The Biochemical Role of Glutamine 188 in Human Galactose-1-phosphate Uridyltransferase*

The substitution of arginine for glutamine at amino acid 188 (Q188R) ablates the function of human galactose-1-phosphate uridyltransferase (GALT) and is the most common mutation causing galactosemia in the white population. GALT catalyzes two consecutive reactions. The first reaction binds UDP-glucose (UDP-Glu), displaces glucose-1-phosphate (glu-1-P), and forms the UMP-GALT intermediate. In the second reaction, galactose-1-phosphate (gal-1-P) is bound, UDP-galactose (UDP-Gal) is released, and the free enzyme is recycled. In this study, we modeled glutamine, asparagine, and a common mutation arginine at amino acid 188 on the three-dimensional model of the Escherichia coli GALT-UMP protein crystal. We found that the amide group of the glutamine side chain could provide two hydrogen bonds to the phosphoryl oxygens of UMP with lengths of 2.52 and 2.82 Å. Arginine and asparagine could provide only one hydrogen bond of 2.52 and 3.02 Å, respectively. To test this model, we purified recombinant human Gln188-, Arg188-, and Asn188-GALT and analyzed the first reaction in the absence of gal-1-P by quantitating glu-1-P released using enzyme-linked methods. Gln188-GALT displaced 80 ± 7.0 nmol glu-1-P/mg GALT/min in the first reaction. By contrast, both Arg188- and Asn188-GALT released more glu-1-P (170 ± 8.0 and 129 ± 28.4 nmol/mg GALT/min, respectively). The overall, double displacement reaction was quantitated in the presence of gal-1-P. Gln188-GALT produced 80,030 ± 5,910 nmol glu-1-P/mg GALT/min, whereas the mutant Arg188- and Asn188-GALT released only 600 ± 71.2 and 2960 ± 283.6 nmole glu-1-P/mg GALT/min, respectively. We conclude from these data that glutamine at position 188 stabilizes the UMP-GALT intermediate through hydrogen bonding and enables the double displacement of both glu-1-P and UDP-Gal. The substitution of arginine or asparagine at position 188 reduces hydrogen bonding and destabilizes UMP-GALT. The unstable UMP-GALT allows single displacement of glu-1-P with release of free GALT but impairs the subsequent binding of gal-1-P and displacement of UDP-Gal.

The enzyme galactose-1-phosphate uridyltransferase (GALT) 1 (EC 2.7.7.12) catalyzes the conversion of UDP-glucose (UDP-Glu) and galactose-1-phosphate (gal-1-P) to form glucose-1-phosphate (glu-1-P) and UDP-galactose (UDP-Gal) in the evolutionarily conserved Leloir pathway of galactose metabolism (1). The enzymology of GALT has been intensively studied using GALT purified from Escherichia coli (2,3). Both E. coli and human GALT enzymes catalyze the conversion of UDP-Glu and gal-1-P to UDP-Gal and glu-1-P via a double displacement mechanism (2-4) (Fig. 1). Under normal physiological conditions, UDP-Glu binds to GALT to form a GALT-UDP-Glu intermediate. Glu-1-P is subsequently released, whereas the GALT enzyme remains bound to UMP. This GALT-UMP intermediate has been isolated and crystallized (5)(6)(7)(8). Gal-1-P then reacts with the GALT-UMP complex to form UDP-Gal, freeing the GALT enzyme for continued catalysis. GALT does not use nucleoside di-or triphosphates as nucleotidyl donor substrates, making it unique among nucleotidyltransferases that utilize phosphates as acceptor groups (9). The GALT proteins are evolutionarily conserved; with an overall amino acid identity between the human and E. coli GALT proteins of 46%. There is near 100% conservation of the amino acid sequence in the catalytic domain studied here (10,11).
