High Resolution X-ray Structure of Tyvelose Epimerase from Salmonella typhi *

Tyvelose epimerase catalyzes the last step in the bio-synthesis of tyvelose by converting CDP- D -paratose to CDP- D -tyvelose. This unusual 3,6-dideoxyhexose occurs in the O -antigens of some types of Gram-negative bacteria. Here we describe the cloning, protein purification, and high-resolution x-ray crystallographic analysis of tyvelose epimerase from Salmonella typhi complexed with CDP. The enzyme from S. typhi is a homotetramer with each subunit containing 339 amino acid residues and a tightly bound NAD (cid:1) cofactor. The quaternary structure of the enzyme displays 222 symmetry and can be aptly described as a dimer of dimers. Each subunit folds into two distinct lobes: the N-terminal motif responsible for NAD (cid:1) binding and the C-terminal region that harbors the binding site for CDP. The analysis described here demonstrates that tyvelose epimerase belongs to the short-chain dehydrogenase/reductase superfamily of enzymes. Indeed, its active site is reminiscent to that observed for UDP-galactose 4-epimerase, an enzyme that plays a key role in galactose metabolism. Unlike UDP-galactose 4-epimerase where the conver-sion of configuration occurs about C-4 of the UDP-glucose or UDP-galactose substrates, in the reaction catalyzed by tyvelose epimerase, the inversion of stereochemistry occurs at C-2. On the

Carbohydrate epimerases are found in all living organisms and play crucial roles in numerous biochemical pathways as recently reviewed (1). The ability of an organism to invert the stereochemistry at a specific chiral center of a carbohydrate can determine its potential to utilize a particular carbon source, form a functional cell wall, and/or act pathogenically toward a host (2). Because of the diverse nature of both the chemical composition of carbohydrates and their various metabolic roles, several different epimerization mechanisms have evolved. These mechanisms can be broadly classified into two categories: those that function on activated carbon centers lying ad-jacent to carbonyl, carboxylic acid, or ester functional groups, and those that epimerize unactivated carbon centers.
Perhaps the best characterized member of the latter group, both kinetically and crystallographically, is UDP-galactose 4-epimerase (3,4). This enzyme plays a key role in the Leloir pathway for galactose metabolism by interconverting UDPgalactose and UDP-glucose. The proposed catalytic mechanism of this epimerase involves several key features: 1) transfer of the hydride from C-4 of the sugar to NAD ϩ to form NADH and abstraction of the 4Ј-hydroxyl hydrogen of the sugar by an enzymatic base to yield a 4Ј-ketopyranose intermediate; 2) rotation of this intermediate by ϳ180°about the phosphoryloxygen bond connecting the UDP and sugar moieties; and 3) transfer of the hydride from NADH back to the opposite face at C-4 of the keto-sugar intermediate and protonation of the resulting hydroxyl group. The enzymatic base in UDP-galactose 4-epimerase is thought to be a conserved tyrosine residue (4).
UDP-galactose 4-epimerase belongs to a subclass of NAD(P)dependent enzymes referred to as the short-chain dehydrogenases/reductases (5). Members of this family contain a characteristic Tyr-XXX-Lys motif in which the conserved tyrosine is thought to play a key role in catalysis. Other proteins belonging to this subfamily include dTDP-D-glucose 4,6-dehydratase (6), GDP-mannose 4,6-dehydratase (7,8), and ADP-L-glycero-Dmanno-heptose 6-epimerase (9), among others. Alignments of the amino acid sequences for these enzymes with the Swiss-Prot data bank indicate that they share at least 20 and 40% sequence identity and similarity, respectively. On the basis of such amino acid sequence alignments another member of the superfamily has recently been identified, namely CDP-tyvelose 2-epimerase (10).
