Conformational Change in Aspartate Aminotransferase on Substrate Binding Induces Strain in the Catalytic Group and Enhances Catalysis*

Aspartate aminotransferase has been known to un-dergo a significant conformational change, in which the small domain approaches the large domain, and the residues at the entrance of the active site pack together, on binding of substrates. Accompanying this conformational change is a two-unit increase in the p K a of the pyridoxal 5 (cid:1) -phosphate-Lys 258 aldimine, which has been proposed to enhance catalysis. To elucidate how the conformational change is coupled to the shift in the aldimine p K a and how these changes are involved in catalysis, we analyzed structurally and kinetically an enzyme in which Val 39 located at both the domain interface and the entrance of the active site was replaced with a bulkier residue, Phe. The V39F mutant enzyme showed a more open conformation, and the aldimine p K a was lowered by 0.7 unit compared with the wild-type enzyme. When Asn 194 had been replaced by Ala in advance, the V39F mutation did not decrease the aldimine p K a , showing that the domain rotation controls the aldimine p K a via the Arg 386 -Asn 194 -pyridoxal 5 (cid:1) -phos-phate linkage system. The maleate-bound V39F enzyme showed the aldimine p K a 0.9 unit lower than that of

The homodimeric structures have been solved for cytosolic (2), mitochondrial (3)(4)(5), and Escherichia coli (6,7) enzymes. Each subunit is composed of large and small domains, and the active site containing the PLP-Lys 258 aldimine is located between the two domains. The ␣-carboxylate group of the dicarboxylic substrates binds to Arg 386 located at the small domain and the -carboxylate group to Arg 292 * located at the large domain. An asterisk indicates that the residue comes from the neighboring subunit. On the basis of structural, steady-state (1), and transient (8,9) kinetic studies, the reaction mechanism of AspAT has been proposed (10,11) and refined (9) (see Scheme I).
The PLP-Lys 258 aldimine (internal aldimine) of the unliganded AspAT has an imine pK a value of ϳ6.5, which is strikingly lower than those of the PLP-amine aldimines with protonated pyridine N (ϳ10; see Ref 13). We have shown that the principle factor that decreases the aldimine pK a value of AspAT is the imine-pyridine torsion of the PLP-Lys 258 aldimine (as represented by the torsion angle C3-C4 -C4Ј-N, expressed as , in the panel of Scheme I), rather than the historically accepted electrostatic effect of the neighboring positive charges of Arg 292 * and Arg 386 . The torsion of the aldimine, the angle of which spans the range from ϳ35°(protonated aldimine) (2) to ϳ90°(unprotonated aldimine) (7), causes a decrease of 3 units in the aldimine pK a (14), whereas the positive charges of the arginine residues decreases the pK a by only 0.7 unit (15). The catalytic significance of the aldimine torsion is considered to be that it increases the energy level of the protonated form of the aldimine in the unliganded enzyme, thereby decreasing the 1 The abbreviations used are: AspAT, aspartate aminotransferase (aspartate, 2-oxoglutarate aminotransferase, EC 2.6.1.1); PLP, pyridoxal 5Ј-phosphate; PMP, pyridoxamine 5Ј-phosphate; V39F AspAT, mutant AspAT in which the residue Val 39 has been replaced with a phenylalanine residue (other mutant AspATs are expressed in the same way); WT, wild-type; MES, 4-morpholineethanesulfonic acid; TAPS, 3-{[2-hydroxy-1,1-bis(hydroxymethyl)ethyl]amino}-1-propanesulfonic acid. free energy gap between the starting state (the unliganded enzyme plus the free substrate) and the transition state leading to the first irreversible step, i.e. the release of the product oxaloacetate (see Scheme I). This decrease in the free energy gap is estimated to be 16 kJ mol Ϫ1 and increases the k cat /K m value for amino acid substrates by 10 3 -fold (14).
The intrinsic pK a of the external aldimine, i.e. the aldimine of PLP and the substrate aspartate, has been estimated to be Ͼ11 (14). The elevated pK a is considered to be caused by the aldimine being fixed to a near planar conformation ( ϭ Ϫ25°) (7) and the deprotonated form being destabilized (14). Accordingly, the external aldimine reduces the fraction of the dead end species H ϩ E L ϭS. The shift from H ϩ E L ϭS to E L H ϩ ϭS is favorable for the progress of the 1,3-prototropic shift (E L H ϩ ϭS 3 K in Scheme I), the first rate-determining step of the entire catalytic reaction (16).
