FbpA-Dependent Biosynthesis of Trehalose Dimycolate Is Required for the Intrinsic Multidrug Resistance, Cell Wall Structure, and Colonial Morphology of Mycobacterium smegmatis

Bacterial strains, plasmids, and media.

All strains and plasmids used in this study are listed in Table 1. Wild-type M. smegmatis strain MC²155 (35) and its transposon-derived mutants were grown in 7H9 liquid and 7H10 (Difco) or LB agar medium supplemented with 0.5% Tween 80. Kanamycin was used at a final concentration of 50 μg ml⁻¹. Hygromycin was used at 100 μg ml⁻¹ and 75 μg ml⁻¹ for Escherichia coli and mycobacteria, respectively. Genomic DNA from M. smegmatis was isolated using the DNAzol kit (MRC). Transformation was carried out as described elsewhere (7).

Mariner transposon-based library screen of antibiotic-sensitive mutants.

The pMycoMar vector that carries a Himar1 transposon was used to make the mutation library (33). Wild-type M. smegmatis MC²155 was transformed with pMycoMar. Transformed bacteria were cultivated at 28°C overnight to recover and amplify the library before plating on LB agar plates containing 50 μg ml⁻¹ kanamycin. After incubation for 3 to 5 days at 40°C, single colonies were picked and spotted in arrays on kanamycin-containing plates. These plates were used as “master plates” to replicate to NE plates (26) containing different antibiotics. Colonies which grew on kanamycin NE plates but failed to grow on selected antibiotic plates were subjected to antibiotic disk tests to confirm their sensitivity profile.

Arbitrary PCR identification of drug-sensitive transposants.

The identification of transposon mutants by using arbitrary PCR was carried out as described previously (28). A first round of PCR was done using the Roche Expand long-template PCR system with the random annealing primers ARB1/ARB6 and the pMycoMar-specific primers MarExt1 and pMarExt2 (Table 1). Cells from colonies grown on kanamycin plates were directly used as template for the PCR. Annealing temperature was set at 45°C. Products of the first-round PCR were used as template for the second-round PCR, which used Taq polymerase (Roche) and the primers ARB2 and MarInt1/MarInt2. PCR products from the second round were cleaned up using a QIAGEN PCR purification kit and sequenced. PCR using primers flanking the identified open reading frame or Southern blot were used to identify the precise insertion site.

Cloning of fbp genes and complementation.

The GC-rich PCR system (Roche) was used to clone genes from M. smegmatis genomic DNA. The fbpA_MS gene was PCR amplified using the primers MS_FbpA.EB and MS_FbpA.HXb. fbpD_MS was amplified using the primers MS_FbpD.EB and MS_FbpD.PHXb (Table 1). PCR products were ligated to the pGEM-T Easy vector (Promega) and sequenced. DNA fragments were then removed by digestion with EcoRI and HindIII and cloned into pMV361 (36) cut with the same enzymes to fuse them with the heat shock promoter hsp60. Genes fused with the hsp60 promoter were then excised from pMV361 using NheI/XbaI and ligated into pMycVec2 (18) digested with SpeI. These plasmids were electroporated into the MAR1 mutant, and transformants were selected on hygromycin-containing LB agar plates. The M. tuberculosis fbpA_MTB gene was PCR amplified using the primers TB_FbpA.EB and TB_FbpA.ClXbE and cloned into pMV361 using EcoRI/ClaI. The fused hsp60-fbpA_MTB was cloned into pMycVec2 as described above. To construct the vector expressing fbpA_MS and fbpD_MS, the NheI/XbaI DNA fragment carrying hsp60-fbpA_MS was filled in with Klenow polymerase I and blunt-end ligated into the EcoRV site of pMycVec2-hsp60-fbpD_MS.

The mutated allele fbpA_MS(S171A) was made using the primers S171fbpA_MS-up and S171fbpA_MS-down in combination with MS_FbpA.EB and MS_FbpA.HXb in a two-round PCR protocol. This changed the TCG (serine) codon into GCG (alanine), corresponding to amino acid residue 171.

