|Home | About | Journals | Submit | Contact Us | Français|
Mycobacterium tuberculosis parasitizes host macrophages and subverts host innate and adaptive immunity. A number of cytokines elicited by the tubercle bacilli have been recognized as mediators of mycobacterial clearance or pathology in tuberculosis. Surprisingly, interleukin-1β (IL-1β), a major pro-inflammatory cytokine activated by processing upon assembly of a specialized protein complex termed the inflammasome, has not been implicated in host-pathogen interactions in tuberculosis. Here, we show that M. tuberculosis prevents inflammasome activation and IL-1β processing, and that a functional M. tuberculosis zmp1 gene is required for this process. Infection of macrophages with the zmp1 null M. tuberculosis triggered activation of caspase-1/IL-1β inflammasome, resulting in increased secretion of IL-1β, enhanced mycobacterial phagosome maturation into phagolysosomes, improved mycobacterial clearance by macrophages, and lower bacterial burden in the lungs of aerosol-infected mice. Thus, we uncovered the previously masked role for IL-1β in control of M. tuberculosis, and the existence of a mycobacterial system that prevents IL-1β/inflammasome activation.
Despite the notoriously high morbidity and mortality associated with active tuberculosis, claiming annually millions of people worldwide, primary infections with Mycobacterium tuberculosis are usually controlled in immunocompetent individuals. The bacterium becomes latent and persists in macrophages within granulomas, but can be reactivated under immunosuppressive and other debilitating conditions (Flynn and Chan, 2001). Given these relationships, the already high global burden of tuberculosis has been further exacerbated by AIDS (Aaron et al., 2004), making tuberculosis one of the leading causes of death from an infectious agent in the world (WHO Fact sheet; http://www.who.in/mediacentre/factsheets/fs104/en/). The role of innate and adaptive immunity in tuberculosis and the associated cytokine profiles have been studied in great detail (Flynn and Chan, 2001). Surprisingly, IL-1β has been lacking in such reports, despite being a major pro-inflammatory cytokine (Dinarello, 1992). The present dearth of information on IL-1β in tuberculosis is further underscored by a recent study implicating IL-1 receptor IL-1R1 signaling in antimycobacterial defenses during acute infection (Fremond et al., 2007)
The generation of active IL-1β, as well as IL-18 and IL-33, from their pro-protein precursors, is a tightly regulated process dependent on a multiprotein complex referred to as the inflammasome (Mariathasan and Monack, 2007; Martinon et al., 2002; Martinon et al., 2006). There are several well characterized inflammasomes: the IPAF inflammasome, with IPAF interacting directly with pro-caspase-1 via its CARD domain; and a number of NALP-containing inflammasomes, based on the pyrin-domain containing Nod-like receptors (NLR) NALP-1 and NALP-3, which require the pyrin- and CARD domain-containing adaptor ASC to bridge the NLRs to pro-caspase-1 (Kanneganti et al., 2007; Mariathasan and Monack, 2007; Martinon and Tschopp, 2007; Suzuki et al., 2007). Caspase-1 is the most extensively studied of the inflammatory caspases with one of its substrates being pro-IL-1β, the precursor to IL-1β (Dinarello, 2005a; Ghayur et al., 1997; Nadiri et al., 2006). Caspase-1 normally exists in its inactive pro-form. Various inflammatory triggers, e.g., bacterial RNA, uric acid crystals, Salmonella flagellin, anthrax lethal toxin, Staphylococcus aureus, Legionella pneumophilia, Listeria monocytogenes, and Shigella lead to inflammasome assembly and hence caspase-1 activation (Cordoba-Rodriguez et al., 2004; Franchi et al., 2006; Kanneganti et al., 2006; Mariathasan et al., 2004; Mariathasan et al., 2006; Martinon et al., 2004; Miao et al., 2006; Molofsky et al., 2006; Ren et al., 2006; Suzuki et al., 2007). In conjunction with increased pro-IL-1β production via TLR activation, and inflammasome activation, pro-caspase-1 and pro-IL1-β are sequestered into secretory lysosomes, followed by Ca2+-dependent exocytosis of active IL-1β (Andrei et al., 2004; Kahlenberg et al., 2005). Released IL-1β leads to production of reactive oxygen and nitrogen intermediates, and activation of other pro-inflammatory cytokines resulting in a vigorous host response to the pathogens and their products (Dinarello, 2005b). Conspicuously, M. tuberculosis does not trigger this common innate defense.
