Initiation of Methylglucose Lipopolysaccharide Biosynthesis in Mycobacteria

Background Mycobacteria produce two unique families of cytoplasmic polymethylated polysaccharides - the methylglucose lipopolysaccharides (MGLPs) and the methylmannose polysaccharides (MMPs) - the physiological functions of which are still poorly defined. Towards defining the roles of these polysaccharides in mycobacterial physiology, we generated knock-out mutations of genes in their putative biosynthetic pathways. Methodology/Principal Findings We report here on the characterization of the Rv1208 protein of Mycobacterium tuberculosis and its ortholog in Mycobacterium smegmatis (MSMEG_5084) as the enzymes responsible for the transfer of the first glucose residue of MGLPs. Disruption of MSMEG_5084 in M. smegmatis resulted in a dramatic decrease in MGLP synthesis directly attributable to the almost complete abolition of glucosyl-3-phosphoglycerate synthase activity in this strain. Synthesis of MGLPs in the mutant was restored upon complementation with wild-type copies of the Rv1208 gene from M. tuberculosis or MSMEG_5084 from M. smegmatis. Conclusions/Significance This is the first evidence linking Rv1208 to MGLP biosynthesis. Thus, the first step in the initiation of MGLP biosynthesis in mycobacteria has been defined, and subsequent steps can be inferred.


Introduction
Mycobacteria produce two cytoplasmic polymethylated polysaccharides (PMPS) of intermediate size in which many of the sugar units are partially O-methylated, thus conferring on the molecules a slight hydrophobicity [1]. One class is known as the 3-O-methylmannose polysaccharides (MMPs) [2][3] and the other as the 6-O-methylglucose lipopolysaccharides (MGLPs) (Fig. 1) [4][5]. The ability of both PMPS to form in vitro stable complexes with medium-and long-chain fatty acyl chains and acyl-CoAs and to regulate the activity of the fatty acid synthase I (FAS-I) has led to the hypothesis that these polysaccharides are important regulators of fatty acid and mycolic acid metabolism in mycobacteria [6][7][8]. These findings, however, have been derived from in vitro studies using enzyme assays and purified substrates and whether they accurately reflect the physiological function of PMPS in whole cells is not clear. With the ultimate goal of studying the physiological roles of MMPs and MGLPs, we have begun to define their biosynthetic pathways and to generate mycobacterial mutants deficient in different aspects of their biosynthesis [1,9].
Our recent evidence indicates that two clusters of genes are likely to participate in the biosynthesis of MGLPs in M. tuberculosis. One encompasses Rv3030-Rv3037c in the genome of M. tuberculosis H37Rv [10] and carries the a-(1R4)-glucosyltransferase gene (Rv3032) responsible for the elongation of MGLPs [9], a putative acetyltransferase gene (Rv3034c), two putative SAM-dependentmethyltransferase genes (Rv3030 and Rv3037c) and a potential branching gene (Rv3031) likely to be involved in the formation of the a-(1R6)-glycosidic bond linking the first and second D-Glcp residues at the reducing end of the molecule. The finding that a M. tuberculosis H37Rv knock-out mutant deficient in the Rv3032 enzyme still produced residual amounts of MGLPs then led us to identify Rv1212c as the likely compensatory a-(1R4)-glucosyltransferase [11]. Failure to disrupt both the Rv3032 and Rv1212c genes in the same M. tuberculosis H37Rv strain further indicated that bacterial growth required at least one of these two genes to be functional. Whether this physiological requirement is particularly related to the synthesis of glycogen, capsular glucan or MGLPs has not yet been elucidated, since Rv3032 and Rv1212c appear to participate in the elongation of all three molecules and to partially compensate for one another [11]. Interestingly, Rv1212c also belongs to a cluster of genes (Rv1208-Rv1213) encoding putative sugar-modifying enzymes [1,11]. The existence of a putative retaining glycosyltransferase of the recently established CAZy GT-81 family (http://www.cazy.org/), Rv1208, showing weak sequence similarities with the glucosyl-3-phosphoglycerate synthase (GpgS) from Persephonella marina (,24% amino acid identity) in the vicinity of Rv1212c suggested that Rv1208 might catalyze the first glucosyl transfer in MGLP biosynthesis ( Fig. 1) [12]. Supporting this assumption, recombinant forms of the orthologs of Rv1208 from Mycobacterium bovis BCG and M. smegmatis have been shown to display GpgS activity in vitro [13]. Moreover, the three-dimensional structures of Rv1208 and its ortholog in M. avium subsp.
