HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 276, Issue 28, 26430-26440, July 13, 2001

This Article Abstract Full Text (PDF) All Versions of this Article: 276/28/26430    most recent M102022200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Kremer, L. Articles by Besra, G. S. Search for Related Content PubMed PubMed Citation Articles by Kremer, L. Articles by Besra, G. S. Social Bookmarking                      What's this?

Galactan Biosynthesis in Mycobacterium tuberculosis

IDENTIFICATION OF A BIFUNCTIONAL UDP-GALACTOFURANOSYLTRANSFERASE^*

Laurent Kremer^§¶, Lynn G. Dover^§, Caroline Morehouse, Paul Hitchin^**, Martin Everett^**, Howard R. Morris, Ann Dell, Patrick J. Brennan, Michael R. McNeil, Christopher Flaherty, Ken Duncan^**, and Gurdyal S. Besra^¶¶

From the  Department of Microbiology and Immunology, University of Newcastle upon Tyne, Newcastle upon Tyne, NE2 4HH, United Kingdom,  Department of Biochemistry, Imperial College of Science, Technology, and Medicine, London, SW7 2AZ, United Kingdom, ^** GlaxoSmithKline Research and Development, Stevenage SG1 2NY, United Kingdom, and  Department of Microbiology, Colorado State University, Fort Collins, Colorado 80523-1677

Received for publication, March 6, 2001, and in revised form, April 13, 2001

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The cell wall of Mycobacterium tuberculosis and related genera is unique among prokaryotes, consisting of a covalently bound complex of mycolic acids, D-arabinan and D-galactan, which is linked to peptidoglycan via a special linkage unit consisting of Rhap-(13)-GlcNAc-P. Information concerning the biosynthesis of this entire polymer is now emerging with the promise of new drug targets against tuberculosis. Accordingly, we have developed a galactosyltransferase assay that utilizes the disaccharide neoglycolipid acceptors -D-Galf-(15)--D-Galf-O-C_10:1 and -D-Galf-(16)--D-Galf-O-C_10:1, with UDP-Gal in conjunction with isolated membranes. Chemical analysis of the subsequent reaction products established that the enzymatically synthesized products contained both -D-Galf linkages ((15) and (16)) found within the mycobacterial cell, as well as in an alternating (15) and (16) fashion consistent with the established structure of the cell wall. Furthermore, through a detailed examination of the M. tuberculosis genome, we have shown that the gene product of Rv3808c, now termed glfT, is a novel UDP-galactofuranosyltransferase. This enzyme possesses dual functionality in performing both (15) and (16) galactofuranosyltransferase reactions with the above neoglycolipid acceptors, using membranes isolated from the heterologous host Escherichia coli expressing Rv3808c. Thus, at a biochemical and genetic level, the polymerization of the galactan region of the mycolyl-arabinogalactan complex has been defined, allowing the possibility of further studies toward substrate recognition and catalysis and assay development. Ultimately, this may also lead to a more rational approach to drug design to be explored in the context of mycobacterial infections.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Despite more than four decades of effective chemotherapy, tuberculosis has re-emerged as one of the leading causes of death by killing 3 million people annually (1). Prevention efforts and control of tuberculosis is seriously hampered by the appearance of multi-drug-resistant strains of Mycobacterium tuberculosis. Therefore, new approaches to the treatment of tuberculosis are needed. Because, the mycobacterial cell wall is essential for viability, it represents a very attractive target for new anti-mycobacterial agents (2, 3).

The cell wall core is composed of a covalently linked complex of mycolic acids, D-arabinan and D-galactan, attached to peptidoglycan via an -L-Rhap-(13)--D-GlcNAc linkage unit (LU)¹ (Fig. 1) and is often referred to as the mAGP complex. Analysis by gas chromatography mass spectrometry (GC-MS) and fast atom bombardment mass spectrometry (FAB-MS) (4-7) revealed that the primary structure was composed of: 1) arabinose (Ara) and galactose (Gal) residues, which are in the furanose (f) ring form; 2) two or three arabinan chains attached to C-5 of some of the 6-linked -D-Galf glycosyl residues; 3) the galactan, consisting of a linear alternating Gal polymer of around 30 residues possessing both 5-linked -D-Galf and 6-linked -D-Galf glycosyl residues; and 4) the galactan region of arabinogalactan, linked to the C-6 of some of the N-glycoly-muramic acid residues of peptidoglycan via the LU.

View larger version (7K): [in this window] [in a new window]   Fig. 1.   Organization of the galactan region as well as its relationship to other major cell wall components. , 5, 6--D-Galf; , 5--D-Galf; , 6--D-Galf; , t--D-Galf.

Several front-line drugs are known to target essential components of the mycobacterial cell wall. For instance, isoniazid is a potent inhibitor of mycolic acid biosynthesis targeting InhA (8-10) and possibly KasA (11-13). Also, earlier studies demonstrated that administration of ethambutol led to a rapid cessation of mycolic acid transfer to the cell wall and a rapid accumulation of trehalose mono- and dimycolates (14, 15). Subsequent studies have shown that ethambutol disrupts the synthesis of the arabinan component of arabinogalactan by targeting various arabinosyltransferases, embABC (16-18).