Here we report that M. tuberculosis actively suppresses IL-1β production. We have previously proposed that M. tuberculosis Zn2+ metalloproteases (Fratti et al., 2002; Vergne et al., 2003) play a role in intracellular survival of mycobacteria. We now show that one of the three corresponding M. tuberculosis genes Rv0198c (zmp1), plays a critical role in preventing caspase-1-dependent activation and secretion of IL-1β. We demonstrate that the M. tuberculosis zmp1 gene is: (i) essential for prevention of inflammasome activation and IL-1β production, (ii) required for mycobacterial survival in macrophages, and (iii) necessary for full virulence in a murine model of tuberculosis.
A bioinformatic analysis of M. tuberculosis genome revealed three putative Zn2+ metalloproteases, Rv0198c, Rv1977, Rv2869c (Fig. 1A), based on the conserved HExxH active site motif common to all Zn2+ metalloproteases (Fig. 1B). A role for one of the candidate genes Rv2869c has been described in lipid metabolism (Makinoshima and Glickman, 2005). Since Rv1977 is deleted in Mycobacterium bovis, and hence could not be implicated in virulence, we focused our studies on the remaining candidate, Rv0198c (zmp1). Phylogenetic analysis of Zmp1 revealed that it belonged to the M13 family of Zn2+ metalloproteases implicated in macrophage function, with neprilysin and endothelin-converting enzyme among its members (Turner et al., 2001; Wahl et al., 2005). Knockout mutants of zmp1 in M. tuberculosis H37Rv and in M. bovis BCG (Fig. 1) were constructed and confirmed by Southern blots (Fig. 1C). We raised polyclonal antibodies against the 15-aa C-terminal peptide of Zmp1. M. tuberculosis H37Rv extracts revealed a 63 kD band corresponding to Zmp1 that was absent in the zmp1::Kmr mutant (Fig. 1D). Although Zmp1 primary structure revealed no apparent export signal sequences, Zmp1 was found in culture supernatants (whole cell lysates) of zmp1 strain and standard preparations of total culture filtrate protein found in M. tuberculosis culture supernatant (Fig. 1D), indicating that Zmp1 is released from the tubercle bacilli.
In vitro studies revealed no difference between the zmp1 and zmp1+ strains grown under standard static culture conditions in Middlebrook 7H9 medium, indicating that Zmp1 is not essential in axenic culture (Fig. 1E). However, the zmp1::Kmr mutant showed reduced growth in mice. C57BL/6 mice were aerosol-infected with M. tuberculosis H37Rv and bacterial burden in the lungs quantified (Fig. 2A and B). The zmp1::Kmr mutant strain was significantly attenuated compared to the zmp1+ parental strain, evident as early as 14 days post-infection (Fig. 2A). The initially established difference was maintained during a longer course of infection (Fig. 2B).
We next examined the zmp1::Kmr mutant strain for survival in macrophages. M. bovis BCG zmp1+ and zmp1::Kmr strains were used to infect J774A and RAW264.7 macrophages and bacterial viability quantified as described in Experimental Procedures. There was a significant reduction in colony counts with the zmp1::Kmr mutant relative to its zmp1+ parent (Fig. 2C), although mutant uptake by the macrophages was not diminished (Suppl. Table S1). These results were confirmed and further extended by genetic complementation in virulent M. tuberculosis H37Rv using the zmp1+ parental strain, its zmp1::Kmr derivative, and a zmp1::Kmr(zmp1+) complemented strain (see Fig. 1C for characterization). Inactivation of zmp1 in virulent M. tuberculosis resulted in a 5-fold reduction of viability after 3 days in macrophages. The loss of viability was fully recovered by complementation with the wild type zmp1 gene (Fig. 2D). These findings were validated using human peripheral blood monocyte-derived macrophages. Infected human macrophages were incubated for 2 h, 3 days, and 7 days. Again, a significant reduction in the viability of M. tuberculosis H37Rv zmp1::Kmr mutant relative to its zmp1+ parent was observed (Fig. 2D). The results in the animal model and in murine and human macrophages demonstrate that zmp1 is required for M. tuberculosis virulence.
