Relationship of Conserved Residues in the IMP Binding Site to Substrate Recognition and Catalysis in Escherichia coliAdenylosuccinate Synthetase*

Gln34, Gln224, Leu228, and Ser240 are conserved residues in the vicinity of bound IMP in the crystal structure of Escherichia coli adenylosuccinate synthetase. Directed mutations were carried out, and wild-type and mutant enzymes were purified to homogeneity. Circular dichroism spectroscopy indicated no difference in secondary structure between the mutants and the wild-type enzyme in the absence of substrates. Mutants L228A and S240A exhibited modest changes in their initial rate kinetics relative to the wild-type enzyme, suggesting that neither Leu228 nor Ser240 play essential roles in substrate binding or catalysis. The mutants Q224M and Q224E exhibited no significant change in K m GTP and K m ASP and modest changes in K m IMP relative to the wild-type enzyme. However, k cat decreased 13-fold for the Q224M mutant and 104-fold for the Q224E mutant relative to the wild-type enzyme. Furthermore, the Q224E mutant showed an optimum pH at 6.2, which is 1.5 pH units lower than that of the wild-type enzyme. Tryptophan emission fluorescence spectra of Q224M, Q224E, and wild-type proteins under denaturing conditions indicate comparable stabilities. Mutant Q34E exhibits a 60-fold decrease in k cat compared with that of the wild-type enzyme, which is attributed to the disruption of the Gln34 to Gln224 hydrogen bond observed in crystal structures. Presented here is a mechanism for the synthetase, whereby Gln224 works in concert with Asp13 to stabilize the 6-oxyanion of IMP.

source. Crystal structures of Escherichia coli AMPSase in its ligand-free state (16,17) and its ligated state (4, 18 -20) are in the literature.
The chemical mechanism proposed for the synthetase by Lieberman (21) and Fromm (22) calls for the formation of 6-phosphoryl-IMP (19,23,24) as an intermediate in the catalytic reaction. O-6 of IMP putatively makes a nucleophilic attack on the ␥-phosphorus atom of GTP, whereupon the amino group of L-aspartate (ASP) displaces phosphate from C-6 of the intermediate. The direct observation of 6-thiophosphoryl-IMP in the active site of the synthetase (20) provides additional support for the proposed mechanism, which, on the basis of initial rate kinetics, is rapid equilibrium random ter ter (25). Isotope exchange studies with the rat muscle enzyme imply a preferred pathway for substrate addition, with aspartate adding to the enzyme after the binding of nucleotides (24). Competitive inhibitors for one of the substrates inhibit noncompetitively with respect to the other two substrates, suggesting that substrate binding sites do not overlap (Ref. 1 and references therein). Furthermore, adenylosuccinate (AMPS) is a competitive inhibitor of IMP and a noncompetitive inhibitor of GTP and ASP, a result consistent with overlapping AMPS and IMP binding sites (25).
Residues located in the vicinity of C-6 of IMP should be important in catalysis and/or substrate binding. Directed mutation, kinetics, and molecular modeling (19,26) revealed the significance of Arg 305 to the recognition of ASP and the stabilization of the transition state. The putative binding of Arg 305 to the ␣-carboxylate of ASP may bring the amino group of ASP into close contact with C-6 of IMP (19). However, Arg 305 is probably not a significant factor in the formation of 6-phosphoryl-IMP. Positions corresponding to 224 are occupied either by Gln or Asn in 12 of the 13 known sequences of AMPSase. The side chain of Gln 224 hydrogen bonds to N-7 and O-6 of IMP (19). On the basis of results presented below, Gln 224 is essential for catalysis. Furthermore, the mutation of Gln 34 to glutamate causes a 60-fold decline in k cat , attributed to the disruption of a hydrogen bond between Gln 34 and Gln 224 , that was observed in ligated crystal structures (19). In contrast, alanine mutations of Leu 228 and Ser 240 , which are conserved residues near bound IMP, had little effect on the kinetics of the synthetase.

