- Email: [email protected]
Attenuated Salmonella typhimurium htrA mutants cause fatal infections in mice deficient in NADPH oxidase and destroy NADPH oxidase-deficient macrophage monolayers
Vaccine 22 (2004) 4124–4131
Attenuated Salmonella typhimurium htrA mutants cause fatal infections in mice deficient in NADPH oxidase and destroy NADPH oxidase-deficient macrophage monolayers Mbithe Mutunga a,1 , Sarah Graham a , Raquel Demarco de Hormaeche a,b , Julie A. Musson a , John H. Robinson c , Pietro Mastroeni b , C.M. Anjam Khan a , Carlos E. Hormaeche a,b,∗ a
School of Cell and Molecular Biosciences, The Medical School, University of Newcastle, Framlington Place, Newcastle upon Tyne NE2 4HH, UK b Department of Clinical Veterinary Medicine, Madingley Road, Cambridge CB3 0ES, UK c School of Clinical Medical Sciences (Rheumatology), The Medical School, University of Newcastle, Framlington Place, Newcastle upon Tyne NE2 4HH, UK Received 2 July 2003; accepted 20 October 2003 Available online 10 May 2004
Abstract Salmonella live vaccine strains harbouring mutations in htrA, a stress protein gene, display increased susceptibility to oxidative stress in vitro. This is believed to be connected to their reduced virulence, perhaps due to impaired survival inside phagocytes, although this has never been formally proven. We report that the in vitro phenotype of increased susceptibility to oxidative stress of Salmonella typhimurium htrA mutants newly prepared by transduction is rapidly lost on subculture, with the mutants becoming as resistant as the parent for reasons that remain unclear. However, despite this change, htrA mutants are still attenuated in normal mice. In contrast, they were found to be lethal for gene targeted gp91phox−/− mice deficient in NADPH oxidase, as was a S. typhimurium SPI-2 mutant known to be virulent in gp9lphox−/− mice. Infection with htrA mutants caused little damage to primary bone marrow macrophage cultures from normal mice; conversely, they caused extensive damage to macrophages from gp9lphox−/− mice, with more than 60% reduction in cell numbers 2.5 h after being infected. The parental wild type strain similarly caused extensive damage to macrophages from both normal and gp9lphox−/− mice, whereas an aroA live vaccine strain had no effect on either normal or gp9lphox−/− macrophages. Taken collectively, the present results suggest that htrA is somehow involved in resistance to oxidative stress in vivo, with the avirulence of htrA mutants in mice being due to mechanisms which involve NADPH oxidase and suppression of bacterial growth within macrophages. © 2004 Published by Elsevier Ltd. Keywords: Salmonella typhimurium; Macrophage; Mutant
1. Introduction Production of reactive oxygen species (ROS) is considered an important feature of intracellular bactericidal mechanisms in phagocytes, and the phagocyte NADPH oxidase has been shown to be a key component in the mechanisms by which phagocytes kill virulent salmonellae [1,5,21,22]. To prevent oxidative damage, bacteria have evolved adaptive responses that result in the expression of numerous genes in response to the stresses exerted by the ROS. These adapta∗ Corresponding author. Tel.: +44-1223-764086; fax: +44-1223-337610; mobile: 07867-801115. E-mail address: [email protected] (C.E. Hormaeche). 1 Present address: Department of National Blood Service, Histocompatibility and Immunogenetics, Holland Drive, Barrack Road, Newcastle upon Tyne NE2 4N0, UK
0264-410X/$ – see front matter © 2004 Published by Elsevier Ltd. doi:10.1016/j.vaccine.2003.10.053
tions include production of stress proteins such as HtrA, a periplasmic serine protease [17]. We have reported that Salmonella typhimurium htrA mutants show increased sensitivity to oxidative stress in vitro in a manner similar to that described for Escherichia coli htrA mutants, and further, that they are attenuated in mice [3,11,13,18]. S. typhimurium htrA mutants are effective as live vaccines and can be used for delivery of recombinant antigens [2]. Salmonella typhi strains with several attenuating lesions, including htrA, are currently under evaluation as candidate live vaccines for human typhoid fever [12,19,20]. However, despite their attenuation in other species, htrA mutants were not attenuated when administered orally to calves [23]. In our original description of Salmonella htrA mutants [11], we suggested that the in vitro phenotype of increased
M. Mutunga et al. / Vaccine 22 (2004) 4124–4131
susceptibility to oxidative stress could provide the explanation for their reduced virulence, perhaps due to an inability to withstand oxidative killing mechanisms in macrophages; however, this has never been formally tested. In contrast, we have recently observed that the in vitro phenotype of sensitivity to oxidative stress of S. typhimurium htrA mutants is lost on subculture by an unexplained, apparently adaptive change, but without loss of attenuation in mice (our unpublished observations), suggesting that the reduced virulence of S. typhimurium htrA mutants might perhaps not be related to oxidative stress. Gene targeted mice deficient in NADPH oxidase have impaired phagocyte function and are considered to be a model of human chronic granulomatous disease (CGD) [10,15]. They show increased susceptibility to bacterial infection, and have been used for the study of virulence of salmonellae [21,22]. It has been recently reported that S. typhimurium strains harbouring mutations in genes of pathogenicity island 2 (SPI-2 mutants), which are very attenuated in normal mice, are fully virulent in NADPH oxidase deficient mice [22]. This was taken as an indication that SPI-2 is involved in resistance to oxidative killing mechanisms in vivo [22]. We reasoned that if Salmonella htrA mutants were also attenuated due to impaired oxidative stress resistance, they might similarly be expected to recover virulence in NADPH oxidase deficient mice. This effectively proved to be the case.
