Characterization of Human UDP-glucose Dehydrogenase

UDP-glucose dehydrogenase (UGDH) catalyzes two oxidations of UDP-glucose to yield UDP-glucuronic acid. Pathological overproduction of extracellular matrix components may be linked to the availability of UDP-glucuronic acid; therefore UGDH is an intriguing therapeutic target. Specific inhibition of human UGDH requires detailed knowledge of its catalytic mechanism, which has not been characterized. In this report, we have cloned, expressed, and affinity-purified the human enzyme and determined its steady state kinetic parameters. The human enzyme is active as a hexamer with values for Km and Vmax that agree well with those reported for a bovine homolog. We used crystal coordinates for Streptococcus pyogenes UGDH in complex with NAD+ cofactor and UDP-glucose substrate to generate a model of the enzyme active site. Based on this model, we selected Cys-276 and Lys-279 as likely catalytic residues and converted them to serine and alanine, respectively. Enzymatic activity of C276S and K279A point mutants was not measurable under normal assay conditions. Rate constants measured over several hours demonstrated that K279A continued to turn over, although 250-fold more slowly than wild type enzyme. C276S, however, performed only a single round of oxidation, indicating that it is essential for the second oxidation. This result is consistent with the postulated role of Cys-276 as a catalytic residue and supports its position in the reaction mechanism for the human enzyme. Lys-279 is likely to have a role in positioning active site residues and in maintaining the hexameric quaternary structure. UDP-glucose dehydrogenase (UGDH) catalyzes two oxidations of UDP-glucose to yield UDP-glucuronic acid. Pathological overproduction of extracellular matrix components may be linked to the availability of UDP-glucuronic acid; therefore UGDH is an intriguing therapeutic target. Specific inhibition of human UGDH requires detailed knowledge of its catalytic mechanism, which has not been characterized. In this report, we have cloned, expressed, and affinity-purified the human enzyme and determined its steady state kinetic parameters. The human enzyme is active as a hexamer with values for Km and Vmax that agree well with those reported for a bovine homolog. We used crystal coordinates for Streptococcus pyogenes UGDH in complex with NAD+ cofactor and UDP-glucose substrate to generate a model of the enzyme active site. Based on this model, we selected Cys-276 and Lys-279 as likely catalytic residues and converted them to serine and alanine, respectively. Enzymatic activity of C276S and K279A point mutants was not measurable under normal assay conditions. Rate constants measured over several hours demonstrated that K279A continued to turn over, although 250-fold more slowly than wild type enzyme. C276S, however, performed only a single round of oxidation, indicating that it is essential for the second oxidation. This result is consistent with the postulated role of Cys-276 as a catalytic residue and supports its position in the reaction mechanism for the human enzyme. Lys-279 is likely to have a role in positioning active site residues and in maintaining the hexameric quaternary structure. UDP-glucose dehydrogenase (UGDH) 1The abbreviations used are: UGDH, UDP-glucose dehydrogenase; FPLC, fast protein liquid chromatography. catalyzes the conversion of UDP-glucose to UDP-glucuronic acid, an essential precursor for synthesis of extracellular matrix polysaccharides in numerous diverse species, including humans. The normal cellular functions of UGDH have been studied extensively in various organisms by targeted gene disruptions, and these studies demonstrate its importance in developmental processes. In Drosophila melanogaster, the enzyme is encoded by the sugarless gene and is required for heparan sulfate modification of proteins that control wing formation (1Hacker U. Lin X. Perrimon N. Development. 