Biosynthesis of Bacterial Glycogen MUTAGENESIS OF A CATALYTIC SITE RESIDUE OF ADP-GLUCOSE PYROPHOSPHORYLASE FROM ESCHERICHIA COLI*

Site-directed mutagenesis was used to explore the role of Lys- 195 in ADP-glucose pyrophosphorylase from Escherichia coli. This residue, which is conserved in every bacterial and plant source sequenced to date, was originally identified as a potential catalytic site residue by covalent modification studies. Mutation of Lys-195 to glutamine produces an enzyme whose K,,, for glucose 1-phosphate is 600-fold greater than that measured for the wild-type enzyme. The effect on glucose 1-phosphate is very specific since kinetic constants measured for ATP, Mg2+, and the allosteric ac- tivator, fructose 1,6-bisphosphate, are unchanged relative to those measured for the wild-type enzyme. Furthermore, the catalytic rate constant, kcat, for the glutamine mutant is similar to that of the wild-type enzyme. Taken together, the results suggest a role for Lys- 195 in binding of glucose 1-phosphate and exclude its role as a participant in the rate-determining step(s) in the catalytic reaction mechanism. To further study the effect of charge, shape, size, and hydrophobicity of the amino acid residue at position 195, a series of mutants were prepared including arginine, histidine, isoleucine, and glutamic acid. In every case, the kinetic constants measured for ATP, Mg2+, and fructose 1,6-bisphosphate were similar to wild-type constants, reinforcing the notion that this residue is responsible for pmol or 1.0 pmol of sugar phosphate substrate, 1.0 pmol (wild-type) or 5.0 pmol (K195Q) of MgCI,, 0.3 pmol of fructose 1,6-bisphosphate, 20 pmol of Hepes buffer, pH 7.0, and 1 pg od inorganic pyrophosphatase (0.6 units). Reaction mixtures were pre-equilibrated at 37 “C before addition of a 10-111 aliquot of enzyme to start the reaction. Reactions were terminated at desired time intervals up to 30 min by boiling for 1 min, followed by transfer to ice. Reaction mixtures were analyzed for total ATP content using a hexokinase, glucose-6-phosphate dehydrogenase-coupled assay. 100 pl of each completed reaction was combined with 0.02% glucose, 2.25 mM NADP, 9 units of hexokinase, and 10 units of glucose-6-phosphate dehydrogenase in a final volume of 1.0 ml. After 15 min of incubation at 37 “C, absorbance at 340 nm was measured, using a blank treated as described above without addition of ADP-glucose pyrophosphorylase enzyme and without glucose-6-phosphate dehydrogenase. ATP content at zero time was determined by omitting ADP-glucose pyrophosphorylase only. Under the conditions with glucose 1-phosphate as substrate, loss of ATP was shown to be linear for wild-type enzyme during the first 8 min of reaction in which the starting 260 nmol of ATP decreased to 240 nmol. This corresponds to a decrease in ATP concentration of 10% and a decrease in glucose 1-phosphate concentration of 20%. A difference of 20 nmol of ATP could be reliably determined; the lower limit was a 6-nmol ATP difference corresponding to a of 0.02 in the assay. Reactions which resulted in less than 6 nmol of ATP consumption in a 30-min incubation had less than 0.01 units activity/ mg enzyme; this was the lower limit of detectable activity. Extent of reaction was controlled by varying the amount of enzyme used for assay.

for the first committed step in glycogen biosynthesis. As such, specifics of its active site chemistry are of interest from the point of view of understanding molecular details of allosteric regulation of catalysis.