Recent biochemical studies indicate that cell wall synthesis occurs in conjunction with mycolic acid, arabinogalactan, and peptidoglycan biosynthesis.² Initially, LU synthesis involves the transfer of GlcNAc-1-P and Rha from their respective sugar nucleotides (UDP-GlcNAc and dTDP-Rha) to endogenous polyprenol-P (probably C₅₀-P) to form the polyprenol-P-P-GlcNAc (lipid 1) and polyprenol-P-P-GlcNAc-Rha (lipid 2) precursors involved in LU synthesis (19). These glycolipid intermediates then serve as acceptors first for the sequential addition of Galf from UDP-Galf to generate polyprenol-P-P-GlcNAc-Rha-Gal_x (x is ~25-30 residues) and second for the transfer of Araf to this growing lipid intermediate polyprenol-P-P-LU-galactan using -D-arabinofuranosyl-1-monophosphoryldecaprenol (DPA) and phosphoribosyl pyrophosphate donors to afford polyprenol-P-P-GlcNAc-Rha-Gal_x-Ara_y (x is ~20-30 residues and y is 60-70 residues) (20), which at some point is mycolylated and transglycosylated to peptidoglycan.²

A number of recent studies have defined the genetics and enzymology surrounding the synthesis of nucleotide precursors involved in mAGP assembly. For instance, dTDP-Rha (rmlA (Rv0034), rmlB (Rv3464), rmlC (Rv3465), rmlD (Rv3266c), and UDP-Galf formation (UDP-GlcpUDP-Galp (galE, Rv3634) and UDP-GalpUDP-Galf (glf, Rv3809c)) have been studied in detail (21-24). Similarly, other preliminary evidence has established other open reading frames involved in LU synthesis, notably, rfe (Rv1302) as the decaprenol-monophosphate--N-acetylglucosaminyltransferase and wbbL (Rv3265c) as the rhamnosyltransferase involved in the synthesis of lipid intermediates 1 and 2 (20). Based on the findings that the UDP-GlcNAc transferase is tunicamycin-sensitive (25, 26), wbbL is an essential enzyme,³ and ethambutol and isoniazid target later steps involved in arabinan and mycolic acid biosynthesis, we suggest that the intermediate steps in mAGP synthesis, notably galactan polymerization, would represent novel drug targets. An understanding of the enzymology and genetics of the -D-(15)-Galf and -D-(16)-Galf transferases is therefore warranted.

In this study, we describe the development of a novel mycobacterial neoglycolipid-based acceptor assay for -D-(15)Galf and -D-(16)Galf transferases. Furthermore, based on this simple neoglycolipid acceptor assay and genome mining of the M. tuberculosis data base using hydrophobic cluster analysis (HCA), we describe the cloning and characterization of the gene Rv3808c, which we have termed glfT. It is responsible for mycobacterial galactan polymerization catalyzing both -D-(15)-Galf and -D-(16)-Galf transferases. Thus, GlfT represents a new cell wall drug target to be exploited in drug discovery programs leading to novel chemotherapeutics targeting M. tuberculosis.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Bacterial Strains and Growth Conditions-- All cloning steps were performed in Escherichia coli XL1-Blue (Stratagene, La Jolla, CA). Mycobacterium smegmatis mc²155 was a generous gift from W. R. Jacobs, Albert Einstein College of Medicine, Bronx, NY (27). M. smegmatis mc²155 was transformed as described previously (28), and recombinant clones were selected on Middlebrook 7H10 agar supplemented with oleic acid-albumin-dextrose-catalase enrichment (OADC; Difco, Detroit, MI) containing 25 µg/ml kanamycin (Sigma). Liquid cultures of M. smegmatis pMV261 and M. smegmatis pMV261-Rv3808c were grown at 37 °C in Luria Bertani (LB) broth medium (Difco) supplemented with 25 µg/ml kanamycin and 0.05% Tween 80. Liquid cultures of E. coli pUC8 and E. coli pUC8-Rv3808c were grown in LB broth at 37 °C with 100 µg/ml ampicillin to an A_600 nm = 0.4 and induced for 4 h with 1 mM isopropyl--D-thiogalactopyranose. Large scale cultures of bacteria, as described above, were harvested, washed, with phosphate-buffered saline, and stored at 20 °C until further use.

Plasmids and DNA Manipulation-- The E. coli-mycobacterial shuttle vector pMV261 containing the hsp60 promoter was used as described previously (29). Analysis of plasmids from mycobacteria was achieved by electroduction in E. coli as described previously (30). Restriction enzymes and T4 DNA ligase were purchased from Roche Molecular Biochemicals, and Vent DNA polymerase was purchased from New England Biolabs (Beverly, MA). All DNA manipulations were performed using standard protocols, as described by Sambrook et al. (31).

Expression of Rv3808c-- The Rv3808c open reading frame was cloned into the mycobacterial over-expression vector pMV261 as follows. PCR amplification was performed using the upstream primer P1 5'-TGA GTG AAC TCG CCG CGA GCC TGC TGT C-3' and the downstream primer P2 5'-CGA ATT CAG CCA TGC TCG GGC TCT TG-3', which contains an EcoRI restriction site (underlined). The 1918-base pair PCR product was then digested with EcoRI and cloned into the MluNI/EcoRI-restricted pMV261 giving rise to pMV261-Rv3808c. For expression in E. coli, the blunt-ended PCR fragment was cloned into a pUC8 plasmid cut by SmaI giving rise to pUC8-Rv3808c. DNA sequencing for each construct verified the coding sequence of Rv3808c as well as its junctions with the hsp60 and Plac promoters.

