Phospholipases C play a role in the pathogenesis of several bacteria. Mycobacterium tuberculosis, the causative agent of tuberculosis, possesses four genes encoding putative phospholipases C, plcA, plcB, plcC and plcD. However, the contribution of these genes to virulence is unknown. We constructed four single mutants of M. tuberculosis each inactivated in one of the plc genes, a triple plcABC mutant and a quadruple plcABCD mutant. The mutants all exhibited a lower phospholipase C activity than the wild-type parent strain, demonstrating that the four plc genes encode a functional phospholipase C in M. tuberculosis. Functional complementation of the ΔplcABC triple mutant with the individual plcA, plcB and plcC genes restored in each case about 20% of the total Plc activity detected in the parental strain, suggesting that the three enzymes contribute equally to the overall Plc activity of M. tuberculosis. RT-PCR analysis of the plc genes transcripts showed that the expression of these genes is strongly upregulated during the first 24 h of macrophage infection. Moreover, the growth kinetics of the triple and quadruple mutants in a mouse model of infection revealed that both mutants are attenuated in the late phase of the infection emphasizing the importance of phospholipases C in the virulence of the tubercle bacillus.
Phospholipases are important virulence factors in an increasing number of intra- and extracellular bacterial pathogens including Clostridium perfringens, Corynebacterium pseudotuberculosis, Pseudomonas aeruginosa, Staphylococcus aureus and Listeria monocytogenes (McNamara et al., 1994) (for reviews, see Titball, 1993; Songer, 1997). Phospholipases can be divided into four groups depending on the position of the bond they hydrolyse on the phospholipid substrate: phospholipases A1, A2, C and D. Phospholipases C (Plc) appear to be the most important playing a significant role in bacterial pathogenesis (Songer, 1997). For example, in the Gram-positive intracellular pathogen L. monocytogenes, the inactivation of the plcA gene encoding a phosphatidylinositol-specific Plc, prevents replication within mouse peritoneal macrophages, cell-to-cell spread and propagation in host tissues (Camilli et al., 1991). The α toxin (CPA) from C. perfringens is the most toxic Plc characterized to date. It has haemolytic, lethal, dermonecrotic, vascular permeabilizing and platelet-aggregating properties (Titball, 1993). The haemolytic phospholipase C of P. aeruginosa (PlcHR) suppresses the bacterium-induced neutrophil respiratory burst by interfering with a protein kinase C-specific signalling pathway (Terada et al., 1999).
Tuberculosis remains the leading cause of death due to a single infectious organism in the world. Despite the considerable amount of work devoted to deciphering the molecular basis of Mycobacterium tuberculosis pathogenicity, little is known about the mechanisms enabling this pathogen to resist destruction by the host and to multiply inside mononuclear phagocytic cells. Owing to their role in the pathogenesis of many bacterial patho-gens, phospholipases have been studied in mycobacteria. Phospholipase C (Plc) and phospholipase D (Pld) activities have been described in several mycobacterial species. However, although Pld activity has been de-tected in both virulent and saprophytic species, Plc and sphingomyelinase activities seem to be restricted to pathogenic Mycobacterium subsp. (Johansen et al., 1996). For example, cell extracts of M. tuberculosis and Mycobacterium ulcerans, the causative agents of tuberculosis and Buruli ulcer, respectively, contain both Plc and Pld activities, whereas only Pld activity has been reported in cell extracts of the non-pathogenic Mycobacterium smegmatis (Johansen et al., 1996; Gomez et al., 2000). Despite this striking correlation, the role of phospholipases C in the pathogenicity of M. tuberculosis had not yet been investigated.
The genome sequences of M. tuberculosis H37Rv (Cole et al., 1998) and CDC1551 (http://www.tigr.org) revealed the presence of four highly related genes encoding putative phospholipases, namely plcA, plcB, plcC and plcD. The plcA, plcB and plcC genes are clustered on the chromosome, whereas plcD is located in a different region. The enzymes encoded by these genes share 30% to 40% of overall amino acid identity with the PlcH and PlcN phospholipases C of P. aeruginosa and about 70% amino acid identity between them. Expression of the M. tuberculosis plcA and plcB genes in M. smegmatis conferred upon this bacterium both sphingomyelinase and phospholipase C activities (Johansen et al., 1996). In addition, cell extracts from an Escherichia coli strain producing a recombinant PlcA protein exhibited β-haemolytic activity (Leão et al., 1995). The plcC and plcD genes whose existence was revealed upon sequencing of the M. tuberculosis genome have not yet been functionally characterized. The plcD gene is truncated and interrupted by a copy of the IS6110 insertion sequence in the laboratory strain M. tuberculosis H37Rv, but not in the clinical isolate CDC1551. Comparison of the structure of the plcD region in seven clinical isolates of M. tuberculosis revealed that three of the isolates carried the truncated version of the plcD gene (Gordon et al., 1999). The plc gene polymorphism is not restricted to plcD as shown by the fact that the M. tuberculosis genomic region encompassing the plcABC locus (region of deletion RD5) is absent from Mycobacterium bovis, M. bovis BCG and Mycobacterium microti (Gordon et al., 1999).
As phospholipases C may contribute to the pathogenesis of several important pathogens, we investigated the role of the plc genes in the virulence of M. tuberculosis by constructing and analysing plc mutant strains.
