Mycobacterium marinum Erp Is a Virulence Determinant Required for Cell Wall Integrity and Intracellular Survival
http://www.100md.com
感染与免疫杂志 2006年第6期
Departments of Microbiology Immunology Medicine, University of Washington, Seattle, Washington 98195
ABSTRACT
The Mycobacterium tuberculosis exported repetitive protein (Erp) is a virulence determinant required for growth in cultured macrophages and in vivo. To better understand the role of Erp in Mycobacterium pathogenesis, we generated a mutation in the erp homologue of Mycobacterium marinum, a close genetic relative of M. tuberculosis. erp-deficient M. marinum was growth attenuated in cultured macrophage monolayers and during chronic granulomatous infection of leopard frogs, suggesting that Erp function is similarly required for the virulence of both M. tuberculosis and M. marinum. To pinpoint the step in infection at which Erp is required, we utilized a zebrafish embryo infection model that allows M. marinum infections to be visualized in real-time, comparing the erp-deficient strain to a RD1 mutant whose stage of attenuation was previously characterized in zebrafish embryos. A detailed microscopic examination of infected embryos revealed that bacteria lacking Erp were compromised very early in infection, failing to grow and/or survive upon phagocytosis by host macrophages. In contrast, RD1 mutant bacteria grow normally in macrophages but fail to induce host macrophage aggregation and subsequent cell-to-cell spread. Consistent with these in vivo findings, erp-deficient but not RD1-deficient bacteria exhibited permeability defects in vitro, which may be responsible for their specific failure to survive in host macrophages.
INTRODUCTION
Mycobacteria, typified by Mycobacterium tuberculosis, are intracellular pathogens that cause persistent infections despite vigorous host responses (13), and much of the work aimed at understanding mycobacterial pathogenesis has focused on the interactions between mycobacteria and the immune system. The infectious process can be viewed as a series of sequential steps, in which the bacteria interface with the immune system in multiple complex ways. Infection begins with the recruitment of macrophages to the site of infection and phagocytosis of infecting organisms. These infected macrophages then migrate to deeper tissues, and additional monocytes are recruited to the infected cell. As bacterial growth ensues, macrophages aggregate and differentiate into the characteristic epithelioid macrophages that constitute granulomas. Later, as adaptive immunity develops, other immune cells such as lymphocytes are recruited to the granulomas. The bacterium-host interactions that orchestrate this complex program are not well understood, and the bacterial determinants that permit both early growth and long-term persistence are now being identified.
The last decade has seen significant advancements in the genetic manipulation of mycobacteria to define and characterize mycobacterial virulence factors (13, 24, 32, 41). Mutations that reduce virulence have been identified, and the majority of these affect the structure and/or the function of the highly complex cell wall and outer lipid layer. These include mutations in genes involved in cell wall lipid synthesis, composition, or transport (8, 9, 11, 14, 19, 20, 22, 23, 28); the PGRS cell surface proteins (2, 36, 39); and the RD1 locus that encodes a putative secretion system (21, 25, 26, 31, 35, 45). However, a clear understanding of the underlying mechanisms for these virulence deficiencies has been lacking.
The Mycobacterium protein Erp, another cell surface component, was the first Mycobacterium virulence protein to be characterized by targeted deletion (4). Despite a series of genetic and biochemical characterizations (5, 16, 17, 29, 30), its role in virulence remains poorly understood. This Mycobacterium-specific secreted/cell surface protein is comprised of three domains: a conserved N-terminal domain with a canonical Sec-dependent signal sequence, a repetitive central domain, and a conserved hydrophobic C-terminal domain that appears to be involved in a loose anchoring of Erp to the cell surface (30). Virulence studies with M. tuberculosis and M. bovis erp mutants revealed that the bacteria are growth attenuated in cultured macrophages and in a mouse model of infection (4, 5).
M. marinum, a close genetic relative of M. tuberculosis (43) (http://www.sanger.ac.uk/Projects/M_marinum/), is used as a model for mycobacterial pathogenesis (3, 15, 18, 21, 22, 33, 34, 42). M. marinum is the causative agent of systemic granulomatous infections of ectotherms, such as fish and frogs (1), and peripheral chronic granulomatous disease (fish tank granulomas) in humans (44). Experimental inoculation of leopard frogs (Rana pipiens) results in a chronic subclinical infection much like latent tuberculosis in humans (37). Given that M. tuberculosis and M. marinum cause similar pathology in infected hosts, have similar endocytic trafficking pathways in cultured macrophages (3), and share virulence determinants (21, 45), it is likely that these related pathogens use similar survival strategies in vivo.
We have previously developed a zebrafish embryo model of infection to characterize the earliest events after mycobacterial infection (15). Zebrafish embryos are transparent during the first 2 weeks of life, allowing the infectious process to be visualized in real time. Studies using this model revealed that M. marinum injected into the bloodstream are rapidly taken up by host macrophages, usually within 30 min of injection. Over the next 24 h, infected macrophages exit the bloodstream, migrating into tissues. These infected macrophages next begin to aggregate and further recruit additional uninfected macrophages, some of which ultimately differentiate into the epithelioid macrophages typical of granulomas. The timing of aggregation, although dependent on both the virulence of the strain used and the number of organisms injected, usually occurs 2 to 5 days postinjection (J. M. Davis and L. Ramakrishnan, unpublished results). This system has proven valuable in clarifying the mechanism of action of Mycobacterium virulence determinants. For example, in a previous study we demonstrated that the M. marinum RD1 mutant (45), like its M. tuberculosis counterpart, is growth attenuated when assayed by bacterial counts in bulk-cultured macrophages. However, detailed analysis using the zebrafish embryo infection model showed that RD1-deficient bacteria are capable of growth in individual macrophages in vivo but fail to induce macrophage aggregation and subsequent intercellular spread (45). This finding revealed that intracellular growth in individual macrophages is insufficient to ensure bacterial proliferation and that additional host cells must be recruited and infected.
In the present study, we characterize another M. marinum virulence determinant, the erp locus, using a combination of in vivo and in vitro assays. We found that, like the M. marinum RD1 locus, erp is required for bacterial growth in vivo. However, by using the zebrafish embryo model, we show that, in contrast to the RD1 locus, erp is required for growth and/or survival in individual macrophages. The growth attenuation of this mutant in vivo may be explained at least in part by a permeability defect.
MATERIALS AND METHODS
Bacterial strains and growth conditions. Mycobacterial strains used for the present study were derived from a human clinical isolate, strain M (ATCC BAA-535) and are shown in Table 1. M. marinum was grown in Middlebrook 7H9 broth (Difco) supplemented with 0.5% bovine serum albumin, 0.005% oleic acid, 0.2% glucose, 0.2% glycerol, 0.085% sodium chloride, and 0.05% Tween 80 or on Middlebrook 7H10 agar (Difco) supplemented with 0.5% bovine serum albumin, 0.005% oleic acid, 0.2% glucose, 0.2% glycerol, and 0.085% sodium chloride. Media were supplemented with 20 μg of kanamycin/ml and 50 μg of hygromycin/ml when appropriate. pGFPHYG2 was obtained by inserting the hygromycin resistance cassette of plasmid pRF498H (40) into the NsiI sites of plasmid pMSP12::GFP (10), thus disrupting the kanamycin resistance cassette.
Construction of the M. marinum erp mutant. A 2,857-bp PstI/BglI fragment containing the erp locus was isolated from a M. marinum genomic cosmid library and inserted into pBluescript SK(+) (Stratagene). The erp gene was interrupted by removing a central 316-bp SmaI fragment and replacing it with the aph (kanamycin resistance) cassette of pYUB213 (a gift from W. Jacobs) to generate an erp::aph mutation. The sacB gene was then introduced at the XbaI site of the polylinker, and the resulting plasmid, pERPKO, was used to transform M. marinum strain M. Kanamycin-resistant (Kanr) sucroses merodiploid colonies were isolated and streaked onto medium containing kanamycin and sucrose to select for sucroser double recombinants. One such isolate was saved and named KK33 (Table 1). Complementation of erp::aph was achieved by cloning a 2,284-bp NruI/MscI fragment containing the wild-type allele into pYUB178Hyg to generate pERPcomp1, which was then transformed into KK33 and integrated into the att site to produce strain KK60 (Table 1). The presence of the knockout and the complementing alleles were confirmed by Southern blot analysis (data not shown).
Macrophage infections. Growth in J774 macrophages was determined as described previously (36). Briefly, 5 x 104 J774A.1 murine macrophages in grown in Dulbecco modified Eagle medium (Gibco) supplemented with 10% fetal bovine serum were seeded into the wells of a 24-well plate and allowed to adhere overnight. The following day, the cells were infected with M. marinum for 1 h at 33°C. Infected medium was replaced with 1 ml of fresh medium containing 40 μg of streptomycin/ml. To determine the intracellular bacterial counts at various times postinfection, cells were washed three times with 1x phosphate-buffered saline, lysed with 100 μl of 1% Triton X-100, scraped with a pipette tip, and diluted 10-fold with 1x phosphate-buffered saline. Dilutions were plated onto 7H10 agar and incubated at 33°C for 7 to 14 days.
Frog infection. Wild-caught male Rana pipiens (approximately 10 to 12 cm) were obtained from J. M. Hazen (Alberg, VT). Frogs were infected by intraperitoneal injection as described previously (12). One, two, or twelve weeks later, liver and spleen tissues were harvested, and the bacterial CFU were enumerated by plating on 7H10 agar supplemented with amphotericin B and, in the case of KK33 and KK60, with kanamycin. Frogs were housed and all animal procedures were carried out in accordance with University of Washington IACUC regulations.
Zebrafish procedures. Zebrafish embryos at 32 h postfertilization were infected by microinjection at the caudal vein as previously described (12, 15). To assess virulence of various M. marinum strains, cohorts of embryos hatched at the same time were infected (45 embryos per strain) and monitored daily for mortalities. Each treatment group of 45 embryos was distributed into three tanks of fifteen and scored separately to account for the possible contribution of tank effects. Mean time to death (MTD) was scored for each tank, and no significant differences were found between tanks of fish infected with the same strain. The data for fish infected with a given strain were then pooled, and MTD values between strains were compared by using analysis of variance (ANOVA). For the visualization of bacteria in vivo, fluorescent strains were used (Table 1), and embryos were mounted for observation as described previously (12, 15, 45). When scoring the levels of infection in zebrafish embryos, the individual making this assessment was blinded to the strain used to infect each embryo so as to avoid bias.