1997; 124: 3565-3573Crossref PubMed Google Scholar). In Caenorhabditis elegans, UGDH fuels the production of glycosaminoglycans essential for vulval morphogenesis and embryonic development (2Hwang H.Y. Horvitz H.R. Proc. Natl. Acad. Sci. U. S. A. 2002; 99: 14224-14229Crossref PubMed Scopus (63) Google Scholar). In zebrafish, the enzyme is critical for normal cardiac development (3Walsh E.C. Stainier D.Y. Science. 2001; 293: 1670-1673Crossref PubMed Scopus (251) Google Scholar), and the phenotype of targeted disruption is extremely similar to the murine HAS2 null (4Camenisch T.D. Spicer A.P. Brehm-Gibson T. Biesterfeldt J. Augustine M.L. Calabro Jr., A. Kubalak S. Klewer S.E. McDonald J.A. J. Clin. Investig. 2000; 106: 349-360Crossref PubMed Scopus (719) Google Scholar); both phenotypes are lethal. Human and mouse HAS2 proteins are 97% identical and are responsible for production of hyaluronan at specific developmental stages (5Fraser J.R. Laurent T.C. Laurent U.B. J. Intern. Med. 1997; 242: 27-33Crossref PubMed Scopus (1512) Google Scholar). Because UGDH provides the requisite UDP-glucuronic acid precursor for hyaluronan synthesis, it is not surprising that UGDH null and HAS2 null would exhibit similar developmental impairments. Hyaluronan is an extracellular matrix component that is involved directly in promoting normal cellular growth and migration, but its elevated production has been extensively implicated in the progression of epithelial cancers such as breast (6Auvinen P. Tammi R. Parkkinen J. Tammi M. Agren U. Johansson R. Hirvikoski P. Eskelinen M. Kosma V.M. Am. J. Pathol. 2000; 156: 529-536Abstract Full Text Full Text PDF PubMed Scopus (429) Google Scholar), colon (7Ropponen K. Tammi M. Parkkinen J. Eskelinen M. Tammi R. Lipponen P. Agren U. Alhava E. Kosma V.M. Cancer Res. 1998; 58: 342-347PubMed Google Scholar), and prostate (8Aaltomaa S. Lipponen P. Tammi R. Tammi M. Viitanen J. Kankkunen J.P. Kosma V.M. Urol. Int. 2002; 69: 266-272Crossref PubMed Scopus (70) Google Scholar). Such elevated synthesis of hyaluronan places increased demand on intracellular UDP-glucuronic acid concentrations, which also must be increased to compensate. Inhibition of hyaluronan synthesis recently has been shown to reduce tumor angiogenesis and thereby restrict in vivo growth of human prostate tumors (9Simpson M.A. Wilson C.M. McCarthy J.B. Am. J. Pathol. 2002; 161: 849-857Abstract Full Text Full Text PDF PubMed Scopus (141) Google Scholar). Restriction of precursor availability by inhibition of UGDH may regulate the production of hyaluronan in tumors and therefore is a possible novel avenue for therapeutic intervention. Interestingly, UGDH was identified in breast cancer cells as a gene product dramatically up-regulated in response to elevated androgen (10Lapointe J. Labrie C. Endocrinology. 1999; 140: 4486-4493Crossref PubMed Scopus (19) Google Scholar), so an androgen-secreting organ such as the prostate may have a mechanism for increased UGDH expression. Although a clone of human UGDH has been reported (11Spicer A.P. Kaback L.A. Smith T.J. Seldin M.F. J. Biol. Chem. 1998; 273: 25117-25124Abstract Full Text Full Text PDF PubMed Scopus (133) Google Scholar), kinetic properties and detailed mechanistic characterization of the purified enzyme (a necessary prelude to inhibitor design) are not available yet. Kinetic studies of the bovine (12Nelsestuen G.L. Kirkwood S. J. Biol. Chem. 1971; 246: 3824-3834Abstract Full Text PDF PubMed Google Scholar, 13Ordman A.B. Kirkwood S. Biochim. Biophys. Acta. 1977; 481: 25-32Crossref PubMed Scopus (39) Google Scholar, 14Ordman A.B. Kirkwood S. J. Biol. Chem. 1977; 252: 1320-1326Abstract Full Text PDF PubMed Google Scholar, 15Ridley W.P. Houchins J.P. Kirkwood S. J. Biol. Chem. 1975; 250: 8761-8767Abstract Full Text PDF PubMed Google Scholar, 16Schiller J.G. Bowser A.M. Feingold D.S. Carbohydr. Res. 1972; 25: 403-410Crossref PubMed Scopus (12) Google Scholar, 17Franzen J.S. Ashcom J. Marchetti P. Cardamone Jr., J.J. Feingold D.S. Biochim. Biophys. Acta. 1980; 614: 242-255Crossref PubMed Scopus (28) Google Scholar, 18Franzen J.S. Marchetti P.S. Feingold D.S. Biochemistry. 