CDP-tyvelose 2-epimerase, hereafter referred to simply as tyvelose epimerase, is involved in the synthesis of tyvelose, a 3,6-dideoxyhexose that occurs in the O-antigens of some types of Gram-negative bacteria (11). Tyvelose is produced via a complex biochemical pathway that employs CDP-D-glucose as the starting ligand. The last step in the pathway, the conversion of CDP-D-paratose to CDP-D-tyvelose, is catalyzed by tyvelose epimerase as outlined in Scheme 1. Interestingly, tyvelose is found in the O-antigens of some pathogens such as Salmonella typhi and Yersinia pseudotuberculosis IVA. In that tyvelose is the major immunological determinant in these bacteria, screening for the tyvelose epimerase gene via PCR amplification is being considered for earlier identification and treatment of S. typhi infections (12).
Here we describe the cloning, purification, and high-resolution x-ray crystallographic analysis of tyvelose epimerase from S. typhi. Crystals amenable for a high-resolution x-ray investigation of the enzyme were grown in the presence of CDP. The enzyme is a homotetramer displaying 222 symmetry. Each subunit contains 339 amino acid residues and one tightly bound NAD ϩ moiety. The subunit is distinctly bilobal with an N-terminal region responsible for dinucleotide binding, a Cterminal motif important for CDP positioning, and an active site wedged between these two lobes. Side chain functional groups directly involved in NAD ϩ binding include Asp-32, Asp-58, Ser-123, and Lys-169 whereas Ser-194, Asn-126, Lys-127, Gln-235, Arg-237, Arg-299, and Asp-302 function in anchoring the CDP moiety into the active site. The x-ray analysis described here defines both the quaternary structure of tyvelose epimerase and its active site geometry, and confirms that the enzyme belongs to the short-chain dehydrogenase/reductase superfamily.

EXPERIMENTAL PROCEDURES
Cloning of the Tyvelose Epimerase Gene-The sequence of the tyvelose epimerase gene, tyv, from S. typhi has been reported and allowed for the design of primers for gene amplification from genomic DNA. Genomic DNA samples from S. typhi clinical isolates CDC numbers 87-2059 and CDC 88-2009 were generous gifts from Dr. Stanley Maloy at the University of Illinois, Urbana-Champaign. The tyvelose epimerase gene was PCR amplified from genomic DNA such that the forward primer 5Ј-CATGCCATGGCCAAGCTTTTAATTACCGGTGG-A-3Ј and reverse primer 5Ј-CCGCTCGAGTATAGAACTAGTCCATCA-TACATTTC-3Ј added NcoI and XhoI cloning sites, respectively. To preserve the NcoI recognition sequence of CCATGG, the codon GCC (encoding alanine) was inserted before the second residue of the gene. The gene was PCR amplified with Platinum Pfx DNA polymerase (Invitrogen) according to the manufacturer's instructions and standard cycling conditions. The PCR product was purified with the QIAquick PCR purification kit (Qiagen Inc.), followed by digestion with both NcoI and XhoI at 37°C overnight. The gene was separated from digestion by-products on a 1.0% agarose gel, excised from the gel, and purified with QIAquick gel purification kit (Qiagen Inc.). The purified tyvelose epimerase gene was then ligated into the expression vector pET-28a (Novagen) that was previously cut with the same restriction enzymes. Escherichia coli DH5␣ cells were transformed with the ligation mixture and then plated onto LB media supplemented with 30 g/ml kanamycin. Individual colonies were selected, cultured overnight, and plasmid DNA was extracted with the QIAprep spin miniprep kit (Qiagen Inc.). Plasmids were tested for insertion of the tyv gene by digesting with NcoI and XhoI. Positive clones were sequenced with the ABI Prism TM Big Dye Primer Cycle sequencing kit (Applied Biosystems, Inc.) to confirm that no mutations were introduced during PCR amplification.