Synthesis of -D-Galf-(15)--D-Galf-O-C_10:1 and -D-Galf-(16)- -D-Galf-O-C_10:1 Acceptors-- The synthetic acceptors based on -D-Galf were synthesized using a combination of methods developed by Sugawara et al. (32), Veeneman et al. (33), Wolfrom et al. (34), and Zurmond et al. (35). Briefly, in the case of the -D-(15)-Galf disaccharide, the thioethyl glycoside was initially acetylated, chlorinated, and alkylated with 9-decen-1-ol using mercuric cyanide and mercuric bromide. Following deprotection and selective protection, the 5-OH position of the triprotected monosaccharide derivative was glycosylated using the activated donor chloro-tetra-O-acetyl--D-Galf, which was eventually deprotected to yield the desired disaccharide, as shown in Fig. 2A. The synthesis is convenient because it can be adapted to generate the -D-(16)-Galf disaccharide. Basically, following deprotection and selective protection/deprotection, using the dimethoxytrityl group, the 6-OH position is glycosylated using chloro-tetra-O-acetyl--D-Galf, which is then deprotected to yield the -D-(16)-Galf disaccharide as shown in Fig. 2B. The acceptor products (boxed in Fig. 2, A and B) were purified to homogeneity by silica gel chromatography and analyzed by ¹H and ¹³C nuclear magnetic resonance spectroscopy (NMR) and electrospray mass spectroscopy (ES-MS) as follows: for -D-Galf-(15)--D-Galf-O-C_10:1, melting point 103-105 °C, [_D] 42.8^o (c 1, H₂0), ¹H NMR (300 MHz, CD₃OD) 1.27-1.48 (10H, m), 1.51-1.62 (2H, m), 2.01-2.10 (2H, m), 3.40 (1H, dt, J = 9.5 and 6.5 Hz), 3.61-4.12 (13H, m), 4.81 (1H, d, J = 2Hz), 4.91 (1H, dm, J = 10.3 Hz), 4.98 (1H, dm, J = 16.9 Hz), 5.17 (1H, s), 5.81 (1H, ddt, J = 16.9, 10.3, and 6.7 Hz); ¹³C NMR (75 MHz, CD₃OD) 27.4, 30.3, 30.4, 30.6, 30.7, 30.9, 35.1, 63.0, 64.3, 69.1, 72.3, 77.3, 78.7, 78.9, 82.9, 83.5, 83.8, 84.9, 109.2, 109.4, 114.9, 140.3. m/z (ES-MS) 498.2926 (M⁺ + NH₄); and for -D-Galf-(16)--D-Galf-O-C_10:1, melting point 92-93 °C, []_D 74.8^o (c 1, H₂0), ¹H NMR (300 MHz, CD₃OD) 1.38-1.52 (10H, m), 1.64-1.71 (2H, m), 2.10-2.15 (2H, m), 3.52 (1H, dt, J = 9.5 and 6.5 Hz), 3.65 (1H, dd, J = 7.0 and 10.4 Hz), 3.72 (1H, dd, J = 4.0 and 2.7 Hz), 3.77-3.82 (1H, dt), 3.83-3.87 (1H, m), 3.89 (1H, dd, J = 4.6 Hz) 3.95-4.13 (7H, m), 4.94 (1H, d, J = 1.8 Hz), 4.98 (1H, dm, J = 12.5 Hz), 5.00 (1H, s), 5.05 (1H, dm, J = 17.1 Hz), 5.90 (1H, ddt, J = 17.1, 12.5, and 6.7 Hz); ¹³C NMR (75 MHz, CD₃OD) 27.3, 30.2, 30.3, 30.6, 30.6, 30.8, 34.9, 69.0, 64.5, 70.7, 71.1, 72.6, 78.9, 78.9, 82.9, 83.5, 84.6, 85.0, 109.3, 110.0, 114.9, 140.3.m/z (ES-MS) 498.2917 (M⁺ + NH₄).

View larger version (14K): [in this window] [in a new window]   Fig. 2.   Neoglycolipid acceptors. A, synthesis of -D-Galf-(15)--D-Galf-O-C10:1. B, synthesis of -D-Galf-(16)--D-Galf-O-C10:1. a, EtSH, concentrated HCl, 39%; b, HgO, concentrated HCl, H2O, 43%; c, Ac2O, pyr, 87%; d, Cl2, CCl4, 84%; e, HgCN2, HgBr2, MeCN, decen-1-ol, 56%; f, NH3, MeOH, 89%; g, Me2C(OMe)2, camphorsulfonic acid, DMF, 86%; h, BzCl, pyr, 91%; i, acetic acid, H2O, 76%; j, PivCl, pyr, 78%; k, HgCN2, HgBr2, MeCN, 59%; l, NH3, MeOH, 90%; m, TrCl, pyr, 72%; n, BzCl, pyr, 89%; o, acetic acid, 58%; p, HgCN2, HgBr2, MeCN, 70%; q, NH3, MeOH, 85%.

The complete synthesis and full description of all intermediates leading to -D-Galf-(15)--D-Galf-O-C_10:1 and -D-Galf-(16)--D-Galf-O-C_10:1 will be documented separately.⁴

Preparation of Membrane and Cell Wall Enzyme Fractions-- M. smegmatis pMV261, M. smegmatis pMV261-Rv3808c, E. coli pUC8, and E. coli pUC8-Rv3808c were grown as described earlier, harvested, washed with phosphate-buffered saline, and stored at 20 °C until required. Mycobacterial cells (10 g wet weight) were washed and resuspended in 30 ml of buffer A, containing 50 mM MOPS (adjusted to pH 8.0 with KOH), 5 mM -mercaptoethanol, and 10 mM MgCl₂ at 4 °C and subjected to probe sonication (Soniprep 150, MSE Sanyo Gallenkamp, Crawley, Sussex, UK; 1-cm probe) for a total time of 10 min in 60-s pulses with 90-s cooling intervals between pulses. E. coli cells were disrupted in a similar fashion using 30-s pulses and 45-s cooling intervals between pulses. The sonicates were centrifuged at 27,000 × g for 60 min at 4 °C. The resulting mycobacterial cell wall pellets were resuspended in buffer A. Percoll (Amersham Pharmacia Biotech, Uppsala, Sweden) was added to yield a 60% suspension and centrifuged at 27,000 × g for 1 h at 4 °C. The upper, particulate, diffuse cell wall enzymatically active (P60) band was collected and washed three times with buffer A and resuspended in buffer A at a final protein concentration of 10 mg/ml. Membrane fractions were obtained by centrifugation of the 27,000 × g supernatant at 100,000 × g for 1 h at 4 °C. The supernatant was carefully removed and the membranes gently resuspended in buffer A at a protein concentration of 20 mg/ml. Protein concentrations were determined using the BCA protein assay reagent kit (Pierce).