2. Materials and methods 2.1. Bacterial strains and bacteriophages S. typhimurium C5 is a prototropic mouse virulent strain [9]. Strain C5 046, a S. typhimurium CS htrA::TnphoA insertion mutant, has been previously described [11,13]. Fresh C5 htrA mutants were made by transduction from C5 046 using P22 HT HT1 05/1 int [16]. SL3261 is a well characterised S. typhimurium SL 1344 aroA live vaccine strain [8]. The S. typhimurium SPI2 mutant, P2D6 ssaV::mTn5, was a gift from David Holden, Imperial College; the strain is attenuated in normal mice, but is known to be virulent in NADPH deficient mice [22]. 2.2. DNA extraction, oligonucleotides and PCR Transduction was confirmed by PCR, using primer pairs developed for the htrA gene. Genomic DNA was extracted from C5, C5 046 and newly transduced CS htrA strains by boiling colonies in 50 l solution containing 1% Triton X-100, 20 mM Tris pH 8.5 and 2 mM EDTA. Primers were designed using the published DNA sequence of the htrA gene [11]. The sequence of the oligonucleotides used were as shown below: • Forward; TGATAACGCCAGCGTGATTA • Reverse; ATACCGATCAGCTCACCGTT
4125
2.3. Sensitivity to oxidizing agents A panel of oxidizing agents (Sigma, Poole, Dorset, UK) were utilised for the phenotypic tests to compare the sensitivity of passaged and newly transduced htrA strains to oxidative stress as compared to the wild type. All solutions of oxidising agents were prepared immediately before use. In liquid medium, exponentially growing cultures (OD600 of 0.1–0.15) were treated with 0.2 mM cumene hydrogen peroxide, 0.2 mM hydrogen peroxide or 0.2 mM menadione as described by Skorko-Glonek et al. [17]. Growth inhibition was determined by measuring the OD600 of treated and untreated cells every 20 min over a period of 2 h. Growth inhibition was calculated as a percentage of the untreated cells and compared between the strains. Growth inhibition was also determined on solid media using minimal agar and LB agar. Approximately 5 ml of 1×107 ml−1 bacteria were used to inoculate a plate by flooding. The entire medium was then removed and four blank paper discs placed on the plate and immediately loaded with the oxidizing agent (300 g hydrogen peroxide and 320 g cumene hydrogen peroxide) as described by Skorko-Glonek [17]. The plates were then incubated overnight at 37 ◦ C. The diameter of the area of growth inhibition was measured and expressed as a mean of four determinations. Each experiment was repeated at least three times. Statistical significance was determined by Student’s t-test; Mann–Whitney analysis of the growth curve data gave similar results. 2.4. Inoculations of NADPH-deficient mice gp91phox−/− mice lack the gp9l component of NADPH oxidase, which is the main cellular source of reactive oxygen species [15]. The gp9lphox−/− mice were obtained from Bantin and Kingman, Hull, UK as a gift from Gordon Dougan, Imperial College, London. Mice of the same genetic background, C57BL/6, were obtained from Bantin and Kingman and were used as controls. The gp9lphox−/− mice were kept in negative pressure isolators and fed on autoclaved food and water. Both gp9lphox−/− and C57BL/6 mice were left to acclimatise for 2 weeks before infection. For inoculation, organisms were grown as stationary 37 ◦ C overnight cultures in LB broth. Animals were injected intraperitoneally with 0.2 ml of serial 10-fold dilutions in PBS ranging from 10−3 to 10−6 . The inoculum dose was confirmed by plating on LB agar. For the pilot experiment, the cfu count of the neat broth of C5 046 was 1 × 109 . For the second experiment, the cfu counts were C5 046–8.4 × 108 ; SL 3261–5.7 × 108 ; and for P2D6 ssaV::mTn5, the SPI2 mutant −1.2 × 109 cfu/ml. Animal deaths were recorded for 4 weeks. 2.5. Generation and maintenance of bone marrow macrophages Bone marrow macrophages were generated from femoral and tibial bone marrow from both gp9lphox−/− and
4126
M. Mutunga et al. / Vaccine 22 (2004) 4124–4131
C57BL/6 mice. Bone marrow cells were teased into cell suspension and 106 cells per bacteriological Petri dish were cultured for 8 days in RPMI 1640 medium supplemented with 2.0 mM l-glutamine, 1 mM sodium pyruvate, 0.05 mM, 10 mM HEPES buffer, 10% FBS, 5% horse serum and 40 g/ml gentamicin (all Sigma). Macrophage differentiation was selectively promoted by the addition of 10% of a culture supernatant from the L929 fibroblast cell line as a source of M-CSF and the resultant cells were shown to be at least 90% macrophages based on F4/80 expression assessed by flow cytometry. 2.6. Infection of bone marrow macrophages Bone marrow macrophages were harvested from the Petri dishes using a cell scraper and resuspended in RPMI 1640 medium supplemented with 2.0 mM l-glutamine, 0.05 mM 2-mercaptoethanol, 10% FBS and 40 g/ml gentamicin, and cell viablity assessed using trypan blue. The macrophages were seeded in 24 well plates at 106 cells per well in 1 ml of medium and incubated overnight at 37 ◦ C in 5% CO2 to allow cells to adhere. The cells were then infected with 10 cfu/ml of C5 WT, C5 046 or SL 3261 and incubated to allow invasion of the cells for 60 min at 37 ◦ C in 5% CO2 . The cells were washed three times with PBS before adding medium containing 100 g/ml gentamicin for 90 min to kill the extracellular bacteria. The cells were then washed three times with PBS, fixed with 10% formal saline for 10 min and stained with crystal violet. The cells were viewed using a light microscope and adherent cells counted at 10 random fields. The results were expressed as mean counts plus or minus standard deviation.
3. Results
Fig. 1. Effect of 2 mM cumene hydrogen peroxide on the growth of a serially passaged (P) C5 htrA strain (C5 046) and a newly transduced non-passaged C5 htrA (NP) strain as compared to the parent strain, C5 WT. Strains were grown with and without 2 mM cumene hydrogen peroxide in LB medium and the growth recorded every 20 min over 2 h. Growth inhibition is shown as a percentage of the untreated cultures. Inhibition of the non-passaged strain was significantly different from both the wild type strain (P < 0001) and the passaged C5 046 (P < 0001).
However, sensitivity of freshly transduced mutants appeared to decrease with subculture until they showed no difference with the parental wild type. Plate assays with hydrogen peroxide showed a double halo of inhibition in the newly transduced C5 htrA strain, which was not seen in the passaged C5 046 or the wild type, suggesting a process of active change already occurring in this mutant (Fig. 2). To ensure that the selected colonies were representative of the transductants, 10 colonies were picked per plate after transduction and the sensitivity to oxidative stress checked for each individually. All displayed similar resistance to oxidative stress.
3.1. Confirmation of TnphoA insertion in C5
3.3. The passaged C5 046 htrA mutant is virulent in gp91phox−/− mice but not in normal control mice
DNA extracted from C5 WT and the C5 htrA mutants was amplified by PCR as described in Section 2. The results showed an increase in size of the htrA gene from 450 to 950 bp as expected (results not shown).
A pilot experiment was performed by infecting two animals each of gp9lphox−/− and normal C57BL/6 controls with approx. 102 , 103 , 104 and 105 cfu of the passaged C5
3.2. Sensitivity of passaged and non-passaged htrA strains to oxidative stress We found that a freshly prepared, non-passaged htrA mutant was more sensitive to cumene hydrogen peroxide than the parent or the passaged strain (Fig. 1); there were no significant differences between the parent and the passaged strain. Growth inhibition of the non-passaged strain reached 50%, whereas inhibition of the parent and the passaged strain remained below 40%. On disc sensitivity assays, the newly transduced strains also showed increased sensitivity to cumene hydrogen peroxide (Table 1).
Table 1 Sensitivity of S. typhimurium C5 wild type, the stock C5 046 htrA and the non-passaged C5 htrA strains to menadione and cumene hydrogen peroxide Mean area of inhibition (mm) on exposure to cumene hydrogen peroxide Strain
Area of inhibition (mm)
C5 wild type C5 046 C5 htrA non-passaged
11.3 12.3 19.8
Assays were performed on minimal agar plates as described in Section 2. Results are the means of five experiments each. Inhibition of the non-passaged strain was significantly different from both C5 wild type (P < 0.0001) and C5 046 (P < 0.002).