Construction of plcA, plcB, plcC, plcD, plcABC and plcABCD mutants of M. tuberculosis
We used the clinical isolate of M. tuberculosis Mt103 for all gene inactivation experiments. Because of the polymorphism affecting the plcABC and plcD genomic regions in the different species of the M. tuberculosis complex, we first confirmed that the plcA, plcB, plcC and plcD genes were all present in this isolate. The four full-length genes were amplified by polymerase chain reaction (PCR) by use of the following primer pairs: 15/16 (plcA), 17/18 (plcB), 19/20 (plcC) and 21/22 (plcD) (Table 1). In each case, an amplification product of the expected size (approximately 1.5 kb) was obtained and sequenced. For plcA, plcB and plcC, the sequence was 100% identical to that of M. tuberculosis H37Rv. In contrast to the situation in H37Rv, a non-truncated and undisrupted copy of the plcD gene was present in Mt103. Its sequence was 100% identical to that of M. bovis BCG and CDC1551. Therefore, strain Mt103 contains four full-length phospholipase C genes.
Table 1. . Primers used in this study.
Screening of the M. tuberculosis Mt103 transposon mutant library for plcA, plcB and plcC mutants
Construction of the ΔplcABC mutant
Construction of the plcD and plcABCD mutants
Amplification of plcA, plcB, plcC and plcD with Turbo pfu DNA polymerase
Two independent approaches were used to isolate insertional mutants of Mt103 deficient in the expression of the plc genes. First, we used PCR to screen 6912 M. tuberculosis clones from a transposon mutant library for the presence of transposon insertions within the plc genes as described by Jackson and colleagues (Jackson et al., 1999). The sequences of the primers used (1, 2, 3, 4, 5 and OP) are shown in Table 1. Primers 1–5 are 25 mers designed to anneal different regions of the plcABC locus (Fig. 1). Primer OP is specific for the inverted repeats (IR) of IS1096 and is directed outward the transposon. Combinations of primers (1/OP, 2/OP, 3/OP, 4/OP and 5/OP) were used to amplify the DNA regions between the respective primers and putative transposon insertion sites within the plc genes. This method enabled us to isolate mutants harbouring transposon insertions in the plcA, plcB and plcC genes (Fig. 1). The presence of a transposon in each of these genes was further confirmed by Southern blotting (data not shown). The sequence of the PCR amplification products determined that the trans-poson was inserted 18 bp downstream from the plcA start codon in the plcA mutant, 1168 bp downstream from the plcB start codon in the plcB mutant and 813 bp downstream from the plcC start codon in the plcC mutant. The plcA, plcB and plcC mutants were named MYC1555, MYC1556 and MYC1557 respectively (Table 2).
Table 2. . Plasmids and strains used in this study.
A plcABC triple mutant was constructed by allelic exchange using the Ts/sacB vectors described by Pelicic and collaborators (Pelicic et al., 1997). An inactive copy of the plcABC cluster was generated by PCR. The PCR product contained the plcABC genomic region lacking a 2.6 kb internal DNA fragment encompassing the last 600 bp of the plcA gene, the entire plcB gene and the first 200 bp of the plcC gene (Fig. 2A). A kanamycin resistance cassette was cloned into the deleted plcABC locus, yielding the disrupted allele plcABC::km. This allele and the xylE gene were then cloned into pPR27, a plasmid carrying a temperature-sensitive origin of replication and the sacB counter-selectable marker (Pelicic et al., 1997). The resulting plasmid, p27PKX, was used to achieve allelic replacement at the plcABC locus of M. tuberculosis Mt103. When selection procedure was applied, 100% of the Kanr, Sucr colonies selected were plcABC allelic exchange mutants, as confirmed by Southern blotting (Fig. 2A). One plcABC mutant was selected and named MYC1558 (Table 2).
Following the same procedure, a disrupted copy of the plcD gene (plcD::hyg) was constructed and cloned into the Ts/sacB vector pPR23 (Pelicic et al., 1997) yielding plasmid pCR8, the vector used to achieve allelic exchange at the plcD locus (Fig. 2B). M. tuberculosis Mt103 and MYC1558 were transformed with this plasmid, and allelic exchange mutants were selected using sucrose counter-selection at a non-permissive temperature. Both plcD and plcABCD mutants were obtained as confirmed by PCR analysis and Southern blotting (Fig. 2B). The plcD-deficient mutant was named MYC2508 and the quadruple plcABCD mutant was named MYC2509 (Table 2).
Mt103 and the various mutant strains all exhibited the same colony morphology on plates and the same apparent growth rate in Middlebrook 7H9 medium (Fig. 3A). Likewise, the growth rates were similar when the triple mutant (MYC1558) was complemented with one or all of the plc genes (Fig. 3B).
The genetic organization of the plcABC cluster suggests that these three genes are co-transcribed. Therefore, we performed reverse transcription (RT)-PCR experiments to investigate whether the transposon insertions within plcA and plcB had a polar effect on the transcription of the downstream plc genes (Fig. 4). In each of the MYC1555, MYC1556 and MYC1557 mutants, transcripts corresponding to the intact plc genes were detected. These results suggest that in all three mutants, the non-targeted Plc proteins are produced, although the amounts of each protein may differ from those found in the wild-type Mt103 strain.
