The K258R Mutant of Aspartate Aminotransferase Stabilizes the Quinonoid Intermediate”

Lys-258 of aspartate aminotransferase forms a Schiff base with pyridoxal phosphate and is responsi-ble for catalysis of the lY3-prototropic shift central to the transamination reaction sequence. Substitution of arginine for Lys-258 stabilizes the otherwise elusive quinonoid intermediate, as assessed by the long wavelength absorption bands observed in the reactions of this mutant with several amino acid substrates. The external aldimine intermediate is not detectable during reactions of this mutant with amino acids, although the inhibitor a-methylaspartate does slowly and stably form this species. These results suggest that external aldimine formation is one of the rate-determining steps of the reaction. The pyridoxamine-5’-phosphate-like enzyme form (330-nm absorption maximum) is un-reactive toward keto acid substrates, and the coenzyme bound to this species is not dissociable from the protein. and ease

' The abbreviations used are: PLP, pyridoxal-5'-phosphate; PMP, pyridoxamine-5'-phosphate; wild type, the wild type form of Escherichia coli aspartate aminotransferase; K258R and K258A, aspartate aminotransferase in which Lys-258 has been changed to arginine or alanine by site-directed mutagenesis; K258R-PLP and K258R-PMP, PLP and PMP forms of K258R, respectively; HEPES, N-2-hydroxyethylpiperazine-N"2-ethanesulfonic acid. al., 1987). For example, the reaction of erythro-0-hydroxyaspartate with aspartate aminotransferase produces a strong, long-lived quinonoid absorption band (Jenkins, 1964;Toney and Kirsch, 1991). Aspartate aminotransferase does not display a significant build up of the quinonoid intermediate in reactions with aspartate or glutamate (Fasella and Hammes, 1967;Gehring, 1986;Gehring et al., 1987) but does with the reactive substrate cysteine sulfinate.' It is possible for the central 1,3-prototropic shift to proceed without the formation of a quinonoid intermediate. Double kinetic isotope effect studies suggest that this occurs in the reaction of cytosolic aspartate aminotransferase with aspartate (Julin and Kirsch, 1989). The suggested mechanism is that Lys-258 donates a proton to C4' simultaneously with Ca proton abstraction.
Previous site-directed mutagenesis studies have conclusively demonstrated that the Schiff base-forming Lys-258 provides general base catalysis of the 1,3-prototropic shift central to the transamination reaction (Toney and Kirsch, 1989;Ziak et al., 1990;Toney and Kirsch, 1992). The substitution of arginine for Lys-258 produces a catalytically inactive enzyme (Morino et al., 1990) which, as demonstrated herein, undergoes half-reactions with amino acids in which the quinonoid intermediate is readily observed.

EXPERIMENTAL PROCEDURES
General"K258R was constructed as described previously for K258M; Toney and Kirsch, 1992. The protein was purified according to Cronin and Kirsch (1988). L-Cysteine sulfinic acid was purchased from Aldrich. All other chemicals were from Sigma.
Absorbance spectra and kinetics were measured with a Uvikon 860 spectrophotometer (Kontron Instruments) interfaced to an IBMcompatible personal computer. Curve fitting was performed with the nonlinear regression program ENZFITTER (R. J. Leatherbarrow, Biosoft Publishing Co.). All experiments were conducted at 25 "C in 0.1 M HEPES-KOH, pH 7.5, except as noted.
Kinetic Analysis of K258R Reactions-The spectral changes in the coenzyme were monitored in single turnover transamination halfreactions under pseudo-first order conditions with -10 @M K258R subunits and amino acid concentrations in large excess. Ionic strength varied from 0.05 to 0.07 M.
The reactions of keto acid substrates were monitored by periodic spectrophotometric scans of the coenzyme region over the course of the experiments, which lasted for 2 days (oxalacetate) to 2.5 months (a-ketoglutarate).
a-Methylaspartate Aldimine Formation-The spectrum of the K258R-a-methylaspartate aldimine was obtained after 10 min of reaction with 0.1 M a-methyl-DL-aspartate. The kinetics of aldimine formation with 50 mM a-methyl-DL-aspartate were monitored at 460 nm.
Resolution of K258R"Dissociation of 330 nm absorbing coenzyme from K258R was attempted by three methods: 1) precipitation with ammonium sulfate, pH 2, and allowing the precipitate to stand for 20 h, as described for K258A (Toney and Kirsch, 1992), 2) partial unfolding in 1 or 2 M guanidine hydrochloride (Herold and  1990) with excess PLP (the unfolding buffer consisted of 0.1 M HEPES-KOH, pH 7.5, 2 M guanidine hydrochloride, 1 mM dithiothreitol, 0.5 mM EDTA, 50 p~ PLP, and 0.5 mg/ml K258R in a total volume of 10 ml. After 2 h of reaction, the protein was dialyzed against 6 liters of 2 mM HEPES-KOH, pH 7.5, 1 mM dithiothreitol, 0.5 mM EDTA at 4 "C to refold the protein), and 3) dialysis (2 ml of 1 mg/ml K258R) against 2 liters of 0.1 M sodium hydroxide at 4 "C for 15 h.