Protein Expression-For protein expression, E. coli Rosetta-(DE3)pLysS (Novagen) cells were transformed with the pET28a-tyv plasmid and plated onto LB media supplemented with 30 g/ml kanamycin. After ϳ16 h, the plates were scraped and the cells were resuspended in LB media to use for the inoculation of 6ϫ 2-liter baffled flasks containing 500 ml of LB plus 30 g/ml kanamycin and 30 g/ml chloramphenicol. The cells were grown at 37°C with aeration to an A 600 of ϳ0.4, at which time they were transferred to a 30°C shaker for 30 min before inducing with 1 mM isopropyl-1-thio-␤-D-galactopyranoside. The cells were allowed to grow for an additional 6 h at 30°C before harvesting by centrifugation at 6000 ϫ g for 8 min. The cell paste was frozen in liquid nitrogen and stored at Ϫ80°C.
Expression of the Selenomethionine-labeled Protein-E. coli Rosetta-(DE3)pLysS (Novagen) cells were transformed with the pET28a-tyv plasmid and plated onto LB media supplemented with kanamycin. Approximately 24 h later, several large colonies were selected to inoculate 500 ml of M9 minimal media supplemented with 30 g/ml kanamycin and 30 g/ml chloramphenicol for growth overnight at 37°C. Subsequently 10 ml of the overnight culture was used to inoculate 3ϫ 2-liter baffled flasks containing 500 ml of M9 minimal media with 5 g/ml thiamine, 30 g/ml kanamycin, and 30 g/ml chloramphenicol. Cultures were grown at 37°C to an A 600 of ϳ0.2. At that point, the temperature of the incubator was adjusted to 30°C for the remainder of the growth. Cultures were grown to an A 600 of ϳ0.75, before each flask was supplemented with 50 mg each of L-lysine, L-threonine, and Lphenylalanine, and 25 mg each of L-leucine, L-isoleucine, L-valine, and L-selenomethionine (13). After 20 additional minutes of growth, the cells were induced with 1 mM isopropyl-1-thio-␤-D-galactopyranoside and allowed to grow for 6 h. Cultures were harvested by centrifugation at 6000 ϫ g for 8 min, and the cell paste was frozen in liquid nitrogen for storage at Ϫ80°C.
Protein Purification-All protein purification steps were carried out on ice or at 4°C. The cell paste was thawed in 100 ml of cold Ni-NTA 1 lysis buffer containing 50 mM NaH 2 PO 4 , 300 mM NaCl, and 10 mM imidazole at pH 8.0. Cells were lysed on ice by four cycles of sonication (30 s) separated by 3 min of cooling. Cellular debris was removed by centrifugation at 4°C for 25 min at 20,000 ϫ g. The clarified lysate was loaded onto a 10-ml Ni-NTA-agarose (Qiagen Inc.) column pre-equilibrated with Ni-NTA lysis buffer. After loading, the column was washed with about 60 ml of Ni-NTA wash buffer (50 mM NaH 2 PO 4 , 300 mM NaCl, 20 mM imidazole, pH 8.0), followed by gradient elution of the protein from 20 to 300 mM imidazole in Ni-NTA lysis buffer. Proteincontaining fractions were pooled based on SDS-PAGE, and dialyzed against 20 mM HEPPS (pH 8.0) with 200 mM NaCl. The dialyzed protein was concentrated to ϳ23 mg/ml, based on an extinction coefficient of ϳ1.3 ml/(mg cm) as calculated by the program Protean (DNAstar, Inc., Madison, WI).
Molecular Weight Determination-Analytical ultracentrifugation was performed by the University of Wisconsin, Madison Biochemistry Instrumentation Facility. Three samples of the purified native tyvelose epimerase at 0.25, 0.5, and 1.0 mg/ml in 10 mM HEPPS (pH 8.0) and 200 mM NaCl were provided for analysis. The experiments were performed with a Beckman Optima XL-A centrifuge and an An 60-Ti rotor at 20°C using 12-mm double sector charcoal-filled Epson centerpieces. Rotor speeds of 3,000, 7,600, 11,000, and 13,500 rpm were utilized while monitoring the radial position of the sample by absorption of light at 280 nm. The molecular weight from a global fit of 9 data sets containing 1246 points was 154,000 indicating that the quaternary structure of tyvelose epimerase is tetrameric.