Drug resistance measurements.

Antibiotic susceptibilities of strains were measured on NE medium (26) using a diffusion assay with antibiotic disks (Pasteur Diagnostics). MICs were determined using E-test antibiotic strips (AB Biodisk) following instructions of the manufacturer. Exponential cultures grown in Middlebrook 7H9 broth supplemented with 0.5% Tween 80 were harvested after 20 h. Cultures were normalized to an optical density at 600 nm of 1.5, and 100 μl was seeded in top agar on NE medium. E-test strips were applied, and MICs were scored after 60 h at 37°C. Values were read from the scale (in μg ml⁻¹) at the point of intersection between the inhibition ellipse edge and the E-test strip.

Extraction and analyses of MAMEs.

MAMEs were extracted as described previously (6). Briefly, 2 ml of 15% tetrabutylammonium hydroxide was added to 50 mg of either defatted cells or whole cells of M. smegmatis in an 8.5-ml tube and heated at 100°C, overnight. After cooling, the reaction mixture was diluted with 2 ml water followed by addition of 1 ml dichloromethane and 250 μl iodomethane and stirred for 30 min. The upper layer was then removed and the lower organic layer was washed with 3 ml of 1 M hydrochloric acid, followed by 3 ml of water. The samples were dried under a stream of nitrogen using a sand bath at 37°C. The crude MAMEs thus obtained were dissolved in 0.5 ml of dichloromethane, transferred to a polypropylene microcentrifuge tube, and evaporated to dryness under nitrogen. The residue was dissolved in a mixture of 0.2 ml toluene and 0.1 ml acetonitrile and a further addition of 0.2 ml acetonitrile. Samples were incubated for 1 h at 4°C. The precipitated MAMEs were removed from the supernatant by centrifugation and resuspended in dichloromethane prior to thin-layer chromatography (TLC) analysis. These MAMEs were analyzed by TLC using a solvent system containing petroleum ether-diethyl ether (95:5) (6). Individual MAMEs were revealed by charring at 110°C for 15 min using 5% ethanolic molybdophosphoric acid.

Extraction and analyses of TDM.

Extraction of TDM was performed as described elsewhere (16). Bacteria were vortexed vigorously in petroleum ether for 2 min, followed by 5 min of incubation at room temperature. The culture was centrifuged at 500 × g for 10 min. The supernatant was removed, and the extraction process was repeated twice. The petroleum ether extracts of mycobacteria contain primarily TDM (>95% of the total extract), with relatively small quantities of free mycolic acid glycerides, menaquinones, and hydrocarbons present. The resulting TDM was analyzed by TLC (Merck 5554-silica gel 60F₂₅₄; 6.6 cm by 6.6 cm) using chloroform-methanol (9:1) as the solvent system. The spots were visualized by spraying with 5% sulfuric acid containing 10% α-naphthol followed by charring for 10 min.

The TDM of M. smegmatis strains were quantified by radioactive labeling. Briefly, 5 Ci [¹⁴C]acetate (2.22 GBq mmol⁻¹; Amersham) was added to 100 ml of exponential-phase bacterial cultures, which were incubated for 3 to 5 h. Labeling was stopped by centrifugation, and the cell pellets were extracted with petroleum ether as described above. The crude TDM (>95% pure) thus obtained was purified by TLC and quantified by scintillation counting. A minimum of three sequential determinations was performed from separate preparations. Infrared (IR) and matrix-assisted laser desorption ionization-time of flight (MALDI-TOF) analyses were carried out to verify the identity of the TDM (see below).

Infrared spectroscopic analysis of TDM.

The identity of the extracted TDM of M. smegmatis was confirmed by IR spectroscopy. IR spectra of samples were recorded using a Perkin-Elmer (Spectrum One) FT-IR spectrometer. Mycobacterium tuberculosis TDM (Sigma) was used as the standard. The purified TDM was dissolved in chloroform to obtain the IR spectra.

MALDI-TOF mass spectrometry analysis of TDM and MAMEs.