Phospholipase C activity of the mutant strains
We used a spectrophotometric assay to monitor the phospholipase C activity of cell extracts from wild-type M. tuberculosis Mt103 and plc mutants. Cell extracts consisting of a crude preparation of cytosols, membranes and cell walls were used, rather than whole cells, to enable us to standardize the amount of total proteins used in the assays (500 μg). This assay detects the hydrolysis of a chromogenic derivative of phosphatidylcholine (PC), p-nitrophenylphosphorylcholine (pNPPC) (Kurioka and Matsuda, 1976). pNPPC is a specific substrate of phospholipases C that releases upon hydrolysis p-nitrophenol that absorbs the light at 410 nm. Measurements were performed in triplicates. The Plc activities of the cell extracts were measured after 1, 6, 14, 18 and 36 h of incubation with pNPPC (Fig. 5A). The Plc activity of the various strains increased for the first 18 h before reaching a plateau. After 18 h, the Plc activity of the single mutants MYC1555, MYC1556, MYC1557 and MYC2508 was significantly reduced (25% decrease for MYC2508 and more than 50% decrease for MYC1555, MYC1556 and MYC1557). Interestingly, MYC1555, MYC1556 and MYC1557 cell extracts had similar Plc activities. As expected, the triple (MYC1558) and quadruple (MYC2509) mutants exhibited the most drastic decreases in Plc activity (70%–80% and 85%–90% respectively) (Fig. 5). Despite the fact that pNPPC is considered to be a specific substrate for phospholipases C, it is possible that some phospholipase D or other lipase activities account for the residual activity detected in the cell extracts of the quadruple mutant MYC2509.
To address the functionality of each Plc enzyme in M. tuberculosis, we carried out a complementation analysis. Each of the plcA, plcB or plcC genes was introduced individually into the mycobacterial expression vector, pMIP12H, under control of the pBlaF* promoter (see Table 2). In addition, a 5497 bp blunt-ended BlpI–MseI restriction fragment from cosmid MTCY98 carrying the entire plcABC cluster and upstream region was inserted into pIPX59, a mycobacterial integrative vector harbouring a kanamycin and a hygromycin resistance gene (Berthet et al., 1998). MYC1558 was transformed with each of the four constructs and phospholipase C activities of the recombinant strains were assayed as described previously. Complementation of the triple mutant with each of the individual genes plcA, plcB or plcC restored 20% of the Plc activity, whereas complementation with the entire plcABC cluster restored full Plc activity (Fig. 5B). The Student's t-test confirmed that the Plc activity was significantly higher in each of the complemented strains than in MYC1558 (p < 0.05). These experiments demonstrate that all four Plc enzymes are involved in the Plc activity of M. tuberculosis and that all four enzymes are functional. Moreover, phosphatidylcholine is not only a potential substrate for PlcA and PlcB as reported earlier (Johansen et al., 1996) but also for PlcC and PlcD.
Subcellular localization of M. tuberculosis phospholipases C
Most bacterial phospholipases C are secreted proteins. The M. tuberculosis plc genes encode proteins with putative signal sequences, suggesting that they are secreted. To investigate the subcellular localization of the phospholipases C in M. tuberculosis, we checked for the presence of recombinant His-tagged PlcA, PlcB and PlcC proteins in different cell fractions. Each of the three plc genes was placed in a Mycobacterium/E. coli shuttle plasmid under control of the pBlaF* promoter and fused to a short sequence encoding a six-histidine tag (Le Dantec et al., 2001). Mt103 and MYC1558 were transformed with these plasmids, and the resulting transformants were grown to exponential phase in Sauton medium before fractionation. Proteins from the culture filtrate, cell wall and cytosol plus membrane fractions were analysed by Western blot and the recombinant Plc proteins were revealed using an anti-His antibody. The analysis performed in the Mt103 and MYC1558 recombinant strains yielded similar results. The three recombinant enzymes were found to be associated with the cell wall fraction of M. tuberculosis Mt103 (Fig. 6). Interestingly, none of the Plc enzymes was detected in the culture filtrate, suggesting that phospholipases C remain associated with the cell envelope rather than being released into the culture medium.
In vitro and in vivo expression of the plc genes
The apparent redundancy among the four Plc enzymes, which share a high degree of sequence identity and have similar enzymatic activities, led us to question the role played by these four enzymes in the biology of M. tuberculosis during in vitro growth and host infection. We used a semiquantitative RT-PCR assay to determine the level of transcription of these genes under several conditions. The amount of cDNA produced, which is proportional to the amount of the specific transcript present in the original RNA sample, was measured.