What is currently known about this enzyme's catalytic site chemistry and mechanism can be briefly summarized. From steady-state kinetics and equilibrium binding studies substrates are known to bind in an ordered mechanism (1,2). ATP binds first with positive cooperativity, followed by noncooperative binding of glucose-1-P (2). Product release is ordered with pyrophosphate released first, followed by ADPglucose (1). The enzyme apparently obeys a rapid equilibrium mechanism in which substrates bind in rapidly reversible binding steps, followed by slower chemical steps to achieve the transition state since the value measured for Glc-1-PI K, from steady-state kinetics (3) is equal to the measured Glc-l-P dissociation constant, Kd in equilibrium binding experiments (2). Involvement of a covalent enzyme intermediate seems unlikely for this enzyme since Sheu et al. (4) demonstrated for the yeast enzyme, UDP-glucose pyrophosphorylase, that inversion of configuration of the a-phosphate of U T P occurs during formation of the product UDP-glucose. Also, no exchange of glucose 1-phosphate with ADP-glucose or pyrophosphate with ATP occurs in the absence of the second substrate. ' Little is known about catalytic site residues involved in substrate binding or catalysis for this enzyme. Covalent modification studies (5,6) originally identified two potential catalytic site residues, Lys-195 and Tyr-114, the latter shown to be involved in ATP and ADP-glucose binding (7,8). Reductive labeling of Lys-195 with pyridoxal phosphate resulted in loss of catalytic activity; ADP-glucose plus Mg2+ afforded protection against loss of activity (9). The importance of this residue to proper functioning of the enzyme is underscored by its conservation in every bacterial and plant source sequenced to date (10)(11)(12)." A number of roles can be envisioned for a lysine residue at this particular catalytic site. Lysine may assist in transition state stabilization as, for example, has been suggested for adenylate kinase from Escherichia coli (13). Alternatively, lysine may be necessary in the initial binding of either substrate, or it may function in concert with its neighboring amino acid residues to properly fold the protein so as to provide the correct geometry of catalytic residues. Site-directed mutagenesis experiments were therefore undertaken to attempt to distinguish between these possibilities. ' T h e abbreviations used are Glc-1-P, glucose 1-phosphate; Fru 1,6-P2, fructose 1,6-bisphosphate; DTT, dithiothreitol; Hepes, 4-(2-hydroxyethy1)-1-piperazineethanesulfonic acid.
Site-directed Mutagenesis-The gene for the native E. coli K12 ADP-glucose pyrophosphorylase was subcloned from the pOP12 plasmid (14) into M13mp18RF. The fragment used, the 1.9-kilobase HincII fragment of pOP12, contains the complete coding region for ADP-glucose pyrophosphorylase as well as the 474 base upstream region which encodes promoter site(s) necessary for expression of the gene (15).
Kunkel (16,17), which utilizes a dut-ung-strain of E. coli (CJ236) Site-directed mutagenesis was performed using the method of for synthesis of the M13 single-stranded template. Following in uitro synthesis and ligation of the mutagenic strand using mutant oligonucleotide as primer, it was transfected into an ung+ host (MV1193), and the resulting phage were screened by dideoxy sequencing in the region of the desired mutation. Mutant phage DNA were each sequenced through the entire coding region of the pyrophosphorylase gene to verify that the desired mutation was obtained without undesired changes in the remainder of the gene. The mutant oligonucleotides used are shown in Fig. 1. Expression and Purification of Mutant and Wild-type Enzymes-The glutamine mutant gene (K195Q) was subcloned into pUC19 a t the SmaI and PstI restriction sites. For expression of the K195Q enzyme, this plasmid was used to transform E. coli G6MD3, a mutant host with a genomic deletion of the glycogen biosynthetic genes. Cells were grown a t 37 "C to stationary phase in 5 liters of enriched medium containing 0.2% glucose and 50 pg/ml diaminopimelic acid and harvested by centrifugation to yield 20 g of cell paste.