paratuberculosis, MAP2569c, in their apo forms and in complex with UDP, UDP-glucose, and both UDP and D-3-phosphoglycerate have been solved allowing the classification of these enzymes as GT-A-type glycosyltransferases and the molecular determinants for substrate recognition and catalysis to be established [14][15]. Direct evidence linking this enzyme to the biogenesis of MGLPs in mycobacteria was, however, lacking. Rv1208 was predicted to be an essential gene of M. tuberculosis by high-density mutagenesis [16]. We thus undertook to analyze the effects of disrupting MSMEG_5084, the ortholog of Rv1208 in M. smegmatis, on the GpgS activity and MGLP synthesis of this bacterium. An assay was developed which allowed the formation of the early precursors of MGLPs to be monitored in mycobacterial cell-free extracts for the first time.

Results and Discussion
Effects of knocking-out MSMEG_5084 on the biosynthesis of MGLP in M. smegmatis The ortholog of the Rv1208 gene in M. smegmatis mc 2 155, MSMEG_5084, was disrupted by homologous recombination using standard protocols [17]. The product of MSMEG_5084 (303 amino acids) shares 74% identity (84% similarity) with its M. tuberculosis counterpart (324 amino acids) on a 302 amino acid overlap. Allelic replacement at the MSMEG_5084 locus was confirmed by PCR and Southern hybridization ( Fig. 2 and data not shown). Complemented mutant strains were obtained by transforming mc 2 DMSMEG_5084 either with pVV2Rv1208, expressing a wild-type copy of Rv1208 from M. tuberculosis H37Rv or pVV16MSMEG_5084, expressing a wild-type copy of MSMEG_5084 from M. smegmatis mc 2 155.
As compared to its wild-type parent M. smegmatis mc 2 155, the MSMEG_5084 mutant displayed a significantly reduced growth rate at 37uC. Mutant colonies typically appeared 3 to 4 days later than wild-type colonies on 7H11-OADC, and 1 to 2 days later than mc 2 155/pVV16MSMEG_5084 colonies. Thus, growth was partially restored in the complemented mutants.
Analysis of the MGLPs produced by different culture batches of the wild-type and mutant strains metabolically labeled with [ 14 Cmethyl]-L-methionine [9] revealed a virtual elimination of the de novo production of these molecules in the mutant (Fig. 3). MGLP production was restored in the mutant complemented with wildtype copies of either Rv1208 or MSMEG_5084 indicating that the two genes are functional orthologs (Fig. 3). The total amount of radioactivity incorporated into the MGLPs of the mutant was only about 20% of that for the wild-type parent. MMP synthesis, in contrast, appeared relatively unaffected in mc 2 DMSMEG_5084 indicating that a deficiency in MGLP production does not affect other PMPS (Fig. 3). These results clearly confirm the primary role of Rv1208 in the initiation of MGLP synthesis, however, since residual amounts of MGLPs were still produced in the mutant, we conclude that another enzyme displaying glucosyl-3-phosphoglycerate synthase or glucosylglycerate synthase activity contributed to the pool.