EXPERIMENTAL PROCEDURES
Materials-GTP, IMP, L-aspartate, adenylosuccinate, phenylmethylsulfonyl fluoride, and bovine serum albumin were obtained from Sigma. A site-directed mutagenesis kit was obtained from Amersham Corp. Restriction enzymes were obtained from Promega. E. coli strain XL-1 blue was obtained from Stratagene. An E. coli purA Ϫ strain H1238 (thr-25, tonA49, argF58, purA54, argI61) and an E. coli purB Ϫ strain H680 (fhuA2, lacY1, tsx-70, glnV44(AS), gal-6, -, purB51, trpC45, his-68, tyrA2, rspL125(strR), malT1( R ), xylA7, mtlA2, thi-1) were obtained from Dr. B. Bachman, Genetic Center, Yale University. 0.45-m polyvinylidene difluoride membranes were obtained from Millipore Corp. Other reagents and chemicals were obtained from Sigma. * This research was supported in part by National Science Foundation Grants MCB-9218763 and MCB-9316244 and National Institutes of Health, U.S. Public Health Service Grant NS10546. This is journal paper number J-17328 of the Iowa Agriculture and Home Economics Experiment Station (Ames, IA), Project 3191. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Site-directed Mutagenesis-Recombinant DNA manipulation employed standard procedures (27). The plasmid containing a 1.8-kilobase pair BamHI-HindIII fragment from PMS204, ligated into PUC118, was used in mutagenesis. The mutagenesis primers were as follows: 5Јtgaccgccctcgtagcgtacaa-3Ј for Q34E, 5Ј-atccagcagcgtacccatcgcaccttc-3Ј for Q224M, 5Ј-atccagcagcgtaccctccgcaccttc-3Ј for Q224E, 5Ј-gatatccagcgccagcgtacc-3Ј for L228A, and 5Ј-gtgttggaagcagttacgtacg-3Ј for S240A. The underlined bases are changes from the original sequence. The primers used to sequence DNA in the vicinity of residues 224 -240 and 34 are 5Ј-ATGGCTGTTGCCGACATC-3Ј and 5Ј-GACGAAGGTA-AAGGTAAG-3Ј, respectively. All primers were synthesized on a Bioresearch 8570EX automated DNA synthesizer at the nucleic acids facility at Iowa State University. Mutagenesis was carried out according to the protocol provided by Amersham. The mutations were confirmed by DNA sequencing using the chain termination method (28) at the nucleic acids facility. The 1.8-kilobase pair BamHI-HindIII fragment with the right mutation was ligated back into PMS204, which was used in the transformation of XL-1 blue cells. The plasmids isolated from that strain were used to transform E. coli purA Ϫ H1238, the strain used for protein expression.
Protein Assay-Protein concentration, defined in terms of monomers, was determined by the method of Bradford (29), using bovine serum albumin as the standard. Molecular mass was determined by matrixassisted laser desorption/ionization time of flight mass spectroscopy (30).
Purification of Wild-type and Mutant AMPSase-The wild-type and mutant enzymes were purified using ammonium sulfate precipitation, phenyl-Sepharose CL-4B chromatography, Cibacron blue affinity chromatography, and DEAE-TSK high pressure liquid chromatography, as described elsewhere (31,32), but subject to the following modifications. Enzyme was eluted from a phenyl-Sepharose CL-4B column by a series of potassium phosphate-buffered solutions: 0.6 M, 0.4 M, and 0.2 M (NH 4 ) 2 SO 4 in 50 mM potassium phosphate (pH 7.0). The wild-type and most of the mutant enzymes appeared in the 0.2 M (NH 4 ) 2 SO 4 solution. A large fraction (80%) of the Q224E mutant, however, appeared in the 0.4 M (NH 4 ) 2 SO 4 solution. Enzyme purity was monitored by SDS-PAGE (33).
Western Blot-Because of the slower growth rate of transformed cells and the elevated expression of the Q224E mutant relative to the wildtype enzyme, we performed Western blots on wild-type, Q224M, and Q224E AMPSases to verify that our isolated mutant enzymes were indeed modifications of AMPSase. Two sets of 0.5 g of Q224M, Q224E, and wild-type AMPSases and 1 g of adenylosuccinate lyase (AMPS lyase) were separated by electrophoresis on a 12% SDS-polyacrylamide gel and then transferred to a polyvinylidene difluoride membrane. A rabbit polyclonal antibody (1:5000) and the standard detection techniques (AuroProbe BLPlus and intenSE BL system, Amersham) were used to visualize the signals.