M. Mutunga et al. / Vaccine 22 (2004) 4124–4131
4127
Fig. 2. Effect of hydrogen peroxide on strains C5 wild type (top), the passaged C5 046 htrA (middle), and the freshly prepared C5 htrA (bottom). A clear double halo of inhibition is seen around the latter.
4128
M. Mutunga et al. / Vaccine 22 (2004) 4124–4131
046 htrA mutant intra-peritoneally as described in Section 2; all the normal controls survived, whereas only one of the gp9lphox−/− injected with the lowest dose survived. The experiment was repeated using similar doses with groups of two mice. As a control, gp9lphox−/− mice were also infected with SL3261, an aroA S. typhimurium vaccine strain which is known to be avirulent in NADPH deficient mice, and a SPI-2 mutant which is known to be very attenuated in normal mice, but which is virulent in NADPH deficient gp91phox−/− mice [22]. All gp91phox−/− mice
injected with either C5 046 or with the SPI-2 mutant died, whereas all C57BI/6 mice injected with these strains survived. Mice died early, with similar times to death for both the C5 046 mutant and for the SPI-2 mutant: 105 3–6 days; 104 6 days; 102 and 103 days 9, 10, 11 for C5 046 and days 7–14 for the SPI-1 mutant. There were no deaths among any mice injected with the aroA SL 3261. Thus, the htrA mutant was virulent for NADPH deficient mice, but not for normal mice, in a manner similar to that described for the SPI-2 mutant [22].
Fig. 3. Appearance of bone marrow macrophage monolayers from gp91phox−/− (left) and C57BL/6 (right) mice after infection with C5 wild type (top), C5 046 htrA (middle) and the SL 3261 strain (bottom).
M. Mutunga et al. / Vaccine 22 (2004) 4124–4131
4129
Fig. 4. Survival of bone marrow macrophages from gp91phox−/− (open bars) and C57BL/6 (solid bars) after infection with S. typhimurium htrA (C5 046), C5 wild type and an attenuated aroA vaccine strain SL 3261. The graph shows the mean number of cells still adherent to the plate after 2.5 h incubation with the bacteria.
3.4. The passaged C5 046 htrA is cytotoxic for gp91phox−/− bone marrow macrophages Bone marrow macrophages derived from gp9lphox−/− and control mice were infected with either the passaged S. typhimurium C5 046 htrA strain, the C5 wild type and the SL 326l aroA vaccine strain as described in Section 2. After infection with C5 046, more than half of gp9lphox−/− macrophages had detached from the plates by 2.5 h, while the normal cells were unaffected. Both normal and gp91phox−/− macrophages suffered similar damage when infected with the wild type strain, whereas the SL3261 aroA strain had no effect on either gp9lphox−/− or normal cells (Figs. 3 and 4).
4. Discussion The present results show that the S. typhimurium htrA mutant C5 046 is attenuated in normal mice, but regains virulence in gp91phox−/− mice. The latter are defective in mounting an oxidative burst due to their lack of the gp91 component of the NADPH oxidase [14]. Infection of primary bone marrow macrophages isolated from gp91phox−/− and control mice confirmed the in vivo findings in that C5 046 was deleterious for gp91phox−/− macrophages in a similar manner to the C5 wild type, but not for macrophages isolated from the control mice. These observations lend support to the view that the low virulence of S. typhimurium htrA mutants in mice may be linked an increased susceptibility to oxidative stress inside host phagocytes. The increased susceptibility to oxidative stress of htrA mutants has been previously described in E. coli [17], and we observed a similar increased sensitivity of salmonella htrA mutants in our original description of the phenotype in salmonellae [11], prompting us to suggest that this might
account for the reduced virulence of the mutants. The htrA mutation has been incorporated into S. typhi live attenuated human typhoid vaccine candidate strains [19,20]. However, we recently observed that the increased susceptibility to oxidative stress in vitro of Salmonella htrA mutants was unstable, and was lost by subculture with the mutant behaving essentially like the parent. Our stock S. typhimurium C5 046 htrA mutant, which we have used for many years, was found to be no more sensitive than the wild type to a panel of oxidising agents. However, susceptibility to different oxidising agents was variable; we found no differences in susceptibility to ferrous sulphate and menadione, and the results with paraquat were inconsistent (results not shown). We found that increased susceptibility was greatest in a strain that had been freshly prepared by transduction into a wild type recipient with a P22 lysate prepared from the stock C5 046 strain, which is itself phenotypically as resistant as the parent. The reversion appears quickly after passage in vitro, and can sometimes be seen as a double halo around a disc of hydrogen peroxide. The reason for this rapidly occurring apparent phenotypic adaptation is unclear, but a similar effect has also been observed in E. coli htrA mutants [[15]; Lipinska, personal communication]. Vazquez-Torres et al. [21,22] have reported that salmonellae harbouring mutations in genes of pathogenicity island 2, which are avirulent in normal mice, are virulent in gene targeted mice deficient in NADPH oxidase similar to those used in this study, indicating that genes within SPI-2 are involved in salmonella virulence by preventing killing by mechanisms involving oxidative mechanisms. The present study similarly indicates that htrA is involved in resistance to oxidative stress. In vivo studies were confirmed by infection of bone marrow derived macrophages from normal and gp9lphox−/− mice in vitro, which showed the htrA mutant causing a reduction in viable gp9lphox−/− macrophages just as rapidly as the wild type affected normal macrophages.
4130
M. Mutunga et al. / Vaccine 22 (2004) 4124–4131
In this study, we recorded the effect of salmonellae on cultured macrophages as a reduction in the numbers of macrophages remaining in the cultures after infection. This is different to the procedure used by Vazquez-Torres et al. [21], who measured the numbers of viable bacteria inside macrophages to assess differences in the rate of intracellular bacterial killing. We found that bacterial counting gave variable results, presumably as a result of the progressive decrease in the numbers of macrophages remaining after washing the monolayers prior to counting. The difference between our results and those of Vazquez-Torres et al. [21] is probably due to the fact that whereas the latter authors used periodate elicited, interferon-gamma stimulated peritoneal macrophages, we used unstimulated, bone marrow derived macrophages which are perhaps less resistant to infection in vitro in these experimental conditions. We used a similar method to that described here for assessing macrophage survival in our earlier study on the role of the Ity gene (now designated Nramp1) in macrophages [7], in which a clear difference was observed in the survival of Salmonella-infected Kupffer cell monolayers from innately resistant and susceptible mice. The numbers of mice used in the in vivo virulence experiments presented here were small, but we did not feel justified in increasing the size of the groups when some were confirmatory controls, and the result of the pilot experiment was so clear. Over a 4 log dose range, virtually all mice either died or survived. Larger numbers would have added little to this data, which clearly showed the virulence of htrA mutants for gp91phox−/− mice. The possible pathways involved in the role of htrA and oxidative stress in salmonella remain unclear. An htrA mutant of Klebsiella pneumoniae was found to be avirulent, and also displayed increased sensitivity to oxidative stress [4] as did htrA mutants of streptococci [6], Bacillus subtillis [14] and S. typhi human typhoid vaccine candidate strains [20], whereas a Yersinia htrA mutant showed only a slight increase in sensitivity [24]. The Yersinia enterocolitica HtrA-like GsrA protein is induced by oxidative stress [25,26]. The exact mechanism by which htrA participates in resistance to oxidative stress remains to be elucidated.