In comparison to the case of the external aldimine, less is understood about the mechanism of controlling the aldimine pK a and its role in catalysis in the Michaelis complex, where PLP still forms the aldimine with Lys 258 (6). The intrinsic pK a of the aldimine has been estimated to be 8.8, by spectroscopic analysis of the maleate-bound enzyme, and it has been explained that the increase in the pK a is because of the electrostatic effect of the anionic ligand (17,18). However, taking into account the relatively weak electrostatic effect (0.7 pH unit) of Arg 292 * and Arg 386 on the aldimine pK a value as described above, we must consider other mechanisms for the upward pK a shift in this complex, such as alterations in the torsion angle of the aldimine and the hydrogen bond network surrounding it. Furthermore, although the increase in the aldimine pK a in the Michaelis complex has been claimed to accelerate the transaldimination step (E L H ϩ ⅐S 3 E L H ϩ ϭS in Scheme I) by shifting the equilibrium from E L ⅐SH ϩ to E L H ϩ ⅐S in the Michaelis complex (10), its exact role in catalysis is obscure, because the transaldimination step is not rate-determining at all in the catalytic reaction (16).
To solve these problems associated with the Michaelis complex, we set out to carry out structural and mechanical analysis of the complex. Upon formation of the Michaelis complex, AspAT undergoes a conformational change (2)(3)(4)(5)(6)(7), in which the small domain approaches the large domain. Together with the domain rotation, the side chains of the residues aligned at the entrance of the active site are reorganized and correctly packed to form a lid over the active site. These changes are expected to cause alterations in the conformation of the aldimine and the hydrogen-bonding pattern around the aldimine in the Michaelis complex. Therefore, we have chosen the mutation of Val 39 (19), which exists at the interface of the large and small domains, to a more bulky residue Phe, with the anticipation of hampering the domain rotation and side-chain packing. This mutant enzyme was analyzed together with the wild-type enzyme, to explore the effect of the substrate-induced conformational change on the important catalytic groups.

EXPERIMENTAL PROCEDURES
Materials-The medium used for bacterial growth contained 0.5% yeast extract, 1% peptone, and 0.5% NaCl, pH 7.0. All other chemicals were of the highest grade commercially available.
Site-directed Mutagenesis-The V39F (19) and N194A (18) mutants of pUC19 containing the E. coli AspAT gene (aspC) were prepared as described. The double mutant V39F/N194A was constructed using the QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA). The V39F plasmid was used as a template. The primers for mutagenesis used were 5Ј-CAT GGC TGC TGC CAT GCC CCA ACC GGT ATC-3Ј and 5Ј-GAT ACC GGT TGG GGC ATG GCA GCA GCC ATG-3Ј. The obtained mutant plasmids were expressed as described previously (20).
Spectroscopic Measurements-Absorption spectra were measured using a Hitachi U-3300 spectrophotometer (Tokyo, Japan) at 298 K. The buffer solution contained 50 mM buffer component(s) and 0.1 M KCl. The buffer components used were MES-NaOH, HEPES-NaOH, and TAPS-NaOH. The concentration of the AspAT subunit in solution was determined spectrophotometrically using the apparent molar extinction coefficient of ⑀ app ϭ 4.7 ϫ 10 4 M Ϫ1 cm Ϫ1 for the PLP form of the enzyme at 280 nm (20). The pK a value of the aldimine N of the PLP-Lys 258 Schiff base was calculated by fitting the data as follows in Equation 3.