The human GALT gene encodes 379 amino acids, contains 11 exons, and spans 4 kilobase pairs of genomic DNA (10). Exon VI is the most conserved domain, is composed of 19 amino acids, and is involved in these catalytic reactions (10). To date, 131 sequence changes have been identified in GALT genes from patients with galactosemia (12)(13)(14)(15). The most common mutation is a substitution of arginine for glutamine at amino acid 188 (Q188R). This mutation is caused by an A to G transition, changing CAG to CGG in codon 188 of the highly conserved exon VI (10,12). This Q188R mutation has a frequency of 70% among white newborns with galactosemia, and homozygosity for the Q188R mutation has been associated with poor prognosis of the disorder (16). The precise pathobiology of this mutation remains unknown. The Gln 188 is highly conserved and adjacent to three other conserved amino acids, histidine-proline-histidine at amino acids 184 -186, which provide a binding domain and source for the initial nucleophilic attack in the E. coli GALT (17).
A few human GALT mutations, including the Q188R mutation, have been characterized by various in vitro expression systems (12, 19 -21, 36). These studies described some kinetic parameters (apparent K M or V max ) of purified mutant proteins but did not explore the normal and mutant mechanisms involved in position 188 at the biochemical level. In this study, we used purified, recombinant human GALT proteins, computer modeling, biochemical assays, and kinetic analyses to clarify the biochemical and molecular mechanisms by which glutamine 188 contributes to the overall reaction and how the Arg 188 and Asn 188 substitutions alter human GALT function.

EXPERIMENTAL PROCEDURES
Computer Modeling of GALT Mutations-The computer software SYBYL (Tripos Associates) was used to visualize and model human GALT at the active site of uridylated-GALT crystal of E. coli (6 -8 ) gal dcm (DE3)] used in this expression study was purchased from Novagen. The gal operon of BL21(DE3) was disrupted by transposon mutagenesis to produce a "GALT-less" background for human GALT expression and purification in the transfected bacteria.
The human wild type Gln 188 -GALT cDNA was cloned into vector pGEM3zf(Ϫ) (Promega) (22) and was amplified by polymerase chain reaction with primers GALTPRO5 and GALTPRO3 (see Table I). The amplified polymerase chain reaction product was digested with restriction enzymes BamHI and HindIII and subcloned into the bacterial expression vector pTrcHisA (Invitrogen), which allowed high level of expression of the cloned Gln 188 -GALT cDNA in E. coli. Ampicillin resistance was conferred to the transformed bacteria E. coli BL21(DE3) strain, which enabled 100% selection. The resultant recombinant vector was named pKL1. To ensure that no mismatches were introduced during polymerase chain reaction, we sequenced the cloned Gln 188 -GALT cDNA twice in both forward and reverse orientations using protocols for the ABI Prism 310 TM Genetic Analyzer Sequencing System (Applied Biosystems). To amplify the entire GALT cDNA, we used primers EXON2FOR, EXON3REV, EXON5FOR, EXON6FOR, EXON6REV, EXON9FOR, and EXON10FR (see Table I).
We introduced the Q188R and Q188N mutations into the wild type Gln 188 -GALT cDNA sequence in pKL1 by the polymerase chain reaction-based site directed mutagenesis protocol developed by Stratagene (Quick Change TM ) and the corresponding mutagenesis primers (see Table I). The change from glutamine to asparagine required changes in two bases in the codon 188 (CAG 3 AAC). This was first achieved by site-directed mutagenesis of the wild type Gln 188 -GALT cDNA cloned in pKL1 to form pACWK using mutagenesis primers Q188KP1 and Q188KP2 (see Table I). It was then followed by the site-directed mutagenesis of the mutant GALT cDNA in pACWK to form pACWN using mutagenesis primers Q188NP1 and Q188NP2 (Table I).
After mutagenesis, the parental wild type plasmids were digested with DpnI, leaving only the mutant plasmids. Both the wild type and mutant plasmids were transformed into competent E. coli BL21(DE3) by the CaCl 2 method (23). The presence of the Arg 188 or Asn 188 mutation in the mutant plasmid was confirmed by restriction analyses and direct DNA sequencing of plasmid DNA. Both mutations eliminated the natural BstNI restriction site at codon 188.
Induction of GALT cDNA Expression-To induce expression of recombinant human GALT in transformed E. coli BL21(DE3), we cultured bacteria in 5 ml of LB broth containing 50 g/ml ampicillin overnight. The next morning 1 ml of the overnight culture was inoculated into 1 liter of fresh, identical medium. The culture was allowed to grow at 37°C until A 600 ϭ 1.0. Then, isopropyl-D-thiogalactoside (Sigma) was added at a final concentration of 1 mM to induce GALT synthesis, and the culture was incubated for an additional 3 h. Cells were then harvested by centrifugation at 6000 ϫ g for 10 min.