The wild-type and mutant enzymes were purified according to previous procedures that were slightly modified (18,19). Cells were suspended in 100 ml 0.05 M glycylglycine, 5 mM DTT, 1 mM EDTA, pH 7.0, on ice. Cells were disrupted by sonication and centrifuged 7200 X g for 10 min in a refrigerated Sorvall RC-5B centrifuge. Potassium phosphate (1.5 M), pH 7.0, was added to the supernatant to give a final phosphate concentration of 30 mM. The supernatant was heated in a 70 "C water bath to a solution temperature of 60 "C, which was maintained for 5 min before cooling on ice. The sample was centrifuged at 10,000 X g a t 4 "C. Supernatant was loaded onto a DEAE-Sepharose column (13.5 X 2.7 cm), equilibrated with buffer A, 0.05 M potassium phosphate buffer, pH 7.0, containing 5 mM DTT, 1 mM EDTA, and 10% glycerol. The column was washed with 60 ml of 0.1 M potassium phosphate buffer, p H 7.5, containing 5 mM DTT and 10% glycerol. A linear gradient containing 200 ml of the above 0.1 M potassium phosphate buffer in the mixing chamber and 200 ml 0.2 M potassium phosphate buffer, pH 7.0, containing 0.5 M KC1, 5 mM DTT, 1 mM EDTA, and 10% glycerol was used to elute the enzyme from the column. Fractions containing activity were pooled, and solid ammonium sulfate was added to 70% saturation. Active enzyme, which precipitated under these conditions, was resuspended in 0.05 M Tris-HC1 buffer, pH 7.5, containing 1 mM EDTA, 0.5 mM DTT, and 10% glycerol, and dialyzed uersus the same. Dialyzed enzyme was loaded onto an Amicon Green A column ( 1.5 X 12 cm) equilibrated with 25 mM Hepes buffer, pH 7.0, 0.1 mM DTT, and 1 mM EDTA. The column was washed with three bed volumes of the equilibration buffer containing 20% (w/v ethylene glycol (buffer B). A step gradient beginning with four volumes of buffer B containing 0.3 M NaC1, followed successively with four volumes of buffer B containing 0.5, 1.0, and 2.0 M KC1 was used to elute enzyme. Active fractions were pooled, concentrated in an Amicon ultrafiltrator, and dialyzed uersus 0.05 M Tris-HC1, pH 7.5, containing 1 mM EDTA, 0.5 mM DTT, and 10% glycerol. Enzyme was then loaded onto a Mono Q H R 5/5 column using the Pharmacia fast protein liquid chromatography system. The column was equilibrated with 0.05 M Tris-HC1, p H 7.5, containing 1 mM EDTA and 0.5 mM DTT. Elution of the enzyme was achieved using a gradient of KC1 from 0 to 1 M in the equilibration buffer. At a flow rate of 0.5 ml/min, for the first two ml, the gradient used was 0-200 mM KC1. A gradient of 200-500 mM was used for the next 20 ml. The column was washed with 2 ml 500 mM KC1, followed by 1 M KCl. Active enzyme fractions were pooled and dialyzed uersus 0.05 M Tris-HC1 buffer, pH 7.5, containing 1 mM EDTA, 0.5 mM DTT, and 10% gycerol. The purification is summarized in Table I. Wild-type enzyme was purified from pOP12-transformed G6MD3 cells following essentially the same protocol. The purified and mutant and wild-type enzymes were judged to be homogeneous as only one protein band was detected on sodium dodecyl sulfate-polyacrylamide gel electrophoresis, having a molecular mass of about 50,000.
For expression of K195R, K195H, K1951, and K195E mutant enzymes, G6MD3 cells were grown a t 37 "C to a cell density of 5 X 10' cells/ml and inoculated with mutant phage at a multiplicity of infection of 10. Cultures were grown an additional 17 h a t 37 "C when cells were harvested by centrifugation. Purification of these mutant enzymes was taken through the DEAE-Sepharose step as described.
Enzyme Kinetics-Enzymatic activity in the ADP-glucose synthesis direction at 37 "C was measured according to the method of Preiss et al. (3). For assay of wild-type enzyme reaction mixtures (final volume 200 pl) contained 0.1 mmol of ['4C]glucose 1-phosphate (specific activity 500-1000 cpm/nmol), 0.3 pmol of ATP, 1.0 pmol of MgCl,, 0.3 pmol of fructose 1,6-bisphosphate, 20 pmol of Hepes buffer, pH 7.0, 100 pg of bovine serum albumin, and 1 pg (0.6 unit) of inorganic pyrophosphatase. For assay of mutant enzymes, the reaction components used were identical to wild-type, with the exception that 5.0 pmol of MgClz rather than 1.0 pmol was used. Fig. 2 illustrates the MgC1, saturation curve for both the wild-type and K195Q mutant enzymes. It is clear that a MgC1, concentration-dependent inhibition occurs in the wild-type but not in the K195Q case. MgC12 saturation curves for the other four mutants were virtually identical to that of the K195Q enzyme. In order to measure catalytic activity under optimal conditions, a MgC1, concentration of 5 mM was used for wildtype enzyme measurements, and 25 mM MgC12 was used for all reported mutant enzyme measurements. Control studies in which mutant enzymes were assayed a t 5 mM MgC1, yielded kinetic constants which were identical within the range of error to those obtained at 25 mM MgC1,; V,,, measurements were lower at 5 mM MgC12.