Here, ⑀ E and ⑀ EH represent the molar absorptivity of the basic (E L ) and the acidic (E L H ϩ ) forms of the PLP form of the enzyme, respectively, at a fixed wavelength. Crystallization of the Mutant Enzymes-Crystallization of V39F AspAT was performed by the hanging drop vapor diffusion method. A 5-l drop containing 40 mg of protein/ml, 10 mM potassium phosphate, pH 7.0, 10 M PLP, and 0.3 mM NaN 3 was added to 5 l reservoir solution containing 10 mM potassium phosphate, pH 7.0, and 40% saturated ammonium sulfate and was equilibrated against 400 l of reservoir solution at 293 K. After 2 days, the drop was seeded with a small crystal of the wild-type (WT) AspAT obtained previously. Crystals with a size (about 0.8 ϫ 0.5 ϫ 0.4 mm) suitable for x-ray experiments were grown for 7 days. Crystals of V39F/N194A AspAT were obtained using the same procedure for V39F AspAT except that the reservoir solution contained 45% saturated ammonium sulfate. Crystals of maleate-bound V39F AspAT were obtained by a co-crystallization method, using the same procedure as that for V39F AspAT except that 1 l of 1 M maleate was mixed with 5 l of enzyme solution, and a small crystal of V39F AspAT was used for seeding. The diffraction data for V39F AspAT were collected on the BL6A station at the Photon Factory, High Energy Accelerator Research Organization, Tsukuba, Japan using an x-ray beam of wavelength 1.00 Å at 293 K and Fuji imaging plates with a screenless Weissenberg camera for macromolecular crystallography (21). Data collection was performed using two crystals. The diffraction data for V39F/N194A AspAT and maleate-bounded V39F AspAT were collected with a Rigaku R-AXIS IIc image plate detector mounted on a Rigaku RU-200 rotating anode generator operated at 40 kV and 100 mA with monochromatized CuK␣ radiation at room temperature. Each data collection was performed using one crystal. All data were processed and scaled using the programs DENZO and SCALEPACK (22) (see Table I). The diffraction data for V39F and V39F/N194A AspATs showed good isomorphism with those of WT AspAT; therefore, the coordinates of 1ARS (7) were used as the first model for refinement. The crystals of maleate-bounded V39F AspAT were isomorphous with those of WT AspAT and the 2-methylaspartate complex; thus, the coordinates of 1ART (7) were used as the first model for refinement. Simulated annealing and several subsequent rounds of least squares refinement using X-PLOR (23) were carried out. After each round of refinement, the model obtained was refitted to an electron density map using program O (24). The mutated residues and solvent molecules were modeled on the basis of 2F o Ϫ F c and F o Ϫ F c electron density maps. The solvent molecules whose thermal factors were above 100.0 Å 2 after refinement were removed from the model. The quality of the model for each structure was evaluated using PROCHECK (25). Table I contains a summary of the refinement statistics. The mean positional error of the atoms from the Luzzati plot was 0.32 Å for both V39F and WT AspATs.

RESULTS
Aldimine pK a Values-The absorption spectra of the V39F mutant AspAT (V39F) show an absorption maximum at 358 nm in the alkaline pH region and at 430 nm in the acidic pH region (data not shown), which are identical to those of the WT AspAT and correspond to the unprotonated (E L ; see Scheme I) and protonated (E L H ϩ ; see Scheme I) structures of the PLP-Lys 258 aldimine, respectively (13,17). The pH dependence of the apparent molar absorptivity at 430 nm is shown in Fig. 1. The aldimine pK a value of V39F is obtained by fitting the data to Equation 3 to be 6.1, which is 0.7 unit lower than that of WT (6.8; see Refs. 9 and 27). Because the aldimine pK a value is closely related to the presence of the hydrogen bond between PLP O3Ј and Asn 194 (14,15), the effect of the V39F mutation was studied on the N194A mutant AspAT (Fig. 1). The aldimine pK a value of the V39F/N194A double mutant AspAT is 8.7, essentially identical to that of the N194A mutant AspAT (pK a ϭ 8.6; see Ref. 18).
The binding of maleate to WT causes an increase in the 430-nm absorption band with a concomitant decrease in the 360-nm band, reflecting the increase in the aldimine pK a value (17). Similar spectral changes are observed for the binding of maleate to V39F (data not shown). The aldimine pK a value of the maleate-bound V39F (V39F⅐maleate) is 7.9 ( Fig. 1), which is 0.9 unit lower than the value 8.8 (18) of the maleate-bound WT (WT⅐maleate). The aldimine pK a values of WT and mutant AspATs in the presence and absence of maleate are summarized in Table II, together with the aldimine torsion angle () obtained from the crystallographic analysis.