To compare the structures of the MGLPs produced by the wildtype, mutant and complemented mutant strains, they were purified by reverse phase chromatography, deacylated, peracetylated and analyzed by MALDI-TOF mass spectrometry. The mass spectrum of the MGP fraction of wild-type M. smegmatis mc 2 155 mainly showed four series of ions separated by 260 mass units and differing by 28 and 16 mass units attributable to the peracetylation of the samples (i.e., replacement of a methyl group by an acetyl group) and potassium adducts of the pseudomolecular ions, respectively (Fig. 4). The observed clustered ions thus reflect the variability of MGPs in terms of their degree of glycosylation, Omethylation and salt (sodium, potassium) adducts. In the wild-type strain, the two most intense ions at m/z 5253 and 5513 were assigned, respectively, to the [M-H+2Na] + pseudomolecular ions of MGP 19 20,13 were detectable in the mutant strain, the abundance of these was clearly much less than in the wild-type strain (Fig. 4). Thus, disrupting MSMEG_5084 in M. smegmatis resulted primarily in simple, less polymerized structures, and also diminished yields. Complementation of the mutant strain with pVV16MSMEG_5084 partially restored the synthesis of the mature forms of MGPs (MGP 19,12 , MGP 20,12 ) in the cells (Fig. 4).
No qualitative or quantitative differences were found between the mutant and wild-type strains in terms of fatty acid and mycolic acid contents (Fig. S1). Thus, altered biosynthetic patterns of MGLPs had no significant impact on fatty acid metabolism. This result is consistent with observations reported by Dr. Ballou's laboratory on a spontaneous mutant of M. smegmatis with defects in MMP and MGLP synthesis [18] and our own observations on a Rv3032 knock-out mutant of M. tuberculosis H37Rv and a MSMEG_2350 knock-out mutant of M. smegmatis mc 2 155, both of which were found to be significantly impaired in MGLP biosynthesis but not in that of fatty acids and mycolates [9]. Although the presence of MMPs (which are thought to play similar regulatory functions as MGLPs on fatty acid synthesis) in

Effects of knocking-out MSMEG_5084 on the biosynthesis of MGLP precursors by M. smegmatis cellfree extracts
To directly correlate Rv1208 and glucosyl-3-phosphoglycerate synthase activity to the biosynthesis of MGLPs in mycobacterial cells, whole cell lysates prepared from the wild-type and mutant strains of M. smegmatis provided enzyme sources in assays aimed at monitoring the formation of glucosylglycerate and MGLP precursors in vitro. UDP-D-[U-14 C]Glc and D-3-phosphoglycerate served as the donor and acceptor substrates, respectively. Timedependent formation of glucosyl-3-phosphoglycerate (GPG) was clearly visible in the cell-free extracts of wild-type M. smegmatis (Fig. 5). As expected, this product was progressively dephosphorylated by an unknown endogenous phosphatase, GlgP (Fig. 1), to yield glucosylglycerate (GG) (Fig. 5). In contrast, barely detectable amounts of GPG and no GG were detected in the assays using mc 2 DMSMEG_5084 extracts (Fig. 5), even after prolonged incubation times.
In conclusion, the disruption of MSMEG_5084 in M. smegmatis results in an 80% decrease in the production of MGLPs directly attributable to a drastic if not complete loss of GPG synthesis. The dispensability of the MSMEG_5084 gene for the growth of M. smegmatis while its ortholog in M. tuberculosis H37Rv, Rv1208, is predicted to be essential may be accounted for by the existence of residual amounts of MGLPs in mc 2 DMSMEG_5084 or by the production by M. smegmatis mc 2 155 but not M. tuberculosis H37Rv of MMPs thought to display the same physiological functions as MGLPs [1]. Altogether, our results indicate that Rv1208 and its ortholog in M. smegmatis encode glucosyl-3-phosphoglycerate synthases involved in the transfer of the first glucosyl residue of MGLPs. Results also confirm the participation of two gene clusters in the biosynthesis of these lipopolysaccharides in the genomes of mycobacteria. The existence of endogenous enzymes with weak glucosyl-3-phosphoglycerate synthase or glucosylglycerate synthase activity probably account for the residual synthesis of MGLPs in the MSMEG_5084 mutant of M. smegmatis.
