Aspartate aminotransferase with the pyridoxal-5'-phosphate-binding lysine residue replaced by histidine retains partial catalytic competence

The active site residue lysine 258 of chicken mitochondrial aspartate aminotransferase was replaced with a histidine residue by means of site-directed mutagenesis. The mutant protein was expressed in Escherichia coli and purified to homogeneity. Addition of 2-oxoglutarate to its pyridoxamine form changed the coenzyme absorption spectrum (lambda max = 330 nm) to that of the pyridoxal form (lambda max = 330/392 nm). The rate of this half-reaction of transamination (kcat = 4.0 x 10(-4)s-1) is five orders of magnitude slower than that of the wild-type enzyme. However, the reverse half-reaction, initiated by addition of aspartate or glutamate to the pyridoxal form of the mutant enzyme, is only three orders of magnitude slower than that of the wild-type enzyme, kmax of the observable rate-limiting elementary step, i.e. the conversion of the external aldimine to the pyridoxamine form, being 7.0 x 10(-2)s-1. Aspartate aminotransferase (Lys258----His) thus represents a pyridoxal-5'-phosphate-dependent enzyme with significant catalytic competence without an active site lysine residue. Apparently, covalent binding of the coenzyme, i.e. the internal aldimine linkage, is not essential for the enzymic transamination reaction, and a histidine residue can to some extent substitute for lysine 258 which is assumed to act as proton donor/acceptor in the aldimine-ketimine tautomerization.     

Reversible modification of amino groups in aspartate aminotransferase.

Amino groups in the pyridoxal phosphate, pyridoxamine phosphate, and apo forms of pig heart cytoplasmic aspartate aminotransferase (L-aspartate: 2-oxoglutarate aminotransferase, EC .2.6.1.1) have been reversibly modified with 2,4-pentanedione. The rate of modification has been measured spectrophotometrically by observing the formation of the enamine produced and this rate has been compared with the rate of loss of catalytic activity for all three forms of the enzyme. Of the 21 amino groups per 46 500 molecular weight, approx. 16 can be modified in the pyridoxal phosphate form with less than a 50% change in the catalytic activity of the enzyme. A slow inactivation occurs which is probably due to reaction of 2,4-pentanedione with the enzyme-bound pyridoxal phosphate. The pyridoxamine phosphate enzyme is completely inactivated by reaction with 2,4-pentanedione. The inactivation of the pyridoxamine phosphate enzyme is not inhibited by substrate analogs. A single lysine residue in the apoenzyme reacts approx. 100 times faster with 2,4-pentanedione than do other amino groups. This lysine is believed to be lysine-258, which forms a Schiff base with pyridoxal phosphate in the holoenzyme.     

Substitution of an arginyl residue for the active site lysyl residue (Lys258) of aspartate aminotransferase

The active site lysyl residue (Lys258) of E. coli aspartate amino transferase was substituted for an arginyl residue by oligonucleotide-directed, site-specific mutagenesis. The mutant enzyme was obviously unable to form an aldimine bond with pyridoxal 5'-phosphate but firmly bound the coenzyme. The finding that the mutation did not lead to entire loss in the enzymic activity suggests that Lys258 may not be essential but auxiliary for enzymic catalysis. It is also conceived that the positive charge provided by Arg258 may contribute to the enzymic catalysis.     

Neurospora tryptophan synthase. Characterization of the pyridoxal phosphate binding site

Tryptophan synthase, which catalyzes the final step of tryptophan biosynthesis, is a multifunctional protein that requires pyridoxal phosphate for two of its three distinct enzyme activities. Tryptophan synthase from Neurospora crassa, a homodimer of two 75-kDa subunits, was shown to bind 1 mol of pyridoxal phosphate/mol of subunit with a calculated dissociation constant for pyridoxal phosphate of 1.1 microM. The spectral properties of the holoenzyme, apoenzyme, and reconstituted holoenzyme were characterized and compared to those previously established for the heterotetrameric (alpha 2 beta 2) enzyme from Escherichia coli. The Schiff base formed between pyridoxal phosphate and the enzyme was readily reduced by sodium borohydride, but not sodium cyanoborohydride. The active site residue that binds pyridoxal phosphate, labeled by reduction of the Schiff base with tritium-labeled sodium borohydride, was determined to be lysine by high performance liquid chromatography analysis of the protein hydrolysate. A 5400-dalton peptide containing the reduced pyridoxal phosphate moiety was generated by cyanogen bromide treatment, purified and sequenced. The sequence is 85% homologous with the corresponding sequence obtained for yeast tryptophan synthase (Zalkin, H., and Yanofsky, C. (1982) J. Biol. Chem. 257, 1491-1500); the lysine derivatized by pyridoxal phosphate is located at the same relative position as that in the yeast and E. coli enzymes.     