MIC assays. The agar dilution method was used to determine the MIC for each strain (22). Individual plates of 7H10 agar (supplemented as described above) were prepared (30-ml volume), with twofold dilutions of the antibiotic being tested. Plates were spotted with 10 μl of early-log-phase cultures in triplicate, and all strains were spotted onto the same plate. Plates were incubated at 33°C and scored after 8 and 11 days.
Statistical analyses. Statistical analysis was performed by using In-Stat software (Graphpad Software, Inc.).
RESULTS
Targeted deletion of the M. marinum erp locus. We compared the previously identified M. marinum erp locus (17) (http://www.sanger.ac.uk/Projects/M_marinum/) with its M. tuberculosis homologue and found an overall 76% amino acid identity and 81% similarity. M. marinum erp is located between the glf and csp open reading frames, making its genomic arrangement syntenous with those of other Mycobacterium species. The central repetitive domain of the predicted protein contains two additional PGLTS repeats compared to that of M. tuberculosis (17), making it the closest identified homologue to those of the M. tuberculosis complex. A BLAST search of the completed M. marinum genome sequence failed to identify any additional open reading frames with homology to the M. tuberculosis erp sequence, suggesting that M. marinum contains only one erp allele. To investigate the role of erp in M. marinum pathogenesis, we generated a loss-of-function mutation by introducing a Kanr cassette via homologous recombination near the end of the N-terminal domain (see Materials and Methods). This mutation was complemented by introducing the wild-type erp allele at the att site (Table 1). As in the case of the M. tuberculosis erp mutant (4), growth of the M. marinum erp::aph mutant strain in liquid culture was similar to that of the wild type (data not shown). However, growth on 7H10 agar was mildly compromised since erp::aph mutant colonies appeared 3 days later than did the wild type. A small colony phenotype has been reported for erp-deficient mutants in other species (30).
The M. marinum erp mutant is growth attenuated in cultured macrophages and adult animals. The M. tuberculosis erp locus is required for growth both in cultured macrophages and in a mouse model of infection (4). We found M. marinum erp to be similarly required for growth in cultured macrophages. As shown in Fig. 1A, the erp::aph strain failed to grow in J774A.1 macrophages over an 8-day time course. This growth attenuation was indistinguishable from that observed for a M. marinum mutant that lacks the RD1 virulence locus (45). In contrast, the wild-type and complemented mutant strains grew approximately 50-fold over the 8-day monitoring period.
To examine whether erp-deficient bacteria were also attenuated in vivo, we assayed their growth in a natural host species, the leopard frog, used to study M. marinum pathogenesis in the laboratory (6, 37). Intraperitoneal infection of frogs with wild-type bacteria resulted in an increased tissue burden over a period of 12 weeks (Fig. 1B). In contrast, the number of erp::aph mutant bacteria decreased slightly, and by 12 weeks postinfection it was 1.5 logs below that of the wild type (P < 0.05). RD1 mutant bacteria were similarly attenuated (P < 0.05), as shown previously (45). As was the case in cultured macrophages, complementation of the erp::aph mutation restored growth in frogs. Thus, the M. marinum and M. tuberculosis erp-deficient mutants had comparable phenotypes with respect to growth in culture, macrophage monolayers, and animals. Furthermore, the attenuation characteristics of the M. marinum erp-deficient mutant were indistinguishable from those of the M. marinum RD1 mutant (31, 45).
The M. marinum erp::aph mutant is attenuated in the zebrafish embryo infection model. The M. marinum erp::aph mutant was next analyzed for virulence in the zebrafish embryo infection model (15). In this model, embryos are infected by microinjection with a small number of bacteria (ranging from 10 organisms to 300) at approximately 32 h postfertilization, shortly after the beginning of circulation. Injection directly into the caudal vein results in immediate dispersal of the bacteria via the bloodstream, their subsequent uptake by circulating macrophages, and migration of infected macrophages into the tissues where they aggregate to form granulomas. To assess the virulence of the erp::aph mutant, 45 embryos were infected per strain, and survival was monitored over a 2-week time course. Embryos infected with the erp::aph and RD1 mutants fared better than those infected with wild-type and erp::aph complemented strains (Fig. 2A). The MTD for embryos infected with the wild-type and complemented strains was significantly shorter than for those infected with the mutants (Table 2).
To determine whether these differences in embryo survival resulted from differential growth of the bacteria in vivo, we undertook a microscopic analysis of embryos infected with wild-type and mutant bacteria expressing gfp from a strong constitutive promoter (see Table 1 and Materials and Methods). Ten embryos were injected per strain and were viewed by fluorescence microscopy 1 h later to verify that the inocula were similar and then again at 5 days postinoculation to examine the degree of bacterial proliferation. As shown in Fig. 2B, the overall burden of bacteria was similar for all embryos at 1 h postinfection, but by 5 days postinfection wild-type-infected embryos harbored more bacteria than embryos infected with erp::aph or RD1 bacteria. Injection with either of these strains resulted in low burden infections and a paucity of macrophage aggregates typical of the wild-type infection (Fig. 2B, arrows). We could not quantify the growth of these strains in zebrafish embryos using bacterial counts, since we were unable to recover erp::aph mutant bacteria using the lysis procedure required for this analysis (45). This failure to recover erp::aph mutant bacteria was specific as wild-type and RD1 bacteria were readily recovered (H. Volkman, D. Beery, and L. Ramakrishnan, unpublished data). These experiments revealed that erp::aph and RD1 bacteria were similarly growth attenuated in zebrafish embryos.
Erp-deficient bacteria fail to grow intracellularly in a zebrafish embryo model of infection. The intracellular growth by mycobacteria has traditionally been assayed in cultured cell lines by enumerating the total growth of the bacterial population in a population of macrophages as a surrogate measure of growth in the individual cells. However, recent studies of the RD1 locus (21, 25, 45) have shown that this can be a faulty assumption. In our previous work we found that RD1 bacteria seemingly failed to grow intracellularly (as measured in such an assay); however, closer examination using the zebrafish embryo model revealed that these bacteria are capable of intracellular growth but rather fail to spread among host cells. We concluded that in intracellular growth assays in cultured cell lines, a limited amount of growth may occur intracellularly, but unless bacteria continue to spread to new host cells, no increase in the overall population is observed. Therefore, we were similarly interested to use this infection model to pinpoint the precise nature of attenuation for the erp::aph mutant.
Prior detailed analyses in the zebrafish embryo have revealed that infection by wild-type bacteria proceeds in sequential steps (15). Upon injection of M. marinum, macrophages migrate to the site of infection and phagocytose the bacteria. Infected macrophages then migrate back into tissues and ultimately aggregate into granulomas. In initial experiments with the erp::aph strain we did not observe any obvious defects in macrophage recruitment, phagocytosis of bacteria, or migration into tissues (D. Beery and L. Ramakrishnan, unpublished observations), suggesting that, like the RD1 mutant (45), the erp::aph mutant was competent for these very early steps. We next performed a detailed comparison of embryos infected with wild-type, RD1, and erp::aph bacteria at 5 days postinfection by using fluorescence and differential interference contrast (DIC) microscopy at high magnification. Similar to what was observed at low magnification (Fig. 2B), abundant macrophage aggregates were found in the wild-type-infected embryos, whereas fewer were observed in embryos infected with either mutant strain (data not shown). With respect to the infection burden of individual (nonaggregated) macrophages, wild-type-infected embryos contained macrophages representing a wide spectrum of infection levels, as expected. Single macrophages contained anywhere from one to >10 bacteria (Fig. 3A). However, a striking difference was apparent between embryos infected with the two mutant strains. Macrophages of embryos infected with the RD1 mutant typically harbored large numbers of bacteria (Fig. 3B), whereas those from embryos infected with the erp::aph mutant frequently harbored very few bacteria (Fig. 3C). Furthermore, erp::aph bacteria were often dim (Fig. 3C), suggesting that they are being killed (in contrast, all strains were similarly bright in culture). These observations suggested that erp::aph and RD1 mutant bacteria might be attenuated for distinct reasons, despite their similarities in other assays.
To quantify the observed differences, we studied 10 infected embryos for each of the three strains. For each embryo, the entire region posterior to the end of the yolk extension (Fig. 3D) was surveyed, and individual (nonaggregated) infected macrophages found in this region were scored. Each macrophage was scored based on the number of bacteria it contained; <5 bacteria, 5 to 10 bacteria, or >10 bacteria. For each embryo, we calculated the proportion of its macrophages that were lightly infected (<5 bacteria), moderately infected (5 to 10 bacteria), and heavily infected (>10 bacteria). Consistent with the qualitative findings (Fig. 3A to C), we observed that significantly more infected macrophages fell into the high-burden category (>10 bacteria per cell) when the embryos were infected with RD1 bacteria than when they were infected with erp::aph bacteria (P < 0.01). The converse was also true; significantly more infected macrophages fell into the low-burden category (<5 bacteria per cell) when the embryos were infected with erp::aph bacteria than when they were infected with RD1 bacteria (P < 0.01). No difference was observed in the proportions of moderately infected macrophages. Wild-type-infected embryos had an intermediate number of macrophages in each category. This was due to the fact that during infection by wild-type bacteria, infected macrophages are titrated out into granulomas, which were not scored in this assay. In contrast, the RD1 and erp::aph mutant-infected macrophages were less likely to form aggregates, and the discrepancy in the number of bacteria they contained reveals that these two mutant strains are attenuated for different reasons despite their identical attenuation phenotypes in virulence assays that measure overall organism burden (Fig. 1). erp-deficient bacteria are attenuated primarily because of their lack of intracellular growth and/or survival, whereas RD1 mutant bacteria grow successfully in macrophages but fail to induce subsequent aggregation and intercellular spread (Fig. 4) (45).