References [1] Buchmeier N, Lipps CJ, So MHY, Heffron F. Recombinationdeficient mutants of Salmonella typhimurium are avirulent and sensitive to the oxidative burst of macrophages. Mol Microbiol 1993; 7:933–6. [2] Chabalgoity A, Khan CMA, Nash AA, Hormaeche CE. A Salmonella typhimurium htrA live vaccine expressing multiple copies of a peptide comprising amino acids 8–24 of herpes simplex virus glycoprotein D as a genetic fusion to tetanus toxin fragment C protects mice from herpes simplex virus infection. Mol Microbiol 1996;19:791–801. [3] Chatfield S, Strahan K, Pickard D, Charles I, Hormaeche CE, Dougan G. Evaluation of Salmonella typhimurium vaccine strains harbouring mutations in htrA and aroA as live oral vaccines in the murine salmonellosis model. Microb Pathogen 1992;12:145–51.
[4] Cortes G, de Astorza B, Benedi VJ, Alberti S. Role of the htrA gene in Klebsiella pneumoniae virulence. Infect Immun 2002;70:4772–6. [5] De Groote MA, Ochsner UA, Shiloh MU, Nathan C, McCord JM, Dinauer MC, et al. Periplasmic superoxide dismutase protects Salmonella from products of phagocyte NADPH oxidase and nitric oxide synthase. Proc Natl Acad Sci USA 1997;94:13997–4001. [6] Diaz-Torres ML, Russell RR. HtrA protease and processing of extracellular proteins of Streptococcus mutans. FEMS Microbiol Lett 2001;16:23–8. [7] Harrington KA, Hormaeche CE. Expression of the innate resistance gene Ity in mouse Kupffer cells infected with Salmonella typhimurium in vitro. Microb Pathogen 1986;1:269–74. [8] Hoiseth SK, Stocker BAD. Aromatic-dependent Salmonella typhimurium are non-virulent and effective as live vaccines. Nature 1981;291:238–9. [9] Hormaeche CE. Natural resistance to Salmonella typhimurium in different inbred mouse strains. Immunology 1979;37:311–8. [10] Jackson SH, Gallin JI, Holland SM. The p47phox mouse knock-out model of chronic granulomatous disease. J Exp Med 1995;182: 751–8. [11] Johnson KS, Charles IG, Dougan G, Miller IA, Pickard D, O’Goara P, et al. The role of a stress–response protein in bacterial virulence. Mol Microbiol 1991;5:401–7. [12] Lowe DC, Savidge TC, Pickard D, Eckmann L, Kagnoff MF, Dougan G, et al. Characterization of candidate live oral Salmonella typhi vaccine strains harboring defined mutations in aroA, aroC, and htrA. Infect Immun 1999;67:700–7. [13] Miller IA, Maskell DJ, Hormaeche CE, Pickard D, Dougan G. The isolation of orally attenuated Salmonella typhimurium following TnphoA mutagenesis. Infect Immun 1989;57:2758–63. [14] Noone D, Howell A, Collery R, Devine KM. YkdA and YvtA, HtrA-Iike serine proteases in Bacillus subtilis, engage in negative autoregulation and reciprocal cross-regulation of ykdA and yvtA gene expression. J Bacteriol 2001;183:654–63. [15] Pollock JD, Williams DA, Gifford MA, Li LL, Du X, Fisherman J, et al. Mouse model of X-Iinked chronic granulomatous disease, an inherited defect in phagocyte superoxide production. Nat Genet 1995;9:202–9. [16] Schmieger H. Phage P22 mutants with increased or decreased transduction abilities. Mol Gen Genet 1972;119:75–81. [17] Skorko-Glonek J, Zurawa D, Kuczwara E, Wozniak M, Wypych Z, Lipinska B. The Escherichia coli heat shock protease HtrA participates in defense against oxidative stress. Mol Gen Genet 1999;262:342–50. [18] Strahan K, Chatfield SN, Tite J, Dougan G, Hormaeche CE. Impaired resistance to infection does not increase the virulence of Salmonella htrA live vaccines. Microb Pathogen 1992;12:311–7. [19] Tacket CO, Sztein MB, Wasserman SS, Losonsky G, Kotloff KL, Wyant TL, et al. Phase 2 clinical trial of attenuated Salmonella enterica serovar Typhi oral live vector vaccine CVD 908-htrA in US volunteers. Infect Immun 2000;8:1196–201. [20] Tacket CO, Sztein MB, Losonsky GA, Wasserman SS, Nataro JP, Edelman R, et al. Safety of live oral Salmonella typhi vaccine strains with deletions in htrA and aroC a aroD and immune response in humans. Infect Immun 1997;65:452–6. [21] Vazquez-Torres A, Jones-Carson J, Mastroeni P, lschiropoulos H, Fang FC. Antimicrobial actions of the NADPH phagocyte oxidase and inducible nitric oxide synthase in experimental salmonellosis. I. Effects on microbial killing by activated peritoneal macrophages in vitro. J Exp Med 2000;192:227–36. [22] Vazquez-Torres A, Xu YS, Jones-Carson J, Holden DW, Lucia SM, Dinauer MC, et al. Salmonella pathogenicity island 2-dependent evasion of the phagocyte NADPH oxidase. Science 2000;287: 1655–8. [23] Villarreal-Ramos B, McAulay M, Chance V, Martin M, Morgan J, Howard CJ. Investigation of the role of CD8+ T cells in bovine tuberculosis in vivo. Infect Immun 2003;71:4297–303.
M. Mutunga et al. / Vaccine 22 (2004) 4124–4131 [24] Williams K, Oyston PC, Dorrell N, Li S, Titball RW, Wren BW. Investigation into the role of the serine protease HtrA in Yersinia pestis pathogenesis. FEMS Microbiol Lett 2000;15:281–6. [25] Yamamoto T, Hanawa T, Ogata S, kamiya S. Identification and characterization of the Yersinia enterocolitica gsrA gene, which protectively responds to intracellular stress induced by macrophage
4131
phagocytosis and to extracellular environmental stress. Infect Immun 1996;64:2980–7. [26] Yamamoto T, Hanawa T, Ogata S, Kamiya S. The Yersinia enterocolitica GsrA stress protein, involved in intracellular survival, is induced by macrophage phagocytosis. Infect Immun 1997;65: 2190–6.