Coupled Assay Method-UV difference spectra were obtained on a GBC UV/VIS 918 spectrophotometer equipped with a Peltier-Effect temperature controller, using the program Scan from 300 to 220 nm. The reaction buffer contained 20 mM Hepes (pH 7.7), 300 M GTP, 450 M IMP, 5 mM MgCl 2 , 5 mM ASP, and excess AMPS lyase. 1 g/ml wild-type AMPSase or 300 g/ml Q224E mutant enzyme was used in the scan. The spectra obtained at time 0 were subtracted from the spectra collected at subsequent time points (Fig. 1).
A coupled assay method monitoring the production of AMP employed UV difference spectroscopy. Excess AMPS lyase was used in the reaction mixture to convert all the AMPS into AMP. The increase in absorbance at 275 nm was monitored using a ⌬⑀ ϭ 15.0 ϫ 10 6 cm 2 /mol, which was derived by comparing the k cat values for wild-type enzyme using the assay method described elsewhere (34) and the method developed in this study.
Kinetic Study of the Wild-type and Mutant AMPSase-The concentrations of stock solutions for GTP and IMP were determined using their extinction coefficients at 253 and 248 nm, respectively. Enzyme assays contained 20 mM Hepes (pH 7.7), and 5 mM MgCl 2 . When GTP was the variable substrate, ASP was fixed at 5 mM, and IMP was fixed at 450 M. When IMP was the variable substrate, GTP was fixed at 300 M, and ASP was fixed at 5 mM. When ASP was the variable substrate, GTP was fixed at 300 M, and IMP was fixed at 450 M.
Circular Dichroism Spectrometry-Circular dichroism spectra for the wild-type enzyme and the mutant enzymes were acquired on a JASCO J700. Samples (100 g/ml) were placed in a 1-cm cuvette, and data were collected in increments of 0.1 nm. Each spectrum was corrected for background contributions of the buffer and smoothed. The data were analyzed by software associated with the spectrometer or by PSIPLOT.
Intrinsic Tryptophan Emission Fluorescence Measurements-Fluorescence measurements, using 1-cm quartz cuvettes, were carried out in a SLM 8000C spectrofluorometer. The proteins were excited at 290 nm, and emission spectra were recorded from 300 to 400 nm. Each sample contained 200 g/ml protein in 20 mM Hepes buffer (pH 7.7). Solutions containing 0 -1.5 N guanidium hydrochloride were used in denaturation experiments.
pH-dependent Kinetics-k cat versus pH profiles for the wild-type and the Q224E mutant were studied to gain additional data relevant to our proposals regarding the catalytic mechanism of the synthetase and the role in that mechanism played by the side chain at position 224. Mes and Hepes were chosen as buffers for the pH range 5.5-8.5, where the wild-type enzyme is active. The assay solution contained 300 M GTP, 5 mM MgCl 2 , 5 mM ASP, and 20 mM Mes/Hepes buffer at different pH values. IMP concentrations varied from 25 to 500 M. 300 g/ml Q224E mutant or 1 g/ml wild-type AMPSase with excess AMPS lyase were used in the assays. Longer assay times (5-10 min) compensated for the lower activity of the Q224E mutant enzyme. 34 , Gln 224 , Leu 228 , and Ser 240 -Sequences of AMPSase from different sources were aligned using the PILEUP option of the GCG sequence analysis package. Positions equivalent to 224 are occupied by asparagine in four sequences, glutamine in eight sequences, and arginine in the AMPSase from Brucella abortus, suggesting a significant role for the amide side chain at this position (Table I). Indeed, Gln 224 hydrogen bonds to the 6-oxo and N-7 of IMP in a crystal structure of the ligated synthetase (19). Position 34 is occupied by glutamine in 10 of 13 sequences. In ligated crystal structures of the E. coli enzyme, NE2 of Gln 34 hydrogen bonds to OE1 of Gln 224 , properly orienting the amide side chain of Gln 224 with respect to bound IMP (Fig. 2). Positions 228 and 240 are leucine and serine, respectively, in all known sequences of the synthetase. The side chain of Leu 228 packs against the base of IMP, whereas OG of Ser 240 hydrogen bonds to the 5Јphosphoryl group of IMP by way of a bridging water molecule.