Enzymatic activity in the pyrophosphorylase direction at 37 "C was measured according to Preiss   a One unit of enzyme activity is expressed as the amount of enzyme required to form 1 pm of ATP/min at 37 "C assayed in the pyrophosphorolysis direction as described under "Experimental Procedures." ND, not determined. Kinetic Characterization-Kinetic data were plotted as initial velocity uersw substrate or effector concentration. Data were replotted as double-reciprocal plots and the method of Wilkinson (20) was used to determine V,,,. Sigmoidal plots were replotted as Hill plots to obtain kinetic constants. Kinetic constants from hyperbolic plots were determined by the method of Wilkinson. For sigmoidal data the following expressions for kinetic constants were used A0.5, l0.5, and S,,,, concentration of activator, inhibitor, or substrate, respectively, giving 50% Of maximal activation, inhibition, or maximal velocity. Duplicates were run in each case; kinetic constants are expressed as the mean f the difference from duplicate determinations.
Sugar Phosphate Specificity-The following assay was developed for measuring the activity of wild-type and K195Q mutant enzymes with respect to sugar phosphate substrates other than glucose 1phosphate. In a reaction volume of 200 pl was combined 0.26 pmol of ATP, either 0.1 pmol or 1.0 pmol of sugar phosphate substrate, 1.0 pmol (wild-type) or 5.0 pmol (K195Q) of MgCI,, 0.3 pmol of fructose 1,6-bisphosphate, 20 pmol of Hepes buffer, pH 7.0, and 1 pg od inorganic pyrophosphatase (0.6 units). Reaction mixtures were preequilibrated at 37 "C before addition of a 10-111 aliquot of enzyme to start the reaction. Reactions were terminated at desired time intervals up to 30 min by boiling for 1 min, followed by transfer to ice. Reaction mixtures were analyzed for total ATP content using a hexokinase, glucose-6-phosphate dehydrogenase-coupled assay. 100 pl of each completed reaction was combined with 0.02% glucose, 2.25 mM NADP, 9 units of hexokinase, and 10 units of glucose-6-phosphate dehydrogenase in a final volume of 1.0 ml. After 15 min of incubation at 37 "C, absorbance at 340 nm was measured, using a blank treated as described above without addition of ADP-glucose pyrophosphorylase enzyme and without glucose-6-phosphate dehydrogenase. ATP content at zero time was determined by omitting ADP-glucose pyrophosphorylase only.
Under the conditions with glucose 1-phosphate as substrate, loss of ATP was shown to be linear for wild-type enzyme during the first 8 min of reaction in which the starting 260 nmol of ATP decreased to 240 nmol. This corresponds to a decrease in ATP concentration of 10% and a decrease in glucose 1-phosphate concentration of 20%. A difference of 20 nmol of ATP could be reliably determined; the lower limit was a 6-nmol ATP difference corresponding to a of 0.02 in the assay. Reactions which resulted in less than 6 nmol of ATP consumption in a 30-min incubation had less than 0.01 units activity/ mg enzyme; this was the lower limit of detectable activity. Extent of reaction was controlled by varying the amount of enzyme used for assay.
Thermal Stability-Enzyme samples were diluted to give the same final protein concentration, 0.68 mg/ml. Dilution buffer was 50 mM Tris, 1 mM EDTA, 0.5 rnM DTT, 10% glycerol, pH 7.5. Potassium phosphate, pH 7.0, was added to give a final concentration of 30 mM phosphate. Individual samples (50 pl volume) of wild-type, K195Q, and K195E enzymes were heated simultaneously for 5 min in a water bath equilibrated at the specified temperature, then immediately placed on ice. They were assayed in the synthesis direction as described above.