1980; 19: 6080-6089Crossref PubMed Scopus (28) Google Scholar, 19Franzen J.S. Ishman R. Feingold D.S. Biochemistry. 1976; 15: 5665-5671Crossref PubMed Scopus (16) Google Scholar, 20Franzen J.S. Kuo I. Eichler A.J. Feingold D.S. Biochem. Biophys. Res. Commun. 1973; 50: 517-523Crossref PubMed Scopus (18) Google Scholar), streptococcal (21Campbell R.E. Sala R.F. van de Rijn I. Tanner M.E. J. Biol. Chem. 1997; 272: 3416-3422Abstract Full Text Full Text PDF PubMed Scopus (75) Google Scholar, 22Campbell R.E. Tanner M.E. J. Org. Chem. 1999; 64: 9487-9492Crossref Scopus (31) Google Scholar, 23Campbell R.E. Mosimann S.C. van De Rijn I. Tanner M.E. Strynadka N.C. Biochemistry. 2000; 39: 7012-7023Crossref PubMed Scopus (99) Google Scholar, 24Ge X. Campbell R.E. van De Rijn I. Tanner M.E. J. Am. Chem. Soc. 1998; 120: 6613-6614Crossref Scopus (27) Google Scholar), and plant (25Stewart D.C. Copeland L. Plant Physiol. 1998; 116: 349-355Crossref Scopus (43) Google Scholar, 26Turner W. Botha F.C. Arch. Biochem. Biophys. 2002; 407: 209-216Crossref PubMed Scopus (44) Google Scholar) enzymes have been reported. All forms of the enzyme are known to catalyze the same reaction. Two successive oxidations are performed by the enzyme to convert the 6′-hydroxyl of UDP-glucose to a carboxylate concurrent with reduction of 2 mol of NAD+ to NADH. Early experiments to determine an enzyme mechanism in the bovine system preceded the availability of sequence information, so specific catalytic residues were not identified. However, results of these studies led to the interesting mechanism proposed in Scheme 1 in which we have numbered key residues identified in our experiments. According to analysis of the bovine enzyme, in the first step the 6′-carbon of UDP-glucose is thought to be covalently attached in a Schiff base linkage with a lysine residue at the active site of the protein. There is evidence to suggest that this imine undergoes the first oxidation step yielding an aldimine intermediate that is not released from the protein. The aldimine is hydrolyzed/attacked by an active site cysteine to yield a thiohemiacetal that has been detected in both bovine and streptococcal enzyme systems. This intermediate then is oxidized a second time to produce a thioester, which is spontaneously hydrolyzed to release UDP-glucuronic acid as a product. Hydrolysis of the thioester is thought to be rate-limiting overall. Notably, the streptococcal enzyme is believed to form a covalent intermediate the product of the first oxidation and cysteine the active which is required for the second oxidation. This residue is and to cysteine in the human enzyme. In this report, we have cloned, expressed, and purified human our this is the first characterization of the human enzyme as well as the first UGDH to be purified and by We have determined that the enzyme is active as a hexamer with evidence for in a of as has been proposed for the bovine enzyme. state kinetic constants were and to be with in the bovine In we a model of human UGDH on the crystal coordinates for the streptococcal homolog. this model and sequence of known UGDH we selected Cys-276 and Lys-279 as likely active site catalytic residues and converted them to serine and alanine, respectively. Enzymatic activity of C276S and K279A point mutants was not measurable under normal assay conditions. Rate constants were measured over several hours and demonstrated that K279A continued to turn over, although 250-fold more slowly than wild type enzyme. C276S, however, performed only a single round of oxidation indicating that it is essential for the second oxidation. This result is consistent with the postulated role of Cys-276 as a catalytic residue as shown in Scheme 1 and supports its position in the reaction mechanism for the human enzyme. Lys-279 to in both the active site and quaternary of the enzyme. of for UGDH and from human prostate tumor cells was used as a for were the sequence for human UGDH to an enzyme site to the and a site to the The product was with the enzymes and a a mutants were the the were in All were by the at the of and of and to UGDH and the point