Galactosyltransferase Assay-- -D-Galf(15)--D-Galf-O-C_10:1 and -D-Galf(16)--D-Galf-O-C_10:1 (2 mM), which were stored as 100 mM ethanol stocks, were dried under a stream of argon in a microcentrifuge tube (1.5 ml), which was placed in a vacuum desiccator for 15 min to remove any residual solvent. Both acceptors were then resuspended with the remaining constituents of the galactosyltransferase assay in buffer A. The reaction mixtures for assessing [¹⁴C]Gal incorporation consisted of UDP-[U-¹⁴C]Gal (Amersham Pharmacia Biotech, 327 mCi/mmol, 0.25 µCi, 10 µl), ATP (1 mM, 5 µl), NADH (100 mM, 8 µl), membranes (250 µg, 12.5 µl), and the cell wall fraction (250 µg, 25 µl) in a final reaction volume of 80 µl. In some instances the cell wall fraction was omitted and replaced by buffer A. The reaction mixtures were then incubated at 37 °C for 1 h. A CHCl₃:CH₃OH (1:1, 533 µl) solution was then added to the incubation tubes and the entire contents centrifuged at 18,000 × g. The supernatant was recovered, dried under a stream of argon, resuspended in C₂H₅OH:H₂O (1:1, 1 ml), and loaded onto a pre-equilibrated (C₂H₅OH:H₂O (1:1)) 1-ml Whatman strong anion exchange cartridge, which was washed with 3 ml of ethanol. The eluate was dried and the resulting products partitioned between the two phases arising from a mixture of n-butanol (3 ml) and H₂0 (3 ml). The resulting organic phase was recovered following centrifugation at 3,500 × g, and the aqueous phase was again extracted twice with 3 ml of n-butanol-saturated water; the pooled extracts were back-washed twice with water saturated with n-butanol (3 ml). The n-butanol-saturated water fraction was dried and resuspended in 200 µl of n-butanol. The total cpm of radiolabeled material extractable into the n-butanol phase was measured by scintillation counting using 10% of the labeled material and 10 ml of EcoScintA (National Diagnostics, Atlanta). The incorporation of [¹⁴C]Gal was determined by subtracting counts present in control assays (either incubation of the reaction components in the absence of the acceptors or from membranes prepared from empty pMV261 or pUC8 constructs). Another 10% of the labeled material was subjected to thin-layer chromatography (TLC) in CHCl₃:CH₃OH:NH₄OH:H₂O (65:25:0.5:3.6) on aluminum-backed Silica Gel 60 F₂₅₄ plates (Merck, Darmstadt, Germany). Autoradiograms were obtained by exposing TLCs to x-ray film (Kodak X-Omat) for 4-5 days.

Analysis of Reaction Products-- Large scale reaction mixtures containing unlabeled UDP-Gal (80 mM), -D-Galf(15)--D-Galf-O-C_10:1 or -D-Galf(16)--D-Galf-O-C_10:1 (80 mM), and the other components were prepared and processed as described above. The final water-saturated n-butanol phases were dried, applied to preparative TLC plates along with radiolabeled material (50,000 cpm) to trace the cold enzymatically synthesized products, and developed in CHCl₃:CH₃OH:NH₄OH:H₂O (65:25:0.5:3.6). The plates were then sprayed with 0.01% 1,6-diphenylhexatriene in petroleum ether:acetone (9:1), and the starting material (acceptors) and products were localized under long wave (366 nm) UV light (36). The plate was then redeveloped in toluene to remove the reagent. Autoradiography was performed by exposing the TLC to x-ray film (Kodak X-Omat) for 24 h. The bands corresponding to reaction products were recovered from the plates by re-extraction with n-butanol (3 × 3 ml). The combined n-butanol phases were washed with water saturated previously with n-butanol, and the dried samples were per-O-methylated, subjected to FAB-MS and subsequent mild acid hydrolysis, per-O-ethylation, and glycosyl linkage analysis.

Chemical Derivatization for FAB-MS and GC-MS Analysis-- Per-O-methylation using the sodium hydroxide procedure was performed as described previously (37). After derivatization the reaction products were purified on a Sep-Pak C₁₈ cartridge (Waters) as described (37). Partially per-O-methylated, per-O-ethylated alditols were prepared for GC-MS linkage analysis as follows. Partial hydrolysis of the per-O-methylated sample was achieved using 2 M trifluoroacetic acid at 70 °C for 1 h. Samples were dried and then reduced with NaBH₄ (10 mg/ml) in 2 M NH₃ for 2 h at room temperature. Excess borates were removed by repeated additions (4×) of 10% acetic acid in methanol followed by evaporation. Samples were then per-O-ethylated following the per-O-methylation procedure described previously but using C₂H₅I (37).

GC-MS and FAB-MS Analysis-- GC-MS analysis was carried out on a Fisons Instruments MD800 fitted with a RTX-5 fused silica capillary column (30 m × 0.25 mm internal diameter, Restek Corp.). The partially per-O-methylated, per-O-ethylated alditol acetates were dissolved in hexanes prior to on-column injection at 65 °C. The GC oven was held at 65 °C for 1 min before being increased to 290 °C at a rate of 8 °C/min. FAB mass spectra were acquired using a ZAB-2 S.E. 2 FPD mass spectrometer fitted with a cesium ion gun operated at 30 kV. Data acquisition and processing were performed using VG Analytical Opus software. Solvents and matrices were as described (37).

Electrophoresis Methods-- M. smegmatis cells were washed in phosphate-buffered saline and disrupted by sonication (1-cm probe; Soniprep 150; MSE Ltd., Crawley, Sussex, UK) for 10 cycles of 60-s pulses with 90-s cooling intervals between pulses and fractionated as described previously. Separation of 20 µg of proteins from the membrane or the P60 cell wall fractions was carried out by SDS-10% polyacrylamide gel electrophoresis on a MiniProtean II system (Bio-Rad). Proteins were stained with Coomassie Blue R350 (Amersham Pharmacia Biotech).