Erp is required for the Mycobacterium permeability barrier. As discussed above, we specifically failed to recover erp::aph mutant bacteria extracted from zebrafish embryos. The extraction method requires that embryo lysates be plated on 7H11 agar in the presence of polymixin B, trimethoprim, and carbenecillin to suppress the growth of intestinal and other aquatic flora (H. Volkman and L. Ramakrishnan, unpublished). Since erp::aph bacteria were clearly present (as visualized by fluorescence microscopy, Fig. 2B), we suspected that they might simply be more susceptible to the antimicrobial compounds in the agar. Therefore, we determined the MIC for wild-type and mutant bacteria of a panel of antimicrobial compounds. As a control we included an M. marinum strain that contains a transposon insertion in kasB, a gene required for mycolate synthesis (22). kasB-deficient bacteria are more permeable to the hydrophobic detergent deoxycholate, and are consequently more susceptible to a variety of lipophilic antibiotics such as rifampin and erythromycin, but retain normal susceptibility to hydrophilic compounds. As shown in Table 3, the erp::aph mutant strain was more susceptible to the lipophilic antibiotics rifampin, erythromycin, and chloramphenicol than were the wild-type and complemented strains. The degree of sensitivity was similar to that observed for kasB-deficient strain and was in contrast to the RD1 strain, which demonstrated normal susceptibility. These findings suggest that the M. marinum erp::aph mutant has increased permeability to a variety of hydrophobic agents, which may explain its decreased growth and/or survival in macrophages.
DISCUSSION
In the present study, we have characterized the role of the M. marinum erp gene during infection. Like its M. tuberculosis and M. bovis counterparts (4, 5), we find that M. marinum Erp is a virulence determinant, required for growth in cultured macrophages and in vivo. In this regard, our findings add to a growing body of evidence that the molecular determinants of M. tuberculosis virulence during infection of mammals are also used by M. marinum during infection of its natural hosts (7, 21, 36, 39, 45). That Erp function is conserved among mycobacteria is further supported by studies in M. tuberculosis showing that the erp loci of M. leprae and M. smegmatis are competent for complementation of an M. tuberculosis erp mutation in the context of a mouse infection (16). Such commonalities suggest that the basic mechanisms of Mycobacterium pathogenesis are conserved.
As was the case with the RD1 locus (45), the zebrafish embryo infection model has refined our understanding of the role of the erp locus in Mycobacterium virulence. Using this model we find that Erp-deficient bacteria have an early and intrinsic defect in intracellular growth and/or survival in macrophages. This is in contrast to RD1 bacteria, which proliferate normally in individual macrophages but have diminished and delayed formation of macrophage aggregates and are defective in intercellular spread within these aggregates (45). This difference in intracellular growth was not revealed in the standard assay designed to monitor net bacterial growth in a population of cultured macrophages (Fig. 1A) but rather was observed only when a detailed comparison of the two mutants was made during infection of zebrafish embryos (Fig. 3). The points at which these two determinants impact the infectious process are depicted in Fig. 4. Although both begin to exert their influence in the early stages of infection, it is clear that erp is required earlier than is the RD1 locus. Erp-deficient bacteria do not survive or grow within individual macrophages, so that infection is thwarted prior to macrophage aggregation into granulomas. It is noteworthy that the differences in the capacity of these two mutants for intracellular growth do not result in a differential outcome later in infection (Fig. 1B and 2A and Table 2). This observation reinforces our previous findings with the RD1 mutant that the capacity for intracellular growth alone is not sufficient to confer full virulence and that the ability of mycobacteria to proliferate to their maximum extent also depends on the cell-to-cell spread impacted by the RD1 locus. Furthermore, it would appear that these two aspects of the infectious process (intracellular growth and intercellular spread) are important throughout infection, since both mutants are comparably attenuated in the short-term embryo infection model and in the long-term chronic frog model. That both mutants are capable of establishing infection, albeit at a lower tissue burden than for the wild type, reinforces the notion that mycobacteria use multiple redundant mechanisms to ensure their survival (9, 14, 36, 39).
The finding that the erp::aph mutant is compromised for intracellular growth, coupled with its failure to grow on Middlebrook 7H11 agar replete with antibiotics, suggested a permeability defect that might explain both phenotypes. Indeed, upon further examination we found that the loss of Erp renders bacteria more permeable to a variety of lipophilic compounds. These permeability defects may render the bacteria more susceptible to intracellular host defenses such as reactive oxygen and nitrogen intermediates, degradative enzymes, and defensins. Permeability defects have been postulated as the reason for attenuation for a variety of Mycobacterium mutants. For example, mutations in the M. marinum kasB gene involved in mycolate production rendered bacteria more permeable to deoxycholate and more sensitive to a variety of lipophilic antibiotics (22). On the other hand, mutations in the M. tuberculosis fadD26 and mmpL7 genes, which are involved in PDIM synthesis (8, 9, 14), rendered bacteria more permeable to detergent (38) and yet not more susceptible to hydrophobic antibiotics (19, 38). Clearly, the complexity of the permeability barrier conferred by the Mycobacterium cell wall and outer lipid layer are such that individual components may confer protection against different types of compounds.
That permeability defects are complex is further revealed by a comparison of the attenuation phenotypes of various permeability mutants. In M. tuberculosis, the fadD26 mutant is growth attenuated in activated, but not naive macrophages (38), whereas macrophage activation is not required to restrict growth of erp mutants (4; the present study). Indeed, some permeability mutants do not have a virulence defect, as is the case with the M. tuberculosis fbcC gene, which encodes a mycolyltransferase (27). In M. marinum, the erp and kasB mutants have similar profiles of susceptibility to lipophilic antibiotics (22; the present study) (Table 3), and yet the former is more attenuated for growth in cultured macrophages (22; C. Cosma and L. Ramakrishnan, unpublished data). In our preliminary examination of the kasB mutant in zebrafish embryos, we found that in this model too, kasB-deficient bacteria are less attenuated than erp::aph bacteria (Cosma, Kim, and Ramakrishnan, unpublished). These differences in intracellular growth could be for several reasons. First, features of the Mycobacterium permeability barrier relevant for intracellular survival may simply not be revealed by in vitro membrane permeability assays. Alternatively, the permeability defect conferred by the loss of Erp may be only partially responsible for its virulence defects and Erp may play a second role in intracellular survival, independent of its contribution to the Mycobacterium permeability barrier. Finally, the loss of KasB may have a severe enough effect on the Mycobacterium cell wall or outer lipid layer that genetic or regulatory compensations have occurred to provide better survival both in vitro and in vivo. Whatever the reason, these results underscore the importance of in vivo assays to determine the true impact on the virulence of mycobacterial permeability defects detected in vitro.
By combining a variety of infection models with in vitro assays, we have demonstrated that the Mycobacterium erp locus is required for intracellular growth and survival and that this requirement may be due in part to its role in maintaining integrity of the cell wall and outer lipid layer. The use of M. marinum as a surrogate model in the zebrafish has allowed a more precise determination of this attenuation defect than was previously apparent.
ACKNOWLEDGMENTS
We thank Hannah Volkman for assistance with embryo infections; Muse Davis for assistance with figures; Richard Burmeister for artwork; Eric Brown for M. marinum strains; and Lynn Connolly, Hannah Volkman, David Sherman, and Yong Gao for comments on the manuscript.
This study was supported by NIH grants RO1 AI036396 and RO1 AI54503 to L.R.
Present address: Centers for Molecular Medicine, State University of New York—Stony Brook, Stony Brook, NY 11794-5120.
REFERENCES
1. Aronson, J. D. 1926. Spontaneous tuberculosis in salt water fish. J. Infect. Dis. 39:315-320.
2. Banu, S., N. Honore, B. Saint-Joanis, D. Philpott, M. C. Prevost, and S. T. Cole. 2002. Are the PE-PGRS proteins of Mycobacterium tuberculosis variable surface antigens Mol. Microbiol. 44:9-19.
3. Barker, L. P., K. M. George, S. Falkow, and P. L. Small. 1997. Differential trafficking of live and dead Mycobacterium marinum organisms in macrophages. Infect. Immun. 65:1497-1504.
4. Berthet, F. X., M. Lagranderie, P. Gounon, C. Lauren-Winter, D. Ensergueix, P. Chavarot, F. Thouron, E. Maranghi, V. Pelicic, D. Portnoi, G. Marchal, and B. Gicquel. 1998. Attenuation of virulence by disruption of the Mycobacterium tuberculosis erp gene. Science 282:759-762.
5. Bigi, F., A. Gioffre, L. Klepp, M. P. Santangelo, C. A. Velicovsky, G. H. Giambartolomei, C. A. Fossati, M. I. Romano, T. Mendum, J. J. McFadden, and A. Cataldi. 2005. Mutation in the P36 gene of Mycobacterium bovis provokes attenuation of the bacillus in a mouse model. Tuberculosis 85:221-226.
6. Bouley, D. M., N. Ghori, K. L. Mercer, S. Falkow, and L. Ramakrishnan. 2001. Dynamic nature of host-pathogen interactions in Mycobacterium marinum granulomas. Infect. Immun. 69:7820-7831.
7. Brennan, M. J., G. Delogu, Y. Chen, S. Bardarov, J. Kriakov, M. Alavi, and W. R. Jacobs, Jr. 2001. Evidence that mycobacterial PE_PGRS proteins are cell surface constituents that influence interactions with other cells. Infect. Immun. 69:7326-7333.
8. Camacho, L. R., P. Constant, C. Raynaud, M. A. Laneelle, J. A. Triccas, B. Gicquel, M. Daffe, and C. Guilhot. 2001. Analysis of the phthiocerol dimycocerosate locus of Mycobacterium tuberculosis: evidence that this lipid is involved in the cell wall permeability barrier. J. Biol. Chem. 276:19845-19854.
9. Camacho, L. R., D. Ensergueix, E. Perez, B. Gicquel, and C. Guilhot. 1999. Identification of a virulence gene cluster of Mycobacterium tuberculosis by signature-tagged transposon mutagenesis. Mol. Microbiol. 34:257-267.
