ZmpB, a novel virulence factor of Streptococcus pneumoniae that induces tumor necrosis factor alpha production in the respiratory tract

Abstract

Inflammation is a prominent feature of Streptococcus pneumoniae infection in both humans and animal models. Indeed, an intense host immune response to infection is thought to contribute significantly to the pathology of pneumococcal pneumonia and meningitis. Previously, induction of the inflammatory response following infection with S. pneumoniae has been attributed to certain cell wall constituents and the toxin pneumolysin. Here we present data implicating a putative zinc metalloprotease, ZmpB, as having a role in inflammation. Null mutations were created in the zmpB gene of the virulent serotype 2 strain D39 and analyzed in a murine model of infection. Isogenic mutants were attenuated in pneumonia and septicemia models of infection, as determined by levels of bacteremia and murine survival. Mutants were not attenuated in colonization of murine airways or lung tissue. Examination of cytokine profiles within the lung tissue revealed significantly lower levels of the proinflammatory cytokine tumor necrosis factor alpha following challenge with the Delta zmpB mutant (Delta 739). These data identify ZmpB as a novel virulence factor capable of inducing inflammation in the lower respiratory tract. The possibility that ZmpB was involved in inhibition of complement activity was examined, but the data indicated that ZmpB does not have a significant effect on this important host defense. The regulation of ZmpB by a two-component system (TCS09) located immediately upstream of the zmpB gene was examined. TCS09 was not required for the expression of zmpB during exponential growth in vitro.

FIG. 1.

Regulation of zmpB by TCS09. RNA isolated from cells grown to mid-log phase was used to determine the presence of zmpB transcript in the Δrr09 mutant (bottom panels) compared to the wild-type D39 strain (top panels) during in vitro growth. Primers ZmpIntF1 and ZmpIntR1 produced an expected band size of 235 bp. Primers 16SF and 16SR, specific for 16S RNA, were used as a positive control for RNA detection and resulted in a band of 700 bp. Lanes for all gels: 1, cDNA template from wild-type D39 (top panels) or the Δrr09 mutant (bottom panels); 2, cDNA template prepared in the absence of reverse transcriptase; 3, no DNA template; 4, genomic DNA template. A band of the expected size was observed in lanes 1 and 4 only with zmpB primers and the 16S primers with both wild-type D39 and mutant Δrr09 cDNAs. No bands were observed for the negative controls (lanes 2 and 3).

FIG. 2.

Cell morphology and autolysis. (Top panels) Cell morphology. Bacterial strains were grown in BHI and examined microscopically at various stages of growth. Gram staining revealed gram-positive cocci which occurred singly, in pairs, and in short chains for both the wild-type and Δ739 mutant strains. Samples examined at the mid-exponential phase of growth are shown. No difference in cell morphology was observed at other stages of growth. (Bottom panel) Autolysis with DOC. D39 and Δ739 cultures were grown to mid-log phase in BHI. Each culture was divided into two equal volumes, and DOC (0.04% final concentration) was added to an aliquot of each strain. Untreated aliquots of cultures were used as controls. Viability counts for all cultures were performed at predetermined intervals for 60 min after DOC treatment. The horizontal broken line represents the lower limit of detection.

FIG. 3.

Intranasal challenge. Mice were challenged intranasally with 10⁶ CFU of the Δ739 mutant and D39 parental strains. (Top panel) Survival of mice. The difference in the median survival time of mice infected with the Δ739 mutant compared to the D39 parental strain was significant (P < 0.005). (Bottom panel) Bacteremia. Levels of bacteremia for the Δ739 mutant compared to D39 were statistically lower at 24 h (P < 0.005) (**) and 36 h (P < 0.05) (*) postinfection. Counts for wild-type bacteria were not included at 48 h because some of the mice were culled prior to this time point due to severity of infection. The broken line represents the lower limit of detection. Error bars indicate standard errors of the means.

FIG. 4.

Colonization of airways and lungs. Mice were challenged intranasally with 10⁶ CFU of the Δ739 mutant and D39 parental strains. Bacterial counts within the BALF (top panel) and lung tissue (bottom panel) were determined at predetermined time points postinfection. The broken lines represent the lower limits of detection. Error bars indicate standard errors of the means.

FIG. 5.

Intravenous challenge. Mice were challenged intravenously with 10⁵ CFU of the Δ739 mutant and D39 parental strains. (Top panel) Survival of mice. The difference in the median survival time of mice infected with the Δ739 mutant compared to the D39 parental strain was significant (P < 0.005). (Bottom panel) Bacteremia following challenge. Levels of bacteremia for the Δ739 mutant compared to D39 were statistically significant at 24 h (P < 0.005) (**), 30 h (P < 0.05) (*), and 36 h (P < 0.005) (**) postinfection. The limit of detection (log₁₀ 1.92 CFU ml⁻¹) is below the scale used, so is not represented. Error bars indicate standard errors of the means.

FIG. 6.

Virulence in complement-deficient mice. Complement-sufficient (C3^+/+) (dotted lines) and complement knockout (C3^−/−) (solid lines) mice were challenged intranasally with 10⁶ CFU of the D39 (squares) and Δ739 mutant (diamonds) bacterial strains. Mice were monitored for survival (top panel) and bacteremia at predetermined time points (bottom panel). No significant difference was observed between the bacterial strains in the survival or bacterial counts in the blood of C3^+/+ mice. In the C3^−/− mice, those challenged with mutant bacteria survived significantly longer than those challenged with the D39 parental strain (P < 0.005). Similarly, for bacterial counts in the blood of mice, no difference was seen between bacterial strains in wild-type C3^+/+ mice, but in C3^−/− mice, lower bacterial counts were observed at all time points examined following challenge with mutant bacteria compared to the D39 parental strain. This difference was statistically significant for 24 and 36 h postinfection (P < 0.05). The boldface broken line represents the lower limit of detection for bacterial counts in the blood and has been interrupted to allow visualization of the low bacterial counts in wild-type C3^+/+ mice. Error bars indicate standard errors of the means.

FIG. 7.

TNF-α in lung tissue. Mice were challenged intranasally with 10⁶ CFU of the Δ739 mutant and D39 parental strain. At predetermined time points, lung tissue was removed and snap frozen. Cytokine analysis was performed on samples by using standard ELISA kits. The TNF-α concentration in the lung at 24 h postinfection is shown. Significantly lower levels of TNF-α were observed in the lung tissue of control mice and mice infected with the Δ739 mutant compared to wild-type-infected mice (P < 0.05). The broken line represents the lower limit of detection. Error bars indicate standard errors of the means.