Structure of V39F-The structure of V39F is shown in Fig Val 39 is one of the residues that locate at the entrance of the active site (6) ( Fig. 2A). In V39F, the space between the Tyr 70 * and Asn 69 * side chains cannot accommodate the bulky side chain of Phe 39 , and the phenyl ring is excluded from the space. The phenyl ring has a direct steric effect on the side chain of Ile 37 and pushes it outward from the active site. This displacement causes the movement of the main chain of the N-terminal part of the small domain outward of the active site, providing a more open conformation of V39F as described above ( Fig. 2A). Together with the rotation of the small domain induced by the 2 The amino acid residues are numbered according to the sequence of pig cytosolic aspartate aminotransferase (12). According to this numbering system, the N-terminal residue is Met-5.
where I ϭ observed intensity, and ͗I͘ ϭ average intensity for multiple measurements. SCHEME I V39F mutation, the side chain of Arg 386 is moved by 0.63 Å outward from the active site (see Fig. 2B and Fig. 4A). This movement caused a small shift (0.33 Å) in the position of the side chain of Asn 194 , which is hydrogen-bonded to the guanidinium group of Arg 386 . The entire structure of the V39F/ N194A double mutant AspAT is essentially identical to that of V39F (data not shown; see PDB 1IX8).
Structure of the Maleate-bound V39F-As in the case of WT, maleate forms a complex with V39F with its two carboxylate groups bound to Arg 292 * and Arg 386 , causing a significant rotation of the small domain (Fig. 3). Although the improper packing of the side chains around Phe 39 hampers the complete rotation of the small domain as observed for WT, the difference in the structure around Arg 386 is not significant between V39F⅐maleate and WT⅐maleate. As a result, the guanidinium groups of Arg 386 and Arg 292 *, maleate, and the side chain of Asn 194 of V39F⅐maleate are closely superimposable to those of WT⅐maleate (see Fig. 3 and Fig. 4B). A great difference between the two structures is observed in the region of residues 37-39 (Fig. 4B). On maleate binding to WT, Ile 37 moves 3.1 Å to cover the active site with a rotation of 120° (Fig. 4, comparison of A  and B). This movement drives the carbonyl O of Gly 38 to approach Tyr 225 O, causing the formation of a weak hydrogen bond (3.6 Å) between the two atoms. In V39F, on the other hand, the presence of the bulky phenyl ring of Phe 39 hampers the movement of the side chain of Ile 37 (Fig. 4, comparison of A  and B). The position and conformation of residues Ile 37 -Gly 38 -Phe 39 in V39F⅐maleate remain like those of the unliganded enzymes.
Reaction of V39F with 2-Methylaspartate-To determine the effect of the V39F mutation on the first two steps of the catalytic reaction, the substrate binding and the transaldimination, we analyzed the reaction of V39F with the substrate analogue 2-methylaspartate. The spectral transitions are essentially the same as in the case of WT (14), and the results are analyzed in the same way (data not shown). Briefly, a single exponential process of the spectral change is observed, and extrapolation to t ϭ 0 yielded a spectrum apparently different from that of the unliganded enzyme, indicating the presence of an enzymesubstrate complex rapidly formed from the unliganded enzyme and 2-methylaspartate. The observed single exponential spec-tral transition is considered to reflect the equilibration process between the rapidly formed species and another species. By combining the results and the chemical properties of 2-methylaspartate, which stops the catalytic reaction at the external aldimine (8), we assigned the two species to be the Michaelis complex (MC) and the external aldimine (EA), as shown in Equation 4 (14).
The apparent rate constant for the spectral change (k app ) is fitted to the equation  (19). To analyze the effect of the V39F mutation on the elementary steps, we studied the kinetic isotope effect of the substitution of ␣-1 H with 2 H on the kinetic parameters. The ratio of k cat for aspartate to k cat for [2-2 H]aspartate was 4.6 in V39F, compared with the value 2.0 of WT (20). Essentially no kinetic isotope effect was observed for K m (ratio is 1.0) in either enzyme (20) (this study).