Crystallization of Native and Selenomethionine-labeled Tyvelose Epimerase-Prior to crystallization, the protein was diluted to 13 mg/ml with 20 mM HEPPS (pH 8.0), and 200 mM NaCl and CDP were added to a final concentration of 5 mM. Potential crystallization conditions were examined with a sparse matrix screen composed of 144 conditions at both room temperature and 4°C via the hanging drop method of vapor diffusion. Single crystals were observed at room temperature at pH 5-6 with 10% poly(ethylene glycol) 8000. Refinement of the crystallization conditions led to large single crystals grown at room temperature with precipitant solutions of 4 -6% poly(ethylene glycol) 8000, 250 mM tetramethylammonium chloride, and 100 mM succinate (pH 5.5). Crystals achieved typical dimensions of ϳ0.5 ϫ 0.3 ϫ 0.15 mm in 2 to 3 weeks and belonged to the space group P2 1 with unit cell dimensions of a ϭ 47.9 Å, b ϭ 167.3 Å, c ϭ 89.4 Å, and ␤ ϭ 105.5°. The solvent content of the crystals was ϳ50%, with one homotetramer in the asymmetric unit.
High Resolution X-ray Data Collection-Both the native and selenomethionine-substituted protein crystals were flash-frozen in the same manner. Briefly, crystals were harvested from the hanging drop experiments and soaked for several hours to several days in a synthetic mother liquor composed of 10 -12% poly(ethylene glycol) 8000, 200 mM NaCl, 300 mM tetramethylammonium chloride, 5 mM CDP, and 100 mM succinate (pH 5.5). These crystals were then transferred to 4% ethylene glycol, 14% poly(ethylene glycol) 8000, 325 mM NaCl, 300 mM tetramethylammonium chloride, 5 mM CDP, and 100 mM succinate (pH 5.5) for 30 s, followed by a final transfer to 15% ethylene glycol, 20% poly(ethylene glycol) 8000, 400 mM NaCl, 300 mM tetramethylammonium chloride, 5 mM CDP, and 100 mM succinate (pH 5.5). Subsequently, the crystals were flash-cooled to Ϫ150°C in a stream of nitrogen gas. Frozen crystals were stored under liquid nitrogen until synchrotron beam time became available. X-ray data sets from both the native protein and the selenomethionine-substituted enzyme crystals were collected on a 3 ϫ 3 tiled "SBC3" CCD detector at the Structural Biology Center 19-BM beamline (Advanced Photon Source, Argonne National Laboratory, Argonne, IL). The x-ray data were processed with HKL2000 and scaled with SCALEPACK (14). Relevant x-ray data collection statistics are presented in Table I.
X-ray Structural Analyses-The structure of tyvelose epimerase was solved via multiwavelength anomalous dispersion phasing with x-ray data collected from the selenomethionine-substituted protein crystals. The software package SOLVE was utilized to determine and refine the positions of the selenium atoms (15). Visual inspection of the selenium sites with the graphics program TURBO revealed four clusters of atoms related by non-crystallographic 222 symmetry (16). Subsequent noncrystallographic symmetry averaging and solvent flattening with the software package DM resulted in a readily interpretable electron density map calculated to 2.6-Å resolution (17). The averaged map was used to build an initial model of one subunit of tyvelose epimerase, after which this subunit was rotated back into the unit cell to create the full tetramer. This tetramer served as a search model for molecular replacement with the program AMORE against the native x-ray data (18). Alternate cycles of least-squares refinement with the software package TNT (19) and manual model building reduced the R-factor to 17.9% for all measured x-ray data from 30-to 1.5-Å resolution. Relevant leastsquares refinement statistics are summarized in Table II. A Ramachandran plot for all nonglycinyl main chain , values is displayed in Fig.  1a. In each monomer there are two residues that lie outside of the allowed regions: Ser-163 and Thr-269. The electron density for these residues is unambiguous. Ser-163 is located at the beginning of the sixth ␣-helix in the N-terminal region whereas Thr-269 lies within a Type III turn delineated by Thr-269 to Asn-272. Neither one of these residues is directly involved in ligand binding. The electron density corresponding to the bound ligands, CDP and NAD ϩ in Subunit I is shown in Fig. 1b.