RESULTS
Kinetic Characterization of K195 Mutant Enzyme-In the synthesis direction of assay, the glutamine mutant was characterized by a 600-fold greater glucose 1-phosphate K , than was measured for the wild-type enzyme ( Table 11). In contrast, kinetic constants for substrate ATP, cofactor M$+, and allosteric regulator, Fru l,6-P2 were similar in mutant and wildtype cases; no more than a 2-fold difference was measured for each of these ligands. The catalytic rate constant, kcat, is only 2-fold lower in the mutant. As indicated in the introduction, the Glc-1-P K, for wild-type enzyme is equal to the equilibrium dissociation constant, K d , for Glc-1-P (2). Observations that Glc-1-P K , for the glutamine mutant is dramatically altered while kcat is relatively similar to that of wild-type leads to the conclusion that Lys-195 is specifically involved in glucose 1-phosphate binding, with no role in the rate-determining step in the enzyme's catalytic mechanism. The small changes in ATP, Mg2+, and Fru l,6-P2 kinetic constants also indicate that replacement of Lys-195 with glutamine does not disturb the overall folding integrity of the enzyme, leaving those ligand-binding sites intact. There was an interesting difference between wild-type and K195Q in measurement of MgC12 dependence (Fig. 2). At high concentrations MgC12 acts as an inhibitor to the wild-type enzyme. This inhibition effect is completely absent in the K195Q mutant enzyme. Not pursued further here, the position 195 mutants will be valuable tools in future study of this inhibition effect. Because of this inhibition effect, kinetic constants reported in Tables I1 and IV were obtained under optimal conditions for wild-type (5 mM MgC12) and mutant enzymes (25 mM MgCl,). Identical kinetic constants for mutant enzymes were obtained using either 5 or 25 mM MgC12, indicating a lack of effect of the chosen concentrations of MgC12 on these measurements. V,,, measurements for the K195Q mutant, however, were slightly lower when measured  using 5 mM MgC12 than when measured using 25 mM MgC12. Considering the mutation's large effect on glucose l-phosphate K,, it was expected that ADP-glucose would also exhibit altered binding properties. Further, given the lack of effect of mutation on the ATP kinetic constant, it was not expected that pyrophosphate K , would be altered. Interestingly, both substrates' kinetic constants were increased 6-and 4-fold, respectively, in the glutamine mutant. (Table 11). kcat in the pyrophosphorolysis direction was decreased about 5-fold in the mutant. This is a similar result to the change in mutant kc,, observed in the synthesis direction.

Mutant Lys-195 ADP-Glu
Sugar Phosphates as Substrates for Wilal-type and K195Q Enzymes-Since Lys-195 appeared to be specifically involved in glucose 1-phosphate binding, it was of interest to test other sugar phosphates as substrates for the mutant enzyme. A variety of compounds were used whose sugar moieties differed from glucose stereochemically or by substitution or elimination of hydroxyl groups. From Table 111, all of the substrate analogues tested act extremely poorly as substrates in both wild-type and K195Q mutant enzymes. In no case was there a substrate analogue which showed relative enhanced reactiv-

:cose Pyrophosphorylases
ity with the mutant enzyme as compared to the wild-type enzyme. Thus, replacement of lysine 195 with glutamine does not show a broadened substrate specificity for sugar-1-P. Such a result may have indicated that the basis for lysine binding to glucose 1-phosphate involves specific interaction with a particular hydroxyl group of the sugar. However, the search was limited to those sugar-phosphates available commercially and at present does not represent a systematic analysis of the contribution of each specific hydroxyl group in the binding of substrate to catalytic site, whether mutant or wild-type. Lysine 195 however, may only participate in forming an ionic bond with the negatively charged phosphate of glucose 1phosphate.
Kinetic Characterization of Position 195 Mutants-To further test the effects of charge, shape, size and hydrophobicity on the ability of amino acid 195 to bind glucose 1-phosphate, a series of mutants was prepared, and kinetic constants were measured (Table IV). ATP, M e , and Fru 1,6-P2 kinetic constants are all relatively unaffected by the nature of the amino acid substitution at position 195. However a range of K, values was observed for glucose 1-phosphate which show a trend toward higher K, values as substitutions go from basic to neutral to an acidic amino acid. Substitution of Lys-195 by glutamic acid resulted in the largest glucose 1-phosphate K,,, observed, so large it could not be accurately measured; the largest GlclP concentration which could be used was 18 mM, a concentration which is 10 times lower than the K , for the glutamic acid mutant. Uncharged residues glutamine and isoleucine gave similar and more reliably measured glucose 1phosphate K, values. Of the series, arginine substitution resulted in the lowest glucose 1-phosphate K,. Even so, the arginine mutant K,,, was 100-fold higher than wild-type, indicating that charge alone is insufficient to confer binding strength in this case. Arginine, being slightly larger than lysine, may sterically prohibit proper binding of glucose 1phosphate. Results from histidine substitution were similar to arginine. The pK, of histidine at position 195 is not known, so it is difficult to assess the role of charge in this mutant. Taken together, the results indicate that a combination of effects are responsible for proper binding of glucose l-phosphate at position 195 including size and shape as well as charge. The trend in K , values from basic to neutral to acidic amino acid substitutions supports the hypothesis that an ionic bond between Lys-195 and the phosphate moiety of glucose 1-phosphate forms the basis of optimal binding affinity.