DISCUSSION
Protein Conformation and the Aldimine pK a -The pK a values of dissociation groups at the active site of enzymes are generally considered to be regulated mainly by the polarity of the microenvironment and the electrostatic effects (including that of hydrogen bonds) of the surrounding residues. In this study on AspAT, we observed that the conformational change in the enzyme protein significantly affects the pK a of the PLP-Lys 258 aldimine at the active site. The V39F mutation causes movement of the small domain outward from the active site ( Fig. 2A). Together with this domain rotation, Arg 386 , the small domain residue located near the PLP-Lys 258 aldimine, moves away from the aldimine. Contrary to the expectation that the electrostatic effect of the positive charges of Arg 386 on the aldimine would be attenuated in V39F, the aldimine pK a of V39F is 0.7 unit lower than that of WT (see Fig. 1 and Table II). If Asn 194 , which bridges PLP O3Ј and Arg 386 , is removed in advance, the V39F mutation does not decrease the aldimine pK a (see Fig. 1 and Table II), although V39F/N194A shows an essentially identical conformation as V39F (comparison of 1IX6 and 1IX8). These results indicate that the effect of the displacement of Arg 386 on the V39F mutation is transmitted through Asn 194 to the PLP-Lys 258 aldimine and affects its pK a . A pos- . The ⑀ E value is set to zero, because the unprotonated aldimine has no absorbance over 400 nm (13). sible mechanism for the pK a shift is that the torsion angle of PLP-Lys 258 , which is the main factor controlling the aldimine pK a (14), is altered. However, because the resolution of V39F AspAT is not high (2.2 Å), the change in is not assessed structurally. The modeled structure of V39F AspAT (1IX6) has a value only 10°larger than that of WT AspAT (1ARS); the difference is not significant. However, it should be noted that what determines the internal aldimine pK a value is the value of the protonated PLP aldimine (14). This structure is not obtained under the crystallization conditions used for 1ARS and 1IX6 (pH 7). A model study indicates that the direction of the movement of the residues Arg 386 and Asn 194 increases the value of the protonated internal aldimine, by pulling the O3Ј atom to the re face of the aldimine. 3 This is consistent with the decreased pK a value of the aldimine.
Maleate Binding and the Conformational Change-The above discussion shows the importance of the PLP-Asn 194 -Arg 386 linkage system that controls the aldimine pK a . The structure of V39F⅐maleate shows that Arg 292 *, Arg 386 , Asn 194 , and maleate are almost superimposable on those of WT⅐maleate. Therefore, the contribution of the PLP-Asn 194 -Arg 386 system to the control of the aldimine pK a is the same between V39F⅐maleate and WT⅐maleate. Accordingly, the 0.9unit decrease in the pK a of V39F⅐maleate should be ascribed to alterations in the other part of the structure. The most remarkable change is seen for residues 37-39. In V39F⅐maleate, because of the steric hindrance of the Phe 39 side chain, Gly 38 carbonyl O does not approach Tyr 225 O (Fig. 4B)  ). In E L H ϩ ⅐S, the hydrogen bond lifts the Tyr 225 side chain and allows rotation of the PLP pyridine ring, reducing the aldimine torsion angle and thereby stabilizing the structure. As a consequence, the approach of Gly 38 to Tyr 225 on binding of maleate to WT is considered to destabilize E L ⅐SH ϩ relative to E L H ϩ ⅐S and increases the aldimine pK a in the Michaelis complex. The binding of maleate to V39F will not increase the aldimine pK a as much, because the movement of Gly 38 main chain on the maleate binding is restricted.
Effect of V39F Mutation on the Elementary Steps-For the analysis of the reaction of AspAT with aspartate at high pH, we consider the information shown in Equation 5 below (9), where MC, EA, and K are the Michaelis complex of the PLP form of AspAT and aspartate, the external aldimine, and the ketimine, respectively (for the structures, see Scheme I).
E M ⅐OA is the Michaelis complex of E M and oxalacetate. The study of Goldberg and Kirsch (16) showed that the ratedetermining steps in the forward reaction are the 1,3-prototropic shift between the external aldimine and the ketimine (k ϩ3 ), the ketimine hydrolysis (k ϩ4 ), and the release of oxalacetate (k ϩ5 ). Because the forward transaldimination rate (k ϩ2 ) is considered to be much larger than other rate constants (16), D k cat is expressed as shown in Equation 6, where k ϩ3 Because k ϩ2 and k Ϫ2 are insensitive to isotopic substitution, D k ϩ3 * is equal to D k ϩ3 . Using the observed D k cat value of 2.0 for WT (20) and 4.6 for V39F (this study), and the reported value of D k ϩ3 ϭ 7.3 for the primary kinetic isotope effect of the 1,3-prototropic shift (28), k ϩ3 *H /L is calculated to be 5.3 for WT and 0.75 for V39F. These values indicate that the V39F muta- tion decreased the k ϩ3 *H (and k ϩ3 *D ) value by 7.1-fold. Therefore, using Equation 7 (shown below), the reduction in the k cat value by the V39F mutation is calculated to be 2.0, which roughly coincides with the observed value of 2.5 (19).