10. Chan, K., T. Knaak, L. Satkamp, O. Humbert, S. Falkow, and L. Ramakrishnan. 2002. Complex pattern of Mycobacterium marinum gene expression during long-term granulomatous infection. Proc. Natl. Acad. Sci. USA 99:3920-3925.
11. Converse, S. E., J. D. Mougous, M. D. Leavell, J. A. Leary, C. R. Bertozzi, and J. S. Cox. 2003. MmpL8 is required for sulfolipid-1 biosynthesis and Mycobacterium tuberculosis virulence. Proc. Natl. Acad. Sci. USA 100:6121-6126.
12. Cosma, C. L., J. M. Davis, L. E. Swaim, H. Volkman, and L. Ramakrishnan. Animal models of Mycobacterium marinum infection. Curr. Protocols Microbiol., in press.
13. Cosma, C. L., D. R. Sherman, and L. Ramakrishnan. 2003. The secret lives of the pathogenic mycobacteria. Annu. Rev. Microbiol. 57:641-676.
14. Cox, J. S., B. Chen, M. McNeil, and W. R. Jacobs, Jr. 1999. Complex lipid determines tissue-specific replication of Mycobacterium tuberculosis in mice. Nature 402:79-83.
15. Davis, J. M., H. Clay, J. L. Lewis, N. Ghori, P. Herbomel, and L. Ramakrishnan. 2002. Real-time visualization of Mycobacterium-macrophage interactions leading to initiation of granuloma formation in zebrafish embryos. Immunity 17:693-702.
16. De Mendonca-Lima, L., Y. Bordat, E. Pivert, C. Recchi, O. Neyrolles, A. Maitournam, B. Gicquel, and J.-M. Reyrat. 2003. The allele encoding the mycobacterial Erp protein affects lung disease in mice. Cell. Microbiol. 5:65-73.
17. De Mendonca-Lima, L., M. Picardeau, C. Raynaud, J. Rauzier, Y. O. de la Salmoniere, L. Barker, F. Bigi, A. Cataldi, B. Gicquel, and J. M. Reyrat. 2001. Erp, an extracellular protein family specific to mycobacteria. Microbiology 147:2315-2320.
18. Dionne, M. S., N. Ghori, and D. S. Schneider. 2003. Drosophila melanogaster is a genetically tractable model host for Mycobacterium marinum. Infect. Immun. 71:3540-3550.
19. Domenech, P., M. B. Reed, and C. E. Barry III. 2005. Contribution of the Mycobacterium tuberculosis MmpL protein family to virulence and drug resistance. Infect. Immun. 73:3492-3501.
20. Dubnau, E., J. Chan, C. Raynaud, V. P. Mohan, M. A. Laneelle, K. Yu, A. Quemard, I. Smith, and M. Daffe. 2000. Oxygenated mycolic acids are necessary for virulence of Mycobacterium tuberculosis in mice. Mol. Microbiol. 36:630-637.
21. Gao, L. Y., S. Guo, B. McLaughlin, H. Morisaki, J. N. Engel, and E. J. Brown. 2004. A mycobacterial virulence gene cluster extending RD1 is required for cytolysis, bacterial spreading, and ESAT-6 secretion. Mol. Microbiol. 53:1677-1693.
22. Gao, L. Y., F. Laval, E. H. Lawson, R. K. Groger, A. Woodruff, J. H. Morisaki, J. S. Cox, M. Daffe, and E. J. Brown. 2003. Requirement for kasB in Mycobacterium mycolic acid biosynthesis, cell wall impermeability and intracellular survival: implications for therapy. Mol. Microbiol. 49:1547-1563.
23. Glickman, M. S., J. S. Cox, and W. R. Jacobs, Jr. 2000. A novel mycolic acid cyclopropane synthetase is required for cording, persistence, and virulence of Mycobacterium tuberculosis. Mol. Cell 5:717-727.
24. Glickman, M. S., and W. R. Jacobs, Jr. 2001. Microbial pathogenesis of Mycobacterium tuberculosis: dawn of a discipline. Cell 104:477-485.
25. Guinn, K. M., M. J. Hickey, S. K. Mathur, K. L. Zakel, J. E. Grotzke, D. M. Lewinsohn, S. Smith, and D. R. Sherman. 2004. Individual RD1-region genes are required for export of ESAT-6/CFP-10 and for virulence of Mycobacterium tuberculosis. Mol. Microbiol. 51:359-370.
26. Hsu, T., S. M. Hingley-Wilson, B. Chen, M. Chen, A. Z. Dai, P. M. Morin, C. B. Marks, J. Padiyar, C. Goulding, M. Gingery, D. Eisenberg, R. G. Russell, S. C. Derrick, F. M. Collins, S. L. Morris, C. H. King, and W. R. Jacobs, Jr. 2003. The primary mechanism of attenuation of bacillus Calmette-Guerin is a loss of secreted lytic function required for invasion of lung interstitial tissue. Proc. Natl. Acad. Sci. USA 100:12420-12425.
27. Jackson, M., C. Raynaud, M. A. Laneelle, C. Guilhot, C. Laurent-Winter, D. Ensergueix, B. Gicquel, and M. Daffe. 1999. Inactivation of the antigen 85C gene profoundly affects the mycolate content and alters the permeability of the Mycobacterium tuberculosis cell envelope. Mol. Microbiol. 31:1573-1587.
28. Jain, M., and J. S. Cox. 2005. Interaction between polyketide synthase and transporter suggests coupled synthesis and export of virulence lipid in M. tuberculosis. PLoS Pathogens 1:12-19.
29. Kocincova, D., B. Sonden, Y. Bordat, E. Pivert, L. de Mendonca-Lima, B. Gicquel, and J.-M. Reyrat. 2004. The hydrophobic domain of the mycobacterial Erp protein is not essential for the virulence of Mycobacterium tuberculosis. Infect. Immun. 72:2379-2382.
30. Kocincova, D., B. Sonden, L. de Mendonca-Lima, B. Gicquel, and J.-M. Reyrat. 2004. The Erp protein is anchored at the surface by a carboxy-terminal hydrophobic domain and is important for cell wall structure in Mycobacterium smegmatis. FEMS Microbiol. Lett. 231:191-196.
31. Lewis, K. N., R. Liao, K. M. Guinn, M. J. Hickey, S. Smith, M. A. Behr, and D. R. Sherman. 2003. Deletion of RD1 from Mycobacterium tuberculosis mimics BCG attenuation. J. Infect. Dis. 187:117-123.
32. McKinney, J. D. 2000. In vivo veritas: the search for TB drug targets goes live. Nat. Med. 6:1330-1333.
33. Meijer, A. H., F. J. Verbeek, E. Salas-Vidal, M. Corredor-Adamez, J. Bussman, A. M. van der Sar, G. W. Otto, R. Geisler, and H. P. Spaink. 2005. Transcriptome profiling of adult zebrafish at the late stage of chronic tuberculosis due to Mycobacterium marinum infection. Mol. Immunol. 42:1185-1203.
34. Pozos, T. C., and L. Ramakrishnan. 2004. New models for the study of Mycobacterium-host interactions. Curr. Opin. Immunol. 16:499-505.
35. Pym, A. S., P. Brodin, R. Brosch, M. Huerre, and S. T. Cole. 2002. Loss of RD1 contributed to the attenuation of the live tuberculosis vaccines Mycobacterium bovis BCG and Mycobacterium microti. Mol. Microbiol. 46:709-717.
36. Ramakrishnan, L., N. A. Federspiel, and S. Falkow. 2000. Granuloma-specific expression of Mycobacterium virulence proteins from the glycine-rich PE-PGRS family. Science 288:1436-1439.
37. Ramakrishnan, L., R. H. Valdivia, J. H. McKerrow, and S. Falkow. 1997. Mycobacterium marinum causes both long-term subclinical infection and acute disease in the leopard frog (Rana pipiens). Infect. Immun. 65:767-773.
38. Rousseau, C., N. Winter, E. Pivert, Y. Bordat, O. Neyrolles, P. Ave, M. Huerre, B. Gicquel, and M. Jackson. 2004. Production of phthiocerol dimycocerosates protects Mycobacterium tuberculosis from the cidal activity of reactive nitrogen intermediates produced by macrophages and modulates the early immune response to infection. Cell. Microbiol. 6:277-287.
39. Sassetti, C. M., and E. J. Rubin. 2003. Genetic requirements for mycobacterial survival during infection. Proc. Natl. Acad. Sci. USA 100:12989-12994.
40. Sherman, D. R., M. Voskuil, D. Schnappinger, R. Liao, M. I. Harrell, and G. K. Schoolnik. 2001. Regulation of the Mycobacterium tuberculosis hypoxic response gene encoding alpha-crystallin. Proc. Natl. Acad. Sci. USA 98:7534-7539.
41. Smith, I. 2003. Mycobacterium tuberculosis pathogenesis and molecular determinants of virulence. Clin. Microbiol. Rev. 16:463-496.
42. Solomon, J. M., G. Leung, and R. R. Isberg. 2003. Intracellular replication of Mycobacterium marinum within Dictyostelium discoideum: efficient replication in the absence of host coronin. Infect. Immun. 71:3578-3586.
43. Tonjum, T., D. B. Welty, E. Jantzen, and P. L. Small. 1998. Differentiation of Mycobacterium ulcerans, M. marinum, and M. haemophilum: mapping of their relationships to M. tuberculosis by fatty acid profile analysis, DNA-DNA hybridization, and 16S rRNA gene sequence analysis. J. Clin. Microbiol. 36:918-925.
44. Travis, W. D., L. B. Travis, G. D. Roberts, D. W. Su, and L. W. Weiland. 1985. The histopathologic spectrum in Mycobacterium marinum infection. Arch. Pathol. Lab. Med. 109:1109-1113.