Sequence Conservation of Gln
Mutagenesis, Expression, and Purification of AMPS Synthetases-All the enzymes reported in this study exhibited greater than 95% purity by SDS-polyacrylamide gel electrophoresis. The molecular mass of the synthetase by matrixassisted laser desorption/ionization time of flight mass spectroscopy was 46 kDa, consistent with the calculated molecular mass of its polypeptide chain. Q224E grew much slower on LB agar plates and in liquid media. Cultures yielded a cell mass of 1.5 g/liter, one-third of that produced by the wild-type strain. However, the yield of purified AMPSase was 15 mg/g of cells, a 15-fold increase relative to the wild-type strain. This phenomenon of reduced cell mass and increased enzyme expression has not been observed for any other mutant of AMPSase. No activity was detected using the assay method of Rudolph and Fromm (25) or Kang and Fromm (32). The isolated protein was identified as a mutant of AMPSase by Western blot and complementation experiments using H680 and H1238 as the recipient strains. The wild-type, Q224M, and Q224E enzymes were recognized by the antibody for AMPSase, whereas AMPS lyase was not (data not shown). In M9 minimal medium, the Q224M, and Q224E mutant plasmids complemented adenine auxotrophy of purA Ϫ in H1238 but not purB Ϫ in H680.
Secondary Structure Analysis-The CD spectra of the mutant and wild-type enzymes were superimposable (data not shown) from 200 to 260 nm. These observations indicated the absence of global conformational changes or the disruption of the secondary structure.
Kinetic Characterization of Wild-type and Mutant AMP-Sases-The kinetic parameters for wild-type and mutant AMP-Sases were determined in pH 7.7 Hepes buffer (Table II). No significant change in the K m values for GTP or ASP was observed for Q224E or Q224M enzymes relative to the wild-type enzyme. The Q224M mutant displayed a 5-fold increase in K m IMP , and a 13-fold decrease in k cat relative to the wild-type enzyme. On the other hand, the Q224E mutant, with only marginal changes in K m for all three substrates, exhibited a decrease in k cat of 4 orders of magnitude relative to wild-type AMPSase. Specificity constants (k cat /K m ) for the Q224M mutant decreased 16-, 75-, and 14-fold for GTP, IMP, and ASP, respectively, relative to the wild-type enzyme, compared with an almost 10 4 -fold decrease for specificity constants of the Q224E mutant for all substrates. These findings implicate Gln 224 not only in the recognition of IMP but also in a critical catalytic function. The catalytic role of Gln 224 had not been recognized previously on the basis of ligated crystal structures,  but it is in harmony with the suggestion of Poland et al. (19) that IMP binds to the active site as the 6-oxyanion (see below). The substitution of Gln 34 with glutamate resulted in a 60fold decrease in k cat . Atom N-E2 of Gln 34 hydrogen bonds with O-E1 of Gln 224 , which putatively orients Gln 224 in its interaction with IMP. The Q34E mutant may force a 180°rotation of the amide side chain of Gln 224 (see below). However, in the case of the Q34E mutant, one cannot rule out other explanations for the observed loss of activity, since the K m GTP also increased 20-fold relative to the wild-type enzyme. Gln 34 is about 10 Å away from the GTP site, suggesting a conformational change over a significant distance.
Alanine substitution of Leu 228 and Ser 240 resulted in 3.6-and 1.4-fold decreases in k cat , respectively, relative to the wild-type enzyme. The side chain of Leu 228 may contribute to the proper orientation of the base of IMP; however, Ser 240 apparently plays no significant role in the recognition of IMP or in catalysis.
Intrinsic Tryptophan Emission Fluorescence Measurements-The fluorescence emission spectra of the mutants, Q34E, Q224M, and Q224E, and the wild-type enzyme were collected in the presence of different concentrations of guanidium chloride. The emission maximum was 333 nm for the mutants and wild-type enzymes in the absence of guanidium chloride, and gradually shifted to 350 nm in the presence of increasing concentrations of guanidium chloride up to 1.5 N (data not shown). Mutations at residues 34 and 224, then, did not change protein stability in the presence of denaturants relative to the wild-type enzyme.
