. typhimurium control, oral vaccination of mice with the r-S. typhimurium p30 construct induced partial protection against an intravenous challenge with the intracellular pathogen Mycobacterium tuberculosis, resulting in similar vaccine efficacy comparable to that of the systemically administered attenuated M. bovis BCG strain. The immune response induced by r-S. typhimurium p30 was accompanied by augmented interferon-γ (IFN-γ) and tumor necrosis factor-α (TNF-α) levels produced by restimulated splenocytes. These data suggest that the HlyB/HlyD/TolC-based antigen delivery system with attenuated r-S. typhimurium as carrier is capable of inducing an immune response against mycobacterial antigens.

  • Recombinant Salmonella
  • Tuberculosis vaccine
  • Protection

1 Introduction

Tuberculosis (TB) continues to be a major cause of morbidity and mortality throughout the world, resulting in 8 million new cases and 2 million deaths each year [1]. The only TB vaccine currently available is the attenuated Mycobacterium bovis strain BCG, which has variable protective efficacy ranging from 0 to 85% in different controlled studies [2,3]. Therefore, the development of more potent TB vaccines is urgently required [4]. The causative agent of TB, Mycobacterium tuberculosis, is a facultative intracellular microorganism that misuses macrophages as its preferred habitat. The altered micro-environment in M. tuberculosis-containing phagosomes and the defects in maturation processes along the endosomal—lysosomal continuum cause impaired processing of mycobacterial antigens and diminished cell surface expression of major histocompatibility complex (MHC) class II molecules, which in turn reduces antigen recognition by immune CD4 T cells [58]. Nevertheless, the important role of CD4 T cells in the control of TB is unquestionable. These protective CD4 T cells are of the Th1 type, i.e. they are potent producers of interferon-γ (IFN-γ) (which induces anti-mycobacterial activities in macrophages [9]).

Secreted proteins of M. tuberculosis are recognized early in the course of experimental TB infection of mice and by T cells of TB patients, and they are therefore considered preferential antigenic targets for vaccine development [4,10]. The antigen 85 complex is expressed by nearly all mycobacterial species analyzed so far [11]. More importantly, p32A (antigen 85A or 32-kDa antigen) and p30 (antigen 85B or 30-kDa antigen) stimulate B and T cell responses in TB patients [1214], and immunization of guinea pigs with a p30 antigen preparation together with adjuvant induced protective immunity against TB [15,16]. By means of DNA vaccination, both antigens, p30 and p32A, induced protection against aerosol challenge with M. tuberculosis in mice and guinea pigs, respectively [1719]. The proteins of the antigen 85 complex, including p30 and p32A proteins of M. tuberculosis belong to a family of gene products with fibronectin-binding capacity and mycolyltransferase activity, which is involved in the final stages of mycobacterial cell wall assembly [11,20]. Sequence comparison revealed that p30 and p32A of M. tuberculosis exhibit 72.8% amino acid identity and that the cross-species homology between p30 of BCG and of M. tuberculosis is 99% [2124].

BCG, M. tuberculosis and Salmonella typhimurium microbes persist within the phagosomes of infected macrophages [25]. From an immunological point of view, the intracellular habitat of these microbes influences antigen processing mechanisms of MHC class I and II pathways, ultimately leading to CD8 and CD4 T cell responses, respectively [25]. Intraphagosomal localization preferentially stimulates MHC class II-restricted CD4 T cells whilst egression into the cytoplasm as achieved by Listeria monocytogenes primarily induces MHC class I-restricted CD8 T cells [25]. Yet, the recently described S. typhimurium aroA strains secreting heterologous antigens stimulate CD4 and CD8 T cell-mediated immunity [2527]. Central to this mode of antigen delivery is the secretion of heterologous fusion proteins by means of the plasmid-encoded HlyB/HlyD/TolC translocator of hemolytic and uropathogenic Escherichia coli strains [25,28]. The latter directs protein secretion via recognition of the C-terminal hemolysin (HlyA) signal peptide across the cell wall of Gram-negative E. coli or Salmonella sp. microorganisms [28]. The recombinant (r-) Salmonella-specific antigen delivery system expressing defined antigens of L. monocytogenes induced both CD4 and CD8 T cells, which adoptively conferred protective immunity against listeriosis in mice [26]. This Salmonella-based antigen delivery device is gaining interest as a possible r-carrier against intracellular bacterial pathogens like M. tuberculosis[25,29].

