TEXT

Top Abstract Text References

Superoxide dismutases (SODs) are metalloenzymes that catalyze the dismutation of superoxide radicals to hydrogen peroxide and molecular oxygen. They are initial components of the cellular defense against reactive oxygen intermediates (ROI) resulting from univalent reduction of oxygen, and they contribute to the survival of bacterial pathogens such as Shigella flexneri (10), Campylobacter jejuni (18), Salmonella enterica serovar Typhimurium (8, 9, 26), Yersinia enterocolitica (19), and Neisseria meningitidis (27).

Mycobacterium tuberculosis produces a tetrameric iron-cofactored SOD (FeSOD or SodA) encoded by the sodA gene (5, 29) and a copper and zinc SOD (CuZnSOD or SodC) encoded by the sodC gene (28). FeSOD is among the major extracellular proteins released by M. tuberculosis during growth (2). It is exported in an active form via a signal peptide-independent pathway that has not been fully characterized (12, 29). The CuZnSOD possesses a putative signal peptide and is localized to the periphery of M. tuberculosis (28). It has been hypothesized that the presence of SODs at the periphery of M. tuberculosis and in the extracellular milieu could protect bacteria from superoxides generated exogeneously, e.g., by host phagocytes (12, 28, 29). The killing of M. tuberculosis by host-activated phagocytic cells is mediated to some extent by ROI along with reactive nitrogen intermediates (1, 4, 14, 15).

To investigate the contribution of mycobacterial CuZnSOD to the defense of bacteria against oxidative killing, we constructed isogenic mutants of M. tuberculosis and M. bovis BCG and compared them with their parental strains for sensitivity to ROI in vitro and for survival in murine bone marrow-derived macrophages and in a guinea pig model of infection.

Characterization of CuZnSOD-deficient M. tuberculosis and M. bovis BCG. The M. tuberculosis sodC gene was mutated by allelic exchange (17). A DNA fragment containing sodC and 500 bp of its flanking sequences was generated by PCR using primers SODC0.5-5' (5'-ggtgctgttgtttctcgg-3') and SODC0.5-3' (5'-tcggcatcactttgtgcg-3'). The fragment was cloned into pCR2.1TOPO (Invitrogen) and subcloned into PstI-digested and blunt-ended pSL1180 (Pharmacia), constructing pOD1. pOD4 was created by cloning the PstI-flanked aph gene (kanamycin resistance) of pUC4K into the sodC PstI site of pOD1. The NotI-SpeI fragment of pOD4 containing sodC::aph was blunt ended and cloned into the SmaI site of pXYL4, a plasmid bearing the xylE gene (17), creating pOD6. The 4-kb BamHI fragment containing sodC::aph and xylE was isolated from pOD6 and ligated at the BamHI site of pPR27, a vector which contains the counterselectable sacB gene and the thermosensitive origin of replication of pAL5000 (17), constructing pOD7. To achieve allelic exchange, pOD7 was electroporated (17) into M. tuberculosis H37Rv (14 001 0001; Centre National de Référence des Mycobactéries, Institut Pasteur, France). Transformants were selected at 32°C on 7H11 medium containing kanamycin (20 µg/ml) and then grown in 7H9 broth containing kanamycin. Gene replacement accompanied by plasmid loss was selected for on 7H11-kanamycin-2% sucrose at 39°C (17). Loss of the plasmid was confirmed in 100% of the resultant colonies by spraying with catechol, a chromogenic substrate of XylE (6, 17). Gene replacement of sodC was verified by Southern blotting of genomic DNA from four colonies, using the sodC gene as a probe (Fig. 1A). One mutant clone was designated MTsodC.

View larger version (37K): [in this window] [in a new window]   FIG. 1.   Disruption of sodC. (A) Southern blot analysis of the M. tuberculosis H37Rv parental strain (lane 1) and sodC-defective mutants (lanes 2 to 5). Chromosomal DNAs were digested with EagI and analyzed by Southern blotting with a 32P-labeled probe corresponding to sodC. sodC mutants gave a single 2.2-kb fragment, as expected from double crossover. (B) Southern blot analysis of the M. bovis BCG parental strain (lane 1) and the sodC mutant (lane 2). Chromosomal DNAs were digested with EcoRI and probed with the digoxigenin-labeled SODC1-SODC2 PCR product. The presence of hybridizing fragments of 1.9 and 1.1 kb is consistent with a double crossover and gene replacement. (C) Absence of CuZnSOD in mutant strains. Western blot analysis was performed with whole-cell extracts of M. tuberculosis H37Rv, sodC-defective mutant MTsodC, M. bovis BCG, and sodC-deficient mutant BCGsodC. Total protein (10 µg) from each extract was immunoblotted with an anti-M. tuberculosis SodC polyclonal antibody.

