Thymidine Diphosphate-6-deoxy-l-lyxo-4-hexulose Reductase Synthesizing dTDP-6-deoxy-l-talose fromActinobacillus actinomycetemcomitans *

The serotype c-specific polysaccharide antigen ofActinobacillus actinomycetemcomitans NCTC 9710 contains an unusual sugar, 6-deoxy-l-talose, which has been identified as a constituent of cell wall components in some bacteria. Two genes coding for thymidine diphosphate (dTDP)-6-deoxy-l-lyxo-4-hexulose reductases were identified in the gene cluster required for biosynthesis of serotype c-specific polysaccharide. Both dTDP-6-deoxy-l-lyxo-4-hexulose reductases were overproduced and purified from Escherichia coli transformed with the plasmids containing these genes. The sugar nucleotides converted by both reductases were purified by reversed-phase high performance liquid chromatography and identified by 1H nuclear magnetic resonance and gas-liquid chromatography. The results indicated that one of two reductases produced dTDP-6-deoxy-l-talose and the other produced dTDP-l-rhamnose (dTDP-6-deoxy-l-mannose). The amino acid sequence of the dTDP-6-deoxy-l-lyxo-4-hexulose reductase forming dTDP-6-deoxy-l-talose shared only weak homology with that forming dTDP-l-rhamnose, despite the fact that these two enzymes catalyze the reduction of the same substrate and the products are determined by the stereospecificity of the reductase activity. Neither the gene for dTDP-6-deoxy-l-talose biosynthesis nor its corresponding protein product has been found in other bacteria; this biosynthetic pathway is identified here for the first time.

the O-chain of lipopolysaccharide from Rhizobium loti NZP 2213 has been reported as a homopolymer composed solely of 6-deoxy-L-talose (7). In addition, other serotype-specific polysaccharides of A. actinomycetemcomitans also contain rare sugars as constituents of microbial polysaccharides; examples include D-fucose in serotype b-specific polysaccharide (8) and 6-deoxy-D-talose in serotype a-specific polysaccharide (2).
In 1973, 6-deoxy-L-talose was characterized as an unusual sugar, and the instability of dTDP-6-deoxy-L-talose, which is the activated nucleotide sugar form of 6-deoxy-L-talose, was reported (9). The enzymatic activity of the biosynthetic pathway of this sugar nucleotide in a cell-free extract of Escherichia coli O45 was also characterized in that report. Since then, no reports on either the isolation of the enzymes or the identification of the genes involved in the biosynthesis of dTDP-6-deoxy-L-talose have been published.
§ To whom correspondence and reprint requests should be addressed. Ϫ mB Ϫ )gal) was used as a host strain for pGEX-6P derivatives (Amersham Pharmacia Biotech). E. coli strains were grown aerobically in 2 ϫ TY broth at 37°C (13). Ampicillin and chloramphenicol were used at final concentrations of 50 g/ml and 20 g/ml, respectively.
DNA Manipulations, PCR, 1 and Sequencing Techniques-DNA fragment preparation, agarose gel electrophoresis, DNA labeling, ligation, and bacterial transformation were performed using the methods described by Sambrook et al. (13). PCR amplification was performed using a GeneAmp PCR System 2400 from Perkin-Elmer. Sequencing was performed using an ABI 373 DNA sequencing apparatus (Perkin-Elmer).
Enzyme Purifications-To purify the fusion proteins produced by pTYB derivatives, 200-ml cultures of E. coli harboring the expression plasmids were grown at 30°C to an optical density at 600 nm of 0.7. IPTG was added to a final concentration of 1.0 mM, and cultures were incubated for an additional 4 h at 30°C. After disruption of the cells by ultrasonication, cell extracts were obtained by centrifugation at 12,000 ϫ g for 30 min at 4°C. Binding of the fusion proteins to chitin beads via intein-chitin binding domain (New England Biolabs), cleavage of the fusion proteins (in 20 mM Tris-HCl, pH 8.0, 200 mM NaCl, 0.1 mM EDTA, and 30 mM dithiothreitol at 4°C), and elution of each product were all done according to the manufacturer's instructions. Induction of MalE-Tll production with IPTG, purification of the fusion protein on amylose columns, and its cleavage with Factor Xa were done essentially as described by the manufacturer (New England Biolabs). Production of the RmlD (NCTC 9710)-glutathione S-transferase fusion protein was induced by IPTG and purified according to the manufacturer's instructions. Glycerol was added to the purified proteins to a final concentration of 50%, and they were stored at Ϫ20°C. The purity of the proteins was checked by SDS-PAGE (4% stacking gels and 12.5% separating gels).
