Molecular structure of human galactose mutarotase.

Galactose mutarotase catalyzes the conversion of beta-d-galactose to alpha-d-galactose during normal galactose metabolism. The enzyme has been isolated from bacteria, plants, and animals and is present in the cytoplasm of most cells. Here we report the x-ray crystallographic analysis of human galactose mutarotase both in the apoform and complexed with its substrate, beta-d-galactose. The polypeptide chain folds into an intricate array of 29 beta-strands, 25 classical reverse turns, and 2 small alpha-helices. There are two cis-peptide bonds at Arg-78 and Pro-103. The sugar ligand sits in a shallow cleft and is surrounded by Asn-81, Arg-82, His-107, His-176, Asp-243, Gln-279, and Glu-307. Both the side chains of Glu-307 and His-176 are in the proper location to act as a catalytic base and a catalytic acid, respectively. These residues are absolutely conserved among galactose mutarotases. To date, x-ray models for three mutarotases have now been reported, namely that described here and those from Lactococcus lactis and Caenorhabditis elegans. The molecular architectures of these enzymes differ primarily in the loop regions connecting the first two beta-strands. In the human protein, there are six extra residues in the loop compared with the bacterial protein for an approximate longer length of 9 A. In the C. elegans protein, the first 17 residues are missing, thereby reducing the total number of beta-strands by one.