Sensitivity of CuZnSOD-deficient M. tuberculosis and M. bovis BCG to ROI. The sodC disruption did not affect the growth rate of the bacteria in 7H9 broth. The sensitivity of mycobacterial strains to plumbagin and menadione, which are two superoxide-generating agents, and to hydrogen peroxide was assessed by metabolic labeling of mycobacteria with [³H]uracil (20). Mid-log-phase cultures of mycobacteria were diluted in 7H9 broth at 10⁸ CFU/ml. A 100-µl volume of this suspension was incubated in a 96-well plate at 37°C for 5 h with the addition of 0.5 µCi of [³H]uracil per ml and the stress reagents at a range of concentrations (0 to 25.6 mM plumbagin, 0 to 76.8 mM menadione, and 0 to 25.6 mM hydrogen peroxide). The assay was stopped and mycobacteria were killed by the addition of 50% ethanol. Cultures were recovered on fiberglass filters in a cell harvester, and radioactivity was measured using a liquid scintillation counter. The background radioactivity was subtracted from subsequent determinations. The inhibitory effect of each reagent was measured as a percentage of the [³H]uracil incorporation observed in wells without reagent. Wilcoxon test and t test analyses were performed, and the most significant result is indicated. Both MTsodC and BCGsodC mutant strains were more sensitive to the superoxide-generating agents plumbagin (P = 0.02 and P < 0.0007, respectively) and menadione (P < 0.0001 and P = 0.03, respectively) than their respective parental strains (Fig. 2A and B and 3A and B). The differences were statistically significant, although the data that were obtained at concentrations where plumbagin and menadione are toxic were similar (Fig. 2B and 3A and B). MTsodC and BCGsodC were also significantly more sensitive to hydrogen peroxide (P = 0.0007 and P < 0.0001, respectively) than were their parental strains (Fig. 2C and 3C). This ROI-sensitive phenotype was successfully complemented in MTsodC by reintroduction of sodC at the attB site on the chromosome by using a hyg-containing derivative of pYUB295 (W. R. Jacobs, Albert Einstein College of Medicine, Bronx, N.Y.). ROI resistance was fully restored (data not shown).

View larger version (16K): [in this window] [in a new window]   FIG. 2.   Sensitivity of the M. tuberculosis sodC-defective mutant to ROI. Mycobacteria were diluted in 7H9 medium at 108 CFU/ml and incubated at 37°C for 5 h with various concentrations of menadione (A), plumbagin (B), or hydrogen peroxide (C). The inhibitory effect of these reagents on M. tuberculosis H37Rv (solid circles) and M. tuberculosis MTsodC (open circles) was measured as percentage of [3H]uracil incorporation in wells without reagent. The P values (paired t test) were considered significant (0.02 for panel B) to extremely significant (<0.0001 for panel A and 0.0007 for panel C).

View larger version (16K): [in this window] [in a new window]   FIG. 3.   Sensitivity of the M. bovis sodC-defective mutant to ROI. Mycobacteria were diluted in 7H9 medium at 108 CFU/ml and incubated at 37°C for 5 h with various concentrations of menadione (A), plumbagin (B), and hydrogen peroxide (C). The inhibitory effect of these reagents on M. bovis BCG (solid circles) and M. bovis BCGsodC (open circles) was measured as a percentage of [3H]uracil incorporation in wells without reagent. The P values (Wilcoxon signed-rank test) were considered significant (0.03 for panel A) to extremely significant (0.0007 for panel B and <0.0001 for panel C).