A fusion protein was constructed encompassing p30 of M. bovis BCG and the C-terminal part of HlyA of E. coli, respectively. Here, we report on the in vivo immunogenicity of S. typhimurium p30 carrier strain and on partial protection against TB in C57BL/6 mice vaccinated orally with this r-construct secreting p30 of M. bovis BCG.

2 Materials and methods

2.1 Bacteria and plasmids

The S. typhimurium SL7207 strain (2337–65 (WRAY) hisG46ΔaroA407 (TcsaroA544::Tn10)) was kindly provided by Dr. B.A.D. Stocker (Department of Medical Microbiology, Stanford University, Palo Alto, CA, USA) [30]. The r-S. typhimurium strain pMOhly1 (termed control) has been described previously [26]. M. bovis BCG strain Chicago (ATCC 27289) and M. tuberculosis H37Rv (originally obtained from Dr. J.K. Seydel, Forschungsinstitut Borstel, Germany) were passaged in mice and then in Dubos broth base (Difco) supplemented with 10% Dubos medium albumin (Difco) and stored in aliquots at −70°C. The mycobacteria—E. coli shuttle vector pAB261 was kindly provided by MedImmune (Gaithersburg, MD, USA).

2.2 Construction of recombinant plasmid pAg85Bs

Plasmid pAg85Bs was constructed by the following cloning steps: BglII–HindIII DNA fragment of pANN202–312 (position 3811 and 5341 [31]) was inserted into pAB261 BamHI/HindIII ([21,32]; MedImmune, personal communication), the resulting plasmid was digested with NcoI and SwaI, and the obtained DNA fragment was cloned into pILH-1 vector cut by NcoI and SwaI restriction enzymes [26,33]. The final plasmid pAg85Bs was isolated from E. coli JM109, analyzed and electroporated in S. typhimurium LB5000, a restriction-negative and modification-proficient strain, by a standard protocol for E. coli (Bio-Rad, CA, USA). Subsequently, plasmids purified from ampicillin-resistant r-S. typhimurium LB5000 colonies were introduced into S. typhimurium SL7207 by electroporation. The resulting r-S. typhimurium p30 strain harboring plasmid pAg85Bs was grown in 2×yeast tryptone medium supplemented with 100 µg ml−1 of ampicillin (Sigma), 10 µg ml−1 each of 2,3-dihydroxybenzoic acid (Sigma, Deisenhofen, Germany) and p-aminobenzoic acid (Sigma, Germany) at 37°C. The r-S. typhimurium p30 or control were grown to an optical density (A600) of one. The bacteria were harvested and centrifuged at 3000×g. Each bacterial pellet was then resuspended in phosphate-buffered saline (PBS)/0.05% Tween 80. Acetone (−20°C) was added to the supernatant at a final concentration of 50% and subsequently stored at −20°C for 1 h. After this treatment, the solution was centrifuged for 30 min at 10 000×g and subsequently, the pellet was resuspended in water. The proteins were completely resuspended in water at the concentration of 8 mg ml−1. These antigen preparations of r-S. typhimurium p30 and control were adjusted to 2 µg protein per µl volume. Proteins prepared under reducing conditions of 7.5% sodium dodecyl sulfate (SDS) and separated by 10% SDS—polyacrylamide gel electrophoresis were electrophoretically transferred to a Hybond ECL nitrocellulose membrane, which was subsequently blocked overnight by incubation with PBS containing 1% bovine serum albumin. Immunostaining was performed with rabbit anti-p30 antibodies (1:5000) and anti-rabbit IgG antibodies conjugated with horseradish peroxidase (1:20 000). The Western blot was developed by means of an enhanced chemiluminescence detection kit according to the manufacturer's description (ECL Western blotting detection reagents, Amersham Pharmacia). For the generation of p30 antibodies, Specol adjuvant (Institute for Animal Science and Health, Lelystad, The Netherlands) together with the peptide H2N-GPSSDPAWERNDPT-CONH2 coupled at the N-terminus to keyhole limpet hemocyanin (Eurogentec, Seraing, Belgium) was used to immunize rabbits by a standard protocol (Eurogentec). The animals were immunized with 200 µg peptide per 1 ml adjuvant (10 injections with 100 µl each time point) by the intradermal route at days 0, 14, 28 and 56. Subsequently, the anti-p30 antibodies were enriched by binding to Protein A-Sepharose and then purified further by affinity chromatography on activated Sepharose (Amersham Pharmacia, Freiburg, Germany) conjugated with antigen 85 complex proteins p32A/p32B/p30 of M. tuberculosis H37Rv strain (generously obtained from Dr. J. Belisle, Fort Collins, USA).