mutants C276S and K279A were used to were to an of in and at and then were for by the of to a of The cells were by and in and were by and for at All proteins were to be in the The was with at for and for at to from the The enzyme then was purified by a to the were of purified protein of All were The purified protein was and 1 and protein in the activity assay we the wild type enzyme The was by the protein with 1 for 1 at by and 1 of the was by and proteins were to have identical also were and the although at for to was only at in a were performed on of protein in this Enzymatic and Kinetic activity of the wild type enzyme was determined by the in at that the reduction of NAD+ to NADH. of the assay was performed for at in 1 UDP-glucose and NAD+ The Km and Vmax for UDP-glucose and NAD+ were determined assay conditions. for UDP-glucose as substrate were measured by NAD+ and from to NAD+ kinetic were by UDP-glucose and NAD+ from to 2 values were for and were with Km and Vmax were by the to the and a single site for substrate and Kinetic for the two point mutants were under assay and activity was to be at the time determine activity in these were with cofactor and and an from to was for The values then were over the time of the assay and in by to a with a for C276S and to model the first and second oxidations of of cofactor of enzyme was The wild type protein was to determine the that not at this then was used to for of the point mutants was to and was these and values in the of NAD+ from to values were for enzyme by to a with a single of the active UGDH enzyme complex was determined by chromatography. UGDH wild type and point mutants were on a and by in at a of was by with under the same conditions. The used of were as was to in of of and of UGDH is identical to the Streptococcus pyogenes enzyme for which a crystal has been We human UGDH L.A. J. Biol. 2000; PubMed Scopus Google Scholar), which the human acid sequence the coordinates for S. pyogenes UGDH of the model was with Scholar). The active site was and the model, in with sequence was used to catalytic which are in and the for human UGDH was in cells and the by its kinetic were reported for the enzyme. was to steady state kinetic constants for wild type human UGDH, determine the of the active enzyme key active site and the role of in the enzyme these wild type UGDH and specific point mutants were in a that encoded an for The UGDH in E. and enzyme form was purified to by that the not the of wild type UGDH its the was by and the protein was with protein. Because and proteins identical of enzyme activity not protein was used for Kinetic of catalyzes conversion of a UDP-glucose substrate to a UDP-glucuronic acid two of NAD+ to NADH. The activity of wild type UGDH was by steady state kinetic kinetic constants and for substrate and we measured the steady state of by its at to determine the of the reaction on cofactor were purified UGDH with of NAD+ in the of UDP-glucose substrate of reaction on substrate was measured by UDP-glucose in the of NAD+ were for both conditions. were to the to Km and Vmax for the reaction by the wild type enzyme of a similar Vmax of of of enzyme. The Km for UDP-glucose was and the Km for NAD+ was state kinetic constants for wild type in a of UGDH studies with S. pyogenes UGDH that the enzyme is a but the bovine enzyme is to be determine the state of human UGDH, we performed analysis by with for the bovine human UGDH is as a hexamer from the at a to the single and also are indicating This was over several of the enzyme with so the of to be an of the enzyme. The for the on the of The enzyme was by not which its at of consistent with a hexameric of the of UGDH the of NAD+ cofactor to UGDH of its oxidation in we the of We determined that of the protein at was by the of NAD+ not was as a of NAD+ and to an a single site The was This is in of to the Km for NAD+ determined by catalytic constants for wild type and in a of for human UGDH has not been but the human enzyme is identical to its in S. pyogenes for which a was recently the coordinates of the enzyme with