Phagosomes containing the parental zmp1+ strain and zmp1::Kmr mutant showed differences in their maturation (Fig. 3). RAW264.7 murine macrophages were infected with M. bovis BCG zmp1+ strain and zmp1::Kmr mutant for 1 h prior to processing for confocal microscopy. Transferrin (early endosomal marker) colocalized with 56 % zmp1+ bacilli vs. 12.5 % zmp1::Kmr bacilli (Fig. 3A). Two late endosomal markers CD63 and LysoTracker DND-99 (LT) colocalized more with the zmp1::Kmr mutant than with zmp1+ bacilli (Fig. 3B and C; 25 % vs. 68.5 % for CD63, and 16 % vs. 98 % for LT). These results indicate that absence of Zmp1 breaches, directly or indirectly, the mycobacterial ability to maintain mycobacterial phagosome maturation arrest (Vergne et al., 2004).
To address the mechanism underlying the requirement for zmp1 in mycobacterial virulence, macrophages infected with M. bovis BCG zmp1+ or zmp1::Kmr mutant strains, were treated with a panel of inhibitors affecting macrophage physiology and assayed mycobacterial phagosomes for CD63 colocalization. BAPTA-AM, a Ca2+ chelator used as a positive control, showed an expected reduction in CD63 colocalization with the zmp1::Kmr bacilli (Suppl. Fig. S1), in accordance with the known role of Ca2+ in phagosome maturation (Malik et al., 2000). Of other tested inhibitors, z-VAD-fmk (z-VAD), a pan-caspase inhibitor (Fig. 4A), showed significant reduction in the maturation of phagosomes harboring the zmp1::Kmr mutant. We next determined which of the caspases played a role, leading to the identification of caspase-1 as the caspase responsible for the increased phagosome maturation of the zmp1::Kmr strain. Ac-YVAD-cmk (YVAD), a caspase-1 specific inhibitor (Garcia-Calvo et al., 1998), suppressed the differences between the zmp1+ and zmp1::Kmr strains. Colocalization of CD63 with M. bovis BCG zmp1::Kmr in the presence or absence of 50 mM YVAD revealed a 50 % reduction in phagolysosome biogenesis, similar to that seen with the pan-caspase inhibitor z-VAD (Fig. 4A). These observations were confirmed with M. tuberculosis H37Rv zmp1+ and zmp1::Kmr strains. The zmp1::Kmr mutant showed enhanced phagolysosome biogenesis, as determined by LT colocalization, compared to the zmp1+ strain. The increase in phagosomal maturation observed with the zmp1::Kmr mutant was abrogated either by genetic complementation with zmp1+ or by addition of YVAD (Fig. 4B).
If caspase-1 activation is the mechanism underlying decreased viability and increased phagosomal maturation of the zmp1::Kmr mutant, then an addition of exogenous YVAD should improve its survival in macrophages. Macrophages were infected with zmp1+ or zmp1::Kmr M. bovis BCG, and treated with YVAD. Fig. 4C shows that YVAD restored survival of the zmp1::Kmr strain. Thus, caspase-1 appears to be activated by the zmp1::Kmr strain in contrast to the wild-type zmp1+strain. Activation of caspase-1 was directly tested by comparison of cell lysates from macrophages infected with M. bovis BCG zmp1+ and zmp1::Kmr, obtained 30 min and 4 h post-infection. Upon activation, pro-caspase-1 is cleaved to p20 (active form) and p10 fragments, with the antibody detecting only the p20 fragment on immunoblots. For both time points, there was more p20 fragment detected in cell lysates from macrophages infected with zmp1::Kmr compared to zmp1+ bacteria or to uninfected cells (Fig. 4D and Suppl. Fig. S2). In contrast, caspase-3 (an apoptotic caspase) remained unchanged, indicating that the effect seen with z-VAD was specific to caspase-1 (Suppl. Fig. S2A).