Survival of CuZnSOD-deficient M. bovis BCG in mouse bone marrow macrophages. To assess the role of SodC in intracellular growth and protection against killing by macrophages, the survival rates of the BCGsodC mutant strain and its parental strain were compared during infection of unstimulated and activated mouse bone marrow-derived macrophages. Bone marrow-derived macrophages were obtained from femoral bones of 6- to 8-week-old female C57BL/6 mice and cultivated for 8 to 10 days in Dulbecco modified Eagle medium (Gibco BRL, Glasgow, Scotland) supplemented with 2 mM L-glutamine (Gibco BRL), 10% fetal calf serum (Labtech), 5% horse serum (Labtech), and 15% L-929 culture supernatant. The macrophages were seeded in 24-well culture plates at 2 × 10⁵ cells/well 24 h before infection. For experiments with activated macrophages, 100 U of gamma interferon per ml and 10 ng of E. coli lipopolysaccharide per ml were added 24 h before infection. The macrophage activation status was confirmed before each experiment by measurement of CD54 (ICAM-1) up-regulation and induction of inducible nitric oxide synthase. Macrophages were infected with the mutant and wild-type BCG strains at a multiplicity of infection of 0.5 to 1 bacterium/cell. After 2 to 3 h at 37°C, the cells were washed twice in phosphate-buffered saline, and fresh medium was added. The number of mycobacteria associated with the monolayers was assessed at 0, 1, 2, and 3 days postinfection for activated macrophages and 0, 1, 3, 5, 7, and 9 days for unstimulated macrophages. The cell monolayer was washed once with phosphate-buffered saline, and then 1 ml of 0.1% Triton X-100 was added to lyse the macrophages. Lysates were serially diluted and plated onto 7H11 medium, and CFU were counted after 17 to 21 days. The experiments were performed twice, with three determinations per time point. The BCGsodC strain and its parental strain showed similar kinetics of intracellular growth in nonstimulated bone marrow macrophages (Fig. 4A). In activated macrophages, approximately 90% killing was observed at 3 days for both the BCGsodC mutant and parental strains (Fig. 4B). These data suggest that SodC does not protect M. bovis BCG against killing by macrophages.

View larger version (15K): [in this window] [in a new window]   FIG. 4.   Survival of sodC-defective mutants in macrophages. Unstimulated mouse bone marrow-derived macrophages (A) and macrophages activated by 100 U of gamma interferon per ml and 10 ng of E. coli lipopolysaccharide per ml (B) were infected with the M. bovis BCG sodC mutant strain (open circles) and the parental strain (solid circles) at a multiplicity of infection of 0.5 to 1 mycobacterium per cell. Macrophages were lysed, and the number of mycobacteria associated with macrophages was assessed by plating on 7H11. p.i., postinfection.

Survival of CuZnSOD-deficient M. tuberculosis in guinea pigs. It has been reported that Brucella abortus and S. enterica serovars Typhimurium, Choleraesuis, and Dublin sodC mutants behaved similarly to their respective parental strains within macrophages, although their CuZnSOD was shown to contribute to pathogenicity in vivo (9, 24). Therefore, it was of interest to test the effect of M. tuberculosis sodC disruption on pathogenicity. Outbred female Hartley guinea pigs were injected subcutaneously with 10⁴ viable units of parental and mutant M. tuberculosis strains in 0.2 ml of saline solution (five replicates/strain). Animals were sacrificed 5 weeks after infection, and there were no visible differences in tuberculosis lesions in the spleen, lungs, lymph nodes, or liver or at the site of injection. The lymph nodes draining the site of injection and spleen were homogenized, and serial dilutions were plated onto 7H11 medium. There was no difference between strains in the number of CFU recovered from spleens or lymph nodes (Fig. 5). Thus, SodC does not make an obvious contribution to the pathogenicity of M. tuberculosis. However, its in vitro sensitivity to ROI suggests that it could protect the periphery of the bacilli against ROI at some stage of its life cycle. If CuZnSOD does not form a major component of defense against ROI, FeSOD may be important. Using an identical strategy to that used to interrupt sodC, we have been unable to disrupt sodA under aerobic or microaerophilic conditions. It is not known whether this is due to technical problems, to the fact that FeSOD is essential for the viability of M. tuberculosis, or to detrimental polar effects on the expression of downstream genes. The role of FeSOD in mycobacterial pathogenicity remains an open question.

View larger version (17K): [in this window] [in a new window]   FIG. 5.   Survival of sodC-defective mutants in guinea pigs. Guinea pigs were injected with the M. tuberculosis sodC mutant strain (open bars) and the parental strain (solid bars). Lymph nodes and spleen were collected and homogenized after 5 weeks, and the number of mycobacteria was assessed by plating onto 7H11 medium.