2.3 Mice and infection studies

C57BL/6 mice were kept under specific pathogen-free conditions in isolators and fed autoclaved food and water ad libitum at the central animal facilities of the University of Ulm. In a given experiment, mice were age- and sex-matched. Previous studies on in vivo plasmid stability showed that all r-S. typhimurium SL7207 strains still harbored their pMOhly1-based plasmids by day 21 post infection (p.i.) [26]. C57BL/6 mice at 8–10 weeks of age were immunized per os (p.o.) three times with 5×109 bacteria per 200 ml PBS or intravenously (i.v.) with a single dose (5×105 bacteria) of r-S. typhimurium p30 or control strains at 5 day intervals. Ten weeks after the last injection, vaccinated mice were completely cured of r-S. typhimurium SL7207 infection, as verified by colony forming units (CFU) analysis on brilliant green Salmonella selective agar (Difco). At day 120 after the last vaccination, these mice were challenged i.v. with 106M. tuberculosis H37Rv bacteria. For control purposes, C57BL/6 mice were immunized i.v. with 106 BCG microorganisms for 120 days and challenged together with the group of animals which had received the r-Salmonella strains previously. At the indicated time points p.i., lung, liver and spleen of these mice were removed and homogenized with a lab blender (Seward Medical, London, UK). M. tuberculosis bacteria were enumerated by plating serial 10-fold dilutions in PBS on Middlebrook agar plates (Difco) supplemented with 10% OADC enrichment (Difco). After 3–4 weeks of incubation at 37°C, mycobacterial colonies were counted.

2.4 Stimulation in vitro and IFN-γ and tumor necrosis factor (TNF) production

At the indicated time points p.i., animals were killed and spleens removed. Single—cell suspensions were prepared in Iscove's modified Dulbecco's medium (IMDM) (Seromed, Berlin, Germany) supplemented with 10% FCS (Boehringer Mannheim, Mannheim, Germany), 2 mM glutamine, 100 U ml−1 penicillin/streptomycin (Gibco, Paisley, UK) and 1 µg ml−1 indomethacin (Sigma, St. Louis, MO, USA) (complete IMDM). Cells were seeded into 96-well plates (Nunc, Roskilde, Denmark) at the final concentration of 2×105 per well and cultured at 37°C in 7% CO2 in air. Cells were stimulated with the following mycobacterial antigen preparations (1 µg ml−1): (i) M. tuberculosis PPD (obtained from the Statens Serum Institute, Copenhagen, Denmark) diluted in PBS and (ii) freeze-dried M. tuberculosis H37Ra (obtained from Difco) suspended in IMDM. After culture for 3 days, cell supernatants were removed, sterile-filtered and frozen at −20°C until IFN-γ detection by enzyme-linked immunosorbent assay (ELISA). All experimental groups were analyzed in triplicate. IFN-γ was determined by double-sandwich ELISA using the monoclonal antibodies R4-6A2 and AN18-17.24, which recognize different epitopes of the cytokine as described [34]. Murine r-IFN-γ (generous gift of Dr. G. Adolf, Ernst Boehringer-Institut für Arzeneimittelforschung, Vienna, Austria), with a specificity of 107 U mg−1, was diluted in complete IMDM to obtain a standard curve. The detection limit of the ELISA was 0.05 U ml−1 IFN-γ. TNF was measured in a cytotoxicity assay using a L929-TNF-sensitive cell line [35]. Due to the use of the same TNF receptors (TNFR-I and II) for binding of TNF-α as well as TNF-β on the cell surface of L929 cells, this assay does not discriminate between both types of TNF. Nevertheless, murine r-TNF-α, specificity 8.3×107 U mg−1 (Genzyme), was used to obtain standard curves. Absorbance obtained by ELISA or bioassay was analyzed using the SpectraMax equipment (molecular Devices, Sunnyvale, CA, USA). Quantification of IFN-γ and TNF concentrations was done with SoftmaxPro software (Molecular Devices).