Activation of caspase-1 was also observed when the M. bovis BCG zmp1+ strain was attenuated by an overnight treatment with 30 μg/ml chloramphenicol (Fig. 5A and B). The opposite was seen with untreated, zmp1+ M. bovis BCG: the zmp1+ strain did not activate caspase-1 processing, and was moreover capable of inhibiting inflammasome activation by the prototypical inflammasome agonists ATP and nigericin (Fig. 5A and B).
The importance of caspase-1 in inducing phagosome maturation of the zmp1::Kmr mutant was assessed in Caspase-1−/− bone marrow-derived macrophages infected with M. tuberculosis H37Rv zmp1+ parent, zmp1::Kmr mutant, and a zmp1::Kmr(zmp1+) complemented strain (Fig. 6A). When compared to the mutant (zmp1::Kmr) strain in wild type macrophages, both the parental (zmp1+) and complemented (zmp1::Kmr(zmp1+)) strains showed lower levels of colocalization with Lysotracker Red as a measure of phagosomal maturation (Fig. 6A). The difference in phagosomal maturation between zmp1 mutant and zmp1+ strains was abrogated when the same strains were compared in Caspase-1−/− macrophages (Fig. 6A).
Results similar to those obtained with cell biological tracers in macrophages from caspase-1 knockout mice, were seen when assaying for survival of M. bovis BCG zmp1::Kmr in RAW264.7 cells knocked down for caspase-1 (Fig. 6B). In contrast to caspase-1, a knockdown of caspase-3 did not affect killing of zmp1::Kmr M. bovis BCG (Fig. 6B), despite a nearly complete knockdown of caspase-3 (Suppl. Fig. S3). To determine the type of inflammasome required for reduced survival of the zmp1 mutant, we knocked down IPAF and ASC, the key components of the previously characterized inflammasomes (Kanneganti et al., 2007; Mariathasan and Monack, 2007; Martinon and Tschopp, 2007) and tested survival of M. bovis BCG zmp1+ and zmp1::Kmr by comparing CFU at 1 h and 3 days after infection. The results, displayed in Fig. 6B, show that both IPAF and ASC are important for the killing of the zmp1::Kmr mutant.
We next tested whether caspase-1 activation upon infection of macrophages with the zmp1 mutant resulted in IL-1β activation. Macrophages were infected with M. bovis BCG zmp1::Kmr for 1 h, treated with neutralizing antibody against IL-1β or with isotypic IgG control, and assayed for CD63 distribution. The IL-1β neutralizing antibody reduced phagosome maturation when compared to untreated samples or samples with control antibody (Fig. 7A), showing that IL-1β secreted by zmp1::Kmr-infected macrophages was responsible for the observed increase in phagolysosome biogenesis. The levels of IL-1β were quantified directly in cell culture supernatants, when the macrophages were treated with LPS (250 ng/ml) and ATP (1mM), as a positive control for IL-1β induction (Mariathasan et al., 2006), or infected with zmp1+ or mutant zmp1::Kmr M. bovis BCG for 4 h. A ten-fold increase in IL-1β was detected in the supernatants of macrophages infected with zmp1::Kmr mutant mycobacteria compared to that of the zmp1+ parental strain (Fig. 7B; 250 pg/ml for the zmp1::Kmr mutant vs. 25 pg/ml for the zmp1+ strain). Experiments in vivo revealed elevated levels of IL-1β in sera from C57BL/6 mice collected 2 weeks post-infection (Suppl. Fig. S4).