45. Volkman, H. E., H. Clay, D. Beery, J. C. Chang, D. R. Sherman, and L. Ramakrishnan. 2004. Tuberculous granuloma formation is enhanced by a Mycobacterium virulence determinant. PLoS Biol. 2:1946-1956.(Christine L. Cosma, Kathr)
ABSTRACT
The Mycobacterium tuberculosis exported repetitive protein (Erp) is a virulence determinant required for growth in cultured macrophages and in vivo. To better understand the role of Erp in Mycobacterium pathogenesis, we generated a mutation in the erp homologue of Mycobacterium marinum, a close genetic relative of M. tuberculosis. erp-deficient M. marinum was growth attenuated in cultured macrophage monolayers and during chronic granulomatous infection of leopard frogs, suggesting that Erp function is similarly required for the virulence of both M. tuberculosis and M. marinum. To pinpoint the step in infection at which Erp is required, we utilized a zebrafish embryo infection model that allows M. marinum infections to be visualized in real-time, comparing the erp-deficient strain to a RD1 mutant whose stage of attenuation was previously characterized in zebrafish embryos. A detailed microscopic examination of infected embryos revealed that bacteria lacking Erp were compromised very early in infection, failing to grow and/or survive upon phagocytosis by host macrophages. In contrast, RD1 mutant bacteria grow normally in macrophages but fail to induce host macrophage aggregation and subsequent cell-to-cell spread. Consistent with these in vivo findings, erp-deficient but not RD1-deficient bacteria exhibited permeability defects in vitro, which may be responsible for their specific failure to survive in host macrophages.
INTRODUCTION
Mycobacteria, typified by Mycobacterium tuberculosis, are intracellular pathogens that cause persistent infections despite vigorous host responses (13), and much of the work aimed at understanding mycobacterial pathogenesis has focused on the interactions between mycobacteria and the immune system. The infectious process can be viewed as a series of sequential steps, in which the bacteria interface with the immune system in multiple complex ways. Infection begins with the recruitment of macrophages to the site of infection and phagocytosis of infecting organisms. These infected macrophages then migrate to deeper tissues, and additional monocytes are recruited to the infected cell. As bacterial growth ensues, macrophages aggregate and differentiate into the characteristic epithelioid macrophages that constitute granulomas. Later, as adaptive immunity develops, other immune cells such as lymphocytes are recruited to the granulomas. The bacterium-host interactions that orchestrate this complex program are not well understood, and the bacterial determinants that permit both early growth and long-term persistence are now being identified.
The last decade has seen significant advancements in the genetic manipulation of mycobacteria to define and characterize mycobacterial virulence factors (13, 24, 32, 41). Mutations that reduce virulence have been identified, and the majority of these affect the structure and/or the function of the highly complex cell wall and outer lipid layer. These include mutations in genes involved in cell wall lipid synthesis, composition, or transport (8, 9, 11, 14, 19, 20, 22, 23, 28); the PGRS cell surface proteins (2, 36, 39); and the RD1 locus that encodes a putative secretion system (21, 25, 26, 31, 35, 45). However, a clear understanding of the underlying mechanisms for these virulence deficiencies has been lacking.
The Mycobacterium protein Erp, another cell surface component, was the first Mycobacterium virulence protein to be characterized by targeted deletion (4). Despite a series of genetic and biochemical characterizations (5, 16, 17, 29, 30), its role in virulence remains poorly understood. This Mycobacterium-specific secreted/cell surface protein is comprised of three domains: a conserved N-terminal domain with a canonical Sec-dependent signal sequence, a repetitive central domain, and a conserved hydrophobic C-terminal domain that appears to be involved in a loose anchoring of Erp to the cell surface (30). Virulence studies with M. tuberculosis and M. bovis erp mutants revealed that the bacteria are growth attenuated in cultured macrophages and in a mouse model of infection (4, 5).
M. marinum, a close genetic relative of M. tuberculosis (43) (http://www.sanger.ac.uk/Projects/M_marinum/), is used as a model for mycobacterial pathogenesis (3, 15, 18, 21, 22, 33, 34, 42). M. marinum is the causative agent of systemic granulomatous infections of ectotherms, such as fish and frogs (1), and peripheral chronic granulomatous disease (fish tank granulomas) in humans (44). Experimental inoculation of leopard frogs (Rana pipiens) results in a chronic subclinical infection much like latent tuberculosis in humans (37). Given that M. tuberculosis and M. marinum cause similar pathology in infected hosts, have similar endocytic trafficking pathways in cultured macrophages (3), and share virulence determinants (21, 45), it is likely that these related pathogens use similar survival strategies in vivo.
We have previously developed a zebrafish embryo model of infection to characterize the earliest events after mycobacterial infection (15). Zebrafish embryos are transparent during the first 2 weeks of life, allowing the infectious process to be visualized in real time. Studies using this model revealed that M. marinum injected into the bloodstream are rapidly taken up by host macrophages, usually within 30 min of injection. Over the next 24 h, infected macrophages exit the bloodstream, migrating into tissues. These infected macrophages next begin to aggregate and further recruit additional uninfected macrophages, some of which ultimately differentiate into the epithelioid macrophages typical of granulomas. The timing of aggregation, although dependent on both the virulence of the strain used and the number of organisms injected, usually occurs 2 to 5 days postinjection (J. M. Davis and L. Ramakrishnan, unpublished results). This system has proven valuable in clarifying the mechanism of action of Mycobacterium virulence determinants. For example, in a previous study we demonstrated that the M. marinum RD1 mutant (45), like its M. tuberculosis counterpart, is growth attenuated when assayed by bacterial counts in bulk-cultured macrophages. However, detailed analysis using the zebrafish embryo infection model showed that RD1-deficient bacteria are capable of growth in individual macrophages in vivo but fail to induce macrophage aggregation and subsequent intercellular spread (45). This finding revealed that intracellular growth in individual macrophages is insufficient to ensure bacterial proliferation and that additional host cells must be recruited and infected.
In the present study, we characterize another M. marinum virulence determinant, the erp locus, using a combination of in vivo and in vitro assays. We found that, like the M. marinum RD1 locus, erp is required for bacterial growth in vivo. However, by using the zebrafish embryo model, we show that, in contrast to the RD1 locus, erp is required for growth and/or survival in individual macrophages. The growth attenuation of this mutant in vivo may be explained at least in part by a permeability defect.
MATERIALS AND METHODS
Bacterial strains and growth conditions. Mycobacterial strains used for the present study were derived from a human clinical isolate, strain M (ATCC BAA-535) and are shown in Table 1. M. marinum was grown in Middlebrook 7H9 broth (Difco) supplemented with 0.5% bovine serum albumin, 0.005% oleic acid, 0.2% glucose, 0.2% glycerol, 0.085% sodium chloride, and 0.05% Tween 80 or on Middlebrook 7H10 agar (Difco) supplemented with 0.5% bovine serum albumin, 0.005% oleic acid, 0.2% glucose, 0.2% glycerol, and 0.085% sodium chloride. Media were supplemented with 20 μg of kanamycin/ml and 50 μg of hygromycin/ml when appropriate. pGFPHYG2 was obtained by inserting the hygromycin resistance cassette of plasmid pRF498H (40) into the NsiI sites of plasmid pMSP12::GFP (10), thus disrupting the kanamycin resistance cassette.
Construction of the M. marinum erp mutant. A 2,857-bp PstI/BglI fragment containing the erp locus was isolated from a M. marinum genomic cosmid library and inserted into pBluescript SK(+) (Stratagene). The erp gene was interrupted by removing a central 316-bp SmaI fragment and replacing it with the aph (kanamycin resistance) cassette of pYUB213 (a gift from W. Jacobs) to generate an erp::aph mutation. The sacB gene was then introduced at the XbaI site of the polylinker, and the resulting plasmid, pERPKO, was used to transform M. marinum strain M. Kanamycin-resistant (Kanr) sucroses merodiploid colonies were isolated and streaked onto medium containing kanamycin and sucrose to select for sucroser double recombinants. One such isolate was saved and named KK33 (Table 1). Complementation of erp::aph was achieved by cloning a 2,284-bp NruI/MscI fragment containing the wild-type allele into pYUB178Hyg to generate pERPcomp1, which was then transformed into KK33 and integrated into the att site to produce strain KK60 (Table 1). The presence of the knockout and the complementing alleles were confirmed by Southern blot analysis (data not shown).
Macrophage infections. Growth in J774 macrophages was determined as described previously (36). Briefly, 5 x 104 J774A.1 murine macrophages in grown in Dulbecco modified Eagle medium (Gibco) supplemented with 10% fetal bovine serum were seeded into the wells of a 24-well plate and allowed to adhere overnight. The following day, the cells were infected with M. marinum for 1 h at 33°C. Infected medium was replaced with 1 ml of fresh medium containing 40 μg of streptomycin/ml. To determine the intracellular bacterial counts at various times postinfection, cells were washed three times with 1x phosphate-buffered saline, lysed with 100 μl of 1% Triton X-100, scraped with a pipette tip, and diluted 10-fold with 1x phosphate-buffered saline. Dilutions were plated onto 7H10 agar and incubated at 33°C for 7 to 14 days.
Frog infection. Wild-caught male Rana pipiens (approximately 10 to 12 cm) were obtained from J. M. Hazen (Alberg, VT). Frogs were infected by intraperitoneal injection as described previously (12). One, two, or twelve weeks later, liver and spleen tissues were harvested, and the bacterial CFU were enumerated by plating on 7H10 agar supplemented with amphotericin B and, in the case of KK33 and KK60, with kanamycin. Frogs were housed and all animal procedures were carried out in accordance with University of Washington IACUC regulations.
Zebrafish procedures. Zebrafish embryos at 32 h postfertilization were infected by microinjection at the caudal vein as previously described (12, 15). To assess virulence of various M. marinum strains, cohorts of embryos hatched at the same time were infected (45 embryos per strain) and monitored daily for mortalities. Each treatment group of 45 embryos was distributed into three tanks of fifteen and scored separately to account for the possible contribution of tank effects. Mean time to death (MTD) was scored for each tank, and no significant differences were found between tanks of fish infected with the same strain. The data for fish infected with a given strain were then pooled, and MTD values between strains were compared by using analysis of variance (ANOVA). For the visualization of bacteria in vivo, fluorescent strains were used (Table 1), and embryos were mounted for observation as described previously (12, 15, 45). When scoring the levels of infection in zebrafish embryos, the individual making this assessment was blinded to the strain used to infect each embryo so as to avoid bias.