Purification of the Recombinant His 6 -tagged Human GALT Proteins-The bacterial expression vector, pTrcHisA cotranslates a hexamer of histidines (His 6 ) in frame at the amino terminus of the cloned GALT cDNA. A nickel-charged resin affinity column was used to purify recombinant GALT from bacterial cell lysates using modification of the manufacturer's method (Qiagen) (24,25). Lysates of bacterial pellets were incubated with nickel-charged resins at 4°C for 45 min with gentle shaking. Nonspecific binding of non-His 6 -tagged proteins were eluted with wash buffer containing 20 mM imidazole. The His 6 -tagged, Gln 188 -, Arg 188 -, or Asn 188 -GALT proteins were then eluted with 100 mM imidazole. Eluates were concentrated and desalted using Centricon-300 and resuspended in a small volume of glycine buffer (100 mM, pH 8.7). Protein concentrations were determined by Bio-Rad assay with bovine serum albumin standards.
Western Blot Analysis of Purified His 6 -tagged GALT Proteins-GALT proteins were identified by a rabbit anti-human GALT polyclonal antibody using a Western protocol described previously (26).
Assay of GALT Enzyme Activity-Assays for the overall, double displacement reaction were carried out at 37°C in 1 ml of glycine buffer (100 mM, pH 8.7) containing 0.6 mM UDP-Glu, 5 mM MgCl 2 , 5 mM dithiothreitol, 0.8 mM NADP, 1.2 mM gal-1-P, 5 M glucose-␣-1, 6-diphosphate, phosphoglucomutase (0.5 IU/ml), glucose-6-phosphate dehydrogenase (0.5 IU/ml) (27). The formation of NADPH was quantitated by absorbance change at 340 nm. The quantitative relationship between increase in NADPH production and glucose-1-phosphate released was quantitated using the Beer-Lambert equation: Absorbance FIG. 1. Double displacement reactions of GALT. Both E. coli and human GALT enzymes catalyze the conversion via a double displacement mechanism (2)(3)(4). Under normal physiological conditions, GALT binds to UDP-Glu to form a GALT-UDP-Glu intermediate. glu-1-P is subsequently released, whereas the enzyme remains bound to UMP. gal-1-P then reacts with the enzyme-UMP intermediate to form UDP-Gal, freeing the GALT enzyme for continued catalysis. k n and k Ϫn denote rate constants of the forward and reverse reactions.
ϭConcentration of solutes ϫ Optical path length ϫ Molar coefficient of extinction. In this study, the solute was NADPH, the optical path length was 1 cm, and the Molar coefficient of extinction for NADPH at 340 nm was 6220 M Ϫ1 cm Ϫ1 . We calculated ⌬Abs 340 (change in the absorbance) ϭ ⌬C (change in the concentration) ϫ 1 ϫ 6220.
To quantitate the single displacement reaction solely, gal-1-P was replaced with distilled water. All chemicals and enzymes were purchased from Sigma.
Mathematical Analyses of Rate Constants for the Single Displacement Reaction-The microcomputer program, SSREG.BAS (steadystate rate equation generator) (28) composed and generously provided by Dr. Robert Gunn (Physiology Department, Emory University) was used to analyze the rate constants for the single displacement reaction. This program is based on the algorithm of Indge and Childs (29), which is a modification of King and Altman's graphical method for derivation of rate equations (30).