cofactor and we the human enzyme as a complex with NAD+ cofactor and UDP-glucose substrate and the of the active site The of was to specific residues in the human enzyme likely to in of the streptococcal UGDH was and to reduce enzyme Based on the model of human UGDH and studies of the enzyme, we identified specific residues the active site that were likely to be for catalytic We point at residues Cys-276 human of and Lys-279 to to the enzyme the position of these two residues in the active site of the human UGDH model to NAD+ and shown are of residues and Kinetic of UGDH activity of UGDH point mutants C276S and K279A was measured under reaction and to be C276S, this is consistent with results of the streptococcal point which activity to an the role of residue in and determine the of catalytic we performed a time over enzyme activity by from to of NAD+ to by C276S was to at a to that of that a single round of oxidation in the of this cysteine but that the cysteine is essential for the second oxidation. In cofactor reduction was not in the reaction by the K279A its continued for values at were as a of assay time to determine kinetic constants for the reaction values for C276S were to at an of This to a of to protein. identical K279A continued to turn over an C276S kinetic were to an with a single for of of of enzyme This is from the of NAD+ by wild type enzyme. Cys-276 therefore is not only essential for the second round of it is also involved in of the first for which not in the time of the two The was with the for C276S but was by a of although not directly required for oxidation, Lys-279 to have a role in the catalytic constants for wild type and in a by and that in enzyme activity wild type and point mutants of UGDH were not to a of cofactor we in the of of was to a a single site. constants demonstrate cofactor wild type and UGDH and for wild C276S, and that the of the proteins were by the point of activity was not the result of a in of NAD+ by the by that Lys-279 in cofactor of catalytic activity of the mutants also result from in quaternary the state of the enzymes was for by C276S a hexameric with a as by of its on that of the wild type enzyme K279A is as a This is consistent with the UGDH quaternary a of the hexameric protein. Although the model of UGDH not a for it to be required for of the that must of UGDH is by from to humans. In and its role is in the of In its is and its is linked to several including cancer Specific targeted inhibition of this enzyme may a therapeutic avenue but requires a detailed of its mechanistic studies have on the bovine forms of the enzyme. In the we have human UGDH was cloned, expressed, and purified to We the first characterization of its mechanistic Kinetic properties of the hexameric enzyme are with those reported for UGDH Specific point mutants were on of the enzyme and used to the components of the enzyme reaction Cys-276 is required for the second but not the first of the two successive Lys-279 is not an for but of this residue and the quaternary of the protein. role is to position the active site cysteine for but it may also be an at the Human UGDH is to the enzyme from bovine but is only identical to the S. pyogenes enzyme, both of which catalyze the same oxidation of UDP-glucose to UDP-glucuronic acid, by similar the kinetic activity and quaternary of the wild type human protein are similar to the the kinetic constants we have measured are of those reported for the bovine enzyme, which also functions as a hexamer J.G. Bowser A.M. Feingold D.S. Carbohydr. Res. 1972; 25: 403-410Crossref PubMed Scopus (12) Google Scholar). The S. pyogenes enzyme, in is and has constants for substrate and cofactor to both proteins R.E. Mosimann S.C. van De Rijn I. Tanner M.E. Strynadka N.C. Biochemistry. 2000; 39: 7012-7023Crossref PubMed Scopus (99) Google Scholar). is not role may in enzyme substrate and of the active site in the bovine enzyme suggest that is the in the of that is the first J.S. Ashcom J. Marchetti P. Cardamone Jr., J.J. Feingold D.S. Biochim. Biophys. Acta. 1980; 614: 242-255Crossref PubMed Scopus (28) Google Scholar, 20Franzen J.S. Kuo I. Eichler A.J. Feingold D.S. Biochem. Biophys. Res. Commun. 