MIC assays. The agar dilution method was used to determine the MIC for each strain (22). Individual plates of 7H10 agar (supplemented as described above) were prepared (30-ml volume), with twofold dilutions of the antibiotic being tested. Plates were spotted with 10 μl of early-log-phase cultures in triplicate, and all strains were spotted onto the same plate. Plates were incubated at 33°C and scored after 8 and 11 days.
Statistical analyses. Statistical analysis was performed by using In-Stat software (Graphpad Software, Inc.).
RESULTS
Targeted deletion of the M. marinum erp locus. We compared the previously identified M. marinum erp locus (17) (http://www.sanger.ac.uk/Projects/M_marinum/) with its M. tuberculosis homologue and found an overall 76% amino acid identity and 81% similarity. M. marinum erp is located between the glf and csp open reading frames, making its genomic arrangement syntenous with those of other Mycobacterium species. The central repetitive domain of the predicted protein contains two additional PGLTS repeats compared to that of M. tuberculosis (17), making it the closest identified homologue to those of the M. tuberculosis complex. A BLAST search of the completed M. marinum genome sequence failed to identify any additional open reading frames with homology to the M. tuberculosis erp sequence, suggesting that M. marinum contains only one erp allele. To investigate the role of erp in M. marinum pathogenesis, we generated a loss-of-function mutation by introducing a Kanr cassette via homologous recombination near the end of the N-terminal domain (see Materials and Methods). This mutation was complemented by introducing the wild-type erp allele at the att site (Table 1). As in the case of the M. tuberculosis erp mutant (4), growth of the M. marinum erp::aph mutant strain in liquid culture was similar to that of the wild type (data not shown). However, growth on 7H10 agar was mildly compromised since erp::aph mutant colonies appeared 3 days later than did the wild type. A small colony phenotype has been reported for erp-deficient mutants in other species (30).
The M. marinum erp mutant is growth attenuated in cultured macrophages and adult animals. The M. tuberculosis erp locus is required for growth both in cultured macrophages and in a mouse model of infection (4). We found M. marinum erp to be similarly required for growth in cultured macrophages. As shown in Fig. 1A, the erp::aph strain failed to grow in J774A.1 macrophages over an 8-day time course. This growth attenuation was indistinguishable from that observed for a M. marinum mutant that lacks the RD1 virulence locus (45). In contrast, the wild-type and complemented mutant strains grew approximately 50-fold over the 8-day monitoring period.
To examine whether erp-deficient bacteria were also attenuated in vivo, we assayed their growth in a natural host species, the leopard frog, used to study M. marinum pathogenesis in the laboratory (6, 37). Intraperitoneal infection of frogs with wild-type bacteria resulted in an increased tissue burden over a period of 12 weeks (Fig. 1B). In contrast, the number of erp::aph mutant bacteria decreased slightly, and by 12 weeks postinfection it was 1.5 logs below that of the wild type (P < 0.05). RD1 mutant bacteria were similarly attenuated (P < 0.05), as shown previously (45). As was the case in cultured macrophages, complementation of the erp::aph mutation restored growth in frogs. Thus, the M. marinum and M. tuberculosis erp-deficient mutants had comparable phenotypes with respect to growth in culture, macrophage monolayers, and animals. Furthermore, the attenuation characteristics of the M. marinum erp-deficient mutant were indistinguishable from those of the M. marinum RD1 mutant (31, 45).
The M. marinum erp::aph mutant is attenuated in the zebrafish embryo infection model. The M. marinum erp::aph mutant was next analyzed for virulence in the zebrafish embryo infection model (15). In this model, embryos are infected by microinjection with a small number of bacteria (ranging from 10 organisms to 300) at approximately 32 h postfertilization, shortly after the beginning of circulation. Injection directly into the caudal vein results in immediate dispersal of the bacteria via the bloodstream, their subsequent uptake by circulating macrophages, and migration of infected macrophages into the tissues where they aggregate to form granulomas. To assess the virulence of the erp::aph mutant, 45 embryos were infected per strain, and survival was monitored over a 2-week time course. Embryos infected with the erp::aph and RD1 mutants fared better than those infected with wild-type and erp::aph complemented strains (Fig. 2A). The MTD for embryos infected with the wild-type and complemented strains was significantly shorter than for those infected with the mutants (Table 2).
To determine whether these differences in embryo survival resulted from differential growth of the bacteria in vivo, we undertook a microscopic analysis of embryos infected with wild-type and mutant bacteria expressing gfp from a strong constitutive promoter (see Table 1 and Materials and Methods). Ten embryos were injected per strain and were viewed by fluorescence microscopy 1 h later to verify that the inocula were similar and then again at 5 days postinoculation to examine the degree of bacterial proliferation. As shown in Fig. 2B, the overall burden of bacteria was similar for all embryos at 1 h postinfection, but by 5 days postinfection wild-type-infected embryos harbored more bacteria than embryos infected with erp::aph or RD1 bacteria. Injection with either of these strains resulted in low burden infections and a paucity of macrophage aggregates typical of the wild-type infection (Fig. 2B, arrows). We could not quantify the growth of these strains in zebrafish embryos using bacterial counts, since we were unable to recover erp::aph mutant bacteria using the lysis procedure required for this analysis (45). This failure to recover erp::aph mutant bacteria was specific as wild-type and RD1 bacteria were readily recovered (H. Volkman, D. Beery, and L. Ramakrishnan, unpublished data). These experiments revealed that erp::aph and RD1 bacteria were similarly growth attenuated in zebrafish embryos.
Erp-deficient bacteria fail to grow intracellularly in a zebrafish embryo model of infection. The intracellular growth by mycobacteria has traditionally been assayed in cultured cell lines by enumerating the total growth of the bacterial population in a population of macrophages as a surrogate measure of growth in the individual cells. However, recent studies of the RD1 locus (21, 25, 45) have shown that this can be a faulty assumption. In our previous work we found that RD1 bacteria seemingly failed to grow intracellularly (as measured in such an assay); however, closer examination using the zebrafish embryo model revealed that these bacteria are capable of intracellular growth but rather fail to spread among host cells. We concluded that in intracellular growth assays in cultured cell lines, a limited amount of growth may occur intracellularly, but unless bacteria continue to spread to new host cells, no increase in the overall population is observed. Therefore, we were similarly interested to use this infection model to pinpoint the precise nature of attenuation for the erp::aph mutant.
Prior detailed analyses in the zebrafish embryo have revealed that infection by wild-type bacteria proceeds in sequential steps (15). Upon injection of M. marinum, macrophages migrate to the site of infection and phagocytose the bacteria. Infected macrophages then migrate back into tissues and ultimately aggregate into granulomas. In initial experiments with the erp::aph strain we did not observe any obvious defects in macrophage recruitment, phagocytosis of bacteria, or migration into tissues (D. Beery and L. Ramakrishnan, unpublished observations), suggesting that, like the RD1 mutant (45), the erp::aph mutant was competent for these very early steps. We next performed a detailed comparison of embryos infected with wild-type, RD1, and erp::aph bacteria at 5 days postinfection by using fluorescence and differential interference contrast (DIC) microscopy at high magnification. Similar to what was observed at low magnification (Fig. 2B), abundant macrophage aggregates were found in the wild-type-infected embryos, whereas fewer were observed in embryos infected with either mutant strain (data not shown). With respect to the infection burden of individual (nonaggregated) macrophages, wild-type-infected embryos contained macrophages representing a wide spectrum of infection levels, as expected. Single macrophages contained anywhere from one to >10 bacteria (Fig. 3A). However, a striking difference was apparent between embryos infected with the two mutant strains. Macrophages of embryos infected with the RD1 mutant typically harbored large numbers of bacteria (Fig. 3B), whereas those from embryos infected with the erp::aph mutant frequently harbored very few bacteria (Fig. 3C). Furthermore, erp::aph bacteria were often dim (Fig. 3C), suggesting that they are being killed (in contrast, all strains were similarly bright in culture). These observations suggested that erp::aph and RD1 mutant bacteria might be attenuated for distinct reasons, despite their similarities in other assays.
To quantify the observed differences, we studied 10 infected embryos for each of the three strains. For each embryo, the entire region posterior to the end of the yolk extension (Fig. 3D) was surveyed, and individual (nonaggregated) infected macrophages found in this region were scored. Each macrophage was scored based on the number of bacteria it contained; <5 bacteria, 5 to 10 bacteria, or >10 bacteria. For each embryo, we calculated the proportion of its macrophages that were lightly infected (<5 bacteria), moderately infected (5 to 10 bacteria), and heavily infected (>10 bacteria). Consistent with the qualitative findings (Fig. 3A to C), we observed that significantly more infected macrophages fell into the high-burden category (>10 bacteria per cell) when the embryos were infected with RD1 bacteria than when they were infected with erp::aph bacteria (P < 0.01). The converse was also true; significantly more infected macrophages fell into the low-burden category (<5 bacteria per cell) when the embryos were infected with erp::aph bacteria than when they were infected with RD1 bacteria (P < 0.01). No difference was observed in the proportions of moderately infected macrophages. Wild-type-infected embryos had an intermediate number of macrophages in each category. This was due to the fact that during infection by wild-type bacteria, infected macrophages are titrated out into granulomas, which were not scored in this assay. In contrast, the RD1 and erp::aph mutant-infected macrophages were less likely to form aggregates, and the discrepancy in the number of bacteria they contained reveals that these two mutant strains are attenuated for different reasons despite their identical attenuation phenotypes in virulence assays that measure overall organism burden (Fig. 1). erp-deficient bacteria are attenuated primarily because of their lack of intracellular growth and/or survival, whereas RD1 mutant bacteria grow successfully in macrophages but fail to induce subsequent aggregation and intercellular spread (Fig. 4) (45).