RESULTS
Computer Modeling of Human Gln 188 -, Arg 188 -, and Asn 188 -GALT-To develop an hypothesis for the structural role of the conserved amino acid Gln 188 in human GALT, we modeled E. coli GALT-UMP coordinates at amino acid Gln 168 (Fig. 2). There are 20 more amino acids at the amino terminus of human GALT protein as compared with the E. coli GALT (10, 11). Thus, human Gln 188 is equivalent to amino acid Gln 168 of E. coli GALT (Fig. 2). We found that the amide group of the glutamine side chain provided two hydrogen bonds to the phosphoryl oxygens of UMP with lengths of 2.52 and 2.82Å (Fig.  3A). When arginine was substituted for glutamine at position 188 of the human GALT protein, the model identified only one hydrogen bond of 2.52 Å from the amide side chain of arginine (Fig. 3B). However, one could argue that although arginine substitution disrupted normal formation of the hydrogen bonds with UMP, it has a charge change that might alter the local GALT conformation. To control for the potential effects of this charge change, we substituted asparagine for glutamine at position 188. Asparagine is isofunctional with respect to glutamine, except its side chain is one carbon atom shorter. In addition, when we modeled the asparagine 188 mutation, only one hydrogen bond of 3.02 Å was formed (Fig. 3C). So if glutamine stabilizes and properly orients the GALT-UMP mainly by the two hydrogen bonds, we would expect asparagine substitution to show similar deleterious results as the natural mutation Q188R. Furthermore, if glutamine 188 stabilizes the GALT-UMP intermediate, we would expect that asparagine or arginine substitution for glutamine would interfere with gal-1-P binding and the subsequent release of UDP-Gal in the second displacement reaction but not the release of glucose-1-phosphate in the first reaction. To further investigate these hypotheses, we dissected and analyzed the two displacement reactions in vitro using purified Gln 188 -, Arg 188 -, and Asn 188 -GALT proteins produced in the transformed bacteria. Purified GALT enzymes, instead of whole cell extracts from the bacterial expression system, were necessary because bacterial extracts contained abundant gal-1-P (data not shown). To isolate and examine the first displacement reaction (see Fig. 5A), we must deplete the reaction of gal-1-P. If gal-1-P is present in the reaction, the products of the first displacement reaction will instantaneously react with gal-1-P and complete the second displacement reaction (Fig. 1).
The Bacterial Expression System for Human GALT-We assayed bacterial extracts of transformed ampicillin-resistant, GALT-less E. coli BL21(DE3) for induced GALT activity by measuring the overall double displacement reaction. Bacterial extracts containing wild type Gln 188 -GALT cDNA had an overall reaction rate of 1338.9 Ϯ 132.6 nmol glu-1-P formed/mg cell extracts/min (n ϭ 6) (see Table III). By contrast, when bacteria were transformed with Arg 188 -or Asn 188 -GALT cDNA, the overall reaction rate for their extracts was 3.2 Ϯ 1.18 or 18.75 Ϯ 6.6 nmol glu-1-P formed/mg cell extracts/min (n ϭ 6), respectively (see Table III). The difference between the results from the mutant Arg 188 -and Asn 188 -GALT were significant, with p ϭ 0.00534. These results were similar to the GALT biochemical phenotypes seen in erythrocytes from normal and galactosemic patients who are homozygous for the Q188R mutation (10,12). These results also show that the His 6 epitope tag fused at the amino terminus of the Gln 188 -GALT did not alter the catalytic function, which is in agreement with an earlier report that the amino terminus epitope-tagged, wild type Gln 188 -GALT was fully functional (31).
Purification of the Recombinant Human GALT Proteins-Next we purified the recombinant human Gln 188 -, Asn 188 -, and Arg 188 -GALT proteins by affinity chromatography (24, 25) (Table II). In Fig. 4A, the three left lanes next to the molecular mass markers contained Gln 188 -GALT (pKL1) eluted with increasing concentrations of imidazole (100, 150, and 250 mM). Two prominent Coomassie Blue-stained protein bands are seen in each successive lane. The upper band has the molecular mass of the dimeric form of GALT, and the lower band has the molecular mass of its monomer. These bands were not seen in the three right lanes, which were loaded with eluates prepared FIG. 2. Amino acid alignment of the predicted protein sequences of E. coli GALT (11) and human GALT (10). The alignment was performed by the modification of the method of Needleman and Wunsch (38). There is a 54% homology and a 46% identity between the two sequences. Two dots represent identical amino acids; one dot represents conservative substitutions.