Acylation of aspartate aminotransferase

1. Acetylation of aspartate aminotransferase from pig heart inhibits completely the enzymic activity when the coenzyme is in the amino form (pyridoxamine phosphate) or when the coenzyme has been removed, but not when the coenzyme is in the aldehyde form (pyridoxal phosphate). 2. The group the acylation of which is responsible for the inhibition has been identified with the in-amino group of a lysine residue at the coenzyme-binding site. Moreover, in the pyridoxamine-enzyme the amino group of the coenzyme is also acetylated. 3. The reactivity of the coenzyme-binding lysine residue is greatly different in the pyridoxamine-enzyme and in the apoenzyme, suggesting the possibility of an interaction of its in-amino group with pyridoxamine or with other groups on the protein.     

Stereochemistry of holoaspartate transaminase after modification of the active site Lys-258.

Proton incorporation at position C4 of the substrate-coenzyme Schiff base of aspartate transaminase is a stereospecific process. After carbamylation of the active site Lys-258, the stereospecificity of the reaction in 2H2O is retained. By a correlation method, it is shown that addition occurs from the si side of the complex and the pyridoxamine phosphate produced is deuterated at position pro-S of the pyridoxamine methylene group. These results constitute a demonstration for the stereochemstry of a half-transamination process of the phosphorylated coenzyme under single turnover conditions. They also illustrate that free Lys-258 is not required to maintain stereospecificity and cast doubts on the implication of this residue as a participant in C4 proton addition during catalysis by the native form of this mammalian enzyme.     

Specific labeling of the active site of cytosolic aspartate aminotransferase through the use of a cofactor analogue, N-(Bromoacetyl)pyridoxamine

The cofactor analogue N-(bromoacetyl)pyridoxamine (BAPM) has been employed to inactivate the cytosolic isozyme of apo-aspartate aminotransferase. Inactivation is the result of covalent bond formation in the (bromoacetyl)-pyridoxamine-transferase complex, via the epsilon-amino group of a lysyl residue at the active site. The stoichiometry of this inactivation is one molecule of (bromoacetyl)pyridoxamine per subunit of the transaminase dimer. Trace amounts of inorganic phosphate protect the enzyme from BAPM inactivation. In the absence of phosphate, inactivation demonstrates time, concentration, and pH dependence with an apparent pK for the target group of about 8.5 or higher. A tryptic peptide from the alpha subform has been obtained containing the carboxymethyl derivative of lysine-258, identifying this particular residue as the reactive group in the region of cofactor binding. Evidence is presented indicating that the pK of Lys-258 appears to be highly dependent upon the electrostatic state of neighboring groups in the active site region. Hence, experimentally obtained values vary according to the chemical nature and charge of the modifying agent or probe.     

Identification of amino acid residues modified by pyridoxal 5'-phosphate in Escherichia coli glutamine synthetase

Chemical modification studies with pyridoxal 5'-phosphate have indicated that lysine(s) appear to be at or near the active site of Escherichia coli glutamine synthetase (Colanduoni, J., and Villafranca, J. J. (1985) J. Biol. Chem. 260, 15042-15050; Whitley, E. J., Jr., and Ginsburg, A. (1978) J. Biol. Chem. 253, 7017-7025). Enzyme samples were prepared that contained approximately 1, approximately 2, and approximately 3 pyridoxamine 5'-phosphate residues/50,000-Da monomer; the activity of each sample was 100, 25, and 14% of the activity of unmodified enzyme, respectively. Cyanogen bromide cleavage of each enzyme sample was performed, the peptides were separated by high performance liquid chromatography, and the peptides containing pyridoxamine 5'-phosphate were identified by their absorbance at 320 nm. These isolated peptides were analyzed for amino acid composition and sequenced. The N terminus of the protein (a serine residue) was modified by pyridoxal 5'-phosphate at a stoichiometry of approximately 1/50,000 Da and this modified enzyme had full catalytic activity. Beyond a stoichiometry of approximately 1, lysines 383 and 352 reacted with pyridoxal 5'-phosphate and each modification results in a partial loss of activity. When various combinations of substrates and substrate analogs (ADP/Pi or L-methionine-SR-sulfoximine phosphate/ADP) were used to protect the enzyme from modification, Lys-352 was protected from modification indicating that this residue is at the active site. Under all experimental conditions employed, Lys-47, which reacts with the ATP analog 5'-p-fluorosulfonylbenzoyl-adenosine does not react with pyridoxal 5'-phosphate.     