RESULTS AND DISCUSSION
Quaternary Structure of Tyvelose Epimerase-On the basis of ultracentrifugation experiments and analysis of the crystalline packing arrangement, tyvelose epimerase is a homotetramer as shown in Fig. 2a. The enzyme has overall dimensions of ϳ60 ϫ 100 ϫ 110 Å and displays 222 symmetry. It can be envisioned as a dimer of dimers with the A/B and C/D pairs mimicking the quaternary structural interactions observed in UDP-galactose 4-epimerase (20). In both UDP-galactose 4-epimerase and the A/B or C/D dimers of tyvelose epimerase, there are two parallel ␣-helices contributed by each monomer that provide the structural framework for the subunit:subunit interface. In tyvelose epimerase, these ␣-helices are formed by Pro-93 to Gln-113 and Pro-164 to Phe-183. The buried surface area in this subunit: subunit interface is ϳ2470 Å 2 , as calculated according to the method of Lee and Richards (21) with a probe sphere of 1.4 Å. There are two additional regions per subunit (Asn-33 to Thr-40 and Glu-53 to Arg-67) that participate in monomer:monomer interactions along the A/C and B/D pairs. The polypeptide chain lying between Glu-53 to Arg-67 contains the third ␤-strand and the third ␣-helix of the Rossmann fold. The interfaces between the A/C and B/D pairs are not as extensive with buried surface areas of ϳ 1330 Å 2 . In the A/B and C/D subunit pairs, the adenine rings of the dinucleotides are separated by ϳ27 Å whereas in the A/C and B/D pairs, these rings are positioned at ϳ11 Å.
Tertiary Structure of the Individual Subunit-The crystals employed in this investigation contained a complete tetramer in the asymmetric unit. The ␣-carbon traces for all of the monomers are virtually identical, however, such that they superimpose with root mean square deviations of 0.4 Å or less. For the sake of simplicity the following discussion will only refer to Subunit I of the coordinate file. As depicted in Fig. 2b, the overall molecular architecture of the tyvelose epimerase subunit can be envisioned as two lobes: the N-terminal region formed by Ala-2 to Gln-200 and Ala-242 to Gly-268 and the C-terminal motif delineated by Phe-201 to His-241 and Thr-269 to Ile-339. The N-terminal motif is dominated by a seven stranded parallel ␤-sheet defined by Ala-2 to Thr-7, Asp-27 to Asp-32, Glu-53 to Gly-57, Ser-76 to Leu-80, Asn-119 to Ser-123, Asn-186 to Tyr-196, and Ala-263 to Ile-266. There are two    Ala-242 to Ala-254) and numerous reverse turns. The C-terminal motif is less complicated with three strands of mixed ␤-sheet (Val-236 to Leu-240, Ser-273 to Ser-275, and Arg-304 to Phe-306), two strands of parallel ␤-sheet (Phe-227 to Gly-231 and Thr-294 to Leu-296), and four major ␣-helices (Trp-208 to Glu-218, Leu-276 to Tyr-286, Lys-311 to Ala-316, and Ala-325 to Ile-339). The overall dimensions of the subunit are ϳ50 ϫ 50 ϫ 50 Å.