FIG. 3. Computer modeling of human GALT mutations in the E. coli
GALT-UMP crystal. Computer software SYBYL was used to model human Q188R mutation in the highly conserved E. coli UMP-GALT protein crystal. The UMP moiety in the uridylated GALT was complexed to histidine 166, which is the equivalent of histidine 186 of the human GALT (see text). In A, the phosphorus of the UMP moiety was colored orange-yellow, and the oxygens were red. The side chain of glutamine 168, which was colored in orange-yellow, formed two hydrogen bonds of lengths 2.52 and 2.82 Å with the phosphoryl oxygens of the intermediate. In B, the amino acid glutamine was replaced with arginine and its side chain was colored red. The arginine side chain could only form one hydrogen bond of length 2.52 Å with the phosphoryl oxygen atoms of the intermediate. In C, the amino acid glutamine was replaced with asparagine, and its side chain was colored blue.
from bacteria transformed with control plasmid pTrcHisA containing no GALT cDNA inserts (the negative controls).
To confirm the identity of the purified proteins, we blotted an identical SDS-PAGE gel to a nytran membrane and performed Western blot analysis using an anti-human GALT polyclonal antibody (26) (Fig. 4B). Two protein species that had molecular masses of the two protein bands on SDS-PAGE reacted with our anti-human GALT polyclonal antibody (26). Because the anti-human GALT polyclonal antibody was raised in rabbits injected with freshly prepared GALT protein isolated from human red blood cells, our Western blot results suggest that recombinant human Gln 188 -GALT protein shares at least two similar epitopes with native human GALT from human erythrocytes (26). In subsequent kinetic analyses, we assume that both dimeric and monomeric forms of GALT are active (32,33). The Arg 188 -or Asn 188 -GALT were purified using the same procedure and gave identical results regarding protein mass by SDS-PAGE with immunoblotting (data not shown).
The Specific Activity of the Purified GALT Proteins-Purified GALT proteins were assayed for their capacity to catalyze the overall double displacement reactions in the presence of 600 M UDP-Glu and 1.2 mM gal-1-P. We found that the specific activity of the wild type Gln 188 -GALT was 80,300 Ϯ 5,910 nmol glu-1-P/mg GALT/min (n ϭ 3) (Table III). By contrast, the specific activities of mutant Arg 188 -and Asn 188 -GALT were 600 Ϯ 71.2 (n ϭ 3) and 2960 Ϯ 283.6 (n ϭ 3) nmol glu-1-P/mg GALT/min, respectively (Table III). Apparently arginine substitution had a more deleterious effect than asparagine substitution on bi-bi molecular GALT reaction.
The First Displacement Reaction-We devised a strategy to separate the first displacement reaction from the overall double displacement reactions through omitting gal-1-P in the reaction. By depleting the reactions of gal-1-P, we initially predicted that each molecule of GALT enzyme would only catalyze the release of the gal-1-P and form the GALT-UMP intermediate. If no gal-1-P was available for the second displacement reaction, the overall reaction would stall without recycling the GALT enzyme for another round (Fig. 5A). If this prediction were true, only one molecule of glucose-1-phosphate would be displaced by each molecule of GALT. Because glucose-1-phosphate was removed from the reaction by our enzyme linked assay, quantitation of glucose-1-phosphate released and NADPH subsequently produced should be stoichiometric. We would then expect a ⌬Abs 340 ϭ 1.0 to result from a change of NADPH concentration ϭ 1.0/6220 M ϭ 161 M in the reaction mixture, as dictated by the Beer-Lambert equation. Because this would predict only 1 mol of NADPH produced from each mole of glucose-1-phosphate released and because each mole of GALT could only displace 1 mol of glucose-1-phosphate, we would have to add 1.6 ϫ 10 Ϫ7 ϫ 47,000 (molecular mass of His 6 -tagged GALT) ϭ 7.56 mg of GALT enzyme to produce a change in absorbance of 1.0. However, if the GALT-UMP complex could dissociate to free UMP and GALT molecules and the free GALT enzyme could recycle only the first reaction with another UDP-Glu molecule, the single displacement of glucose-1-phosphate could continue (Fig. 5B). It would be possible under this second hypothesis to achieve a ⌬Abs 340 ϭ 1.0 with less than 7.56 mg of GALT enzyme.