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Sequence Comparisons of Various Glycosyltransferase Sequences-- Similarity between glycosyltransferases is often very low and precludes their straightforward grouping by standard sequence alignment algorithms (38, 39). The sensitive HCA method has been used successfully in several cases for the grouping of proteins of very low sequence similarity (40). This method relies upon a two-dimensional representation of protein sequences, in which hydrophobic clusters are determined and then used for sequence comparisons, thus allowing a visual comparison and detection of conserved structural features (see Fig. 3B).

View larger version (55K): [in this window] [in a new window]   Fig. 3.   Putative galactosyltransferases. A, genomic organization around glf and Rv3808c locus in M. tuberculosis. B, HCA of Rv3808c and a number of processive inverting -glycosyltransferases (from Saxena et al. (60)): AcsAB, cellulose synthase from Acetobacter xylinum (1-4; GenBankTM accession no. X54676); HasA, hyaluronan synthase from Streptococcus pyogenes (1-3, 1-4; GenBankTM accession no. L21187); NodC, nodulation factor synthesis from Azorhizobium caulinodans (1-4; GenBankTM accession no. L18897); Chs1, chitin synthase 1 from Saccharomyces cerevisiae (1-4; GenBankTM accession no. M14045); GlfT, galactofuranosyltransferase from M. tuberculosis H37Rv (1-5, 1-6; Rv3808c); Ppm1, interdomain hinge region of polyprenol monophosphomannose synthase from M. tuberculosis H37Rv (Rv2051c).

Wiggins and Munro (41) have recently shown that the amino acid DXD motif was found to be conserved in several glycosyltransferase families from both prokaryotes and eukaryotes, even though these families do not show any other obvious sequence relationships. In almost all cases, the pair of aspartic acid residues are flanked by four hydrophobic residues on the N-terminal side, with the third of these often being an aromatic residue. This motif has also been reported to be crucial for substrate recognition and/or catalytic activity of several glycosyltransferases including galactosyltransferases (41-44).

The M. tuberculosis H37Rv genome was analyzed using a combination of BLAST (45) and HCAs identifying several (Rv0539, Rv1208, Rv1500, Rv3631, Rv1518, Rv1541, Rv2957, Rv3631, Rv3782, Rv3808c) putative -glycosyltransferases. Notably, these analyses identified the gene Rv3808c, which being located immediately downstream from the UDP-Galp mutase glf (Rv3809c) (Fig. 3A), represented a strong candidate as a galactofuranosyltransferase. The first four nucleotides of Rv3808c and the last four of Rv3809c overlapped, suggesting that Rv3808c was possibly a -galactosyltransferase transcriptionally coupled to the glf gene. BLAST analysis also revealed that Rv3808c possessed homology with glycosyltransferases from Methanobacterium thermoautotrophicum, Pyrococcus abyssi, Streptococcus pneumoniae, and RfbE (Rv3782), another putative glycosyltransferase from M. tuberculosis. The HCAs presented in Fig. 3B illustrate residues and patterns of hydrophobic clustering that are highly conserved in Rv3808c and other -glycosyltransferases.

Over-expression of Rv3808c in M. smegmatis and E. coli-- To establish the relationship between the M. tuberculosis open reading frame Rv3808c and galactan polymerization, we adopted a two-step strategy. First, we over-expressed Rv3808c in M. smegmatis and E. coli. Briefly, the gene was amplified by PCR, and the PCR product was ligated into the mycobacterial expression vector pMV261 under the control of the mycobacterial hsp60 promoter, resulting in pMV261-Rv3808c. For expression in E. coli, the blunt-ended PCR fragment was cloned into a pUC8 plasmid under the control of a Plac promoter, giving rise to pUC8-Rv3808c. Subsequently, constructs (including the empty pMV261 and pUC8 plasmids) were transformed into M. smegmatis and E. coli. SDS-polyacrylamide gel electrophoresis (Fig. 4A) illustrated the levels of Rv3808c expression in M. smegmatis harboring the pMV261-Rv3808c construct, consistent with expected size of Rv3808c (68 kDa) in both the P60 and membrane fractions, which was also characteristic in terms of cellular localization of biochemical activity (19, 20). Moreover, a similar product of apparent molecular mass 68 kDa was also observed within the P60 and membrane fractions of E. coli pUC8-Rv3808c and was absent in E. coli pUC8 (Fig. 4B). Interestingly, over-expression using pET28a also led to over-expression of Rv3808c but led to the recombinant protein being recovered entirely within the cell pellet with no measurable enzymatic activity detected within the cell pellet or the cytosolic or membrane fractions (data not shown). The second part of the proposed strategy was to develop a convenient assay system to measure galactosyltransferase activities involved in mycobacterial galactan polymerization, i.e. 5-linked -D-Galf and 6-linked -D-Galf transferase activities.

View larger version (34K): [in this window] [in a new window]   Fig. 4.   Over-expression of Rv3808c in M. smegmatis (A) and E. coli (B). Crude lysates of strains carrying either the control plasmid (pMV261 and pUC8) or the Rv3808c-over-expressing plasmids were fractionated as described under "Materials and Methods." Proteins (20 µg) were then subjected to a 10% SDS-polyacrylamide gel electrophoresis and visualized by Coomassie Blue staining. The numbers on the right (A) or the left (B) margin indicate the molecular size standards, and the arrows indicate the over-expressed Rv3808c protein.