Detection of Sugar Nucleotides by Reversed-phase HPLC-Conversion of sugar nucleotides was confirmed using reversed-phase HPLC as described by Tonetti et al. (14). Samples (10 l) diluted 10-fold with distilled water were injected onto a TSKgel ODS-80Ts column (0.46 ϫ 15 cm; Tosoh) with 0.5 M KH 2 PO 4 as the mobile phase at a flow rate of 1.0 ml/min at 40°C. The eluate was monitored with a UV detector at 260 nm.
NMR Spectroscopy-Approximately 2 mg of sugar nucleotides were pooled from several HPLC runs on an ODS-80Ts. The fraction at each run was immediately cooled on ice, and 4 volumes of cold ethanol was added to the solution to avoid degradation of dTDP-6-deoxy-L-talose in 0.5 M KH 2 PO 4 at room temperature. After removing the excess phosphate by adding ethanol, the solution was concentrated by evaporation and lyophilized. The samples were dissolved in D 2 O and used for NMR analysis. The NMR analysis was performed within 2 days after purification of the dTDP-hexose, and the sample was stored at Ϫ30°C. 1 H NMR spectra were recorded with a Bruker AM400 spectrometer. The measurement was made at 298 K. The chemical shifts were referenced to 3-(trimethylsilyl)propanesulfonic acid at 0.0 ppm. The 1 H spectra of 128 scans were recorded with presaturation of HOD resonance at 4.72 ppm. Two-dimensional COSY measurement was performed for signal assignments.
Gas-Liquid Chromatography Analysis-dTDP-6-deoxy-L-talose and dTDP-L-rhamnose samples were obtained by the same method as described in Ref. 12. About 2 mg of dTDP-6-deoxy-L-talose or dTDP-Lrhamnose was dissolved in 300 l of 0.1 M HCl. Ampoules containing the solutions were sealed under vacuum and heated at 80°C for 1 h to hydrolyze the dTDP-sugars; the water and HCl were then evaporated. The pellets were converted into the corresponding D-(ϩ)-2-octyl glycoside acetate by the method of Leontein et al. (15). Each product was characterized by gas-liquid chromatography (model GC-14B; Shimadzu Works) with a fused silica capillary column (CP Sil-88, 0.25 mm ϫ 50 m; Chrompack Inc.) at 200°C. Approximately 1 l of the sample was injected, and the split ratio was 1:20. Helium was used as a carrier gas at a flow rate of 0.9 ml/min.