In vitro , all of the plc genes were expressed ( Fig. 4 ) which is consistent with the Plc activity assays shown in Fig. 5 . The addition of phospholipids to the growth medium is known to stimulate phospholipase activity in bacteria. For example, specific acyl-hydrolysing phospholipase activities are six to 15-fold higher in cell extracts from Mycobacterium microti and Mycobacterium avium grown in the presence of phospholipids than in extracts from the same mycobacteria grown in the absence of phospholipids ( Wheeler and Ratledge, 1992 ). Accordingly, the Plc activity detected in crude extracts of M. tuberculosis Mt103 and H37Rv increased four to 10-fold upon the addition of liposomes composed of phosphatidylcholine (PC) to the culture medium ( Fig. 7A ). We further investigated whether this increased Plc activity was related to an induction of the expression of the plc genes following the addition of PC. Total RNA was extracted from M. tuberculosis Mt103 and H37Rv that had been grown in 7H9 medium with or without PC. We used a semiquantitative RT-PCR assay to compare the amount of mRNA corresponding to the plc genes recovered from M. tuberculosis H37Rv and Mt103 grown in the two culture conditions. The mean induction ratios ranged from 3 to 9 for the four plc genes ( Fig. 7B ), strongly suggesting that the induction of the expression of the plc genes in M. tuberculosis grown in the presence of PC is responsible for the higher enzymatic activity detected ( Fig. 7A ). There was no statistically significant difference between the induction ratio in M. tuberculosis H37Rv and that in Mt103 (Student's t -test, p < 0.05).
Phospholipids are major components of eukaryotic cell membranes. M. tuberculosis might be in close contact with high concentrations of these compounds when growing in the phagosomal compartment. Therefore, the plc genes are probably induced following the infection of host cells. To study the regulation of the plc genes within host cells, THP1 macrophages were infected with M. tuberculosis H37Rv (multiplicity of infection (MOI) = 10:1 bacteria per macrophage). The macrophages were lysed at different times after infection and the mRNAs corresponding to the plc genes were recovered. The recovered mRNAs were amplified using semiquantitative RT-PCR and their amounts were compared with those obtained from bacteria grown in axenic conditions (7H9 medium). In all experiments, sigA was used as an internal standard and the results are expressed relative to the amount of sigA transcripts. sigA is an essential housekeeping sigma factor in M. tuberculosis, and the amount of sigA mRNA remains constant in different growth conditions (Manganelli et al., 1999) and during macrophage infection (Manganelli et al., 2001). The expression of the M. tuberculosis H37Rv plcA, plcB and plcC genes was strongly induced immediately after infection (Fig. 8). This high level of induction was maintained for about 24 h and then rapidly decreased finally reaching the same level as detected in vitro. Interestingly, the maximum levels of induction occurred at different times for each gene, suggesting that the roles of the three genes are not redundant. plcB was significantly more induced than the other plc genes after 1 h of infection, whereas the plcC and plcA genes were significantly more induced after 24 h of infection. Thus, the expression of the plc genes (relative to that of sigA) is greatly upregulated during the infection of host cells and their induction is transient. The differences between the induction ratios measured in vitro (using commercial PC) and in vivo may be due to differences in the quality or quantity of the phospholipid substrates or/and in the stability of the mRNAs.
The plc mutants do not show reduced growth in human macrophages
The finding that the expression of the plc genes was upregulated in M. tuberculosis after the infection of THP-1 macrophages suggests that these proteins contribute to the intracellular survival of the tubercle bacillus. To test this hypothesis, we infected the human monocytic cell line THP-1 with wild-type M. tuberculosis Mt103, the triple mutant (MYC1558) and the quadruple mutant (MYC2509) (Fig. 9). Macrophages were lysed at various times after infection and the number of intracellular viable M. tuberculosis colony-forming units (cfu) was determined. The mutants did not show any reduced virulence in this model.
Virulence of the mutant strains in the mouse model of infection
The virulence phenotypes of the triple (MYC1558) and quadruple (MYC2509) mutant strains were compared with that of the wild-type M. tuberculosis Mt103 strain in mice infected via the aerosol route. MYC1558 and MYC2509 were significantly attenuated for growth in the lungs of mice (Fig. 10A). At 26 days post infection, the mice infected with the mutants contained 10-fold less cfu than mice infected with the wild-type strain (Fig. 10A). The same reduction was observed in the spleen (data not shown). This difference increased to about 1.5 log units at later time-points (42 d), suggesting that phospholipases C have an important role in the persistence of infection. The quadruple mutant was not more attenuated than the triple mutant in this model. This suggests either that plcD does not contribute to virulence or that it acts in association with the other phospholipases. Infection experiments were also carried out with the MYC1558 strain complemented with the plcA, plcB or plcC genes carried on multicopy plasmids (Fig. 10B). Complementation with the plcB gene significantly increased the virulence of MYC1558 on days 26 and 42, whereas complementation with plcA increased the virulence of this strain on day 10. Complementation with plcC seemed to partially restore the virulence of the triple mutant on day 26, although the difference between the cfu counts of the two strains was not statistically significant, as determined by the Student's t-test (t95). These results and the fact that the plcABC cluster is distant from other ORFs on the M. tuberculosis chromosome (there are 850 bp between the plc cluster and the upstream open reading frame (ORF), and 195 bp between the plc cluster and the downstream ORF) make it unlikely that the decreased virulence of the MYC1558 strain is due to polar effect of the mutation affecting the expression of neighbouring genes. Instead, these results demonstrate that plcA, plcB and plcC all contribute to the virulence of M. tuberculosis.
The completion of the M. tuberculosis genome sequence revealed several genes thought to be involved in the virulence of the tubercle bacillus. These genes included the four putative phospholipase C genes. Two of these genes, plcA and plcB, were identified before the publication of the genome sequence of M. tuberculosis (Leão et al., 1995). The other two, plcC and plcD, were identified after the completion of the M. tuberculosis genome sequence (Cole et al., 1998). Despite the well known role of phospholipases C in the virulence of several intracellular bacterial pathogens, the involvement of these enzymes in the pathogenicity of mycobacteria had not yet been investigated.