Development of a Simple Mycobacterial Galactosyltransferase Assay-- Lee et al. (46) described the synthesis of a variety of diarabinoside- and triarabinoside-based acceptors designed for the development of an in vitro mycobacterial arabinosyl transferase assay using the synthetic donor DP[1-¹⁴C]A. Following careful chemical analysis of the enzymatically synthesized products, it was confirmed that DPA was the donor for the mycobacterial enzymes DPA:arabinan (15) arabinosyltransferase and DPA:arabinan (12) arabinosyltransferase. Based on this rationale of the use of specific neoglycolipid acceptors, the -D-Galf-(15)--D-Galf-O-C_10:1 and -D-Galf-(16)--D-Galf-O-C_10:1 acceptors were synthesized as shown in Fig. 2, A and B, respectively, corresponding to the two major structural motifs found within the galactan of arabinogalactan. Assays performed in the presence of membranes and the cell wall enzymatic fraction P60 resulted in excellent [¹⁴C]Galf incorporation from UDP-[¹⁴C]Galp following endogenous conversion to UDP-[¹⁴C]Galf and transferase activity for both -D-Galf-(15)--D-Galf-O-C_10:1 and -D-Galf-(16)--D-Galf-O-C_10:1 acceptors. A concentration of 4 mM of both acceptors resulted in maximum galactosyltransferase activity with concentrations of >10 mM leading to significant inhibition of the [¹⁴C]Galf transferase activity, presumably because of the detergent-like properties of the acceptor adversely affecting enzymatic activity at these higher concentrations. Typically, -D-Galf-(15)--D-Galf-O-C_10:1, which behaved as a poorer substrate for UDP-[¹⁴C]Galf and the respective galactosyltransferase(s), yielded 16,000-30,000 cpm/assay, whereas the more efficient -D-Galf-(16)--D-Galf-O-C_10:1 acceptor afforded 50,000-80,000 cpm/assay. A key feature of the assay appeared to be the inclusion of NADH, which when omitted resulted in a deleterious effect. This effect has recently been attributed a co-factor for the UDP-[¹⁴C]Galp mutase glf gene (47). Interestingly, assays performed with P60 alone resulted in very poor [¹⁴C]Galf incorporation using both acceptors; however, it provided a synergistic effect (0.5-fold increase) when added with membranes as compared with membranes alone (data not shown). This is attributable to the higher specific activity observed for UDP-Galp (glf) mutase activity within P60 preparations, resulting in a greater pool of UDP-Galf for the subsequent galactosyltransferase(s). As a consequence, assays were always supplemented with P60 and NADH.

TLC/autoradiography clearly demonstrated the enzymatic conversion of both the disaccharide acceptors to their corresponding trisaccharide products (Fig. 5A, lane 2, -D-Galf-(16)--D-Galf-(15)--D-Galf-O-C_10:1, and lane 3, -D-Galf-(15)--D-Galf-(16)--D-Galf-O-C_10:1). The -D-Galf-(16)--D-Galf-O-C_10:1 gave rise to a second, slower migrating band, which, based on relative migration profiles, would be anticipated to be a tetrasaccharide product (-D-Galf-(16)--D-Galf-(15)--D-Galf-(16)--D-Galf-O-C_10:1) resulting from further elongation of the trisaccharide precursor (Fig. 5A, lane 3, -D-Galf-(15)--D-Galf-(16)--D-Galf-O-C_10:1). It is clear from the absence of radioactivity in the control assay (Fig. 5A, lane 1) that the disposable anion exchange cartridge, followed by the water/n-butanol partitioning steps successfully removed any unused UDP-[¹⁴C]Gal, [¹⁴C]Gal and other polyprenol-P-based lipid precursors involved in arabinogalactan biosynthesis. Another advantage of the partitioning steps was the removal of any salts, which would otherwise hinder the resolution of the enzymatically synthesized products.

View larger version (25K): [in this window] [in a new window]   Fig. 5.   An autoradiogram of reaction products produced through inclusion of acceptors -D-Galf-(15)--D-Galf-O-C10:1 and -D-Galf-(16)--D-Galf-O-C10:1 at 4 mM with mycobacterial membrane preparations and UDP-[14C]Galp. A, membranes prepared from M. smegmatis carrying pMV261. Lane 1, control (no acceptor); lane 2, -D-Galf-(15)--D-Galf-O-C10:1 [G5G]; lane 3, -D-Galf-(16)--D-Galf-O-C10:1 [G6G]. B, membranes prepared from M. smegmatis carrying either pMV261 or pMV261-Rv3808c. Lane 1, pMV261 -D-Galf-(15)--D-Galf-O-C10:1 [G5G]; lane 2, pMV261-Rv3808c, -D-Galf-(15)--D-Galf-O-C10:1 [G5G]; lane 3, pMV261, -D-Galf-(16)--D-Galf-O-C10:1 [G6G]; lane 4, pMV261-Rv3808c, -D-Galf-(16)--D-Galf-O-C10:1 [G6G]; lane 5, pMV261; lane 6, pMV261-Rv3808c. TLC/autoradiography was performed using CHCl3:CH3OH:NH4OH:H2O (65:25:0.5:3.6), and products were revealed through exposure to Kodak X-Omat film for 4 days.

Mycobacterial Galactosyltransferase Assays and M. smegmatis pMV261-Rv3808c and E. coli pUC8-Rv3808c-- M. smegmatis transformed with pMV261-Rv3808c or empty pMV261 was examined for galactosyltransferase activity using the -Galf-(15)--D-Galf-O-C_10:1 and -D-Galf-(16)--D-Galf-O-C_10:1 neoglycolipid acceptor cell-free assay described above. Analysis of the reaction products by TLC/autoradiography (Fig. 5B) clearly indicate that the overproduction of Rv3808c in the recombinant M. smegmatis strain resulted in a higher overall incorporation of [¹⁴C]Gal from the nucleotide precursor into both acceptors in comparison to the empty pMV261 strain. The level of activity varied between preparations, but pMV261-Rv3808c consistently enhanced activity by 50-70% in comparison with the empty pMV261 plasmid (data not shown). Interestingly, the formation of the resulting trisaccharide product from both acceptors, -D-Galf-(15)--D-Galf-O-C_10:1 and -D-Galf-(16)--D-Galf-O-C_10:1 was increased (Fig. 5B, lane 2) as was the tetrasaccharide product for -D-Galf-(16)--D-Galf-O-C_10:1 (Fig. 5B, lane 4).