From the results exemplified in Fig. 6, only microgram quantities of Gln 188 -GALT were required to quantitate the single displacement reaction. Each data point in Fig. 6 represents the mean value of three independent experiments where only 8 g of purified Gln 188 -GALT protein was used to assay the first displacement reaction. There was a steady, significant release of glucose-1-phosphate over 3 h as compared with the "negative control" (Fig. 6). The negative control was column eluate of the bacterial extracts prepared from bacteria, which were transformed with the expression vector without any GALT cDNA insert. According to the Beer-Lambert equation, a ⌬Abs 340 of 0.697 at 175 min would equal to a ⌬C ϭ 11.2 ϫ 10 Ϫ5 M. For ⌬C ϭ 11.2 ϫ 10 Ϫ5 M; glucose-1-phosphate produced in the reaction ϭ 11.2 ϫ 10 Ϫ8 mol ϭ 112 nmol. The 8 g of purified GALT protein added to the reaction was equivalent to only 170 pmol of GALT monomeric subunits (8 ϫ 10 Ϫ6 /47,000 ϭ 1.70 ϫ 10 Ϫ10 mol ϭ 170 pmol). Thus, some of the GALT-UMP intermediate must dissociate to free UMP and GALT and recycle a single displacement reaction releasing glucose-1-phosphate in the absence of gal-1-P. The rate of release of glucose-1-phosphate in the first displacement reaction was directly proportional to the amount of the wild type and mutant enzymes up FIG. 4. Purification of the recombinant human His 6 -tagged GALT proteins. A, the left three lanes next to the molecular mass markers in the Coomassie blue-stained SDS-PAGE gel were loaded with the eluates eluted by buffers containing increasing concentrations of imidazole (100, 150, and 250 mM). Two prominent protein bands representing the recombinant His 6 -tagged Gln 188 -GALT proteins were seen in these three lanes. The two proteins bands seen on the left three lanes were not seen on the right three lanes, which were loaded with eluates prepared by the same purification procedures, except the bacterial extracts were prepared from a bacteria strain transformed with the control vector (no GALT cDNA insert). B, a SDS-PAGE gel identical to the one shown in A was blotted to a nytran membrane, and Western blot analysis was performed on the nytran blot using a polyclonal anti-GALT antibody and protocol previously described (26). to 10 g in a time course of 5 h (data not shown). Eluate prepared from E. coli transformed with the expression vector alone without any GALT cDNA insert (negative control) had no significant increase in Abs 340 after 3 h of incubation in identical reagents, substantiating the specificity of the reactions (Fig. 6). When we used 14 C-labeled UDP-Glu as substrate in the single displacement reaction and separated the reaction products by DEAE-cellulose chromatography, we found that detectable quantities of [ 14 C]glucose-1-phosphate were formed from micrograms of GALT, supporting our spectrophotometric findings (data not shown). We then examined the first displacement reaction using the purified mutant Arg 188 -and Asn 188 -GALT enzymes. We found that the Asn 188 -and Arg 188 -GALT catalyzed the displacement of glucose-1-phosphate from UDP-Glu. In fact, the mutant GALT enzymes were faster than wild type Gln 188 -GALT (Fig.  6). Gln 188 -GALT produced 80 Ϯ 7.0 nmol glu-1-P/mg GALT/ min, whereas the Arg 188 -and Asn 188 -GALT enzymes displaced 170 Ϯ 8.0 and 129 Ϯ 28.4 nmol glu-1-P/mg GALT/min, respectively (Table III).
The Double Displacement Reactions-When gal-1-P was added to the wild type Gln 188 -GALT to complete the overall double displacement reactions, there was a 1,000-fold increase of glucose-1-phosphate released within a minute (Fig. 6 and Table III). The release of glucose-1-phosphate in the overall reaction was 80,030 Ϯ 5,910 nmol glu-1-P/mg GALT/min, whereas only 80 Ϯ 7.0 nmol glu-1-P/mg GALT/min was formed in the absence of gal-1-P (Table III). This suggested that the bi-bi molecular double displacement reaction was 3 orders of magnitude faster than the isolated single displacement reaction.