Active caspase-1 cleaves pro-IL-1β yielding a smaller (17 kD) active IL-1β, but as most of the IL-1β activation occurs during its exocytosis, the active fragment is not readily visible on immunoblots of cellular lysates. Instead, levels of pro-IL-1β, which go up concomitantly with IL-1β activation (Mariathasan et al., 2006), can be monitored. At 30 min post-infection, there was no significant difference in pro-IL-1β levels. However, there were significantly higher levels of pro-IL-1β by 4 h in zmp1::Kmr infected macrophages compared to control and zmp1+ infected macrophages (Suppl. Fig. S2B). Pro-IL-1β levels in macrophages infected with zmp1+ parental bacteria were significantly lower than in control cells, suggesting that not only did mycobacteria prevent IL-1β activation but they also had the ability to down-regulate synthesis of pro-IL-1β (Suppl. Fig. S2B).
We next tested macrophage activation using nitroblue tetrazolium (NBT) in an assay for detection of superoxide. Macrophages were infected with M. bovis BCG zmp1+ and zmp1::Kmr strains for 1 h in the presence of NBT and assayed using confocal and phase contrast microscopy. There was a significant increase in superoxide levels accumulated in macrophages infected with zmp1::Kmr mycobacteria when compared to the zmp1+ bacilli (Fig. 7C). This effect was inhibited by diphenyliodonium (DPI), an inhibitor of the NADPH oxidase, and Rottlerin, a protein kinase C inhibitor. DPI also inhibited colocalization of zmp1::Kmr mycobacteria with CD63 (Suppl. Fig. S1). Macrophage activation in this process was independent of TNF-α action, as TNF-α levels in culture supernatants did not differ between the wild-type and zmp1 mutant strains (Suppl. Fig. S5).
We next examined whether IL-1β knockdown in macrophages can alter consequences for mycobacteria associated with zmp1 inactivation. The effects of pro-IL1β and pro-caspase-1 siRNA knockdowns were tested by monitoring both zmp1::Kmr phagosome maturation (Fig. 7D) and mycobacterial survival (Fig. 6B). Macrophages were transfected with scrambled siRNA control, caspase-1 siRNA, or IL-1β siRNA for 24 h prior to infection with mutant zmp1::Kmr M. bovis BCG. Following infection for 1 h, CD63 distribution was quantified using confocal microscopy. Transfection with caspase-1 or IL-1β siRNA significantly reduced phagolysosomal localization of M. bovis BCG zmp1::Kmr (Fig. 7D). A knockdown of IL-1β increased survival of zmp1::Kmr bacteria to the level seen with the zmp1+ strain (Fig. 6B).
If increased phagolysosome biogenesis and killing of zmp1::Kmr mutant mycobacteria are a direct result of IL-1β activation then addition of exogenous IL-1β should have comparable effects on zmp1+ mycobacteria. To test this hypothesis, macrophages were infected with zmp1+ M. bovis BCG and incubated for 1 h without or with recombinant IL-1β, and assayed for colocalization of mycobacterial phagosomes with CD63. A dose-dependent increase in mycobacterial phagolysosome biogenesis was seen when compared to untreated controls (Fig. 7E), showing that IL-1β was sufficient to induce mycobacterial phagosome maturation. Thus, zmp1 was not required in defense against IL-1β-induced effectors in macrophages, but instead was necessary to pre-empt IL-1β activation. These results demonstrate that zmp1 is required by mycobacteria to prevent activation of the caspase-1/IL-1β inflammasome, with critical downstream consequences for M. tuberculosis intracellular survival and virulence.