Synthesis of dTDP-6-deoxyhexoses from D-Glucose 1-Phosphate and dTTP-Conversion of D-glucose 1-phosphate and dTTP to dTDP-sugars was detected by reversed-phase HPLC (Fig. 3). The elution profile of the reaction mixture containing dTTP, D-glucose 1-phosphate, and the gene product of the rmlA homologue from A. actinomycetemcomitans NCTC 9710 (Fig.  3B) and that of the reaction mixture containing dTTP, D-glucose 1-phosphate, and the gene products of the rmlA and rmlB homologues (Fig. 3C) are shown. These elution profiles were in agreement with those of the mixtures reacted with the rmlA and rmlB gene products (glucose-1-phosphate thymidylyltransferase and dTDP-D-glucose-4,6-dehydratase, respectively) from strain Y4 instead of the corresponding gene products from strain NCTC 9710 (data not shown). Considering these results in combination with high homologies of their amino acid sequences, we decided to use the products of the rmlA and rmlB homologues from strain NCTC 9710 as glucose-1-phosphate thymidylyltransferase and dTDP-D-glucose-4,6-dehydratase, respectively, in further investigations. Addition of either the rmlC, rmlD (Y4), rmlD (NCTC 9710), or tll gene product to the reaction mixture containing D-glucose 1-phosphate, dTTP, and the rmlA and rmlB products did not change its elution profile (data not shown). These results agree with the report that the dTDP-6-deoxy-L-lyxo-4-hexulose intermediate is not released from dTDP-6-deoxy-D-xylo-4-hexulose-3,5-epimerase in the dTDP-L-rhamnose biosynthetic pathway in E. coli (16). The rmlD (Y4) product converted dTDP-6-deoxy-D-xylo-4-hexulose to dTDP-L-rhamnose in conjunction with the rmlC (NCTC 9710) product (Fig. 3D). When the gene product of the rmlD homologue from strain NCTC 9710 was added to the reaction mixture instead of the strain Y4 rmlD product, the elution profile exhibited the same pattern as that of the reaction mixture containing the strain Y4 rmlD product (Fig. 3, D and E). This result also suggests that the rmlA, rmlB, and rmlC genes from strain NCTC 9710 code for glucose-1-phosphate thymidylyltransferase, dTDP-D-glucose-4,6-dehydratase, and dTDP-6deoxy-D-xylo-4-hexulose-3,5-epimerase, respectively. The elution profile of a reaction mixture containing D-glucose 1-phosphate, dTTP, NADPH, the rmlABC products, and the tll product exhibited only one peak, and its retention time agreed with that of authentic NADP (data not shown). When NADH instead of NADPH was added to the reaction mixture, two major peaks were observed and the retention time of the second peak (26.5 min) was in agreement with that of authentic NAD  Samples were injected onto a TSKgel ODS-80Ts column. A, no enzyme was added to the reaction mixture. B, the purified rmlA gene product was added to the reaction mixture. C, the purified rmlA and rmlB gene products were added to the reaction mixture. D, NADPH and the purified rmlA, rmlB, rmlC, and rmlD (NCTC 9710) gene products were added to the reaction mixture. E, NADPH and the purified rmlA, rmlB, rmlC, and rmlD (Y4) products were added to the reaction mixture. F, NADH and the purified rmlA, rmlB, rmlC, and tll products were added to the reaction mixture. (Fig. 3F). The peaks of putative dTDP-L-rhamnose and dTDP-6-deoxy-L-talose were collected and analyzed by gas-liquid chromatography and NMR.
Identification of Sugars in the Products by Enzymatic Reactions-The D-(ϩ)-2-octyl glycoside acetates derived from the sugar components of the dTDP-sugars were analyzed by gasliquid chromatography (Fig. 4). The peaks of the dTDP-sugar derivatives produced by the putative dTDP-L-rhamnose-producing dTDP-6-deoxy-L-lyxo-4-hexulose reductases forming from strain Y4 and NCTC 9710 were in agreement with that of authentic L-rhamnose (Fig. 4, A, B, and C). Because authentic 6-deoxytalose is not commercially available, hydrolysates of serotype-specific polysaccharides purified from strain NCTC 9710 (serotype c) and ATCC 29523 (serotype a) were used as standards for 6-deoxy-L-talose and 6-deoxy-D-talose, respectively (2), (Fig. 4, D and F). The profile of a derivative of the dTDP-sugar synthesized by the tll gene product was in agreement with that of 6-deoxy-L-talose, a hydrolysate of serotype c-specific polysaccharide (Fig. 4E).
