Most drug-resistant clinical isolates of the tubercle bacillus are resistant to isoniazid, a first-line antituberculous drug. This antibiotic was shown to act on Mycobacterium tuberculosis by inhibiting a 2-trans-enoyl-acyl carrier protein reductase, called InhA. However, the exact role played by InhA in mycobacteria remained unclear. A mycobacterial enzyme fraction containing InhA was isolated. It displays a long-chain fatty acid elongation activity with the characteristic properties described for the FAS-II (fatty acid synthetase II) system. Inhibition of this activity by InhA inhibitors, namely isoniazid, hexadecynoyl-CoA or octadecynoyl-CoA, showed that InhA belongs to the FAS-II system. Moreover, the InhA inhibitors also blocked the biosynthesis of mycolic acids, which are major lipids of the mycobacterial envelope. The data strongly suggest that isoniazid acts on the mycobacterial cell wall by preventing the FAS-II system from producing long-chain fatty acid precursors for mycolic acid biosynthesis.

Phenol-phthiocerol glycolipids have been found previously in Mycobacterium leprae, M . kansasii, M. bovis and M . marinum, but not in M . tuberculosis. A search for such glycolipids in this latter species showed that the Canetti strains of M . tuberculosis synthesize a major triglycosyl phenol-phthiocerol, accompanied by minor amounts of other glycolipids with a similar aglycone moiety.The triglycoside moiety has the following structure: 2,3,The aglycone moiety consists in phenol-phthiocerol (two homologs). Its two secondary alcohol functions are esterified by mycocerosic acids (homologs with 26-32 carbon atoms and with 2 -4 methyl branches). The proposed structure differs on several points from the M . leprae glycolipids, but presents some analogy with the major glycolipid of M. kansasii.A minor monoglycosyl phenol-phthiocerol was also studied. Its overall structure is very similar to that of M . bovis, with 2-0-methyl rhamnose as sugar moiety.In the last few years, it has been realized that glycolipids are major surface antigens of pathogenic mycobacteria [l], and important applications are expected from this: the detection and identification of bacteria in patients and the elaboration of semi-synthetic vaccines.Three phenol-phthiocerol trisaccharides isolated from Mycobacterium leprae have attracted much attention, since they were the first antigens to be chemically defined in this important pathogen [l -31.Long ago, phenol-phthiocerol monosaccharides were identified in M . bovis [4], M. marinum [5] and M . kansasii [6]; the structure of the glycolipid of the latter species has been revised recently [7, 81. The four mycobacteria quoted above, as well as M . tuberculosis, are pathogenic for man and have the same kind of mycolic acids in common [9]. However no phenol-phthiocerol glycolipid had been found previously to this paper in the most important pathogen of this genus, M . tuberculosis. We searched for glycolipids in various strains belonging to this species, and a major phenol-phthiocerol trisaccharide, accompanied by minor amounts of a phenol-phthiocerol monosaccharide, was detected in the four Canetti strains of M. tuberculosis tested. The structure of the triglycosyl glycolipid is described here. It differs on several points from the glycolipids of M . leprae, and presents some analogy with the major glycolipid of M. kansasii.