MALDI-TOF mass spectra (in the positive mode) were acquired on a Voyager-DE STR mass spectrometer (PerSeptive Biosystems) equipped with a pulsed nitrogen laser emitting at 337 nm. Samples were analyzed in the reflectron mode using an extraction delay time of 100 ns and an accelerating voltage operating in positive ion mode of 20 kV (21). To improve the signal-to-noise ratio, 150 single shots were averaged for each mass spectrum and, typically, four individual spectra were accumulated to generate a summed spectrum.

Isolation of a multidrug-sensitive mutant of M. smegmatis from a mariner transposon library.

We used the Himar1 transposon system (20) to mutagenize wild-type M. smegmatis MC²155. This library of 15,000 isolated insertion mutants was replica plated and screened for increased sensitivity to various antibiotics, including erythromycin, spectinomycin, spiramycin, and chloramphenicol (see Materials and Methods).

A mutant (MAR1), found in a screen for erythromycin sensitivity (MIC less than 8 μg ml⁻¹), displayed additional sensitivities to many other antibiotics. Disk diffusion assays showed that MAR1 was more sensitive to many clinically important antibiotics with diverse chemical structures and functions. These included hydrophilic and hydrophobic antibiotics targeted to the ribosome (macrolides, lincosamides, streptogramins, spectinomycin, and chloramphenicol), cell wall assembly enzymes (vancomycin, imipenem, and daptomycin), and RNA polymerase (rifampin). MICs for the four representative antibiotics were quantified using E-test diffusion strips; MAR1 was 24-, 85-, ≫8-, and 12-fold more susceptible to erythromycin, imipenem, rifampin, and vancomycin, respectively (Fig. 1).

MAR1-a fbpA mutant.

Arbitrary PCR first localized the transposon mutation in MAR1 to a gene encoding a protein of the “antigen 85 complex,” also known as “fibronectin binding proteins” (Fbps). The amino acid sequence of the mutated gene (fbpA_MS) had highest similarities to M. tuberculosis fbpA (fbpA_MTB) and homologs in other mycobacteria; the locus displayed synteny with the M. tuberculosis and Mycobacterium leprae genomes (Fig. 2A). All three genomes encode a fbpA-like gene upstream of a fbpD-like gene (Fig. 2A). Interestingly, the M. smegmatis locus has an additional fbp-like gene upstream of fbpA_MS (fbpE_MS) (Fig. 2A), which was less homologous to fbpA.

Alignment of amino acid sequences of these proteins with other Fbp proteins revealed that FbpE_MS and FbpA_MS contained a carboxyesterase domain with a conserved serine 171 (Fig. 2D) that is essential for in vitro mycolyltransferase activity (5). Phylogenic analyses also supported the homology of FbpA and FbpD proteins in the three genomes (not shown). FbpE_MS was equally similar to M. tuberculosis FbpA, FbpB, and FbpC.

The disruption was further confirmed by recovering the mutant locus by PCR using primers flanking the putative open reading frame. The mutant gene generated a larger fragment corresponding to the inserted transposon. Insertion of the transposon resulted in a decreased mobility of the PCR product on an agarose gel (Fig. 2B). Sequencing of the junction region from these PCR products identified the insertion site at a dinucleotide TA (Fig. 2A), thus introducing a stop codon after the triplet encoding the Leu79 residue. To confirm loss of the fbpA_MS gene product, Western blot assays were done using an antibody against the M. tuberculosis antigen 85 complex HYT27 (Antibodyshop A/S, Gentofte, Denmark). There were two protein bands that cross-reacted with HYT27 in MC²155 culture filtrate; the higher-molecular-weight, weaker band was missing in the MAR1 filtrate (Fig. 2C).

Complementation of MAR1 with the fbpA_MS gene transcribed from the heat shock promoter hsp60 (see Materials and Methods) only partially restored resistance to antibiotics (Fig. 1). Expression of the M. tuberculosis fbpA gene in MAR1 provided a lower level of antibiotic resistance (Fig. 1).