Attenuated Salmonella typhimurium htrA mutants cause fatal infections in mice deficient in NADPH oxidase and destroy NADPH oxidase-deficient macrophage monolayers Mbithe Mutunga a,1 , Sarah Graham a , Raquel Demarco de Hormaeche a,b , Julie A. Musson a , John H. Robinson c , Pietro Mastroeni b , C.M. Anjam Khan a , Carlos E. Hormaeche a,b,∗ a
School of Cell and Molecular Biosciences, The Medical School, University of Newcastle, Framlington Place, Newcastle upon Tyne NE2 4HH, UK b Department of Clinical Veterinary Medicine, Madingley Road, Cambridge CB3 0ES, UK c School of Clinical Medical Sciences (Rheumatology), The Medical School, University of Newcastle, Framlington Place, Newcastle upon Tyne NE2 4HH, UK Received 2 July 2003; accepted 20 October 2003 Available online 10 May 2004
Abstract Salmonella live vaccine strains harbouring mutations in htrA, a stress protein gene, display increased susceptibility to oxidative stress in vitro. This is believed to be connected to their reduced virulence, perhaps due to impaired survival inside phagocytes, although this has never been formally proven. We report that the in vitro phenotype of increased susceptibility to oxidative stress of Salmonella typhimurium htrA mutants newly prepared by transduction is rapidly lost on subculture, with the mutants becoming as resistant as the parent for reasons that remain unclear. However, despite this change, htrA mutants are still attenuated in normal mice. In contrast, they were found to be lethal for gene targeted gp91phox−/− mice deficient in NADPH oxidase, as was a S. typhimurium SPI-2 mutant known to be virulent in gp9lphox−/− mice. Infection with htrA mutants caused little damage to primary bone marrow macrophage cultures from normal mice; conversely, they caused extensive damage to macrophages from gp9lphox−/− mice, with more than 60% reduction in cell numbers 2.5 h after being infected. The parental wild type strain similarly caused extensive damage to macrophages from both normal and gp9lphox−/− mice, whereas an aroA live vaccine strain had no effect on either normal or gp9lphox−/− macrophages. Taken collectively, the present results suggest that htrA is somehow involved in resistance to oxidative stress in vivo, with the avirulence of htrA mutants in mice being due to mechanisms which involve NADPH oxidase and suppression of bacterial growth within macrophages. © 2004 Published by Elsevier Ltd. Keywords: Salmonella typhimurium; Macrophage; Mutant
1. Introduction Production of reactive oxygen species (ROS) is considered an important feature of intracellular bactericidal mechanisms in phagocytes, and the phagocyte NADPH oxidase has been shown to be a key component in the mechanisms by which phagocytes kill virulent salmonellae [1,5,21,22]. To prevent oxidative damage, bacteria have evolved adaptive responses that result in the expression of numerous genes in response to the stresses exerted by the ROS. These adapta∗ Corresponding author. Tel.: +44-1223-764086; fax: +44-1223-337610; mobile: 07867-801115. E-mail address: [email protected] (C.E. Hormaeche). 1 Present address: Department of National Blood Service, Histocompatibility and Immunogenetics, Holland Drive, Barrack Road, Newcastle upon Tyne NE2 4N0, UK
0264-410X/$ – see front matter © 2004 Published by Elsevier Ltd. doi:10.1016/j.vaccine.2003.10.053
tions include production of stress proteins such as HtrA, a periplasmic serine protease [17]. We have reported that Salmonella typhimurium htrA mutants show increased sensitivity to oxidative stress in vitro in a manner similar to that described for Escherichia coli htrA mutants, and further, that they are attenuated in mice [3,11,13,18]. S. typhimurium htrA mutants are effective as live vaccines and can be used for delivery of recombinant antigens [2]. Salmonella typhi strains with several attenuating lesions, including htrA, are currently under evaluation as candidate live vaccines for human typhoid fever [12,19,20]. However, despite their attenuation in other species, htrA mutants were not attenuated when administered orally to calves [23]. In our original description of Salmonella htrA mutants [11], we suggested that the in vitro phenotype of increased
M. Mutunga et al. / Vaccine 22 (2004) 4124–4131
susceptibility to oxidative stress could provide the explanation for their reduced virulence, perhaps due to an inability to withstand oxidative killing mechanisms in macrophages; however, this has never been formally tested. In contrast, we have recently observed that the in vitro phenotype of sensitivity to oxidative stress of S. typhimurium htrA mutants is lost on subculture by an unexplained, apparently adaptive change, but without loss of attenuation in mice (our unpublished observations), suggesting that the reduced virulence of S. typhimurium htrA mutants might perhaps not be related to oxidative stress. Gene targeted mice deficient in NADPH oxidase have impaired phagocyte function and are considered to be a model of human chronic granulomatous disease (CGD) [10,15]. They show increased susceptibility to bacterial infection, and have been used for the study of virulence of salmonellae [21,22]. It has been recently reported that S. typhimurium strains harbouring mutations in genes of pathogenicity island 2 (SPI-2 mutants), which are very attenuated in normal mice, are fully virulent in NADPH oxidase deficient mice [22]. This was taken as an indication that SPI-2 is involved in resistance to oxidative killing mechanisms in vivo [22]. We reasoned that if Salmonella htrA mutants were also attenuated due to impaired oxidative stress resistance, they might similarly be expected to recover virulence in NADPH oxidase deficient mice. This effectively proved to be the case.