3 Results

The attenuated S. typhimurium aroA SL7207 strain combined with the HlyB/HlyD/TolC export machinery of hemolytic E. coli isolates was used for delivering p30 of BCG as a secreted 63–kDa fusion protein. S. typhimurium p30 microorganisms were constructed, which harbored the pAg85Bs secretion vector encoding the p30 fusion protein (Fig. 1A). This p30 polypeptide consists of the N-terminal 44 residues of listeriolysin-HlyA encoded by pILH-1[33], followed by 14 random amino acids and the complete p30 sequence with 322 residues of pAB261 (MedImmune, personal communication; [21]) and ending with three linker residues and 190 amino acids of HlyA encoded by plasmid pANN202-312 downstream of the BglII site at position 3811 [31] (Fig. 1B). The 63-kDa p30 fusion protein is detectable in the supernatant of r-S. typhimurium p30 by immunostaining with affinity-purified rabbit anti-peptide p30 antibodies (Fig. 1C). This formally proves export of p30 by the r-S. typhimurium p30 antigen delivery system and suggests that the antigen is readily available for immune recognition.

Figure 1

Construction of the recombinant plasmid pAg85Bs and characterization of r-S. typhimurium p30. (A) The plasmid pAg85Bs encoding p30 protein of M. bovis BCG and the genes (hlyC, hlyA, hlyB, hlyD) necessary for HlyA expression by E. coli or attenuated S. typhimurium strains are shown. (B) Gene fusion encoding the p30-HlyA (p30) fusion protein secreted by S. typhimurium p30 and length of the respective fusion partners. (C) Immunoblot of supernatant proteins developed with rabbit anti-p30 antibodies. Samples of r-S. typhimurium control or p30 were prepared and analyzed as described in Section 2. The position of p30 of r-S. typhimurium p30 is indicated. Abbreviations: hly, listeriolysin; hlyA, hemolysin A.

In general, IFN-γ and TNF-α represent two prominent cytokines which contribute to anti-mycobacterial defence [36,37]. C57BL/6 mice were immunized with 5×105S. typhimurium p30 or control i.v., and at indicated time points, mice were killed and splenocytes were subsequently monitored for IFN-γ and TNF production. Splenocytes were restimulated with 1 µg ml−1M. tuberculosis H37Ra antigen preparation for 3 days in vitro. Splenocytes of r-S. typhimurium p30-vaccinated mice produced more than 2-fold higher concentrations of IFN-γ in comparison with spleen cells of mice that had received the r-Salmonella control strain (Fig. 2). Similarly, TNF release by splenocytes from mice immunized with r-S. typhimurium p30 was augmented. Nevertheless, after restimulation with M. tuberculosis H37Ra lysate, a strong cross-reactive immune response against the r-Salmonella control strain was observed (Fig. 2). We conclude that the r-S. typhimurium p30 construct enhanced the immune reactivity against p30 antigen compared to the r-Salmonella control strain.

Figure 2

IFN-γ and TNF production by spleen cells of mice immunized with r-S. typhimurium p30 or control. At the indicated time points, IFN-γ (A) and TNF (B) secretion by spleen cells of r-S. typhimurium p30 (■) or control (▲) immunized mice were monitored by ELISA (IFN-γ) or bioassay (TNF), respectively, after in vitro restimulation with M. tuberculosis H37Ra antigen preparation (1 µg ml−1) for 3 days. Splenocytes of three mice per group were pooled and results represent geometric means±S.D. of triplicate cultures from one representative experiment of three. Medium controls never exceeded 0.1 U ml−1 or 0.1 pg ml−1 IFN-γ or TNF, respectively.

For analyzing the protective efficacy of r-S. typhimurium p30, C57BL/6 mice were vaccinated p.o. and then challenged i.v. with M. tuberculosis H37Rv 120 days later. Sixty days post challenge, the bacterial burden was determined in different organs. For control purposes, groups of mice were immunized with the r-Salmonella control or M. bovis BCG and challenged with M. tuberculosis H37Rv. Animals were killed at day 60 post challenge and CFU were determined by plating serial dilutions of organ homogenates. As shown in Fig. 3, protection was induced by systemic BCG or mucosal S. typhimurium p30 vaccination, and was comparable in spleens and lungs of mice challenged with M. tuberculosis. At this time point, splenocytes of r-S. typhimurium p30-immunized mice showed significantly increased production of IFN-γ and TNF after in vitro restimulation with M. tuberculosis H37a lysate or PPD underlining, the enhanced anti-TB immune response compared with that in BCG or r-Salmonella control immunized mice (Fig. 4). We conclude from these results that immunization with r-S. typhimurium p30 primed a Th1-biased cytokine profile.