Energetic Interpretation of the Acid-Base Chemistry and Kinetic Parameters-The results of the acid-base chemistry and kinetics of WT and V39F are interpreted in the three-dimensional free energy profile shown in Fig. 5B. In this free energy profile, the energy levels of all the reaction intermediates are arranged two-dimensionally; one coordinate displays the catalytic step, and the other shows the chemical species with different protonation states. The advantage of this expression is that we can understand energetically how the pK a of a dissociation group is involved in catalysis (14). For comparison of the WT-and V39F-catalyzed reactions, the energy level of the transition state is aligned in Fig. 5B.
In the Michaelis complex of WT, E L ⅐SH ϩ and E L H ϩ ⅐S have similar energy levels (14). On the V39F mutation, the free energy level of E L ⅐SH ϩ is decreased by 6.9 kJ⅐mol Ϫ1 whereas that of E L H ϩ ⅐S is decreased by only 1.7 kJ⅐mol Ϫ1 . The energy level of the external aldimine is unchanged. The obtained free energy diagram can explain 1) the 0.9-unit decrease in the aldimine pK a in the Michaelis complex, 2) the 9-fold equilibrium shift from the external aldimine to the Michaelis complex, and 3) the increase in the free energy gap between the ground state (the rapid-equilibrium mixture of the Michaelis complex and the external aldimine) and the transition state of the 1,3-prototropic shift, which results in the 7-fold decrease in k ϩ3 * . Thus, a change in essentially a single energy level (E L ⅐SH ϩ ) accounts for all the changes in the acid-base and kinetic parameters caused by the V39F mutation. This conclusion is consistent with the comparative structures of WT⅐maleate and V39F⅐maleate, the model for the Michaelis complex. That is, if the ligand-induced conformational change of residues 37-39 covering the active site and the concomitant approach of Gly 38 main chain to Tyr 225 is blocked by the V39F mutation, the enzyme is unable to destabilize the species E L ⅐SH ϩ . The reorientation of the PLP ring in the external aldimine lowers the side chain of Tyr 225 , and Tyr 225 O is distant from the carbonyl O of Gly 38 (2, 7) (1ART and 1AJS). We consider that the lack of the Gly 38 -Tyr 225 interaction in the external aldimine is the reason why the V39F mutation does not affect the energy levels of the external aldimine species.
Strain in the Michaelis Complex-Structural and kinetic analyses of the V39F mutant AspAT described as above provide important findings as follows. First, the conformation of As-pAT, i.e. the position of the small domain relative to that of the large domain, affects the aldimine pK a value, and the effect of the conformational change on the aldimine is mediated through the PLP-Asn 194 -Arg 386 linkage system. Second, despite the incorporation of the bulky Phe side chain to the domain interface, which causes outward rotation of the small domain, the position and conformation of the bound maleate and the residues that interact with it, i.e. Arg 292 *, Asn 194 , and Arg 386 , are the same as those in the WT⅐maleate, showing the flexibility of the small domain. Therefore, the lowered aldimine pK a of V39F⅐maleate as compared with that of WT⅐maleate is not because of the outward domain rotation but with the reduction in the Gly 38 -Tyr 225 interaction provided by the ligand-induced 4   Horizontal bars indicate the free-energy levels expressed in kJ⅐mol Ϫ1 . The labels attached to the bars are defined as shown in Scheme I. The horizontal coordinate shows the chemical species sorted by the protonation state. Perpendicular to this one is the reaction coordinate. The broken bars are those of V39F. The energy levels of the transition state are adjusted between WT and V39F. The energy levels are calculated based on the data obtained in this study and Ref. 14, corrected for the difference in the pK a value of the amino acid ␣-amino group (10.6 for MeAsp and 9.6 for aspartate). The rate of the 1,3-prototropic shift of WT is estimated to be 3500 s Ϫ1 , based on the consideration that this step is 16% rate-determining (calculated using the equation ( D k cat Ϫ 1)/ D k ϩ3 Ϫ 1); see Ref. 28) in the half-reaction.
packing of the residues at the active site. Together with the kinetic analysis of V39F, it was proposed that the Gly 38 -Tyr 225 interaction destabilizes the unprotonated aldimine species in the Michaelis complex (i.e. ground-state destabilization), and contributes to the increase in k cat . The role of the substrateinduced strain shown in this study is in contrast to that of the torsional strain of the PLP-Lys 258 aldimine in the unliganded enzyme, which reduces the energy gap between the starting state (unliganded enzyme plus unbound substrate) and the transition state ES ‡ by elevating the free energy level of the starting state and contributes to the increase in k cat /K m (14).