2. Materials and methods 2.1. Bacterial strains and bacteriophages S. typhimurium C5 is a prototropic mouse virulent strain [9]. Strain C5 046, a S. typhimurium CS htrA::TnphoA insertion mutant, has been previously described [11,13]. Fresh C5 htrA mutants were made by transduction from C5 046 using P22 HT HT1 05/1 int [16]. SL3261 is a well characterised S. typhimurium SL 1344 aroA live vaccine strain [8]. The S. typhimurium SPI2 mutant, P2D6 ssaV::mTn5, was a gift from David Holden, Imperial College; the strain is attenuated in normal mice, but is known to be virulent in NADPH deficient mice [22]. 2.2. DNA extraction, oligonucleotides and PCR Transduction was confirmed by PCR, using primer pairs developed for the htrA gene. Genomic DNA was extracted from C5, C5 046 and newly transduced CS htrA strains by boiling colonies in 50 l solution containing 1% Triton X-100, 20 mM Tris pH 8.5 and 2 mM EDTA. Primers were designed using the published DNA sequence of the htrA gene [11]. The sequence of the oligonucleotides used were as shown below: • Forward; TGATAACGCCAGCGTGATTA • Reverse; ATACCGATCAGCTCACCGTT
4125
2.3. Sensitivity to oxidizing agents A panel of oxidizing agents (Sigma, Poole, Dorset, UK) were utilised for the phenotypic tests to compare the sensitivity of passaged and newly transduced htrA strains to oxidative stress as compared to the wild type. All solutions of oxidising agents were prepared immediately before use. In liquid medium, exponentially growing cultures (OD600 of 0.1–0.15) were treated with 0.2 mM cumene hydrogen peroxide, 0.2 mM hydrogen peroxide or 0.2 mM menadione as described by Skorko-Glonek et al. [17]. Growth inhibition was determined by measuring the OD600 of treated and untreated cells every 20 min over a period of 2 h. Growth inhibition was calculated as a percentage of the untreated cells and compared between the strains. Growth inhibition was also determined on solid media using minimal agar and LB agar. Approximately 5 ml of 1×107 ml−1 bacteria were used to inoculate a plate by flooding. The entire medium was then removed and four blank paper discs placed on the plate and immediately loaded with the oxidizing agent (300 g hydrogen peroxide and 320 g cumene hydrogen peroxide) as described by Skorko-Glonek [17]. The plates were then incubated overnight at 37 ◦ C. The diameter of the area of growth inhibition was measured and expressed as a mean of four determinations. Each experiment was repeated at least three times. Statistical significance was determined by Student’s t-test; Mann–Whitney analysis of the growth curve data gave similar results. 2.4. Inoculations of NADPH-deficient mice gp91phox−/− mice lack the gp9l component of NADPH oxidase, which is the main cellular source of reactive oxygen species [15]. The gp9lphox−/− mice were obtained from Bantin and Kingman, Hull, UK as a gift from Gordon Dougan, Imperial College, London. Mice of the same genetic background, C57BL/6, were obtained from Bantin and Kingman and were used as controls. The gp9lphox−/− mice were kept in negative pressure isolators and fed on autoclaved food and water. Both gp9lphox−/− and C57BL/6 mice were left to acclimatise for 2 weeks before infection. For inoculation, organisms were grown as stationary 37 ◦ C overnight cultures in LB broth. Animals were injected intraperitoneally with 0.2 ml of serial 10-fold dilutions in PBS ranging from 10−3 to 10−6 . The inoculum dose was confirmed by plating on LB agar. For the pilot experiment, the cfu count of the neat broth of C5 046 was 1 × 109 . For the second experiment, the cfu counts were C5 046–8.4 × 108 ; SL 3261–5.7 × 108 ; and for P2D6 ssaV::mTn5, the SPI2 mutant −1.2 × 109 cfu/ml. Animal deaths were recorded for 4 weeks. 2.5. Generation and maintenance of bone marrow macrophages Bone marrow macrophages were generated from femoral and tibial bone marrow from both gp9lphox−/− and
4126
M. Mutunga et al. / Vaccine 22 (2004) 4124–4131
C57BL/6 mice. Bone marrow cells were teased into cell suspension and 106 cells per bacteriological Petri dish were cultured for 8 days in RPMI 1640 medium supplemented with 2.0 mM l-glutamine, 1 mM sodium pyruvate, 0.05 mM, 10 mM HEPES buffer, 10% FBS, 5% horse serum and 40 g/ml gentamicin (all Sigma). Macrophage differentiation was selectively promoted by the addition of 10% of a culture supernatant from the L929 fibroblast cell line as a source of M-CSF and the resultant cells were shown to be at least 90% macrophages based on F4/80 expression assessed by flow cytometry. 2.6. Infection of bone marrow macrophages Bone marrow macrophages were harvested from the Petri dishes using a cell scraper and resuspended in RPMI 1640 medium supplemented with 2.0 mM l-glutamine, 0.05 mM 2-mercaptoethanol, 10% FBS and 40 g/ml gentamicin, and cell viablity assessed using trypan blue. The macrophages were seeded in 24 well plates at 106 cells per well in 1 ml of medium and incubated overnight at 37 ◦ C in 5% CO2 to allow cells to adhere. The cells were then infected with 10 cfu/ml of C5 WT, C5 046 or SL 3261 and incubated to allow invasion of the cells for 60 min at 37 ◦ C in 5% CO2 . The cells were washed three times with PBS before adding medium containing 100 g/ml gentamicin for 90 min to kill the extracellular bacteria. The cells were then washed three times with PBS, fixed with 10% formal saline for 10 min and stained with crystal violet. The cells were viewed using a light microscope and adherent cells counted at 10 random fields. The results were expressed as mean counts plus or minus standard deviation.
3. Results
Fig. 1. Effect of 2 mM cumene hydrogen peroxide on the growth of a serially passaged (P) C5 htrA strain (C5 046) and a newly transduced non-passaged C5 htrA (NP) strain as compared to the parent strain, C5 WT. Strains were grown with and without 2 mM cumene hydrogen peroxide in LB medium and the growth recorded every 20 min over 2 h. Growth inhibition is shown as a percentage of the untreated cultures. Inhibition of the non-passaged strain was significantly different from both the wild type strain (P < 0001) and the passaged C5 046 (P < 0001).
However, sensitivity of freshly transduced mutants appeared to decrease with subculture until they showed no difference with the parental wild type. Plate assays with hydrogen peroxide showed a double halo of inhibition in the newly transduced C5 htrA strain, which was not seen in the passaged C5 046 or the wild type, suggesting a process of active change already occurring in this mutant (Fig. 2). To ensure that the selected colonies were representative of the transductants, 10 colonies were picked per plate after transduction and the sensitivity to oxidative stress checked for each individually. All displayed similar resistance to oxidative stress.
3.1. Confirmation of TnphoA insertion in C5
3.3. The passaged C5 046 htrA mutant is virulent in gp91phox−/− mice but not in normal control mice
DNA extracted from C5 WT and the C5 htrA mutants was amplified by PCR as described in Section 2. The results showed an increase in size of the htrA gene from 450 to 950 bp as expected (results not shown).
A pilot experiment was performed by infecting two animals each of gp9lphox−/− and normal C57BL/6 controls with approx. 102 , 103 , 104 and 105 cfu of the passaged C5
3.2. Sensitivity of passaged and non-passaged htrA strains to oxidative stress We found that a freshly prepared, non-passaged htrA mutant was more sensitive to cumene hydrogen peroxide than the parent or the passaged strain (Fig. 1); there were no significant differences between the parent and the passaged strain. Growth inhibition of the non-passaged strain reached 50%, whereas inhibition of the parent and the passaged strain remained below 40%. On disc sensitivity assays, the newly transduced strains also showed increased sensitivity to cumene hydrogen peroxide (Table 1).
Table 1 Sensitivity of S. typhimurium C5 wild type, the stock C5 046 htrA and the non-passaged C5 htrA strains to menadione and cumene hydrogen peroxide Mean area of inhibition (mm) on exposure to cumene hydrogen peroxide Strain
Area of inhibition (mm)
C5 wild type C5 046 C5 htrA non-passaged
11.3 12.3 19.8
Assays were performed on minimal agar plates as described in Section 2. Results are the means of five experiments each. Inhibition of the non-passaged strain was significantly different from both C5 wild type (P < 0.0001) and C5 046 (P < 0.002).