When gal-1-P was added to the reaction mixture containing the mutant Arg 188 -or the Asn 188 -GALT, there was a less significant increase in the overall reaction (Fig. 6), confirming the need for glutamine 188 to stabilize the UMP-GALT intermediate, bind gal-1-P, and accomplish the second displacement reaction rate (Table III).
Mathematical Analysis of Rate Constants for the Single Displacement Reaction-We analyzed the rate constants of the first single displacement reaction for both Gln 188 -and Arg 188 -GALT in the absence of gal-1-P (Fig. 5B). We used a microcomputer program (28) based on the algorithm of Indge and Childs (29) and King and Altman's original graphical method for derivation of rate equations (39). From this analysis, the overall rate of product formation at steady state is as follows.
does not contain the term k Ϫ2 , which is zero because all glucose-1-phosphate released was converted to glucose-6phosphate by phosphoglucomutase. Although slower than the overall reaction, the single displacement reaction was saturable when catalyzed by either Gln 188 -or Arg 188-GALT (Fig. 7). The V max for Arg 188 -GALT was greater than that of Gln 188 -GALT (294.1 nmol glu-1-P/mg GALT/min versus 149.3 nmole glu-1-P/mg GALT/min). From Equation 3, one sees that one criterion for the V max of Arg 188 -GALT to be greater than that of Gln 188 -GALT is that the rate constant k 5 of the conversion of the Arg 188 -GALT-UMP intermediate to free Arg 188 -GALT and free UMP would be greater than the k 5 of the conversion of the Gln 188 -GALT-UMP intermediate to free Gln 188 -GALT and free UMP. Thus by formal mathematical deduction, it takes longer for the Gln 188 -GALT-UMP intermediate to produce free GALT than the Arg 188 -GALT-UMP intermediate, and thus a lower capacity to recycle when the first displacement reaction is isolated. DISCUSSION GALT is of considerable interest to fundamental biochemistry and provides a unique insight into a bi-bi molecular, double displacement reaction (2, 3). Earlier site-directed mutagenesis 18.75 Ϯ 6.6 (n ϭ 6) 129 Ϯ 28.4 (n ϭ 3) 2,960 Ϯ 283.6 (n ϭ 3) FIG. 5. The single displacement model. A, in the absence of gal-1-P, the conventional viewpoint for the first displacement reaction would have predicted that only 1 mol of glucose-1-phosphate could be formed from each mole of GALT. The GALT enzyme will stall at the point of GALT-UMP intermediate. B, however, if free GALT could be released from the GALT-UMP intermediate, it could catalyze further rounds of the first displacement reaction. As a result, more than 1 mol of glucose-1-phosphate would be generated from each mole of GALT. k n and k Ϫn denote rate constants of the forward and reverse reactions, respectively. Glucose-1-phosphate released was quantitated by an enzyme linked assay by which 1 mol of NADPH was produced by each mole of glucose-1-phosphate release. experiments concluded that amino acids His 164 and His 166 of the E. coli GALT were crucial for initiating the first displacement reaction, but little is known regarding the requirements for the second displacement reaction (17). Meanwhile, a prevalent missense mutation, Q188R, found in the GALT genes of many galactosemic patients suggested an important catalytic role played by the highly conserved amino acid glutamine at position 188, which is equivalent to glutamine 168 in E. coli GALT (10,16,18,19).
Our initial examination of the structural role of the amino acid glutamine 168 (equivalent to glutamine 188 in human GALT) in the crystal structure of the E. coli GALT-UMP (5)(6)(7)(8) showed us that Gln 188 in the human GALT could stabilize the GALT-UMP intermediate through two hydrogen bonds formed between the amide side chain of Gln 188 and the phosphoryl oxygen of the UMP moiety (Fig. 3A). Because only one hydrogen bond could be formed in the cases of arginine or asparagine substitutions (Fig. 3, B and C), we subsequently proposed that the stabilization of the GALT-UMP intermediate is essential for the subsequent release of UDP-Gal in the second displacement reaction. This hypothesis was substantiated when we dissected and analyzed the two displacement reactions in vitro using purified recombinant Gln 188 -, Arg 188 -, and Asn 188 -GALT enzymes.