K258R Reactions with Amino Acids Exhibit a Quinonoid Absorption
Band- Fig. 1 shows the coenzyme absorption spectra of K258R before and approximately 30 s after cysteine sulfinate was added to a final concentration of 5 mM. The addition of the amino acid yields a transient increase in absorbance centered a t 525 nm, consistent with a quinonoid intermediate (Kallen et al., 1985;Metzler et al., 1988), which decays to give a 330-nm band as the final product of the reaction. The K258R spectrum in the absence of substrate agrees with that previously reported (Kuramitsu et al., 1987) and is similar to that of freshly isolated K258A, which exhibits maxima at 334 and 395 nm (Toney and Kirsch, 1992).
All four substrates investigated (Table I)    absorbance at 405 nm follows a single exponential decay (Equation 1).

A,OS = (A" -A,)e-" + A,
The curve shown fitted to the 525-nm absorbance data is characteristic of that of an intermediate, B, in a three component, serial mechanism (Equations 2 and 3) (Moore and Pearson, 1981). The adjustable parameters in each of the above regression analyses were the absorbance (A,,, A,, A"'", and A'""", as appropriate) and rate constant values. The values of the rate constants obtained from the curve fitting are given in Fig. 2. Fig. 3 presents spectra of K258R before and 10 min after reaction with 100 mM a-methyl-DL-aspartate. Reaction with this inhibitor converts the initial 405-nm absorbance band to one of 430 nm, characteristic of an external aldimine. The observed pseudo-first order rate constant for formation of this external aldimine is (1.8 k 0.4) X s-I. This is 1100-fold less than the value of 2.0 f 0.15 s" observed with K258A and a-methyl-DL-aspartic acid under the same conditions (Toney and Kirsch, 1992).

K258R Forms a Stable Aldimine with a-Methylaspartate-
K258R Does Not React with a-Keto Acids-No changes in the coenzyme absorption spectrum of K258R-PMP were observed after mixing 20 p M enzyme with 100 mM oxalacetate or a-ketoglutarate and incubation for 2 days or 2.5 months, respectively.
The Coenzyme Form Absorbing at 330 nm Cannot Be Dissociated from the Enzyme-Three methods previously proven successful in removing coenzyme from wild type and other aspartate aminotransferase mutants' (Herold and Kirschner, 1990) failed to dissociate the 330 nm-absorbing form of the coenzyme from K258R, but were effective in dissociating that absorbing at 405 nm. Kuramitsu et al. (1987) reported that K258R denatured in 4 M guanidine hydrochloride retains the 330-nm absorption band after passage over a gel filtration column equilibrated with denaturant.