M. Mutunga et al. / Vaccine 22 (2004) 4124–4131
4127
Fig. 2. Effect of hydrogen peroxide on strains C5 wild type (top), the passaged C5 046 htrA (middle), and the freshly prepared C5 htrA (bottom). A clear double halo of inhibition is seen around the latter.
4128
M. Mutunga et al. / Vaccine 22 (2004) 4124–4131
046 htrA mutant intra-peritoneally as described in Section 2; all the normal controls survived, whereas only one of the gp9lphox−/− injected with the lowest dose survived. The experiment was repeated using similar doses with groups of two mice. As a control, gp9lphox−/− mice were also infected with SL3261, an aroA S. typhimurium vaccine strain which is known to be avirulent in NADPH deficient mice, and a SPI-2 mutant which is known to be very attenuated in normal mice, but which is virulent in NADPH deficient gp91phox−/− mice [22]. All gp91phox−/− mice
injected with either C5 046 or with the SPI-2 mutant died, whereas all C57BI/6 mice injected with these strains survived. Mice died early, with similar times to death for both the C5 046 mutant and for the SPI-2 mutant: 105 3–6 days; 104 6 days; 102 and 103 days 9, 10, 11 for C5 046 and days 7–14 for the SPI-1 mutant. There were no deaths among any mice injected with the aroA SL 3261. Thus, the htrA mutant was virulent for NADPH deficient mice, but not for normal mice, in a manner similar to that described for the SPI-2 mutant [22].
Fig. 3. Appearance of bone marrow macrophage monolayers from gp91phox−/− (left) and C57BL/6 (right) mice after infection with C5 wild type (top), C5 046 htrA (middle) and the SL 3261 strain (bottom).
M. Mutunga et al. / Vaccine 22 (2004) 4124–4131
4129
Fig. 4. Survival of bone marrow macrophages from gp91phox−/− (open bars) and C57BL/6 (solid bars) after infection with S. typhimurium htrA (C5 046), C5 wild type and an attenuated aroA vaccine strain SL 3261. The graph shows the mean number of cells still adherent to the plate after 2.5 h incubation with the bacteria.
3.4. The passaged C5 046 htrA is cytotoxic for gp91phox−/− bone marrow macrophages Bone marrow macrophages derived from gp9lphox−/− and control mice were infected with either the passaged S. typhimurium C5 046 htrA strain, the C5 wild type and the SL 326l aroA vaccine strain as described in Section 2. After infection with C5 046, more than half of gp9lphox−/− macrophages had detached from the plates by 2.5 h, while the normal cells were unaffected. Both normal and gp91phox−/− macrophages suffered similar damage when infected with the wild type strain, whereas the SL3261 aroA strain had no effect on either gp9lphox−/− or normal cells (Figs. 3 and 4).
4. Discussion The present results show that the S. typhimurium htrA mutant C5 046 is attenuated in normal mice, but regains virulence in gp91phox−/− mice. The latter are defective in mounting an oxidative burst due to their lack of the gp91 component of the NADPH oxidase [14]. Infection of primary bone marrow macrophages isolated from gp91phox−/− and control mice confirmed the in vivo findings in that C5 046 was deleterious for gp91phox−/− macrophages in a similar manner to the C5 wild type, but not for macrophages isolated from the control mice. These observations lend support to the view that the low virulence of S. typhimurium htrA mutants in mice may be linked an increased susceptibility to oxidative stress inside host phagocytes. The increased susceptibility to oxidative stress of htrA mutants has been previously described in E. coli [17], and we observed a similar increased sensitivity of salmonella htrA mutants in our original description of the phenotype in salmonellae [11], prompting us to suggest that this might
account for the reduced virulence of the mutants. The htrA mutation has been incorporated into S. typhi live attenuated human typhoid vaccine candidate strains [19,20]. However, we recently observed that the increased susceptibility to oxidative stress in vitro of Salmonella htrA mutants was unstable, and was lost by subculture with the mutant behaving essentially like the parent. Our stock S. typhimurium C5 046 htrA mutant, which we have used for many years, was found to be no more sensitive than the wild type to a panel of oxidising agents. However, susceptibility to different oxidising agents was variable; we found no differences in susceptibility to ferrous sulphate and menadione, and the results with paraquat were inconsistent (results not shown). We found that increased susceptibility was greatest in a strain that had been freshly prepared by transduction into a wild type recipient with a P22 lysate prepared from the stock C5 046 strain, which is itself phenotypically as resistant as the parent. The reversion appears quickly after passage in vitro, and can sometimes be seen as a double halo around a disc of hydrogen peroxide. The reason for this rapidly occurring apparent phenotypic adaptation is unclear, but a similar effect has also been observed in E. coli htrA mutants [[15]; Lipinska, personal communication]. Vazquez-Torres et al. [21,22] have reported that salmonellae harbouring mutations in genes of pathogenicity island 2, which are avirulent in normal mice, are virulent in gene targeted mice deficient in NADPH oxidase similar to those used in this study, indicating that genes within SPI-2 are involved in salmonella virulence by preventing killing by mechanisms involving oxidative mechanisms. The present study similarly indicates that htrA is involved in resistance to oxidative stress. In vivo studies were confirmed by infection of bone marrow derived macrophages from normal and gp9lphox−/− mice in vitro, which showed the htrA mutant causing a reduction in viable gp9lphox−/− macrophages just as rapidly as the wild type affected normal macrophages.
4130
M. Mutunga et al. / Vaccine 22 (2004) 4124–4131
In this study, we recorded the effect of salmonellae on cultured macrophages as a reduction in the numbers of macrophages remaining in the cultures after infection. This is different to the procedure used by Vazquez-Torres et al. [21], who measured the numbers of viable bacteria inside macrophages to assess differences in the rate of intracellular bacterial killing. We found that bacterial counting gave variable results, presumably as a result of the progressive decrease in the numbers of macrophages remaining after washing the monolayers prior to counting. The difference between our results and those of Vazquez-Torres et al. [21] is probably due to the fact that whereas the latter authors used periodate elicited, interferon-gamma stimulated peritoneal macrophages, we used unstimulated, bone marrow derived macrophages which are perhaps less resistant to infection in vitro in these experimental conditions. We used a similar method to that described here for assessing macrophage survival in our earlier study on the role of the Ity gene (now designated Nramp1) in macrophages [7], in which a clear difference was observed in the survival of Salmonella-infected Kupffer cell monolayers from innately resistant and susceptible mice. The numbers of mice used in the in vivo virulence experiments presented here were small, but we did not feel justified in increasing the size of the groups when some were confirmatory controls, and the result of the pilot experiment was so clear. Over a 4 log dose range, virtually all mice either died or survived. Larger numbers would have added little to this data, which clearly showed the virulence of htrA mutants for gp91phox−/− mice. The possible pathways involved in the role of htrA and oxidative stress in salmonella remain unclear. An htrA mutant of Klebsiella pneumoniae was found to be avirulent, and also displayed increased sensitivity to oxidative stress [4] as did htrA mutants of streptococci [6], Bacillus subtillis [14] and S. typhi human typhoid vaccine candidate strains [20], whereas a Yersinia htrA mutant showed only a slight increase in sensitivity [24]. The Yersinia enterocolitica HtrA-like GsrA protein is induced by oxidative stress [25,26]. The exact mechanism by which htrA participates in resistance to oxidative stress remains to be elucidated.