Erp is required for the Mycobacterium permeability barrier. As discussed above, we specifically failed to recover erp::aph mutant bacteria extracted from zebrafish embryos. The extraction method requires that embryo lysates be plated on 7H11 agar in the presence of polymixin B, trimethoprim, and carbenecillin to suppress the growth of intestinal and other aquatic flora (H. Volkman and L. Ramakrishnan, unpublished). Since erp::aph bacteria were clearly present (as visualized by fluorescence microscopy, Fig. 2B), we suspected that they might simply be more susceptible to the antimicrobial compounds in the agar. Therefore, we determined the MIC for wild-type and mutant bacteria of a panel of antimicrobial compounds. As a control we included an M. marinum strain that contains a transposon insertion in kasB, a gene required for mycolate synthesis (22). kasB-deficient bacteria are more permeable to the hydrophobic detergent deoxycholate, and are consequently more susceptible to a variety of lipophilic antibiotics such as rifampin and erythromycin, but retain normal susceptibility to hydrophilic compounds. As shown in Table 3, the erp::aph mutant strain was more susceptible to the lipophilic antibiotics rifampin, erythromycin, and chloramphenicol than were the wild-type and complemented strains. The degree of sensitivity was similar to that observed for kasB-deficient strain and was in contrast to the RD1 strain, which demonstrated normal susceptibility. These findings suggest that the M. marinum erp::aph mutant has increased permeability to a variety of hydrophobic agents, which may explain its decreased growth and/or survival in macrophages.
DISCUSSION
In the present study, we have characterized the role of the M. marinum erp gene during infection. Like its M. tuberculosis and M. bovis counterparts (4, 5), we find that M. marinum Erp is a virulence determinant, required for growth in cultured macrophages and in vivo. In this regard, our findings add to a growing body of evidence that the molecular determinants of M. tuberculosis virulence during infection of mammals are also used by M. marinum during infection of its natural hosts (7, 21, 36, 39, 45). That Erp function is conserved among mycobacteria is further supported by studies in M. tuberculosis showing that the erp loci of M. leprae and M. smegmatis are competent for complementation of an M. tuberculosis erp mutation in the context of a mouse infection (16). Such commonalities suggest that the basic mechanisms of Mycobacterium pathogenesis are conserved.
As was the case with the RD1 locus (45), the zebrafish embryo infection model has refined our understanding of the role of the erp locus in Mycobacterium virulence. Using this model we find that Erp-deficient bacteria have an early and intrinsic defect in intracellular growth and/or survival in macrophages. This is in contrast to RD1 bacteria, which proliferate normally in individual macrophages but have diminished and delayed formation of macrophage aggregates and are defective in intercellular spread within these aggregates (45). This difference in intracellular growth was not revealed in the standard assay designed to monitor net bacterial growth in a population of cultured macrophages (Fig. 1A) but rather was observed only when a detailed comparison of the two mutants was made during infection of zebrafish embryos (Fig. 3). The points at which these two determinants impact the infectious process are depicted in Fig. 4. Although both begin to exert their influence in the early stages of infection, it is clear that erp is required earlier than is the RD1 locus. Erp-deficient bacteria do not survive or grow within individual macrophages, so that infection is thwarted prior to macrophage aggregation into granulomas. It is noteworthy that the differences in the capacity of these two mutants for intracellular growth do not result in a differential outcome later in infection (Fig. 1B and 2A and Table 2). This observation reinforces our previous findings with the RD1 mutant that the capacity for intracellular growth alone is not sufficient to confer full virulence and that the ability of mycobacteria to proliferate to their maximum extent also depends on the cell-to-cell spread impacted by the RD1 locus. Furthermore, it would appear that these two aspects of the infectious process (intracellular growth and intercellular spread) are important throughout infection, since both mutants are comparably attenuated in the short-term embryo infection model and in the long-term chronic frog model. That both mutants are capable of establishing infection, albeit at a lower tissue burden than for the wild type, reinforces the notion that mycobacteria use multiple redundant mechanisms to ensure their survival (9, 14, 36, 39).
The finding that the erp::aph mutant is compromised for intracellular growth, coupled with its failure to grow on Middlebrook 7H11 agar replete with antibiotics, suggested a permeability defect that might explain both phenotypes. Indeed, upon further examination we found that the loss of Erp renders bacteria more permeable to a variety of lipophilic compounds. These permeability defects may render the bacteria more susceptible to intracellular host defenses such as reactive oxygen and nitrogen intermediates, degradative enzymes, and defensins. Permeability defects have been postulated as the reason for attenuation for a variety of Mycobacterium mutants. For example, mutations in the M. marinum kasB gene involved in mycolate production rendered bacteria more permeable to deoxycholate and more sensitive to a variety of lipophilic antibiotics (22). On the other hand, mutations in the M. tuberculosis fadD26 and mmpL7 genes, which are involved in PDIM synthesis (8, 9, 14), rendered bacteria more permeable to detergent (38) and yet not more susceptible to hydrophobic antibiotics (19, 38). Clearly, the complexity of the permeability barrier conferred by the Mycobacterium cell wall and outer lipid layer are such that individual components may confer protection against different types of compounds.
That permeability defects are complex is further revealed by a comparison of the attenuation phenotypes of various permeability mutants. In M. tuberculosis, the fadD26 mutant is growth attenuated in activated, but not naive macrophages (38), whereas macrophage activation is not required to restrict growth of erp mutants (4; the present study). Indeed, some permeability mutants do not have a virulence defect, as is the case with the M. tuberculosis fbcC gene, which encodes a mycolyltransferase (27). In M. marinum, the erp and kasB mutants have similar profiles of susceptibility to lipophilic antibiotics (22; the present study) (Table 3), and yet the former is more attenuated for growth in cultured macrophages (22; C. Cosma and L. Ramakrishnan, unpublished data). In our preliminary examination of the kasB mutant in zebrafish embryos, we found that in this model too, kasB-deficient bacteria are less attenuated than erp::aph bacteria (Cosma, Kim, and Ramakrishnan, unpublished). These differences in intracellular growth could be for several reasons. First, features of the Mycobacterium permeability barrier relevant for intracellular survival may simply not be revealed by in vitro membrane permeability assays. Alternatively, the permeability defect conferred by the loss of Erp may be only partially responsible for its virulence defects and Erp may play a second role in intracellular survival, independent of its contribution to the Mycobacterium permeability barrier. Finally, the loss of KasB may have a severe enough effect on the Mycobacterium cell wall or outer lipid layer that genetic or regulatory compensations have occurred to provide better survival both in vitro and in vivo. Whatever the reason, these results underscore the importance of in vivo assays to determine the true impact on the virulence of mycobacterial permeability defects detected in vitro.
By combining a variety of infection models with in vitro assays, we have demonstrated that the Mycobacterium erp locus is required for intracellular growth and survival and that this requirement may be due in part to its role in maintaining integrity of the cell wall and outer lipid layer. The use of M. marinum as a surrogate model in the zebrafish has allowed a more precise determination of this attenuation defect than was previously apparent.
ACKNOWLEDGMENTS
We thank Hannah Volkman for assistance with embryo infections; Muse Davis for assistance with figures; Richard Burmeister for artwork; Eric Brown for M. marinum strains; and Lynn Connolly, Hannah Volkman, David Sherman, and Yong Gao for comments on the manuscript.
This study was supported by NIH grants RO1 AI036396 and RO1 AI54503 to L.R.
Present address: Centers for Molecular Medicine, State University of New York—Stony Brook, Stony Brook, NY 11794-5120.
REFERENCES
1. Aronson, J. D. 1926. Spontaneous tuberculosis in salt water fish. J. Infect. Dis. 39:315-320.
2. Banu, S., N. Honore, B. Saint-Joanis, D. Philpott, M. C. Prevost, and S. T. Cole. 2002. Are the PE-PGRS proteins of Mycobacterium tuberculosis variable surface antigens Mol. Microbiol. 44:9-19.
3. Barker, L. P., K. M. George, S. Falkow, and P. L. Small. 1997. Differential trafficking of live and dead Mycobacterium marinum organisms in macrophages. Infect. Immun. 65:1497-1504.
4. Berthet, F. X., M. Lagranderie, P. Gounon, C. Lauren-Winter, D. Ensergueix, P. Chavarot, F. Thouron, E. Maranghi, V. Pelicic, D. Portnoi, G. Marchal, and B. Gicquel. 1998. Attenuation of virulence by disruption of the Mycobacterium tuberculosis erp gene. Science 282:759-762.
5. Bigi, F., A. Gioffre, L. Klepp, M. P. Santangelo, C. A. Velicovsky, G. H. Giambartolomei, C. A. Fossati, M. I. Romano, T. Mendum, J. J. McFadden, and A. Cataldi. 2005. Mutation in the P36 gene of Mycobacterium bovis provokes attenuation of the bacillus in a mouse model. Tuberculosis 85:221-226.
6. Bouley, D. M., N. Ghori, K. L. Mercer, S. Falkow, and L. Ramakrishnan. 2001. Dynamic nature of host-pathogen interactions in Mycobacterium marinum granulomas. Infect. Immun. 69:7820-7831.
7. Brennan, M. J., G. Delogu, Y. Chen, S. Bardarov, J. Kriakov, M. Alavi, and W. R. Jacobs, Jr. 2001. Evidence that mycobacterial PE_PGRS proteins are cell surface constituents that influence interactions with other cells. Infect. Immun. 69:7326-7333.
8. Camacho, L. R., P. Constant, C. Raynaud, M. A. Laneelle, J. A. Triccas, B. Gicquel, M. Daffe, and C. Guilhot. 2001. Analysis of the phthiocerol dimycocerosate locus of Mycobacterium tuberculosis: evidence that this lipid is involved in the cell wall permeability barrier. J. Biol. Chem. 276:19845-19854.
9. Camacho, L. R., D. Ensergueix, E. Perez, B. Gicquel, and C. Guilhot. 1999. Identification of a virulence gene cluster of Mycobacterium tuberculosis by signature-tagged transposon mutagenesis. Mol. Microbiol. 34:257-267.