The recent development of insertional mutagenesis tools for pathogenic mycobacteria (Pelicic et al., 1997) allowed us to construct M. tuberculosis mutants inactivated in each of the four plc genes, a triple plcABC mutant and a quadruple plcABCD mutant. Plc activity assays conducted on the non-complemented and complemented mutant strains revealed that the four plc genes encode functional phospholipases C, capable of hydrolysing a phosphatidylcholine-like substrate.
Reverse transcription (RT)-PCR assays provided evidence that the plc genes are induced during the infection of human THP-1-derived macrophages. This was not unexpected as Wheeler and Ratledge (Wheeler and Ratledge, 1992) showed that phospholipase activities are higher in mycobacteria grown in mice than in those grown in the lipid-free Dubos medium. The expression of the plc genes of other pathogenic bacteria is also upregulated during host infection (Agaisse et al., 1999; Marquis and Hager, 2000). The fact that the expression of the M. tuberculosis plc genes is upregulated in macrophages suggests that phospholipases C play a role in host infection. Consistent with this hypothesis, the disruption of the plcABCD or plcABC genes impaired the ability of M. tuberculosis Mt103 to multiply in the lungs and spleen of infected mice. This is the first evidence that phospholipases C are required for the full virulence of M. tuberculosis. The attenuated phenotype of the mutants only became evident after the acute phase of the infection, suggesting that phospholipases C are important during persistent infection. Virulence was partially restored when the triple plcABC mutant was complemented with each of the plc genes, plcA, plcB or plcC. Interestingly, the virulence did not decrease further in the quadruple mutant, suggesting that PlcD does not make a significant contribution to the virulence of M. tuberculosis or that it acts in synergy with the other Plc enzymes. A limited role for PlcD in the virulence of the tubercle bacillus would be consistent with the finding of Gordon and colleagues (Gordon et al., 1999) who showed that three out of seven clinical isolates analysed were deficient in the expression of plcD. Interestingly, the triple and quadruple mutants were not attenuated in infected THP-1 cells despite the fact that the expression of the plc genes was highly induced in this model. The growth kinetics in this cellular model may not have been measured for a sufficient period of time to detect the effects of the mutations on the intracellular multiplication and survival of M. tuberculosis. Furthermore, in vitro macrophage cultures are incomplete models that do not reflect the exact conditions encountered by bacteria in vivo and therefore may not have allowed us to detect certain attenuated phenotypes.
As mentioned earlier, the plcA, plcB and plcC genes are absent from the genomes of M. bovis and M. bovis BCG, which only carry a full-length copy of the plcD gene (Gordon et al., 1999). As PlcD exhibits some phospholipase C activity in M. tuberculosis, this enzyme may account for the Plc activity detected in M. bovis (Johansen et al., 1996). The absence of Plc activity in M. bovis BCG could be due to defects in the expression of the plcD gene. Alternatively, the PlcD enzyme may not be functional in M. bovis and M. bovis BCG, and other unrelated phospholipases C or other enzymes (such as a phospholipase D and phosphatases) may account for the Plc activity detected in M. bovis.
The fact that M. tuberculosis contains three or four phospholipases C that are important for its virulence in mice raises the questions about the functional redundancy of these enzymes. As noted earlier, other organisms, such as L. monocytogenes (Mengaud et al., 1991; Raveneau et al., 1992), contain two Plc enzymes that hydrolyse different substrates. Therefore, the M. tuberculosis Plc enzymes may have different affinities for different phospholipid substrates, thereby increasing their spectrum of action. As suggested by Johansen and colleagues (Johansen et al., 1996), a complementary hypothesis is that the plc genes are regulated differently so that they act at different stages of host infection. The differential temporal patterns of activation of the plcA, plcB and plcC genes during cell infection may support this last hypothesis (Fig. 8).
Phospholipases C expressed within host cells might serve several functions related to virulence. First, they might provide the bacteria with nutrients. Indeed, biochemical studies suggest that in chronically infected lung tissues, fatty acids might be a major source of carbon and energy for M. tuberculosis metabolism (Segal, 1984; Wheeler et al., 1990). The relatively high phospholipase activities detected in mycobacteria harvested from host tissues (Wheeler and Ratledge, 1991; 1992) and the induction of the M. tuberculosis plc genes upon entry into phagocytic cells (Fig. 8) are consistent with these enzymes having a role in the release of fatty acids from host phospholipids. M. tuberculosis may then use the fatty acids as a carbon source through the β-oxidation cycle and the glyoxylate shunt. In this regard, it is interesting to compare the regulation of the plc genes with that of the isocitrate lyase gene (icl) which encodes an essential glyoxylate shunt enzyme that is expected to act downstream of the phospholipases C. In M. tuberculosis, the expression of icl is strongly induced during the first few hours of macrophage infection, before returning to background levels 24 h post infection (McKinney et al., 2000). This is similar to the pattern observed for phospholipases C genes, which suggests that they both have roles in the same pathway. Furthermore, as for the plc genes, disruption of the icl gene impaired the ability of M. tuberculosis to multiply and to persist in mouse organs during persistent infection.