Our analysis of the first displacement reaction also showed that the GALT-UMP intermediate formed could undergo hydrolysis to form free GALT and UMP and recycle UDP-Glu to glucose-1-phosphate (Fig. 5B). Microgram quantities of GALT were capable of catalyzing this first displacement reaction, whereas milligram quantities would have been required if no FIG. 6. Exemplary experiments of first displacement and overall reactions by wild type and mutant GALT. A, the first displacement reaction. Duplicate samples containing 8 g of purified Gln 188 -(represented by closed circles), Arg 188 -(represented by closed triangles) or Asn 188 -GALT (represented by closed squares) were used in these experiments. At time 0, 0.6 mM UDP-Glu was added to the reaction. Glucose-1phosphate released was coupled to the formation of NADPH in the reaction mixtures, which was monitored spectrophometrically by measuring the change in absorbance at 340 nm. B, the overall reaction. The procedures are similar to those described for A except after 46 min, 1.2 mM gal-1-P was added to the samples, and the change in absorbance was monitored continuously. Each data point is the mean of two independent experiments. Negative control (represented by closed diamonds) was column eluate prepared from bacterial cells transformed with vector pTrcHisA alone (no GALT cDNA insert). recycling occurred. These data supported the earlier studies of the E. coli GALT defining nucleophilic His 166 N⑀2 attack by E. coli GALT on the GALT-bound UDP-Glu substrate and consequent displacement of glucose-1-phosphate to form the GALT-UMP intermediate (5)(6)(7)(8). A "covalent" bond was modeled between the UMP moiety and the amino acid His 166 of the E. coli GALT in the uridyl-GALT intermediate (5)(6)(7)(8). These investigators emphasized that there was a substantial positive charge build-up on the His 166 imidazolium ring (5)(6)(7)(8), suggesting that uncatalyzed hydrolysis of imidazolium-UMP between pH 5 and pH 10 could occur, and such uncatalyzed hydrolysis was further enhanced by the positive charge built up in the imidazolium group (37). Because we excluded gal-1-P, it was no longer available to this charged imidazolium group. Furthermore, glucose-1-phosphate was removed by the enzyme-linked reactions. Thus, the covalent bond formed by the zwitterionic UMP-imidazolium group was subjected to uncatalyzed hydrolysis resulting in the formation of free GALT and free UMP. To support this notion, we found that when the mutant Arg 188 -GALT was used, a higher V max for the glucose-1-phosphate release in the first displacement reaction was observed, supporting the thesis that Arg 188 -GALT forms a less stable GALT-UMP intermediate than Gln 188 -GALT, and therefore faster recycling than with Gln 188 -GALT was used (Fig. 7).
Our assay for the first displacement reaction using the purified wild type and mutant GALT also showed that the mechanism of dysfunction caused by the Q188R and Q188N substitutions did not impair binding of UDP-Glu or displacement of glucose-1-phosphate from GALT-UDP-Glu (Fig. 1). These results are in agreement with the crystallization studies of the E. coli GALT (35) and our kinetic analysis of the Arg 188 -GALT and Gln 188 -GALT of humans for the first displacement reaction (Fig. 7). In the E. coli GALT-UMP crystal, it was shown that Gln 168 was not involved in the initial binding of UDP-Glu (35). Our kinetic studies of the human GALT showed that both purified proteins were saturable with UDP-Glu, and although Arg 188 -GALT had a greater V max than Gln 188 -GALT, both enzymes had the similar K M of about 400 M (Fig. 7). Thus by computer modeling and kinetic analyses of the first displacement reaction, reduced hydrogen bonding was a rational explanation for the observed enhanced recycling capacity.
By contrast, when the overall reaction was analyzed through release of glucose-1-phosphate, the substitution of either arginine or asparagine for glutamine at amino acid 188 severely impaired GALT capacity (Table III). It thus appears that to accomplish the overall double displacement reaction, a stable GALT-UMP intermediate is required to bind gal-1-P and release UDP-Gal. This phenomenon is better accomplished by two hydrogen bonds to the phosphoryl moiety of UMP from the amide side chain of glutamine than by the single hydrogen bonds available from arginine or asparagine (Fig. 3).