10. Chan, K., T. Knaak, L. Satkamp, O. Humbert, S. Falkow, and L. Ramakrishnan. 2002. Complex pattern of Mycobacterium marinum gene expression during long-term granulomatous infection. Proc. Natl. Acad. Sci. USA 99:3920-3925.
11. Converse, S. E., J. D. Mougous, M. D. Leavell, J. A. Leary, C. R. Bertozzi, and J. S. Cox. 2003. MmpL8 is required for sulfolipid-1 biosynthesis and Mycobacterium tuberculosis virulence. Proc. Natl. Acad. Sci. USA 100:6121-6126.
12. Cosma, C. L., J. M. Davis, L. E. Swaim, H. Volkman, and L. Ramakrishnan. Animal models of Mycobacterium marinum infection. Curr. Protocols Microbiol., in press.
13. Cosma, C. L., D. R. Sherman, and L. Ramakrishnan. 2003. The secret lives of the pathogenic mycobacteria. Annu. Rev. Microbiol. 57:641-676.
14. Cox, J. S., B. Chen, M. McNeil, and W. R. Jacobs, Jr. 1999. Complex lipid determines tissue-specific replication of Mycobacterium tuberculosis in mice. Nature 402:79-83.
15. Davis, J. M., H. Clay, J. L. Lewis, N. Ghori, P. Herbomel, and L. Ramakrishnan. 2002. Real-time visualization of Mycobacterium-macrophage interactions leading to initiation of granuloma formation in zebrafish embryos. Immunity 17:693-702.
16. De Mendonca-Lima, L., Y. Bordat, E. Pivert, C. Recchi, O. Neyrolles, A. Maitournam, B. Gicquel, and J.-M. Reyrat. 2003. The allele encoding the mycobacterial Erp protein affects lung disease in mice. Cell. Microbiol. 5:65-73.
17. De Mendonca-Lima, L., M. Picardeau, C. Raynaud, J. Rauzier, Y. O. de la Salmoniere, L. Barker, F. Bigi, A. Cataldi, B. Gicquel, and J. M. Reyrat. 2001. Erp, an extracellular protein family specific to mycobacteria. Microbiology 147:2315-2320.
18. Dionne, M. S., N. Ghori, and D. S. Schneider. 2003. Drosophila melanogaster is a genetically tractable model host for Mycobacterium marinum. Infect. Immun. 71:3540-3550.
19. Domenech, P., M. B. Reed, and C. E. Barry III. 2005. Contribution of the Mycobacterium tuberculosis MmpL protein family to virulence and drug resistance. Infect. Immun. 73:3492-3501.
20. Dubnau, E., J. Chan, C. Raynaud, V. P. Mohan, M. A. Laneelle, K. Yu, A. Quemard, I. Smith, and M. Daffe. 2000. Oxygenated mycolic acids are necessary for virulence of Mycobacterium tuberculosis in mice. Mol. Microbiol. 36:630-637.
21. Gao, L. Y., S. Guo, B. McLaughlin, H. Morisaki, J. N. Engel, and E. J. Brown. 2004. A mycobacterial virulence gene cluster extending RD1 is required for cytolysis, bacterial spreading, and ESAT-6 secretion. Mol. Microbiol. 53:1677-1693.
22. Gao, L. Y., F. Laval, E. H. Lawson, R. K. Groger, A. Woodruff, J. H. Morisaki, J. S. Cox, M. Daffe, and E. J. Brown. 2003. Requirement for kasB in Mycobacterium mycolic acid biosynthesis, cell wall impermeability and intracellular survival: implications for therapy. Mol. Microbiol. 49:1547-1563.
23. Glickman, M. S., J. S. Cox, and W. R. Jacobs, Jr. 2000. A novel mycolic acid cyclopropane synthetase is required for cording, persistence, and virulence of Mycobacterium tuberculosis. Mol. Cell 5:717-727.
24. Glickman, M. S., and W. R. Jacobs, Jr. 2001. Microbial pathogenesis of Mycobacterium tuberculosis: dawn of a discipline. Cell 104:477-485.
25. Guinn, K. M., M. J. Hickey, S. K. Mathur, K. L. Zakel, J. E. Grotzke, D. M. Lewinsohn, S. Smith, and D. R. Sherman. 2004. Individual RD1-region genes are required for export of ESAT-6/CFP-10 and for virulence of Mycobacterium tuberculosis. Mol. Microbiol. 51:359-370.
26. Hsu, T., S. M. Hingley-Wilson, B. Chen, M. Chen, A. Z. Dai, P. M. Morin, C. B. Marks, J. Padiyar, C. Goulding, M. Gingery, D. Eisenberg, R. G. Russell, S. C. Derrick, F. M. Collins, S. L. Morris, C. H. King, and W. R. Jacobs, Jr. 2003. The primary mechanism of attenuation of bacillus Calmette-Guerin is a loss of secreted lytic function required for invasion of lung interstitial tissue. Proc. Natl. Acad. Sci. USA 100:12420-12425.
27. Jackson, M., C. Raynaud, M. A. Laneelle, C. Guilhot, C. Laurent-Winter, D. Ensergueix, B. Gicquel, and M. Daffe. 1999. Inactivation of the antigen 85C gene profoundly affects the mycolate content and alters the permeability of the Mycobacterium tuberculosis cell envelope. Mol. Microbiol. 31:1573-1587.
28. Jain, M., and J. S. Cox. 2005. Interaction between polyketide synthase and transporter suggests coupled synthesis and export of virulence lipid in M. tuberculosis. PLoS Pathogens 1:12-19.
29. Kocincova, D., B. Sonden, Y. Bordat, E. Pivert, L. de Mendonca-Lima, B. Gicquel, and J.-M. Reyrat. 2004. The hydrophobic domain of the mycobacterial Erp protein is not essential for the virulence of Mycobacterium tuberculosis. Infect. Immun. 72:2379-2382.
30. Kocincova, D., B. Sonden, L. de Mendonca-Lima, B. Gicquel, and J.-M. Reyrat. 2004. The Erp protein is anchored at the surface by a carboxy-terminal hydrophobic domain and is important for cell wall structure in Mycobacterium smegmatis. FEMS Microbiol. Lett. 231:191-196.
31. Lewis, K. N., R. Liao, K. M. Guinn, M. J. Hickey, S. Smith, M. A. Behr, and D. R. Sherman. 2003. Deletion of RD1 from Mycobacterium tuberculosis mimics BCG attenuation. J. Infect. Dis. 187:117-123.
32. McKinney, J. D. 2000. In vivo veritas: the search for TB drug targets goes live. Nat. Med. 6:1330-1333.
33. Meijer, A. H., F. J. Verbeek, E. Salas-Vidal, M. Corredor-Adamez, J. Bussman, A. M. van der Sar, G. W. Otto, R. Geisler, and H. P. Spaink. 2005. Transcriptome profiling of adult zebrafish at the late stage of chronic tuberculosis due to Mycobacterium marinum infection. Mol. Immunol. 42:1185-1203.
34. Pozos, T. C., and L. Ramakrishnan. 2004. New models for the study of Mycobacterium-host interactions. Curr. Opin. Immunol. 16:499-505.
35. Pym, A. S., P. Brodin, R. Brosch, M. Huerre, and S. T. Cole. 2002. Loss of RD1 contributed to the attenuation of the live tuberculosis vaccines Mycobacterium bovis BCG and Mycobacterium microti. Mol. Microbiol. 46:709-717.
36. Ramakrishnan, L., N. A. Federspiel, and S. Falkow. 2000. Granuloma-specific expression of Mycobacterium virulence proteins from the glycine-rich PE-PGRS family. Science 288:1436-1439.
37. Ramakrishnan, L., R. H. Valdivia, J. H. McKerrow, and S. Falkow. 1997. Mycobacterium marinum causes both long-term subclinical infection and acute disease in the leopard frog (Rana pipiens). Infect. Immun. 65:767-773.
38. Rousseau, C., N. Winter, E. Pivert, Y. Bordat, O. Neyrolles, P. Ave, M. Huerre, B. Gicquel, and M. Jackson. 2004. Production of phthiocerol dimycocerosates protects Mycobacterium tuberculosis from the cidal activity of reactive nitrogen intermediates produced by macrophages and modulates the early immune response to infection. Cell. Microbiol. 6:277-287.
39. Sassetti, C. M., and E. J. Rubin. 2003. Genetic requirements for mycobacterial survival during infection. Proc. Natl. Acad. Sci. USA 100:12989-12994.
40. Sherman, D. R., M. Voskuil, D. Schnappinger, R. Liao, M. I. Harrell, and G. K. Schoolnik. 2001. Regulation of the Mycobacterium tuberculosis hypoxic response gene encoding alpha-crystallin. Proc. Natl. Acad. Sci. USA 98:7534-7539.
41. Smith, I. 2003. Mycobacterium tuberculosis pathogenesis and molecular determinants of virulence. Clin. Microbiol. Rev. 16:463-496.
42. Solomon, J. M., G. Leung, and R. R. Isberg. 2003. Intracellular replication of Mycobacterium marinum within Dictyostelium discoideum: efficient replication in the absence of host coronin. Infect. Immun. 71:3578-3586.
43. Tonjum, T., D. B. Welty, E. Jantzen, and P. L. Small. 1998. Differentiation of Mycobacterium ulcerans, M. marinum, and M. haemophilum: mapping of their relationships to M. tuberculosis by fatty acid profile analysis, DNA-DNA hybridization, and 16S rRNA gene sequence analysis. J. Clin. Microbiol. 36:918-925.
44. Travis, W. D., L. B. Travis, G. D. Roberts, D. W. Su, and L. W. Weiland. 1985. The histopathologic spectrum in Mycobacterium marinum infection. Arch. Pathol. Lab. Med. 109:1109-1113.
45. Volkman, H. E., H. Clay, D. Beery, J. C. Chang, D. R. Sherman, and L. Ramakrishnan. 2004. Tuberculous granuloma formation is enhanced by a Mycobacterium virulence determinant. PLoS Biol. 2:1946-1956.(Christine L. Cosma, Kathr)