The fbpA_MS mutation caused major alterations in colonial morphology and hydrophobicity.

The MAR1 mutant had significant alterations in colonial morphology. The edges of wild-type colonies were irregular, while those of MAR1 were smooth and rounded (Fig. 3A). On the surface, colonies of the wild type formed wrinkled and acne-like structures after a few days of cultivation, while MAR1 colonies were shiny and smooth (Fig. 3A). In addition, while wild-type colonies were dry and fragile, MAR1 colonies were wetter and stickier. This phenotype was partially suppressed when culture media were supplemented with Tween 80 detergent. These observations first suggested that the mutant surface topology might be due to alterations in its surface tension. This was verified by a simple experiment in which droplets of water or oil were applied to wild-type or mutant lawns (Fig. 3B). On a lawn of a wild-type M. smegmatis culture, a drop of water formed a bead while a drop of oil readily spread over the surface. This assay provided a simple visualization of the characteristic hydrophobicity of mycobacteria. In contrast, water spread on a MAR1 lawn while oil formed a bead. This phenotype was suppressed when MAR1 was complemented with fbpA_MS. Colonial morphology and hydrophobicity of MAR1 were partially rescued by expression of fbpA_MTB (not shown). These observations implied that the characteristic hydrophobicity of the mycobacterial cell wall was lost in the MAR1 mutant and was due to fbpA activity.

FbpD_MS expression does not play a role in the phenotypes of MAR1.

To test if fbpD_MS, the gene downstream of fbpA_MS, was involved in the observed phenotypes of MAR1, cloned fbpD_MS was expressed from the heat shock promoter in MAR1 (as described above for fbpA_MS). fbpD_MS expression could not rescue the antibiotic resistance (Fig. 1) or morphological defects (not shown) of MAR1. However, when this fbpD_MS expression plasmid was supplemented with another DNA fragment containing hsp60-fbpA_MS, both the defects were partially rescued (antibiotic resistance results are shown in Fig. 1). This suggested that the MAR1 phenotypes were FbpA_MS specific and did not involve FbpD_MS.

FbpA_MS activity requires the serine esterase domain.

The reduced antibiotic resistance in the fbpA_MS disruption mutant was likely a consequence of a loss of mycolyltransferase activity. Fbp enzymatic activity is dependent on a serine esterase domain (5), which forms a trehalose-binding pocket (32). Mutagenesis studies revealed that the conserved serine 171 in this domain was essential for in vitro mycolyltransferase activity, ligating mycolate to trehalose (5). In order to test whether the FbpA_MS-dependent antibiotic resistance was due to this activity, we constructed a vector expressing a mutated allele of FbpA_MS in which serine 171 was changed to alanine (S171A). This allele, fbpA_MS(S171A), failed to rescue the defects of MAR1 in antibiotic resistance and colonial morphology, suggesting that the MAR1 phenotypes were due to a loss of mycolyltransferase activity (Fig. 1).

The MAR1 cell envelope had a decreased trehalose-dimycolate content.

The two principle mycolate complexes (TDM and MAMEs) were analyzed in the wild type, MAR1, and the complemented strain. Thin-layer chromatography of the products suggested a significant decrease in the TDM content of MAR1 (Fig. 4A). To confirm this result, total lipids of M. smegmatis cultures were labeled by growth in medium supplemented with ¹⁴C-radiolabeled acetate. TDM from equal amounts of labeled cells, as measured by weight, was extracted and purified in organic solvents; its identity was confirmed by IR spectroscopy (Fig. 4C) and MALDI-TOF (not shown). Scintillation counting revealed a 45% reduction of TDM in MAR1 compared to the wild type (Fig. 4B).

MAMEs derived from the mycolates covalently linked to arabinogalactan were, however, unchanged in MAR1. Thin-layer chromatography of radiolabeled extracts showed the content of α-MAME, α′-MAME, and epoxy-MAME were similar in MC²155 and MAR1 (Fig. 5). Detailed MALDI-TOF analysis of MAMEs isolated from TLC plates also showed no difference in their masses (not shown).