NMR Analysis of dTDP-L-rhamnose and dTDP-6-deoxy-L-talose-Approximately 2 mg of dTDP-6-deoxy-L-talose and dTDP-L-rhamnose were pooled from several HPLC runs on an ODS-80Ts (Fig. 3, D and F). After removing the excess phosphate by adding ethanol, the solution was concentrated by evaporation. The concentrated solutions were lyophilized and dissolved in D 2 O. The NMR spectrum of authentic L-rhamnose was also measured (data not shown). The NMR spectra of authentic dTDP and these dTDP-hexoses are shown in Fig. 5. Assignment of these resonances was verified by two-dimensional homonuclear 1 H COSY experiments (Fig. 6). Assigned chemical shifts and coupling constants are summarized in Table I. Signals for the nucleotide moieties in the dTDP-sugars were in good agreement with those of dTDP. Signals for the sugar moiety of dTDP-L-rhamnose were in good agreement with the reported chemical shift values of ␤-L-rhamnose (17) except for H1ЈЈ. The large downfield shift (0.35 ppm) of the H1ЈЈ signal can be attributed to neighboring phosphate groups, which also affect the H2ЈЈ signal, thereby causing the poor resolution. The signal of H3Ј is rather small, perhaps because it is affected by decoupling of water nearby. The coupling constants also supported the orientations of H2ЈЈ, H3ЈЈ, H4ЈЈ, and H5ЈЈ being equatorial, axial, axial, and axial, respectively, except for the J 1,2 value of 8.80 Hz. This value is inconsistent with the chemical shift values indicating the ␤ configuration, and it is reasonable to think that the neighboring phosphate groups FIG. 4. Gas-liquid chromatography analysis of acetylated D-(؉)-2-octyl glycosides obtained from the hydrolysates of the purified dTDP-L-rhamnose and dTDP-6-deoxy-L-talose. A, glycosides of authentic L-rhamnose. B, glycosides of the hydrolysate of purified dTDP-hexose converted from dTTP and D-glucose 1-phosphate by the rmlA, rmlB, rmlC, and rmlD (Y4) products. C, glycosides of the hydrolysate of purified dTDP-hexose converted from dTTP and D-glucose 1-phosphate by the rmlA, rmlB, rmlC, and rmlD (NCTC 9710) gene products. D, glycosides of the hydrolysate of the purified serotype cspecific antigen or 6-deoxy-L-talan. E, glycosides of the hydrolysate of purified dTDP-hexose converted from dTTP and D-glucose 1-phosphate by the rmlA, rmlB, rmlC, and tll gene products. F, glycosides of the hydrolysate of the purified serotype a-specific antigen or 6-deoxy-D-talan. also affect the coupling constant. Because 6-deoxy-L-talose is not commercially available, the values of chemical shifts and coupling constants were compared with those of dTDP-L-rhamnose. The chemical shifts were in good agreement with those of dTDP-L-rhamnose except for H3ЈЈ, H4ЈЈ, and H5ЈЈ. In addition, the J 3,4 and J 4,5 values were small coupling constants in contrast to the dTDP-L-rhamnose values. These results also support this sugar nucleotide being dTDP-6-deoxy-L-talose. DISCUSSION We have previously cloned the gene cluster essential for the biosynthesis of serotype b-specific polysaccharide antigen from A. actinomycetemcomitans Y4 and identified the four genes coding for the enzymes that synthesize dTDP-L-rhamnose from D-glucose 1-phosphate, dTTP, and NADPH (10,12). The genes coding for glucose-1-phosphate thymidylyltransferase, dTDP-D-glucose-4,6-dehydratase, dTDP-6-deoxy-D-xylo-4-hexulose-3,5-epimerase, and dTDP-6-deoxy-L-lyxo-4-hexulose reductase are designated as rmlA, rmlB, rmlC, and rmlD, respectively (18). The gene cluster essential for the biosynthesis of serotype c-specific polysaccharide antigen, 6-deoxy-L-talan, was also cloned from A. actinomycetemcomitans NCTC 9710; four genes whose products exhibited high homology to the rmlA, rmlB, rmlC, and rmlD gene products were found (11). The amino acid sequences of glucose-1-phosphate thymidylyltransferase, dTDP-D-glucose-4,6-dehydratase, and dTDP-6-deoxy-D-xylo-4hexulose-3,5-epimerase of strain Y4 show more than 90% identity with their strain NCTC 9710 homologues, whereas strain Y4 dTDP-6-deoxy-L-lyxo-4-hexulose reductase shares only 58% identity with its corresponding gene product in strain NCTC 9710. We expected the rmlD homologue of strain NCTC 9710 to be the gene coding for dTDP-6-deoxy-L-lyxo-4-hexulose reductase, which synthesizes dTDP-6-deoxy-L-talose. We expected this because both dTDP-6-deoxy-L-talose and dTDP-L-rhamnose are predicted to be synthesized from dTDP-6-deoxy-L-lyxo-4-hexulose by reduction, and the stereoselectivity of the reduction determines the direction of synthesis of these compounds to dTDP-6-deoxyhexoses (11). Contrary to our expectations, the gene product of the rmlD homologue of strain NCTC 9710 converted dTDP-6-deoxy-L-lyxo-4-hexulose to dTDP-L-rhamnose (Fig. 3E).