A second possible role for phospholipases C may be to degrade the phagosomal membrane, thus modifying its permeability or leading to total degradation. However, unlike the situation in other pathogenic bacteria (Marquis et al., 1995), mycobacterial phospholipases C remain associated with the cell envelope (Fig. 6) (Wheeler and Ratledge, 1991) and are not released into the culture medium. This localization appears to contradict the role of these enzymes in the degradation of the phagosomal membrane. Some authors have speculated that this arrangement indicates that mycobacterial phospholipases have a non-aggressive role (Wheeler and Ratledge, 1992) which may ultimately allow the controlled release of fatty acids from the host allowing intracellular mycobacteria to obtain nutrients without causing major damage. This property would be advantageous to mycobacterial agents that cause chronic diseases.
Finally, by activating the arachidonic acid cascade, M. tuberculosis phospholipases C may interfere with signal transduction events in infected cells, thus modulating the host immune responses (Meyers and Berk, 1990; Titball, 1993).
In conclusion, this report provides the first evidence that phospholipases C are involved in the virulence of M. tuberculosis. Based on the similarities between the plc genes and the icl gene in terms of intracellular regulation and their involvement in the late phase of mouse infection, we propose that the major role of phospholipases C in the course of infection is to provide M. tuberculosis with host fatty acids, which are then used as a carbon source through the β-oxidation cycle and the glyoxylate shunt.
Bacterial strains and growth conditions
Mycobacterium smegmatis mc 2 155 ( Snapper et al., 1990 ), Mycobacterium bovis BCG Pasteur (CIPT 140040001) and Mycobacterium tuberculosis 103 (clinical isolate Mt103) were used in this study. Mycobacteria were grown at 32°C, 37°C or 39°C in liquid Middlebrook 7H9 medium (Difco) supplemented with 0.05% Tween 80 and ADC (Becton Dickinson), in Sauton medium or on solid Middlebrook 7H10 or 7H11 medium (Difco) supplemented with OADC (Becton Dickinson). Escherichia coli DH5α, the strain used in the cloning experiments, was grown on Luria–Bertani medium (LB) (Difco). When required, the medium was supplemented with 2% sucrose or the following amounts of anti-biotics: 100 μg ml −1 of ampicillin, 100 μg ml −1 of gentamicin, 20 μg ml −1 of kanamycin, 200 μg ml −1 of hygromycin for E. coli or 50 μg ml −1 for mycobacteria.
Construction of the plcABC, plcD and plcABCD mutants
plcD (MYC2508), plcABC (triple mutant, MYC1558) and plcABCD (quadruple mutant, MYC2509) mutants were constructed by allelic replacement using the Ts/ sacB method described by Pelicic and colleagues ( Pelicic et al., 1997 ). plcABC :: km , the disrupted plcABC allele used in the gene replacement experiment, was generated by polymerase chain reaction (PCR) using the primer pairs 7/8 and 9/10 ( Table 1 ). The amplified PCR fragments (a 1.5 kb fragment containing the first 915 bp of plcA and the upstream region, and a 1.8 kb fragment carrying the last 1330 bp of plcC and the downstream region) were digested with Xho I and ligated to generate a disrupted plcABC cluster lacking a 2.6 kb internal DNA fragment. The disrupted plcABC fragment was then cut with Eco RI and inserted into Eco RI-cut pUC19 yielding pUCP. The kanamycin resistance cassette ( km ) from pUC4K (Amersham Pharmacia Biotech) was cut with Sal I and then introduced into the unique Xho I site of pUCP. plcABC :: km carried on a 4.5 kb Eco RI fragment was then inserted into Eco RI-cut pXYL4 ( Pelicic et al., 1997 ), yielding pPKX. Finally, p27PKX, the vector used for allelic replacement, was obtained by inserting the 5.5 kb Bam HI fragment from pPKX carrying plcABC :: km and xylE into the Bam HI site of pPR27 ( Pelicic et al., 1997 ). plcD :: hyg , the disrupted allele used for allelic replacement at the plcD locus of M. tuberculosis was generated by PCR. A 750 bp DNA fragment designed to carry a Not I and a Hin dIII restriction sites and a 1000 bp fragment designed to carry a Hin dIII and a Spe I restriction sites were amplified by PCR using the 11/12 and 13/14 primer pairs, respectively (see Table 1 ). The two Hin dIII-cut PCR fragments were ligated and the resulting 1.75 kb fragment was digested with Not I and Nde I before insertion into a NotI/ SpeI-cut derivative of pBluescript KS− (Stratagene) devoid of HindIII site yielding pBS(plcD). A hygromycin resistance cassette (hyg) extracted from pUChygro on a HindIII restriction fragment was then ligated into the HindIII-cut pBS(plcD) to yield pBS(plcD::hyg). pCR8, the plasmid used for allelic replacement at the plcD locus of Mt103 and MYC1558 was finally obtained by inserting the NotI–SpeI-cut disrupted plcD::hyg allele from pBS(plcD::hyg) into pPR23 (Pelicic et al., 1997), which had been cut with the same enzymes.
DNA manipulations and PCR amplifications
All nucleic acid manipulations were performed according to standard molecular biology techniques (Sambrook et al., 1989) or to the recommendations of the manufacturers. All transformations were performed by electroporation using a Gene Pulser unit (Bio-Rad). Plasmids were extracted from E. coli using the QIAprep Spin Miniprep kit and DNA fragments were purified using the QIAquick PCR-purification kit and the QIAquick gel Extraction kit (Qiagen).