For complementation studies, the entire coding sequence of Rv1208 was PCR-amplified from M. tuberculosis H37Rv genomic DNA using the primers 59-tataacatatgacagcatcggagctggtc-39 and 59-tataaaagcttcagcgcggccgcatcac-39 and cloned into the NdeI and HindIII restriction sites of the expression vector pVV2, yielding pVV2Rv1208. pVV2 is a shuttle E. coli/Mycobacterium plasmid derived from pMV261 [20]. It harbors kanamycin and hygromycin-resistance markers and allows the constitutive production of N-terminal His 6 -tagged proteins in mycobacteria under the control of the phsp60 promoter. MSMEG_5084 was PCR- amplified using the primers smeg1208.1 and smg1208.2 described above and cloned into the blunted SpeI site of pVV16 [19] for expression from its own promoter.

Purification and analysis of MGLPs
MGLPs were extracted and purified from cold and radiolabeled M. smegmatis Sauton's cultures as previously described [9]. TLC analyses were performed on aluminum-backed silica gel 60precoated plates F 254 (E. Merck, Darmstadt, Germany) using chloroform/methanol/water (56:38:10, by vol.) as the eluent. Radiolabeled MGLPs were visualized by exposure of TLC plates to Kodak X-Omat AR films at 280uC. MGPs were obtained by deacylating dry MGLPs in 1 M NaOH at 37uC for 3 hr as described [9]. Samples (5-10 mg) were peracetylated with 100 ml acetic anhydride in the presence of 100 ml of pyridine for 1 hr at 110uC. After removal of the solvent, peracetylated MGPs were extracted three times with chloroform/water (1:1). The chloroform phases were pooled, washed three times with water, and the dried residue dissolved in 10 ml chloroform prior to Matrix-Assisted Laser Desorption-Ionization Time  Reaction mixtures were incubated at 37uC for 5 to 30 min and terminated by cooling in ethanol/dry ice. Reaction products were separated by TLC on aluminum-backed silica gel 60precoated plates F 254 developed in 1-propanol/ethyl acetate/ water/25% ammonia (50:10:30:10 by vol.) and revealed by autoradiography. The products of the reactions were characterized by co-migration with authentic glucosylglycerate (GG) and glucosyl-3phosphoglycerate (GPG) standards produced in vitro by M. smegmatis mc 2 155 cell free extracts. GPG was purified by preparative TLC and structurally characterized using Electro-Spray Ionisation mass spectrometry (ESI/MS) on a 6220 TOF (Agilent Technologies) in the negative ion mode (Fig. S2). GG was produced by treating purified GPG with 4 U alkaline phosphatase (Sigma) for 30 min at 37uC.

Fatty acid and mycolic acid analysis
Fatty acid and mycolic acid methyl esters were prepared from whole M. smegmatis cells by methanolysis using methanolic-HCl (Supelco). Mycolic acid methyl esters were analyzed by TLC using n-hexane/ethyl acetate (95:5; three developments) as the eluent. Fatty acid methyl esters were analyzed by gas chromatographymass spectrometry (GC-MS) on a Varian CP-3800 gas chromatograph equipped with a Varian 320-MS TQ mass spectrometer using a 5% phenyl-methyl low bleed Factor Four GC column operating at a temperature of 50uC for 1 min followed by programmed increases of 30uC per min to 100uC and 10uC per min to 300uC. Figure S1 Comparative analysis of the fatty acid and mycolic acid compositions of wild-type M. smegmatis mc2155 and mc2(delta)MSMEG_5084. Wild-type mc2155 and mc2(delta)MS-MEG_5084 were grown in Sauton's medium as surface pellicles at 37uC. A) Mycolic acid methyl esters (MAMEs) were analyzed by TLC using n-hexane/ethyl acetate (95:5; three developments) as the eluent and revealed by charring with cupric sulfate (10% in a 8% phosphoric acid solution); B) Fatty acid methyl esters (FAMEs) were analyzed by gas chromatography-mass spectrometry. Shown are the relative percentages of each fatty acid in the strains.