Innate and adaptive cell-mediated immunity plays a major role in the host response against M. tuberculosis (Flynn and Chan, 2001). Among the cytokines that mediate immunological resistance to tuberculosis TNF-α, IL-12, and IFN-γ play essential roles (Altare et al., 1998; Kamijo et al., 1993; Keane et al., 2001), whereas during immunological recall response, Th17 cells, producing IL-17, are important in recruiting Th1 cells to the periphery in the lung (Khader et al., 2007). Less is known about the early innate immune response and initial control of M. tuberculosis, although the cytokines such as IL-12 and TNF-α can induce early IFN-γ production and appear essential in controlling the disease (Flynn and Chan, 2001). In this work, we have demonstrated that IL-1β, a very early component of innate immunity responses, is an effective anti-tuberculosis agent when induced or exogenously supplied. Complementary to our findings, a recent report has indicated that IL-1 receptor may play a role in the control of M. tuberculosis (Fremond et al., 2007). Furthermore, we show here that M. tuberculosis inhibits inflammasome activation to prevent processing and of caspase-1 and IL-1β. Either a genetic inactivation of the zmp1 gene or a more general attenuation of the bacilli can bring about inflammasome and IL-1β activation, normally pre-empted by the wild type M. tuberculosis.
The present study indicates that a significant aspect of tubercle bacillus pathogenicity is its ability to suppress inflammasome activation in host macrophages. The process of inhibiting host cell inflammasome activation depends on the M. tuberculosis zmp1 gene. The inflammasomes activated in the absence of zmp1 depend on IPAF and ASC. In the simplest rendition of inflammasome organization, IPAF and ASC belong to different inflammasomes (Martinon and Tschopp, 2007). For example, only IPAF, but not ASC, and caspase-1 are implicated in Legionella phagosome maturation through IPAF-dependent recognition of Legionella flagellin (Amer et al., 2006). Although IPAF can directly interact with Caspase-1 via its CARD domain, ASC can homo-ologomerize (Fernandes-Alnemri et al., 2007), and, since it contains both the pyrin and the CARD domains, it may also participate as recently pointed out (Kanneganti et al., 2007; Mariathasan and Monack, 2007) in linking IPAF with Caspase-1. The above relationships may underlie both our observations with mycobacteria and those by Nuñez and colleagues with Shigella (Suzuki et al., 2007). In the latter case, shigellae, which are like mycobacteria nonflagellated microorganisms, induce caspase-1 activation and IL-1β processing in a process dependent on both IPAF and ASC (Suzuki et al., 2007). Alternatively to the above scenario of IPAF and ASC engagement in a common inflammasome, a temporal succession of IPAF- and ASC-dependent inflammasomes may be involved, similar to the time-resolved relationships implicating both IPAF and ASC inflammasomes in response to Salmonella (Mariathasan et al., 2004).
Activation of the caspase-1/IL-1β inflammasome is an important first line of defense against microbes (Drenth and van der Meer, 2006; Martinon et al., 2002; Petrilli et al., 2005; Saleh, 2006; Scott and Saleh, 2006). However, its strong potential in protection against mycobacteria has hitherto not been recognized, albeit activation of caspase-1 has been observed for mycobacterial lysates (Netea et al., 2006) and a recent report has indicated the importance of IL-1R-mediated signaling in tuberculosis immunity (Fremond et al., 2007). Our findings that M. tuberculosis blocks inflammasome activation underscore the strong but unrealized potential of IL-1β action against mycobacteria, and are in keeping with the observations that certain pathogenic microorganisms modulate inflammasome activation in ways best suited to their infectious cycles. Similarly to M. tuberculosis, the cowpox virus inhibits inflammasome activation via a direct block of caspase-1 activation (Johnston et al., 2005). In contrast, Shigella and Salmonella activate the inflammasome and utilize the ensuing inflammation for invasion and disease progression (Lara-Tejero et al., 2006; Monack et al., 2000; Schroeder and Hilbi, 2006). Based on our work, it is possible to conclude that M. tuberculosis dampens this early host response by limiting the production and activation of IL-1β. By using a mycobacterial mutant that is incapable of suppressing IL-1β production, we have demonstrated that phagosome maturation block, intracellular survival of mycobacteria, and full virulence in an animal model of tuberculosis depend on active suppression of inflammasome activation by the tubercle bacilli.