Figure 3

Protection of r-S. typhimurium p30, control and M. bovis BCG-immunized C57BL/6 mice against M. tuberculosis H37Rv infection. Mice were left untreated (open bars) or were immunized with r-S. typhimurium control p.o. (cross-hatched bars), M. bovis BCG i.v. (horizontal-hatched bars) or r-S. typhimurium p30 vaccination p.o. (solid bars) for 120 days before challenge infection (i.v.) with M. tuberculosis H37Rv. After another 60 days, CFU in spleens and lungs were determined. Each column represents the geometric mean of five animals per group, and vertical bars indicate S.D. CFU differences between r-S. typhimurium p30 and M. bovis BCG immunized mice are not statistically significant (P>0.05; Student's t test), whereas the differences between r-S. typhimurium p30 and control vaccinated or non-immune mice are significant (P<0.05; Student's t test).

Figure 4

IFN-γ and TNF production by spleen cells of mice immunized with r-S. typhimurium p30, control or M. bovis BCG and challenged with M. tuberculosis H37Rv. At day 60 post challenge, IFN-γ (A) and TNF (B) secretion of spleen cells of p.o. r-S. typhimurium p30 (solid bars), p.o. r-S. typhimurium control (cross-hatched bars) or i.v. M. bovis BCG-vaccinated (horizontal-hatched bars) were monitored by ELISA or bioassay, respectively, after in vitro restimulation with PPD or M. tuberculosis H37Ra antigen preparation (1 µg ml−1) for 3 days. Results represent geometric means of triplicate cultures from one representative experiment of three, vertical bars indicate S.D. Medium controls never exceeded 0.1 U ml−1 or 0.1 pg ml−1 IFN-γ or TNF, respectively.

4 Discussion

Our report describes protection against M. tuberculosis by oral immunization of mice with a r-S. typhimurium p30 carrier strain secreting the naturally secreted antigen p30 of M. bovis BCG. Furthermore, our data reveal that display of secreted antigens by the r-S. typhimurium carrier enhanced immune responses against M. tuberculosis as indicated by augmented IFN-γ and TNF production prior to and after challenge infection. This finding is consistent with the assumption that secreted antigens are central to the induction of protective immunity against intracellular pathogens [3840].

From an immunological point of view, the secretion of p30 by the anti-TB vaccine M. bovis BCG as well as by r-S. typhimurium p30 represents a similar mode of antigen delivery. Both strains introduce antigen p30 from a vacuolar compartment within antigen presenting cells into antigen processing pathways. Therefore, it possibly should be noted that a single immunodominant antigen expressed by r-S. typhimurium possesses a similar protective efficacy against M. tuberculosis as does M. bovis BCG with its multitude of antigens recognized by M. tuberculosis-specific lymphocytes.

So far, only vaccinia virus has been used as a heterologous carrier for M. tuberculosis antigens such as the lipo-glycoproteins of molecular mass 38 or 19 kDa [41]. Taken together, our data show the feasibility of r-S. typhimurium carrier strains for inducing immune responses against TB, at least when an immunodominant naturally secreted antigen of M. bovis BCG with high homology to that of M. tuberculosis is delivered as secreted fusion protein.


This work was supported by the German Science Foundation, project Ka 573/3–1/2 to J.H. and S.H.E.K., and the BMBF project ‘Mycobacterial Infections’ to S.H.E.K., I.G. and W.G. received support from Grant Go168/16-3 of the German Science Foundation and FORGEN L1. We thank D. Miko and C. McCoull for expert help and we are grateful to MedImmune and Dr. B.A.D. Stocker for donating plasmids and S. typhimurium SL7207, respectively, to Drs. J. Belisle and E. Adolf for additional helpful reagents.


  • 1 GlaxoWellcome, Verona, Italy.


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
  22. [22].
  23. [23].
  24. [24].
  25. [25].
  26. [26].
  27. [27].
  28. [28].
  29. [29].
  30. [30].
  31. [31].
  32. [32].
  33. [33].
  34. [34].
  35. [35].
  36. [36].
  37. [37].
  38. [38].
  39. [39].
  40. [40].
  41. [41].
View Abstract