pH Effects on AMPSase Activity-Acting as hydrogen donors or acceptors, the amide side chains of glutamine and asparagine do not ionize, whereas glutamate and aspartate may undergo ionization. Replacing the amide group of Gln 224 with a carboxylate group resulted in a decrease in k cat of 4 orders of magnitude, whereas substrate affinity, as measured by K m , underwent only a modest change. We decided to study the pH-dependent kinetics for the wild-type and Q224E mutant enzymes to correlate enzyme activity with the state of protonation of Glu 224 . Titration curves for mutant and wild-type enzyme are typically bell-shaped for k cat versus pH profiles (Fig. 3). K m IMP for mutant and wild-type enzymes showed a maximum 2-fold change over the pH range 5.5-8.5. The optimum pH for wild-type AMPSase is 7.7, and that for the Q224E mutant is 6.2. At pH 6.2, only a 20-fold decrease in k cat was observed for the Q224E mutant relative to the wild-type enzyme. Protonation of the side chain of Glu 224 may be responsible for the increase in relative activity of the Q224E mutant as the pH decreases from 7.7 to 6.2. It is not possible, however, to rule out the suggestion that mutation of Gln 224 to Glu results in a localized environmental change that affects the pK value of another critical residue. DISCUSSION The properties of Q224M, Q224E, and Q34E mutants of AMPSase can be understood in relation to the crystal structure of the IMP⅐NO 3 Ϫ complex (19). For the wild-type enzyme (Fig.  4A), Asp 13 and Gln 224 putatively work in concert to stabilize the 6-oxyanion of IMP. Donor-acceptor distances between relevant atoms of Gln 34 , Gln 224 , Asp 13 , and IMP are consistent with the hydrogen bonding of Fig. 4A. Furthermore, a D13A mutant has no observed activity (35), consistent with its proposed role as a catalytic base.
The Q34E mutant could disrupt the hydrogen bond between Gln 224 and O-6 of IMP by enforcing a conformational change on Gln 224 (Fig. 4B). However, there is some uncertainty in this interpretation because the Q34E mutant has no effect on K m IMP but does significantly increase K m GTP . Thus, the decrease in k cat observed for the Q34E mutant may stem from conformational changes in addition to those represented in Fig. 4B.
As the K m values for substrates are not significantly perturbed by the mutation Q224M, the decrease in k cat most likely emanates from the loss of hydrogen bonds to N-7 and O-6 of IMP, the latter probably having the largest impact on catalysis (Fig. 4C). The greater reduction in k cat for the Q224E mutant than for the Q224M mutant may stem from the localization of a negative charge in proximity to O-6 of IMP (Fig. 4D). Partial restoration of activity at pH 6.2, then, is due putatively to the protonation of Glu 224 , which reforms the critical hydrogen bond to O-6 of IMP (Fig. 4D). The failure to restore the Q224E mutant to 100% of the wild-type enzyme activity at pH 6.2 may be a consequence of (i) tautomeric states that put the proton onto O-6 ( Fig. 4E) and/or (ii) the absence of the N-7 to Gln 224 hydrogen bond, an absence that may undermine the stability of the hydrogen bond between Glu 224 and O-6 of IMP.
The physiological importance of Gln 224 is supported further by the phenomenon of reduced cell proliferation in combination with enhanced protein expression for the mutant Q224E. Mutation of essential catalytic residues severely inhibits the growth of E. coli in the case of D13A and H41N mutants (35). The kinetic data imply that a disruption of the purine nucleotide pathway may impair the reproductive capacity of the organism. The enhanced level of expression of the Q224E mutant may be a self-rescuing strategy of E. coli, to compensate for the FIG. 3. k cat versus pH profiles of wild-type and Q224E mutant AMPSase. The k cat values for both enzymes were normalized for the clarity of comparison. The maximum k cat value is 1.59 s Ϫ1 for the wild-type enzyme at pH 7.7 and 9.31 ϫ 10 Ϫ3 s Ϫ1 for Q224E mutant enzyme at pH 6.2. impaired AMPSase and to promote adenylate synthesis by the purine nucleotide salvage pathway. It is unclear, however, why this mechanism of recovery is not operative in the case of the D13A or H41N mutant. Perhaps these mutants of AMPSase impair adenylate biosynthesis so severely that E. coli cannot sustain transcription and translation processes for overproduction of a given protein.