1973; 50: 517-523Crossref PubMed Scopus (18) Google Scholar). This may a for human and bovine Although it is possible that the activity of our K279A is to its we have a point that also is but activity with that of wild type not Lys-279 is shown by the model to be in a of to the the has been that the enzyme preceded the forms (11Spicer A.P. Kaback L.A. Smith T.J. Seldin M.F. J. Biol. Chem. 1998; 273: 25117-25124Abstract Full Text Full Text PDF PubMed Scopus (133) Google Scholar). The then likely would be species, and the would be The quaternary of our lysine mutants are consistent with of the and that at in by the is a with which to the of residues an enzyme active site. only such has been cysteine of S. pyogenes UGDH was converted to a and the activity but were X. Campbell R.E. van De Rijn I. Tanner M.E. J. Am. Chem. Soc. 1998; 120: 6613-6614Crossref Scopus (27) Google Scholar). site residues in the bovine enzyme have been identified by We to mechanistic for the human enzyme by a of activity and Although only identical to UGDH, the sequence of the human enzyme well the crystal coordinates determined for the enzyme. the protein in this we were to generate an of the the active site of a complex of human UGDH, NAD+ and UDP-glucose substrate of this model, with of residues in a sequence of UGDH diverse and of for the bovine and the of two key acid residues likely to have a role in the catalytic Cys-276 and The a of the active site that may to the complex and Lys-279 to the of the NAD+ of that residue on our NAD+ and forms an with in a at the active thereby to the of the and This may be a a component of the reaction has been that the would as a base for the first oxidation step in the reaction R.E. Mosimann S.C. van De Rijn I. Tanner M.E. Strynadka N.C. Biochemistry. 2000; 39: 7012-7023Crossref PubMed Scopus (99) Google Scholar). at the active site then would be to and would on by Lys-279 and but our not this The cysteine is of the 6′-hydroxyl of UDP-glucose and is for The mechanism proposed for UGDH the production of as an intermediate (21Campbell R.E. Sala R.F. van de Rijn I. Tanner M.E. J. Biol. Chem. 1997; 272: 3416-3422Abstract Full Text Full Text PDF PubMed Scopus (75) Google Scholar). then is postulated to formation of a thiohemiacetal Cys-276 and this which is oxidized to a thioester to yield the carboxylate product. demonstrate that Cys-276 is of a single NAD+ which is consistent with this However, the intermediate has been from the the bovine reaction of the bovine system postulated a reaction intermediate that was covalently linked to the enzyme and not released oxidations A.B. Kirkwood S. J. Biol. Chem. 1977; 252: 1320-1326Abstract Full Text PDF PubMed Google Scholar). and Kirkwood A.B. Kirkwood S. J. Biol. Chem. 1977; 252: 1320-1326Abstract Full Text PDF PubMed Google Scholar) a Schiff base intermediate that would be linked an active site demonstrate this and the enzyme residue these the active lysine with and with thereby a at the active site. studies suggest that the enzyme first forms a Schiff base linkage with the 6′-carbon of the to the first oxidation, as we have in Scheme C276S also are consistent with this Interestingly, although Lys-279 to be for its role is to the position of an essential catalytic residue the active site we have an J. J. J. and M. A. in which that is an essential for the first oxidation of UDP-glucose and it as a for the Schiff base The of the studies to the reaction mechanism as in Scheme 1 with key residues we have identified in The reaction sequence is by the of by forms a covalent linkage to the of the is as and the is oxidized to a Schiff base and NADH. The of Cys-276 as the Schiff a thiohemiacetal that is oxidized by reduction of a second NAD+ to the Hydrolysis of the thioester the UDP-glucuronic acid product. The mechanism that specific may be at the of Schiff base thiohemiacetal formation these would the of covalent by the enzyme and be mechanistic We of for experiments.