This increase was surprising because we anticipated that Rv3808c would encode a single -D-galactosyltransferase (5-linked or 6-linked). The level of [¹⁴C]Gal incorporation and TLC profiles suggests three possibilities. First, Rv3808c could encode for a regulatory protein, and increased expression would lead to enhanced activity for both acceptors as observed. Second, the acceptors could possess very poor specificity and both recognize either 5-linked -D-Galf or 6-linked -D-Galf transferases. An alternative and more attractive scenario would be that Rv3808c is a bifunctional enzyme and performs both 5-linked -D-Galf and 6-linked -D-Galf activities. This is an hypothesis that could be determined first by expression of enzymatically active Rv3808c in a heterologous host, such as E. coli, and then by isolation of the corresponding products and complete chemical characterization to determine their structures.

E. coli transformed with pUC8-Rv3808c or pUC8 was examined for galactosyltransferase activity using the -D-Galf-(15)--D-Galf-O-C_10:1 and -D-Galf-(16)--D-Galf-O-C_10:1 neoglycolipid acceptor cell-free assay described above. Again TLC/autoradiography of the reaction products demonstrated that the control assays (without acceptor) are devoid of any radiolabeled products, illustrating the efficiency of the strong anion exchange columns (Fig. 6, lanes 1 and 2). Interestingly, the control E. coli pUC8 (Fig. 6, lane 3) possessed a minor product, presumably a trisaccharide, in relation to -D-Galf-(15)--D-Galf-O-C_10:1 and a series of products, possibly tri-, tetra-, and pentasaccharide with -D-Galf-(16)--D-Galf-O-C_10:1 (Fig. 6, lane 5). This would imply that the acceptors are recognized by the endogenous E. coli galactosyltransferases, which utilize UDP-Galf in O-antigen biosynthesis, even though the acceptors do not conform to a motif present in the O-antigen side chain (48). However, it is clear from TLC/autoradiography (Fig. 6, lanes 4 and 6) that membrane preparations from E. coli transformed with pUC8-Rv3808c and in conjunction with both of the neoglycolipid acceptors produced profiles distinct from E. coli pUC8 but characteristic and similar to M. smegmatis and M. smegmatis pMV261-Rv3808c. These results obtained in a heterologous host eliminated the possibility that Rv3808c encoded a regulatory protein and suggested that Rv3808c was in fact a galactosyltransferase transferring [¹⁴C]Gal from UDP-[¹⁴C]Galf to both acceptors. The issue of which linkages, 5- or 6-linked, were formed in these assays was established through detailed product characterization (see below).

View larger version (35K): [in this window] [in a new window]   Fig. 6.   An autoradiogram of reaction products produced through inclusion of acceptors -D-Galf-(15)--D-Galf-O-C10:1 and -D-Galf-(16)--D-Galf-O-C10:1 at 4 mM with E. coli membrane preparations and UDP-[14C]Gal. Membranes prepared from E. coli pUC8 and E. coli pUC8-Rv3808c. Lane 1, pUC8, no acceptor; lane 2, pUC8-Rv3808c, no acceptor; lane 3, pUC8, -D-Galf-(15)--D-Galf-O-C10:1 (G5G); lane 4, pUC8-Rv3808c, -D-Galf-(15)--D-Galf-O-C10:1 (G5G); lane 5, pUC8, -D-Galf-(16)--D-Galf-O-C10:1 (G6G); lane 6, pUC8-Rv3808c, -D-Galf-(16)--D-Galf-O-C10:1 (G6G). TLC/autoradiography was performed using CHCl3:CH3OH:NH4OH:H2O (65:25:0.5:3.6), and products were revealed through exposure to Kodak X-Omat film for 5 days.