DISCUSSION
Appearance of the Quinonoid Intermediate in Reactions of K258R with Amino Acids-The quinonoid has been postulated to be an obligatory intermediate in all PLP-dependent reactions of amino acids (Metzler et al., 1954;Dunathan, 1971). The quinone-like structure of this species (Scheme 1) is strongly supported by the position (450-550 nm) and shape of its absorption bands (Metzler et al., 1988). The structure of the quinonoid formed from diethylaminomalonate and pyridoxal has been confirmed by NMR and IR spectroscopy (Abbott and Bobrik, 1973). The spectra presented in Fig. 1 clearly demonstrate the formation of a transient, long wavelength-absorbing intermediate, the quinonoid, in the reaction M. D. Toney and J. F. Kirsch, submitted for publication. of K258R with cysteine sulfinate.
Two mechanisms for quinonoid formation need consideration. The close proximity of the positively charged guanidino group of Arg-258 to Ca of the external aldimine intermediate might facilitate proton abstraction by solvent (which must occur in the reactions of K258A and K258M with amino acids in the absence of exogenous amines Kirsch, 1989, 1992). Alternatively, the free base form of Arg-258 might directly abstract the Ca proton. The independence of the value of kc,, for wild type on pH between 5 and 10 indicates that the pK, of the Lys-258 c-amino group in the external aldimine complex is less than 5 (Kiick and Cook, 1983), although the c-amino group of free lysine has a pK, value of 10.5. The Arg-258 guanidino group (pK, = 12.5 for free arginine) might have a similar environmentally induced -6 unit decrease in pK, value. Thus, at pH 7.5 (the pH employed) a significant fraction of the Arg-258 guanidino group might be in the reactive free base form. This latter mechanism of direct proton abstraction by Arg-258 is favored due to the similarity to the wild type reaction.
The higher stability of the quinonoid formed with K258R compared to wild type is explained by the greater basicity of the arginine versus the lysine side chain functional group. The quinonoid intermediate is an ion pair constituted by the positively charged functional group at position 258 and the carbanionic pyridoxyl moiety. Considered simply, the 100fold greater basicity of the Arg-258 guanidino group, compared to the Lys-258 €-amino group, will enhance the stability of this ion pair.
The kinetic traces shown in Fig. 2 demonstrate that the quinonoid species formed from K258R and cysteine sulfinate is kinetically competent (ie. it is formed and decays with rates that are no slower than the overall rate of conversion of aldimine to ketimine at any time point). This suggests, but does not prove, that the quinonoid is directly on the reaction pathway. An alternative kinetic mechanism is one where the quinonoid is formed off the reaction pathway in equilibrium with an intermediate on the pathway (Julin and Kirsch, 1989). The on-the-pathway intermediate in equilibrium with the quinonoid would most likely be the external aldimine since the ketimine absorbs near 330 nm (Toney and Kirsch, 1992) where a lag phase is observed. The equilibrium constant would, in this case, greatly favor the quinonoid over the external aldimine since no absorption for the latter (430 nm) is observed during the reactions.
The Reaction of K258R with a-Methyl-DL-aspartic Acid- The reactions of K258R with amino acid substrates do not generate detectable transient absorbance at 430 nm characteristic of external aldimines. Rather, the 405-nm absorbance of K258R-PLP converts directly to the quinonoid (525 nm) absorption band, although the chemical mechanism dictates an external aldimine precursor. The spectra of Fig. 3 demonstrate that the inhibitor a-methyl-DL-aspartic acid does form a stable external aldimine with the characteristic 430-nm absorption band (Kallen et al., 1985). The observed rate constant for this reaction is 1100-fold smaller than that measured for K258A, which is also unable to form an internal Schiff base linkage ("internal aldimine") with PLP, and is thus likewise incapable of employing the facile transimination mechanism utilized by wild type enzyme. The x-ray structure of K258R (Kamitori et al., 1990) suggests that this is due to a mispositioning of the PLP aldehyde function due to steric interactions between PLP and the arginine side chain. Thus the rate of decrease of A,,,, probably reflects rate-determining external aldimine formation and not Ca proton abstraction.