Structure of Tyvelose Epimerase
Binding Sites for NAD ϩ and CDP-Shown in Fig. 3 is a close-up view of the active site for the tyvelose epimerase subunit. There are 15 well ordered water molecules located within 3.2 Å of the CDP or NAD ϩ moieties. The cytosine ring of CDP is hydrogen bonded to the protein via the carbonyl oxygen of Tyr-204 and the peptidic NH of Ser-230. An ordered water molecule is located at 2.9 Å from the amino group of the pyrimidine ring. A strikingly aromatic patch of residues, com-posed of Tyr-204, Trp-208, Trp-211, and Phe-212, surrounds one side of the cytosine ring. The CDP ribose adopts the C 2 Јendo conformation and its 3-hydroxyl group is anchored to the protein via the side chains of Asp-302 and Gln-235. Both Arg-237 and Arg-299, through their side chain guanidinium groups, are involved in electrostatic interactions with the ␣-phosphoryl oxygens. Additionally, N ⑀ of Arg-237 lies within 3 Å of one of the ␤-phosphoryl oxygens. Asn-126 and Ser-194 form hydrogen bonding interactions with the other two ␤-phosphoryl oxygens. The ␤-phosphoryl group of CDP is situated at ϳ5 Å from the nicotinamide ring of NAD ϩ .
The nicotinamide ring of the dinucleotide adopts the synconformation as typically observed for B-side specific dehydrogenases. Both riboses of the dinucleotide adopt C 2 Ј-endo conformations. The adenine ring of NAD ϩ forms hydrogen bonding interactions with the carboxylate group of Asp-58 and the back- bone peptidic NH group of Ile-59. Asp-32 bridges the 2-and 3-hydroxyl groups of the adenine ribose. The ribose 2-hydroxyl group also lies within hydrogen bonding distance to the peptidic NH group of Arg-36. There are four water molecules located within 3.2 Å of the phosphoryl oxygens of the NAD ϩ ligand. Additionally, the peptidic NH groups of Phe-12 and Leu-13 form hydrogen bonds with OP2 and OP5 of the NAD ϩ , respectively. These residues reside in the first ␣-helix of the Rossmann fold. The 3-hydroxyl group of the nicotinamide ribose lies with 3.2 Å of the backbone carbonyl oxygen of Leu-80, a water molecule, and N of Lys-169. This interaction between Lys-169 and the dinucleotide is conserved among the members of the short-chain dehydrogenase/reductase family. The 2-hydroxyl group of the nicotinamide ribose is positioned near a water molecule that, in turn, lies within 3 Å of the nicotinamide ring nitrogen. In the short-chain dehydrogenase/reductase family, it has been postulated that Lys-169 activates Tyr-165 that is the ultimate base in the reaction mechanism. The distance between Lys-169 and Tyr-165 in tyvelose epimerase is 4.9 Å, a value that is comparable with that observed in other members of the superfamily. Interestingly, the distance between the nicotinamide ring nitrogen, which carries the positive charge in the oxidized form of the cofactor, and O of Tyr-165 is 4.3 Å. It is possible that the chemical role of Lys-169 is in proper dinucleotide positioning and not in activation of the conserved tyrosine residue. Regarding the carboxamide moiety of the nicotinamide ring, there is an intramolecular hydrogen bond between its nitrogen and OP4 of the phosphate backbone (2.9 Å). The carbonyl oxygen of the carboxamide group hydrogen bonds with a water molecule and the backbone peptidic NH group of  Comparison of Tyvelose Epimerase with UDP-galactose 4-Epimerase-Tyvelose epimerase is highly homologous to both UDP-galactose 4-epimerase from E. coli (27% identity and 43% similarity), and human UDP-galactose 4-epimerase (23% identity and 41% similarity). In contrast to tyvelose epimerase, however, both the human and bacterial forms of UDP-galactose 4-epimerase are homodimers. Shown in Fig. 4a is a superposition of the polypeptide chain backbones for the human UDPgalactose 4-epimerase and the tyvelose epimerase subunits. There are only four major differences in conformations between these two enzymes as indicated by the letters A-D in Fig. 4a.