Chemical Analysis of Reaction Products-- The newly synthesized products resulting from Rv3808c expression were further characterized from assays containing both acceptors -D-Galf-(15)--D-Galf-O-C_10:1 (resulting trisaccharide ?) and -D-Galf-(16)--D-Galf-O-C_10:1 (resulting trisaccharide and tetrasaccharide ?) and cold UDP-Gal. The per-O-methylated -D-Galf-(15)--D-Galf-O-C_10:1 acceptor revealed strong signals at m/z 601 and 709 in Fig. 7A, corresponding to the sodiated molecular ion, and an adduct with the monothioglycerol matrix giving rise to a pseudomolecular ion, respectively. The corresponding trisaccharide product Fig. 7B shows the sodiated molecular ion for the trisaccharide product at m/z 805 and the matrix-derived pseudomolecular ion at m/z 913. The signal at m/z 601 in Fig. 8A is the sodiated molecular ion for the per-O-methylated -D-Galf-(16)--D-Galf-O-C_10:1 acceptor. The sodiated per-O-methylated trisaccharide product is shown in Fig. 8B at m/z 805, with the matrix-derived pseudomolecular ion 108 mass units higher at m/z 913. Fig. 8C shows the per-O-methylated tetrasaccharide product at m/z 1009 and the matrix-derived pseudomolecular ion at m/z 1117. The FAB-MS analysis of the per-O-methylated products provided direct evidence for the addition of an extra hexosyl unit to -D-Galf-(15)--D-Galf-O-C_10:1 with the sequential addition of two such units giving rise to a tri- and tetrasaccharide products in relation to -D-Galf-(16)--D-Galf-O-C_10:1. Further evidence for the addition of new Galf units to both acceptors was sought through glycosyl linkage analysis. Initial GC-MS analysis of the partially per-O-methylated, per-O-acetylated alditol acetates produced non-stoichiometric amounts of each linkage, thus hindering their assignment. A second strategy was adopted where the per-O-methylated products were partially hydrolyzed with acid, reduced and per-O-ethylated (PMAE) according to Fig. 9, and analyzed by GC-MS to determine what new glycosyl linkages were catalyzed by Rv3808c. This strategy in combination with high performance liquid chromatography fractionation of PMAE cleavage products was used previously to establish the structure of mycobacterial arabinogalactan (4). The purified PMAE cleavage products were subsequently hydrolyzed, reduced, and per-O-acetylated, and following GC-MS analysis, they provided a finger-print profile in terms of relative retention times and fragmentation ions characteristic of individual PMAE cleavage products (4). In this study, the -D-Galf-(15)--D-Galf-O-C_10:1 and -D-Galf-(16)--D-Galf-O-C_10:1 acceptors generated, as expected, a terminal 15-linked PMAE cleavage product (Table I, retention time 26.92 min, cleavage product A [t-Gal1-5Gal-ol] and a terminal 16-linked PMAE cleavage product (Table I, retention time 28.14 min, cleavage product B [t-Gal1-6Gal-ol]), respectively, consistent with our previous studies based on retention times of these PMAE fragments (4). The enzymatically synthesized trisaccharide product from the -D-Galf-(15)--D-Galf-O-C_10:1 acceptor produced two cleavage products (Table I). Cleavage product C [6Eth-Gal1-5Gal-ol], with a retention time of 27.00 min, corresponded to an internal cleavage product, which based on the per-O-ethylation pattern suggested that the new Galf unit was 6-linked (Table I). More importantly, the cleavage product at 27.99 min based on retention time corresponded to a terminal 16-linked PMAE cleavage product [t-Gal1-6Gal-ol] implying a new 16 linkage. The two cleavage products together provide the complete structure of the enzymatically synthesized product as -D-Galf-(16)--D-Galf-(15)--D-Galf-O-C_10:1. The enzymatically synthesized trisaccharide product from the -D-Galf-(16)--D-Galf-O-C_10:1 acceptor also produced two cleavage products (Table I) with retention times of 28.24 min corresponding to the internal cleavage product D [5Eth-Gal1-6Gal-ol] suggesting that the new Galf unit was 5-linked and, more importantly, cleavage product A at 26.80 min, which corresponds to a terminal 15-linked PMAE cleavage product [t-Gal1-5-Gal-ol], thus implying a new 15 linkage. The two cleavage products together provide the complete structure of the enzymatically synthesized product as -D-Galf-(15)--D-Galf-(16)--D-Galf-O-C_10:1. A similar analysis of PMAE cleavage products produced with the enzymatically synthesized tetrasaccharide product from the -D-Galf-(16)--D-Galf-O-C_10:1 acceptor produced PMAE cleavage products B [t-Gal1-6Gal-ol, retention time 28.09 min], C [6Eth-Gal1-5Gal-ol], retention time 27.32 min, and D [5Eth-Gal1-6Gal-ol], retention time 28.24 min (Table I), thus establishing the identity of the enzymatically synthesized tetrasaccharide product as -D-Galf-(16)--D-Galf-(15)--D-Galf-(16)--D-Galf-O-C_10:1.

View larger version (30K): [in this window] [in a new window]   Fig. 7.   FAB-MS analysis of per-O-methylated -D-Galf-(15)--D-Galf-O-C10:1 acceptor (A) and the putative trisaccharide product (B). A, the signal at m/z 687 is the protonated pseudomolecular ion. The signals at m/z 731 and 815 are matrix derived. B, the signals at m/z 783 and 891 are the protonated molecular ion and pseudomolecular ion, respectively. The signals at m/z 579 and 687 are the protonated molecular ion and pseudomolecular ion of the starting acceptor. Signals marked with an X are solvent impurities.

View larger version (24K): [in this window] [in a new window]   Fig. 8.   FAB-MS analysis of per-O-methylated -D-Galf-(16)--D-Galf-O-C10:1 acceptor (A) and the putative trisaccharide product (B) and tetrasaccharide product (C). The signal at m/z 731 in panel A is matrix-derived. Signals marked with an X are solvent impurities.

View larger version (20K): [in this window] [in a new window]   Fig. 9.   An illustration of the sequence of reactions used to produce partially per-O-methylated, reduced, and partially per-O-ethylated monoglycosyl alditols.

View larger version (28K): [in this window] [in a new window]   Fig. 10.   Cleavage products.

Thus, the chemical analysis is consistent with the galactosyltransferase activity encoded by Rv3808c, alternating between 15 and 16 linkages. In summary, the biochemical evidence shows that Rv3808c, which we have now termed GlfT, is a bifunctional enzyme that catalyzes the polymerization of the galactan region of mAGP through -D-(15) and -D-(16)-galactofuranosyltransferases activities.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The D-galactan region of arabinogalactan from M. tuberculosis is based upon a backbone structure of repeat units of 5)--D-Galf-(16)--D-Galf-(1. Neoglycolipid glycosyltransferase acceptors have been utilized in a variety of glycosyltransferase assays and in particular in studies related to arabinan biosynthesis in M. tuberculosis (46). In this current report a novel galactosyltransferase assay was developed that utilized synthetic neoglycolipid acceptors -D-Galf-(15)--D-Galf-O-C_10:1 and -D-Galf-(16)--D-Galf-O-C_10:1, with UDP-Galf as the Galf donor presented in the form of UDP-Galp and UDP-Galp mutase (the glf gene product) and isolated membrane preparations from M. smegmatis and E. coli. Chemical analysis of the reaction products demonstrated that the -D-Galf-(15)--D-Galf-O-C_10:1 acceptor resulted in a trisaccharide product (-D-Galf-(16)--D-Galf-(15)--D-Galf-O-C_10:1), whereas the -D-Galf-(16)--D-Galf-O-C_10:1 acceptor yielded a trisaccharide (-D-Galf-(15)--D-Galf-(16)--D-Galf-O-C_10:1) and a tetrasaccharide (-D-Galf-(16)--D-Galf-(15)--D-Galf-(16)--D-Galf-O-C_10:1) consistent with the expected alternating linkage profile of the galactan region of arabinogalactan.

The "one enzyme, one linkage" hypothesis of Hagopian and Eylar (49) was one of the early dogmas of glycobiology. However, several enzymes consisting of a single polypeptide chain have been demonstrated to catalyze the transfer of a least two distinct sugars. For instance, work on the E. coli K5 enzyme KfiC that synthesizes t