The inhibitor AMP shows a similar trend in Ki measurements to that of the glucose 1-phosphate K , measurements, although the magnitudes of the differences in AMP Ki values are smaller than were seen for Glc-1-P K, values (Table IV).
The observations that AMP but not ATP binding is affected by mutations at position 195 will be addressed under "Discussion." Thermal Stabilities of Wild-type, K195Q, and K195E Proteins-The thermal stability experiment (Fig. 3) was done under nonreversible conditions which do not allow thermodynamic parameters to be calculated; however, the data was gathered simultaneously for all three enzymes thereby allowing direct comparison of the three. Clearly, even for the worst case substitution in terms of effect on Glc-1-P K, (glutamic acid for lysine), the protein retains the same heat inactivation profile as the wild-type protein. Position 195 is obviously not critical to the stability of the native folded state. Evidence of this was also observed in the relatively uniform kinetic constants obtained for ATP, Mg"', and Fru 1,6-P2 for the five mutants, indicating a lack of effect of amino acid replacement at position 195 on these ligand-binding sites.

DISCUSSION
From the data presented here, Lys-195 binds glucose 1phosphate, probably in an ionic interaction between the negatively charged phosphate and its positively charged e-amino group. The possibility of its having a role in transition state stabilization is unlikely given the catalytic rate constant ( kcat) results which indicate that lysine is not critical to the ratelimiting step in the catalytic mechanism. From the thermal stability data and kinetic constants for ATP, M$+, and Fru 1,6-P2, Lys-195 is also not responsible for the maintenance of the native conformation of enzyme.
Kinetic results indicate that the nature of the substitution at position 195 affected AMP-or ADP-glucose binding, yet had virtually no effect on ATP binding. ATP may bind such that the triphosphate portion of the molecule anchors the aphosphate oxygen in an orientation away from the t-amino group of lysine 195. In such an orientation, the ATP"$+ complex does not interact with Lys-195 at all. In contrast, the a-phosphate of AMP or ADP-glucose may be rotated up into the region of Lys-195 so as to allow interaction between a negatively charged phosphate oxygen and positively charged c-amino group of lysine 195. The adenine portion of AMP and of ADP-glucose could occupy the same site on the protein that ATP binds. This agrees well with earlier kinetic and covalent modification work which suggested that the adenine moieties of ATP, ADP-glucose, and AMP share the same binding site but that the sugar-phosphate portions of these ligands may occupy unique sites (2,(21)(22)(23). It is also in agreement with steady-state kinetic data which shows AMP to act noncompetitively with respect to glucose 1-phosphate, indicating that both compounds can bind to the enzyme simultaneously. Moreover, earlier binding studies (2)  wild-type enzyme has shown that ATP binding exhibits halfsite reactivity with only two molecules of substrate binding to one molecule of the homotetrameric enzyme. However, ADP-glucose and AMP show full site binding suggesting a binding conformation/site for AMP and ADP-glucose different than that for ATP. The large increases in glucose 1-phosphate K,,, when position 195 is replaced by residues other than lysine explains the absolute conservation of lysine in all pyrophosphorylases sequenced to date. The shape, size, and charge properties of lysine are clearly necessary for proper functioning of the enzyme under physiological concentrations of glucose l-phosphate. This is an interesting example of a highly specific function for an amino acid residue at the catalytic site of an enzyme. There is a high degree of specificity for the size, shape, and charge of one amino acid, lysine, to adequately bind substrate. However, there is also a large degree of tolerance for amino acid substitutions in terms of the ability of the protein to maintain its native folding properties including maintenance of a closely neighboring binding site (ATPbinding site). Efforts are in progress to crystallize the pyrophosphorylase enzyme in both its activated and unactivated conformations. Structural analysis of these enzyme forms should provide more information on the feasibility of the conclusions drawn here, as well as highlight other catalytic site residues critical to binding and catalysis.