The exact details of how the zmp1 gene product acts are not known. Its predicted sequence indicates that it is a Zn2+ metalloprotease. Zn2+ metalloproteases are ubiquitous, found both in prokaryotes and eukaryotes where they carry out diverse functions. Some, like the matrix metalloproteases, can be promiscuous acting as gelatinases or collagenases. Others have much stricter specificities, and can be highly specialized with very narrow and unique substrates. Zn2+ metalloproteases from pathogenic microorganisms include clostridial botulinum and tetanus toxins that cleave SNAREs and prevent neurotransmitter exocytosis; anthrax lethal factor that proteolytically inactivates p38MAPK; and Pseudomonas aeruginosa elastase that degrades surfactant protein D (Alcorn and Wright, 2004; Lalli et al., 2003; Park et al., 2002). The putative mycobacterial Zn2+ metalloprotease, Zmp1, possibly when released from M. tuberculosis may act to prevent the activation of the inflammasome. However, indirect mechanisms cannot be excluded at present. Regardless of the details of its action, Zmp1 through this work, has been identified as an important virulence determinant and represents a potentially useful drug target. The finding that IL-1β, when activated, can sway the balance towards M. tuberculosis phagosome maturation and elimination of the bacilli, represents an important new insight that should allow development of novel strategies for prevention of tuberculosis.
DMEM was from Gibco, fetal bovine serum (FBS) from Hyclone, L-glutamine from Biowhittaker, Histopaque, chloramphenicol and kanamycin from Sigma, and ADC (10% bovine serum albumin fraction V, glucose and catalase) from Difco. IL-1β and TNFα ELISA kits, and recombinant mouse IL-1β were from R & D Systems. Anti-CD63 antibody was from Santa Cruz Biotech, anti-caspase-1 and anti-caspase-3 antibody were from Cell Signaling. Neutralizing antibody to IL-1β was from Abcam, control rat IgG1 was from Serotec. The Alexa-488 and Alexa-568 secondary antibodies, and lysotracker were from Molecular Probes. Texas Red-labeled transferrin was from Jackson Immuno Research. The siRNAs were from Dharmacon. Diphenyliodonium salt, rottlerin, and nitroblue tetrazolium salt were from Sigma, z-VAD-fmk and Ac-YVAD-cmk were from Alexis Biochemicals, and BAPTA-AM was from Calbiochem. C57Bl/6 mice were from Jackson Laboratory. M. tuberculosis culture filtrate protein (CFP) was obtained from J. Beslisle
Mouse macrophage cell-lines J774A and RAW 264.7 were maintained in DMEM, 4 mM L-glutamine, and 10% FBS. Human peripheral blood monocytes were prepared from normal individual donors by density gradient centrifugation (400 × g for 30 min) through a Ficoll-Hypaque gradient (Sigma). Adherent monocytes were matured into macrophages by incubating for 5 days in RPMI containing 10% human AB serum. Bone marrow macrophages were cultured from femurs of C57Bl/6 WT and Caspase-1−/− mice and maintained in media containing DMEM, 4 mM L-glutamine, 20% FBS, 30% Macrophage-colony stimulating factor (L-cell conditioned media). Supernatants from one week old L-cell cultures were aliquoted and stored at −80°C until use M. tuberculosis var. bovis BCG 1721 wild-type and zmp1::Kmr mutant strains and M. tuberculosis H37Rv 1424 wild-type, zmp1::Kmr mutant and zmp1::Kmr(zmp1+) complemented strains were grown until mid-exponential phase in 7H9 broth supplemented with 0.5% Tween 80, 0.2% glycerol and ADC/OADC at 37°C. When appropriate, antibiotics were added at the following concentrations; kanamycin at 50 μg/ml, streptomycin at 100 μg/ml and hygromycin at 50 μg/ml. Mycobacteria either expressed GFP or were fluorescently labeled with 5 mg/ml Texas red-X or 0.1 mg/ml Alexa 488 in PBS for 1 h. M. tuberculosis H37Rv 1424 (Sander et al., 2004), a derivative of M. tuberculosis H37Rv (ATCC), or M. bovis BCG 1721, a derivative of M. bovis BCG Pasteur, carrying a non-restrictive rpsL alteration, were used to generate isogenic deletion mutants deficient in zmp (Rv0198c). For all experiments, mycobacteria were homogenized to remove clumps and opsonized in DMEM supplemented with 10% FBS for 30 min prior to infection. All manipulations with live M. tuberculosis H37Rv were carried out under Biosafety Level 3 conditions.