In comparing the strain NCTC 9710 cluster essential for 6-deoxy-L-talan synthesis with the gene cluster producing serotype b-specific polysaccharide antigen, it was found that ORF7, ORF8, and ORF9 ( Fig. 1) were specific to the former cluster. The amino acid sequence of ORF8 product exhibited weak identity (22.6%) with E. coli UDP-glucose-4-epimerases and a GXXGXXG motif of a NAD-binding domain is present at the N terminus of the ORF8 product (19). UDP-glucose-4-epimerase catalyzes both the conversion of UDP-galactose to UDP-glucose and the reverse reaction; the target of the reaction is the C-4 position of the nucleotide-activated hexoses. L-Rhamnose and 6-deoxy-L-talose differ in the stereochemistry of the C-4 carbon. As serotype c-specific polysaccharide antigen does not contain L-rhamnose, we expected that dTDP-L-rhamnose would serve as a precursor for the synthesis of dTDP-6deoxy-L-talose and that the ORF8 product would be dTDP-Lrhamnose-4-epimerase for the conversion of dTDP-L-rhamnose to dTDP-6-deoxy-L-talose.  Again contrary to our expectations, the ORF8 product did not convert dTDP-L-rhamnose to dTDP-6-deoxy-L-talose (data not shown), but dTDP-6-deoxy-L-lyxo-4-hexulose to dTDP-6-deoxy-L-talose (Fig. 3F). NMR and gas-liquid chromatography analyses also showed that the product of the reaction was dTDP-6deoxy-L-talose and corresponded to the component of serotype c-specific polysaccharide antigen (Figs. 4 and 5). This gene product was therefore identified as a dTDP-6-deoxy-L-lyxo-4hexulose reductase for synthesizing dTDP-6-deoxy-L-talose, and the gene was designated tll.
In fact, the formation of dTDP-6-deoxy-L-talose by the enzymatic activity of dTDP-6-deoxy-L-lyxo-4-hexulose reductase was detected in E. coli O45 27 years ago, but no characterization of the enzyme has since been reported (9). Bacterial cell surface polysaccharides consisting solely of 6-deoxy-L-talose are rare, although this hexose has been found as a component of some bacterial polysaccharides (5)(6)(7). The type II O-antigenic polysaccharide of Burkholderia pseudomallei lipopolysaccharide contains 6-deoxy-L-talose and the cluster of 15 genes required for its production has been identified and sequenced (20). The rml genes were also found in this gene cluster, although the B. pseudomallei lipopolysaccharide does not contain L-rhamnose. Three genes, wbiB, wbiG, and wbiI, were viewed as candidates for nucleotide sugar epimerases involved in the conversion of dTDP-L-rhamnose to dTDP-6-deoxy-L-talose. None of these gene products shares homology with the tll gene product except for the consensus NAD-binding domains (GXXGXXG). In B. pseudomallei, dTDP-L-rhamnose may be a precursor of dTDP-6-deoxy-L-talose.