Polymerase Chain Reactions amplifications were carried out in a GenAmp PCR system 9600 machine (Perkin-Elmer). The primers used are described in Table 1. The PCR mixes (GenAmpR PCR core reagents, Roche) were as recommended by the manufacturer with the following modifications: dNTP were included at a final concentration of 100 μM, 1 unit of AmpliTaq Gold DNA polymerase or Turbo Pfu DNA polymerase (Stratagene) and 20 pmol of each primer were used per reaction. The PCR conditions consisted of one denaturation cycle (95°C, 10 min for AmpliTaq Gold or 95°C, 5 min for Turbo Pfu DNA polymerase), followed by 30 cycles of denaturation (95° C, 1 min), annealing (60°C, 1 min), primer extension (72°C, 1.5 min) and a final extension at 72°C for 10 min.
The labelling of the DNA probes with [α-32P]-dCTP and Southern blot analyses were performed as described (Jackson et al., 1997).
Construction of the complementation vectors carrying the plcA, plcB and plcC genes
Wild-type copies of the Mt103 plcA, plcB and plcC genes were PCR-amplified using the 15/16, 17/18 and 19/20 primer pairs respectively (Table 1). The primers were designed to generate PCR products harbouring BamHI and KpnI restrictions sites at their extremities. The amplification products were ligated into the BamHI and KpnI-cut pMIP12 (Le Dantec et al., 2001), a Mycobacterium/E. coli replicative shuttle plasmid that allows genes to be expressed under the control of the pBlaF* promoter and C-terminal six-His-tagged recombinant proteins to be produced. E. coli DH5α transformants harbouring pMIP12 vectors with plcA, plcB or plcC inserts (named pCR1, pCR2 and pCR3 respectively) (Table 2) were screened by colony hybridization as described by Jackson and colleagues (Jackson et al., 1997) with minor modifications. The membranes were washed in high-stringency conditions at 65°C, twice in 2× SSC plus 0.1% SDS, once in 1× SSC plus 0.1% SDS and twice in 0.1× SSC plus 0.1% SDS. The probes used corresponded to the PCR-amplified plcA, plcB and plcC genes described above. For the complementation of MYC1558 with the plcA, plcB and plcC genes, the three structural genes were extracted from pCR1, pCR2 and pCR3 on XbaI–BglII restriction fragments and inserted into XbaI- and BglII-cut pMIP12H, a derivative of pMIP12 carrying a hygromycin resistance cassette. pCR9, the vector carrying the entire plcABC cluster was constructed by inserting a blunt-ended 5497 bp MseI–BlpI restriction fragment from cosmid MTCY98 into the XbaI-cut and blunt-ended pIPX59, a Mycobacterium integrative vector harbouring a hygromycin resistance cassette (Berthet et al., 1998).
Double-stranded plasmid DNA (pCR1, pCR2 and pCR3) and PCR fragments were sequenced using an automated DNA sequencer (Applied Biosystems, model 373) with a dye deoxy terminator cycle sequencing kit (Applied Biosystems).
Extraction of RNA from M. tuberculosis and RT-PCR experiments
Mycobacterium tuberculosis was grown to OD 600 = 0.6. Cells were broken in a solution of Trizol (1 ml) (Life Technologies) with mini glass beads using a Bead Beater apparatus (Polylabo) set at maximum speed. RNA was extracted with 300 μl of chloroform:isoamyl alcohol. After 10 min of centrifugation at 13 000 g , the aqueous phase was transferred to a tube containing 270 μl of isopropanol. Total RNA was then precipitated overnight at 4°C and washed with 1 ml of a 75% ethanol solution before resuspension in diethyl pyrocarbonate (DEPC, Sigma)-treated water. Contaminating DNA was removed by digestion with DNase I according to the manufacturer's instructions (Ambion). The same protocol was used to extract RNA from M. tuberculosis infecting THP-1 cells. At different time-points, infected macrophages were treated with Trizol before RNA extraction. This experiment was repeated twice on two independent stocks of THP-1 cells.
Reverse transcription (RT)-PCR experiments were re-peated at least three times on the RNA extracted from infected THP-1 cells and carried out as described by Manganelli and collaborators (Manganelli et al., 1999) using 2 μg of RNA and specific primers corresponding to each plc gene and to the sigA gene (Table 1). The PCR conditions were the same as those used by Manganelli and colleagues (Manganelli et al., 1999). The plc genes gave single specific amplification products that could be labelled with the general fluorescent probe Sybr Green (Sigma), 25 ng μl−1 final concentration. The amplification of sigA by the Sybr Green technique was less specific and clearly sensitive to RNA contamination from the macrophages. Therefore, a specific fluorescent probe (beacon, 25 ng μl−1 final concentration) was used for sigA (Table 1). The significance of differences was determined by the Student's t-test (t95).