3
During normal galactose metabolism, b-D-galactose is converted to glucose 1-phosphate via the action of four enzymes that constitute the Leloir pathway as shown in Scheme 1 (1). In the first step of this pathway, b-Dgalactose is epimerized to a-D-galactose through the action of galactose mutarotase. The second step involves the phosphorylation of a-D-galactose to galactose-1-phosphate by galactokinase. As indicated in Scheme 1, galactose-1phosphate uridylyltransferase catalyzes the third step by transferring a UMP group from UDP-glucose to galactose 1-phosphate, thereby generating glucose 1phosphate and UDP-galactose. To complete the pathway, UDP-galactose is converted to UDP-glucose by UDP-galactose 4-epimerase.
Mutations in three of the enzymes of the Leloir pathway, namely galactokinase, galactose-1-phosphate uridylyltransferase, or UDP-galactose 4epimerase, have been demonstrated to result in the diseased state known as galactosemia (2,3). The symptoms of this genetic disease include early onset cataracts (typically within the first two years of life) and, in more severe cases, liver, kidney, and brain damage. Cataract formation is believed to be caused by the build up of unmetabolized galactose in the lens of the eye. Aldose reductase catalyzes the reduction of the sugar to galactitol (dulcitol) which is not, in contrast to galactose, readily transported across the plasma membrane (4)(5)(6). As such, the accumulation of this highly osmotically active compound draws excess water into the lens resulting in cataracts by a mechanism which is not fully understood. The more severe symptoms of galactosemia are most likely caused by a combination of these osmotic effects and the accumulation of the by guest on March 23, 2020 http://www.jbc.org/ Downloaded from intermediate a-D-galactose-1-phosphate (7) although the precise mechanism of toxicity is not understood. No known disease-causing mutations in human galactose mutarotase have been identified thus far. However, galactitol is reported to be an inhibitor of mammalian mutarotase (8) raising the possibility that its accumulation would partly block the action of mutarotase resulting in a further build up of unmetabolized galactose.
Galactose mutarotase activity was first reported in Escherichia coli in 1965 (9) and has since been observed in a wide range of organisms including bacteria (9)(10)(11)(12)(13)(14)(15)(16), plants (17,18), fungi (19), and mammals (20)(21)(22)(23). It is present in the cytoplasm of most cells thus suggesting that the in vivo rate of uncatalyzed mutarotation is insufficient for the metabolic needs of the organism. Indeed, deletion of the mutarotase gene from E. coli results in slow growth on minimal media containing phenyl-b-D-galactopyranose as the sole carbon source (11).
The first three-dimensional structure of a dimeric mutarotase, namely that from Lactococcus lactis, was reported in 2002 (24). The enzyme was shown to adopt a b-sandwich motif, similar to that seen in domain 5 of b-galactosidase (25) with the sugar binding site located in a wide, shallow cleft. Five highly conserved residues (Arg 71, His 96, His 170, Asp 243 and Glu 304) were within hydrogen bonding distance to the hydroxyl groups of the bound galactose ligand. Previous kinetic studies suggested that the reaction mechanism for the enzyme proceeds via an acid-base mechanism with transient ring opening (26,27) and the model of the mutarotase from L. lactis implicated Glu 304 as the likely catalytic base and either His 96 or His 170 as the catalytic acid (24). Accordingly, the reaction mechanism of galactose mutarotase is thought to occur via by guest on March 23, 2020 http://www.jbc.org/ Downloaded from abstraction of the proton from the C-1 hydroxyl group by Glu 304 in the L. lactis enzyme and protonation of the ring oxygen of the sugar by His 170, which results in ring opening. Subsequent rotation about the C-1/C-2 bond followed by a reversal of the ring opening events ultimately yields a product with an altered configuration at C-1 (28). Site-directed mutagenesis of these residues (or their equivalents in the E. coli enzyme) have confirmed the role of Glu 304 as the catalytic base and established His 170 as the catalytic acid (28,29).
The gene encoding the human enzyme was identified in 2003 and the protein product characterized (30). From these investigations it was suggested that Glu 307 and His 176 fulfill the roles in acid/base catalysis in the human mutarotase (30). Interestingly, regardless of the source, mutarotases tend to show a preference for galactose over glucose and have high k cat and K m values with this sugar (typically 10 4 -10 5 s -1 and 10 -100 mM respectively) (23,(30)(31)(32)(33).
The preference of these enzymes for galactose over glucose was addressed structurally with the L. lactis mutarotase and, at least in the case of this bacterial protein, is explained by the existence of non-productive binding modes in the active site for glucose and its derivatives (33). With respect to quaternary structure, human mutarotase appears to be monomeric (10,30) whereas the L.
lactis enzyme was shown to be a dimer on the basis of ultracentrifugation experiments (24). The biochemical significance, if any, of these differences in quaternary structure is not understood at the present time.
Here we describe the three-dimensional structure of human galactose mutarotase, both in the unbound state and complexed with its substrate, b-Dgalactose. This study represents the first report of a eukaryotic mutarotase with  Structural Analysis of Apo-Galactose Mutarotase. An x-ray data set was collected to 2.2 Å resolution at 4 o C with a Bruker HISTAR area detector system equipped with Supper "long" mirrors. The x-ray source was CuKa radiation from a Rigaku RU200 x-9 ray generator operated at 50 kV and 90 mA. The x-ray data were processed with XDS (34,35) and internally scaled with XSCALIBRE (Rayment and Wesenberg, unpublished). X-ray data collection statistics are presented in Table I. The structure was solved by molecular replacement with the program AMORE (36) and using as a search model one subunit of the L. lactis enzyme (PDB 1L7J). Two peaks were readily identified in the rotation search. A subsequent translation search and rigid body refinement to 3.0 Å yielded an R-factor of 46 %. To reduce model bias, the electron densities corresponding to the two monomers in the asymmetric unit were averaged with software package AVE (37,38). From this "averaged" electron density map, a model for the complete monomer was constructed and then placed back into the unit cell for subsequent least-squares refinement with TNT (39). Alternate cycles of manual rebuilding and least-squares refinement reduced the R-factor to 17.2 % for all measured x-ray data to 2.2 Å resolution. Refinement statistics are presented in Table II.  Table II.
Quality of the X-ray Models. The models for both the apo-and sugar-bound forms of human galactose mutarotase refined to low R-factors with excellent overall stereochemistry as indicated in Table II The side chain of Arg 78 points away from the active site cleft.

13
When crystals of human galactose mutarotase are grown in the presence of D-galactose, the space group changes from P2 1 to P1 with four molecules in the asymmetric unit. These four monomers are, however, virtually identical such that their a-carbons superimpose with root-mean-square deviations of between 0.17 Å and 0.18 Å. The Gly-His motif at the N-terminus is visible in the electron density for two of the four monomers in the asymmetric unit. Again, for the sake of simplicity, the following discussion will refer only to Monomer I in the P1 asymmetric unit. Within experimental error there are no significant structural changes in the enzyme that occur upon galactose binding. All main chain plus C-b atoms for the apo-and substrate-bound forms of galactose mutarotase correspond with a root-mean-square deviation of 0.19 Å. The sugar moiety is situated in a shallow active site cleft with its 3-, 4-, and 6-hydroxyl groups somewhat solvent exposed as shown in Figure 2b.