Phthienoic acids constitute a family of dextro-rotary odd-numbered unsaturated fatty acids isolated exclusively from virulent strains of human and bovine tubercle bacilli. In the bacterial cell they are not free and a search for their linked form in complex wall lipids of Mycobacterium tuberculosis (strain Canetti) showed that they esterified trehalose.Structural elucidation of the major phthienoyl trehalose showed the occurrence of five acyl residues located at 2, 2', 3', 4 and 6' positions of trehalose. The acyl substituents were mainly 2,4,6-trimethyl tetracos-2-enoic acid (C27 phthienoic acid) accompanied by its homologs. In addition to these branched fatty acids, straight-chain CI6 and C,* acyls composed about 20% of the substituents. The proposed structure is a new one, both for the mycobacterial-specific glycolipid and for the substituted positions on trehalose.Other minor acyl trehaloses were detected in M . tuberculosis (strain Canetti), differing from the major component by the occurrence of an additional hydroxy fatty acid (3-hydroxy-2,4,6-trimethyl tetracosanoic acid) or by the number of acyl substituents. The major glycolipid presented a weak activity in vitro on mitochondria1 oxidative phosphorylation. These glycolipids and phthienoic acids could serve as virulence indicators.Mycobacterial cell walls are endowed with large amounts of glycolipids belonging to three classes, according to their structure: glycopeptidolipids called mycosides C, phenolic glycolipids, and trehalose-containing glycolipids (for recent reviews see [I -31). While dimycoloyl trehalose (called cord factor) is widespread among mycobacteria 111, the other glycolipids exhibit a more specific distribution in the genus Mycobacterium, allowing identification of species or subspecies by means of the presence of specific antigenic determinants [3]. In the course of establishing such a principle for identification of mycobacterial species, non-tuberculous mycobacteria have been examined, skirting the most serious pathogen Mycobacteriurn tuberculosis [3]. Seeking specific glycolipids of the tubercle bacillus, we recently discovered phenolic glycolipids in the Canetti strains [4]. In a detailed examination of the lipid content of these strains, we found unusual glycolipids and the structure of the major one is described here. This glycolipid is a polyacylated trehalose, where the acyl substituents are mainly phthienoic acids (Fig. 1) isolated so far only from virulent human and bovine strains of the tubercle bacillus [S, 61. MATERIALS AND METHODS Mycohacterial strains and culturesM . tuberculosis (H3,Ra, H37Rv, Canetti) and M . bovis (ATCC 19210 and BCG) were obtained from the Institut Pasteur (Paris). Cultures were grown for 3 weeks on Sauton medium at 37°C and were harvested by filtration during the stationary growth phase. Lipid extractionThe bacteria were kept in CHC13/CH30H (l/l, by vol.) for several days at room temperature to kill them. The cells were extracted twice with CHC13/CH30H (2/1, by vol.) and the crude extracts were po...

Mycobacteria are known to acquire resistance to the antituberculous drug pyrazinamide (PZA) through mutations in the gene encoding pyrazinamidase (PZase), an enzyme that converts PZA into pyrazinoic acid, the presumed active form of PZA against bacteria. Additional mechanisms of resistance to the drug are known to exist but have not been fully investigated. Among these is the non-uptake of the pro-drug, a possibility investigated in the present study. The uptake mechanism of PZA, a requisite step for the activation of the pro-drug, was studied in Mycobacterium tuberculosis. The incorporation of [14C]PZA by the bacilli was followed in both neutral and acidic environments since PZA activity is known to be optimal a t acidic pH. By using a protonophore (carbonyl cyanide m-chlorophenylhydrazone; CCCP), valinomycin, arsenate and low temperature, it was shown that an ATP-dependent transport system is involved in the uptake of PZA. Whilst the structurally analogous compound nicotinamide inhibited the transport system of PZA, other structurally related compounds such as pyrazinoic acid, isoniazid and cytosine did not. Acidic conditions were also without effect. Based on diffusion experiments in liposomes, it was found that PZA diffuses rapidly through membrane bilayers, faster than glycerol, whilst the presence of OmpATb, the porin-like protein of M. tuberculosis, in proteoliposomes slightly increased the diffusion of the drug. This finding may explain why the cell wall mycolate hydrophobic layer does not represent the limiting step in the diffusion of PZA, as judged from comparative experiments using a M. tuberculosis strain and its isogenic mutant elaborating 40 O/ O less covalently linked mycolates. PZase activity, and PZA uptake and susceptibility in different mycobacterial species were compared. M. tuberculosis, a naturally PZA-susceptible species, was the only species that exhibited both PZase activity and PZA uptake; no such correlation was observed with the four naturally resistant species examined. Mycobacterium smegmatis possessed a functional PZase but did not take up PZA; the reverse was true for the PZase-negative strain of Mycobacterium avium used, with PZA uptake comparable to that of M. tuberculosis. Mycobacterium bovis BCG and Mycobacterium kansasii exhibited neither a PZase activity nor PZA uptake. These data clearly demonstrate that one of the mechanisms of resistance to PZA resides in the failure of strains to take up the drug, indicating that susceptibility to PZA in mycobacteria requires both the presence of a functional PZase and a PZA transport system. No correlation was observed between the occurrence and cellular location of PZase and of nicotinamidase in the strains examined, suggesting that one or both amides can be hydrolysed by other mycobacterial amidases.