References [1] Buchmeier N, Lipps CJ, So MHY, Heffron F. Recombinationdeficient mutants of Salmonella typhimurium are avirulent and sensitive to the oxidative burst of macrophages. Mol Microbiol 1993; 7:933–6. [2] Chabalgoity A, Khan CMA, Nash AA, Hormaeche CE. A Salmonella typhimurium htrA live vaccine expressing multiple copies of a peptide comprising amino acids 8–24 of herpes simplex virus glycoprotein D as a genetic fusion to tetanus toxin fragment C protects mice from herpes simplex virus infection. Mol Microbiol 1996;19:791–801. [3] Chatfield S, Strahan K, Pickard D, Charles I, Hormaeche CE, Dougan G. Evaluation of Salmonella typhimurium vaccine strains harbouring mutations in htrA and aroA as live oral vaccines in the murine salmonellosis model. Microb Pathogen 1992;12:145–51.
[4] Cortes G, de Astorza B, Benedi VJ, Alberti S. Role of the htrA gene in Klebsiella pneumoniae virulence. Infect Immun 2002;70:4772–6. [5] De Groote MA, Ochsner UA, Shiloh MU, Nathan C, McCord JM, Dinauer MC, et al. Periplasmic superoxide dismutase protects Salmonella from products of phagocyte NADPH oxidase and nitric oxide synthase. Proc Natl Acad Sci USA 1997;94:13997–4001. [6] Diaz-Torres ML, Russell RR. HtrA protease and processing of extracellular proteins of Streptococcus mutans. FEMS Microbiol Lett 2001;16:23–8. [7] Harrington KA, Hormaeche CE. Expression of the innate resistance gene Ity in mouse Kupffer cells infected with Salmonella typhimurium in vitro. Microb Pathogen 1986;1:269–74. [8] Hoiseth SK, Stocker BAD. Aromatic-dependent Salmonella typhimurium are non-virulent and effective as live vaccines. Nature 1981;291:238–9. [9] Hormaeche CE. Natural resistance to Salmonella typhimurium in different inbred mouse strains. Immunology 1979;37:311–8. [10] Jackson SH, Gallin JI, Holland SM. The p47phox mouse knock-out model of chronic granulomatous disease. J Exp Med 1995;182: 751–8. [11] Johnson KS, Charles IG, Dougan G, Miller IA, Pickard D, O’Goara P, et al. The role of a stress–response protein in bacterial virulence. Mol Microbiol 1991;5:401–7. [12] Lowe DC, Savidge TC, Pickard D, Eckmann L, Kagnoff MF, Dougan G, et al. Characterization of candidate live oral Salmonella typhi vaccine strains harboring defined mutations in aroA, aroC, and htrA. Infect Immun 1999;67:700–7. [13] Miller IA, Maskell DJ, Hormaeche CE, Pickard D, Dougan G. The isolation of orally attenuated Salmonella typhimurium following TnphoA mutagenesis. Infect Immun 1989;57:2758–63. [14] Noone D, Howell A, Collery R, Devine KM. YkdA and YvtA, HtrA-Iike serine proteases in Bacillus subtilis, engage in negative autoregulation and reciprocal cross-regulation of ykdA and yvtA gene expression. J Bacteriol 2001;183:654–63. [15] Pollock JD, Williams DA, Gifford MA, Li LL, Du X, Fisherman J, et al. Mouse model of X-Iinked chronic granulomatous disease, an inherited defect in phagocyte superoxide production. Nat Genet 1995;9:202–9. [16] Schmieger H. Phage P22 mutants with increased or decreased transduction abilities. Mol Gen Genet 1972;119:75–81. [17] Skorko-Glonek J, Zurawa D, Kuczwara E, Wozniak M, Wypych Z, Lipinska B. The Escherichia coli heat shock protease HtrA participates in defense against oxidative stress. Mol Gen Genet 1999;262:342–50. [18] Strahan K, Chatfield SN, Tite J, Dougan G, Hormaeche CE. Impaired resistance to infection does not increase the virulence of Salmonella htrA live vaccines. Microb Pathogen 1992;12:311–7. [19] Tacket CO, Sztein MB, Wasserman SS, Losonsky G, Kotloff KL, Wyant TL, et al. Phase 2 clinical trial of attenuated Salmonella enterica serovar Typhi oral live vector vaccine CVD 908-htrA in US volunteers. Infect Immun 2000;8:1196–201. [20] Tacket CO, Sztein MB, Losonsky GA, Wasserman SS, Nataro JP, Edelman R, et al. Safety of live oral Salmonella typhi vaccine strains with deletions in htrA and aroC a aroD and immune response in humans. Infect Immun 1997;65:452–6. [21] Vazquez-Torres A, Jones-Carson J, Mastroeni P, lschiropoulos H, Fang FC. Antimicrobial actions of the NADPH phagocyte oxidase and inducible nitric oxide synthase in experimental salmonellosis. I. Effects on microbial killing by activated peritoneal macrophages in vitro. J Exp Med 2000;192:227–36. [22] Vazquez-Torres A, Xu YS, Jones-Carson J, Holden DW, Lucia SM, Dinauer MC, et al. Salmonella pathogenicity island 2-dependent evasion of the phagocyte NADPH oxidase. Science 2000;287: 1655–8. [23] Villarreal-Ramos B, McAulay M, Chance V, Martin M, Morgan J, Howard CJ. Investigation of the role of CD8+ T cells in bovine tuberculosis in vivo. Infect Immun 2003;71:4297–303.
M. Mutunga et al. / Vaccine 22 (2004) 4124–4131 [24] Williams K, Oyston PC, Dorrell N, Li S, Titball RW, Wren BW. Investigation into the role of the serine protease HtrA in Yersinia pestis pathogenesis. FEMS Microbiol Lett 2000;15:281–6. [25] Yamamoto T, Hanawa T, Ogata S, kamiya S. Identification and characterization of the Yersinia enterocolitica gsrA gene, which protectively responds to intracellular stress induced by macrophage
4131
phagocytosis and to extracellular environmental stress. Infect Immun 1996;64:2980–7. [26] Yamamoto T, Hanawa T, Ogata S, Kamiya S. The Yersinia enterocolitica GsrA stress protein, involved in intracellular survival, is induced by macrophage phagocytosis. Infect Immun 1997;65: 2190–6.