Fractionation of M. tuberculosis
Culture filtrates from M. tuberculosis cultures grown in Sauton medium were filtered twice through a 0.2 μm membrane (Millipore) to remove contaminating cells. Bacterial cells were resuspended in water and broken for 3 min with mini glass beads in a Bead Beater apparatus (Polylabo) set at maximum speed. Beads and unbroken cells were removed by centrifugation at 5000 g for 10 min. Supernatants were further centrifuged for 30 min at 15 000 g. The resulting supernatant contained cytoplasmic and membrane com-ponents and the pellet contained cell wall components. The protein concentration of each fraction was measured using a Coomassie blue assay (Bio-Rad). In total, 100 μg of proteins was used for the isocitrate dehydrogenase activity assay as described previously (Raynaud et al., 1998). Isocitrate dehydrogenase is a cytosolic enzyme (Andersen et al., 1991) that can be used as a marker to check that the cell wall and culture filtrate fractions are not contaminated with cytoplasm.
Phospholipase C activity
The substrate used in the phospholipases C assays was the chromogenic derivative of phosphatidylcholine p-nitrophenyl phosphorylcholine (pNPPC, Sigma) (Kurioka and Matsuda, 1976). Assays were performed on crude bacterial extracts consisting of the supernatant recovered after breaking M. tuberculosis cells in a Bead Beater and centrifugation at 5000 g for 10 min. Crude extracts containing 500 μg of proteins were incubated at 37°C for 1, 6, 14, 18 and 36 h in a buffer containing 6 mM Tris (pH 7.2), 5 mM pNPPC and 1.5% sorbitol. The reaction was stopped by the addition of 3 ml of a 1% Na2CO3 solution and the release of p-nitrophenol was measured at 410 nm. Negative controls contained water instead of cell extracts and positive controls contained commercial phospholipase C (Sigma). To avoid bacterial contamination in the reaction mix, chloramphenicol (100 μg ml−1) and gentamicin (100 μg ml−1) (Sigma) were systematically added. The presence of possible microbial contamination was carefully checked after each long incubation period (Raynaud et al., 1998). All enzymatic assays were performed in triplicates. The significance of differences was determined by the Student's t-test (t95).
SDS-polyacrylamide gel electrophoresis (SDS–PAGE) and immunoblotting
SDS–polyacrylamide gel electrophoresis (SDS–PAGE) was carried out on a Mini-Protean apparatus (Bio-Rad) and proteins were transferred to Hybond C membranes (Amersham) with a mini-Transblot apparatus (Bio-Rad) following the recommendations of the manufacturer. The production of recombinant Plc proteins by M. tuberculosis recombinant strains was checked using a mouse monoclonal anti-His antibody (Qiagen) diluted 1:3000 and a sheep anti-mouse IgG HRP conjugated secondary antibody (Amersham Pharmacia Biotech) diluted 1:10 000. Bound antibodies were detected using the ECL system (Amersham Pharmacia Biotech).
Aerosol infection in mice
Six- to eight-week-old female BALB/c mice were infected with approximately 200 cfu of M. tuberculosis Mt103, MYC1558, MYC2509, MYC2501, MYC2502 and MYC2503 via the aerosol route. M. tuberculosis aerosols were generated from bacterial suspensions consisting of 3 × 107 cfu ml−1 in phosphate-buffered saline (PBS) solution (pH 7.4) with 0.05% Tween 80. Mice were exposed to the aerosols for 15 min. Four or five mice were used for each experimental time-point. At various time-points post infection, the lungs and spleens were removed aseptically and homogenized. Serial dilutions of organ homogenates were plated on solid medium 7H11 supplemented with the appropriate antibiotics (Jackson et al., 1999).
Infection of THP-1-derived macrophages with M. tuberculosis
THP-1 cells were obtained from the ATCC collection and grown at 37°C in a 5% CO2 atmosphere. Cells were maintained in RPMI-1640 medium (Life Technologies) containing 2 mM L-glutamine, 1.5 g l−1 of sodium bicarbonate, 4.5 g l−1 of glucose, 1.0 mM sodium pyruvate, 50 μM 2-mercaptoethanol and 10% foetal bovine serum. THP-1 cell suspensions were adjusted to a concentration of 106 cells ml−1 in warm RPMI supplemented with 50 nM phorbol 12-myristate 13-acetate (Sigma), used to seed tissue culture plates (1 ml per well) and allowed to differentiate for 24 h. The medium was then removed and replaced with 1 ml of bacterial suspension in RPMI containing 1 × 105 cfu ml−1 (multiplicity of infection (MOI) = 1:10 cfu per macrophage). After 16 h at 37°C, the medium was removed and the wells were washed twice with RPMI to remove extracellular bacteria. On days 1, 4, 7 and 11, cells were lysed with 500 μl of a 2% saponin solution, and the number of viable intracellular cfu was counted by plating serial dilutions of the lysis solution onto Middlebrook 7H10 agar. This infection experiment was carried out in duplicate.
We thank Pierre Chavarro and Eddy Maranghi for their technical help with the infection of mice. We would also like to thank Dr Eugenie Dubnau, Dr Patricia Fontan and Mr Salvatore Marras for their excellent advice and for their help with the RT-PCR experiments. We are grateful to Dr Nathalie Winter for the gift of pMIP12 and pMIP12H, to Pr. Stewart Cole for cosmid MTCY98, to Dr Jean-Marc Reyrat for excellent discussions and to Dr Philip Draper for critical reading of the manuscript. This work was supported by grants from the European Commission (contract IC18-CT97-0252) and from the European Economic Community (contract QLK2- 2000–01761). I.S. was supported by the NIH grant AI 44856.