M. tuberculosis Zn2+ metalloprotease gene (zmp = Rv0198c; open reading frame: start −236.507; stop 234.519; http:genolist.pasteur.fr/TubercuList) and its flanking region were isolated from a M. tuberculosis BAC clone library (Brosch et al., 1998) and used to generate allelic replacement mutants, as described (Springer et al., 2001). For complementation, a 4423 bp NotI zmp fragment was cloned into plasmid pMV361-hyg linearized with HpaI, transformed into M. tuberculosis zmp1::Kmr and selected on hygromycin. Rabbit polyclonal antibody against a C-terminal 15 amino acids peptide of Zmp1 was raised and affinity purified using commercial sources.
C57/BL6 mice were infected by aerosolization and M. tuberculosis H37Rv lung bacterial burden analyzed as described (Dittrich et al., 2006). Macrophages were infected with various strains for 1–2h lysed in cold water, and bacterial colony counts determined for input, 0 day (1–2h post-infection) and other time-points indicated as described (Kyei et al., 2006; Roberts et al., 2006).
RAW 264.7 cells were resuspended in a nucleoporator buffer at a density of 5×106 with 1.5 μg of cognate siRNA. Cells were nucleoporated (Amaxa) for 24–48 h prior to infection.
IL-1β was from R and D Systems, Inc. 10 μg of rat monoclonal neutralizing Ab to IL-1β or control rat IgG1 was added at the time of infection and maintained during the course of infection. Rottlerin and diphenylene iodonium were used at a final concentration of 5 μM. z-VAD-fmk and Ac-YVAD-CMK were used at a final concentration of 50 μM. Western blotting was done using antibodies to caspase-1 and IL-1β (Cell Signaling).
Macrophages were infected at an MOI of 10:1 with wild-type (zmp1+) or mutant (zmp1::Kmr) BCG-GFP, in the presence of 150 μg/ml NBT. Where indicated, appropriate inhibitors were added. After 30 min, macrophages were washed three times in PBS and fixed with 2% paraformaldehyde. Cells were visualized using phase contrast and confocal microscopy, and inspected for characteristic formazan precipitate indicative of superoxide formation.
BCG lysates were prepared by bead-beating as previously described (Springer et al., 2001). Uninfected, wild-type or zmp1::Kmr BCG-infected macrophages (30 min time point) were lysed for 1 h in lysis buffer (2% Igepal CA630, 1:25 complete EDTA-free protease inhibitors, 2mM PMSF, 2mM EDTA, 10mM Tris, 0.5% Deoxycholic acid, 1mM sodium orthovanadate, 50mM sodium fluoride) and solubilized in SDS sample buffer. Samples were resolved by 12.5% SDS-PAGE and transferred to nitrocellulose. Membranes were blocked in 5% dry milk in PBS containing 0.1% Tween, probed with the corresponding secondary Ab (Pierce) and protein detected via ECL chemiluminescence (Pierce).
Data represented by mean ± SEM were analyzed using student’s t-test. In some cases, data pooled from three independent experiments were analysed using contingency tables with Fisher’s exact least square determination (Graphpad Prism 4).
We thank E. Böttger for collaboration and support during mutant construction, B. Wildmann for cloning, and S. Cole for the BAC clone library. This work was supported by grants from Swiss National Science Foundation (contract 3200-068 488), Swiss Lung Foundation, and Wolfermann-Nägeli Foundation to P.S. and National Institutes of Health grants AI42999 and AI45148 to V.D.