For example, there is a 7-residue deletion in tyvelose epimerase, relative to UDP-galactose 4-epimerase, starting at Arg-36 and extending to His-56. This particular region is involved in the A/C and B/D interfaces in tyvelose epimerase and thus may be one factor for the differing quaternary states between these two proteins. There is a 16-residue insertion in tyvelose epimerase, defined by Leu-132 to Tyr-153, which forms two antiparallel ␤-strands that are connected by a Type I turn (Thr-140 to Arg-143). Interestingly, this region in not involved in subunit:subunit contacts nor located near the NAD ϩ binding site. The functional significance of this insertion in tyvelose epimerase is unknown. The third region of difference occurs as a 6-residue deletion in tyvelose epimerase relative to UDP-galactose 4-epimerase from Gly-197 to Arg-199. This area is not located within the immediate vicinity of the NAD ϩ or CDP ligands. Finally, there is a 6-residue deletion in tyvelose epimerase from Gly-231 to Val-236, which corresponds to the region defined by Gly-227 to Val-238 in UDP galactose 4-epimerase. In both of these enzymes, this area of polypeptide chain is involved in binding the nucleotide portion of the substrate, either CDP or UDP, in tyvelose epimerase or UDPgalactose 4-epimerase, respectively. Indeed, while the positions of the NAD ϩ ligands in both enzymes are nearly identical, the locations of the CDP versus UDP moieties are somewhat different as can be seen in Fig. 4a. The ␣and ␤-phosphorus atoms of the CDP versus UDP moieties are shifted by ϳ2 Å. This shift may be a function of the need for the substrate, CDP-paratose, to position differently in the active site of tyvelose epimerase as compared with UDP-glucose (or UDP-galactose) in the active site of UDP-galactose 4-epimerase. Recall that in the reaction mechanism of UDP-galactose 4-epimerase, the inversion of stereochemistry occurs about C-4 of the hexose rather than C-2 as in tyvelose epimerase. Also, paratose is a 3,6-dideoxysugar. Strikingly, in the model of the human UDP-galactose 4-epimerase, the 2Ј-hydroxyl group of UDP-glucose is hydrogen bonded to Asn-207 whereas the 6Ј-hydroxyl group lies within 2.6 Å of O ␦1 of Asn-187. In tyvelose epimerase, these residues have been replaced by Gly-207 and Ser-194, respectively.
Previous detailed mechanistic investigations of tyvelose epimerase from Y. pseudotuberculosis IVA have suggested that its catalytic mechanism may be reminiscent to that proposed for UDP-galactose 4-epimerase (22). The reaction catalyzed by UDP-galactose epimerase is an epimerization about C-4 of UDP-glucose (or UDP-galactose) as indicated in Scheme 1. Extensive structural investigations in this laboratory have concluded that in the human UDP-galactose 4-epimerase, Tyr-157 functions to initiate the reaction by abstracting the 4Ј-hydroxyl hydrogen from UDP-glucose (or UDP-galactose) with the concomitant transfer of the hydride at C-4 to the nicotinamide ring of NAD ϩ . Additionally, Ser-132 in the human epimerase is thought to facilitate proton transfer from the sugar 4Ј-hydroxyl group to O of Tyr-157 possibly via a low barrier hydrogen bond (4). Tyr-157 in human UDP-galactose 4-epimerase is conserved in tyvelose epimerase as Tyr-165 whereas Ser-132 in the human enzyme has been replaced with Thr-125. Both Tyr-165 and Thr-125 occupy similar positions within the active site of tyvelose epimerase as observed for Tyr-157 and Ser-132 in the UDP-galactose 4-epimerase active site region. Assuming that the reaction mechanisms are similar for these two epimerases as suggested by the mechanistic data, it is possible to speculate on the manner in which CDP-paratose or CDP-tyvelose might bind in the active site of the enzyme. Previous studies on the FIG. 4. Comparison of tyvelose epimerase with UDP-galactose 4-epimerase. Shown in a is a superposition of the ␣-carbons for these two enzymes. The bound ligands for tyvelose epimerase (NAD ϩ and CDP) and UDP-galactose 4-epimerase (NADH and UDP-glucose) are displayed in blue and green, respectively. On the basis of the observed binding for CDP to tyvelose epimerase, models of CDP-paratose and CDP-tyvelose were built into the active site as shown in b. The CDP-paratose moiety is highlighted in yellow-filled bonds, whereas the CDPtyvelose ligand is depicted in red-filled bonds.