Effects of D-serine on bacterial D-amino acid transaminase: accumulation of an intermediate and inactivation of the enzyme

Incubation of pure bacterial D-amino acid transaminase with D-serine or erythro-beta-hydroxy-DL-aspartic acid, which are relatively poor substrates, leads to generation of a new absorbance band at 493 nm that is probably the quinonoid intermediate. The 420-nm absorbance band (due to the pyridoxal phosphate coenzyme) decreases, and the 338-nm absorbance band (due to the pyridoxamine phosphate or some other form of the coenzyme) increases. A negative Cotton effect at 493 nm in the circular dichroism spectra is also generated. Closely related D amino acids do not lead to generation of this new absorption band, which has a half-life of the order of several hours. Treatment of the enzyme with the good substrate D-alanine leads to a small but detectable amount of the same absorbance band. D-Serine but not erythro-beta-hydroxyaspartate leads to inactivation of D-amino acid transaminase, and D-alanine affords partial protection. The results indicate that D-serine is a unique type of inhibitor in which the initial steps of the half-reaction of transamination are so slow that a quinonoid intermediate with a 493-nm absorption band accumulates. A derivative formed from this intermediate inactivates the enzyme.     

The ionization states of the 5'-phosphate group in the various coenzyme forms bound to mitochondrial aspartate aminotransferase

We have carried out a Fourier transform infrared spectroscopic study of mitochondrial aspartate aminotransferase in the spectral region where phosphate monoesters give rise to absorption. Infrared spectra in the above-mentioned region are dominated by protein absorption. Yet, below 1020 cm-1 protein interferences are minor, permitting the detection of the band arising from the symmetric stretching of dianionic phosphate monoesters [T. Shimanouchi, M. Tsuboi, and Y. Kyogoku (1964) Adv. Chem. Phys. 8, 435-498]. The integrated intensity of this band in several enzyme forms (pyridoxal phosphate, pyridoxamine phosphate, and sodium borohydride-reduced, pyridoxyl phosphate form) does not change with pH in the range 5-9. This behavior contrasts that of free pyridoxal phosphate (PLP) and pyridoxamine phosphate (PMP) in solution, where the dependence of the same infrared band intensity with pH can be correlated to the known pK values for the 5'-phosphate ester in solution. The integrated intensity value of this infrared band for the PLP enzyme form before and after reduction with sodium borohydride is close to that given by free PLP at pH 8-9. These results are taken as evidence that in the active site of mitochondrial aspartate aminotransferase the 5'-phosphate group of PLP remains mostly dianionic even at a pH near 5. Thus, it is suggested that the chemical shift changes associated with pH titrations of various PLP forms reported in a previous 31P NMR study of this enzyme [M. E. Mattingly, J. R. Mattingly, and M. Martinez-Carrion (1982) J. Biol. Chem. 257, 8872] are due to the fact that the phosphorus chemical shift senses the O-P-O bond distortions induced by the ionization of a nearby residue. Since no chemical shift changes were observed in pH titrations of the PMP forms (lacking an ionizable internal aldimine) of this isozyme, the Schiff base between PLP and Lys-258 at the active site is the most likely candidate for the ionizing group influencing the phosphorus chemical shift in this enzyme.     

Chemical modification of the lysine residues of bacterial formate dehydrogenase

Inactivation of formate dehydrogenase by formaldehyde, pyridoxal and pyridoxal phosphate was studied. The effects of concentrations of the modifying agents, substrates, products and inhibitors on the extent of the enzyme inactivation were examined. A complete formate dehydrogenase inactivation by pyridoxal, pyridoxal, phosphate and formaldehyde is achieved by the blocking of 2, 5 and 13 lysine residues per enzyme subunit, respectively. The coenzymes do not protect formate dehydrogenase against inactivation. In the case of modification by pyridoxal and pyridoxal phosphate a complete maintenance of the enzyme activity and specific protection of one lysine residue per enzyme subunit is observed during formation of a binary formate-enzyme complex, or a ternary enzyme--NAD--azide complex. One lysine residue is supposed to be located at the formate-binding site of the formate dehydrogenase active center.     

Stereochemistry and mechanism of a new single-turnover, half-transamination reaction catalyzed by the tryptophan synthase alpha 2 beta 2 complex

Tryptophan synthase is a versatile enzyme that catalyzes a wide variety of pyridoxal phosphate dependent reactions that are also catalyzed in model systems. These include beta-replacement, beta-elimination, racemization, and transamination reactions. We now show that the apo-alpha 2 beta 2 complex of tryptophan synthase will bind two unnatural substrates, pyridoxamine phosphate and indole-3-pyruvic acid, and will convert them by a single-turnover, half-transamination reaction to pyridoxal phosphate and L-tryptophan, the natural coenzyme and a natural product, respectively. This enzyme-catalyzed reaction is more rapid and more stereospecific than an analogous model reaction. The pro-S 4'-methylene proton of pyridoxamine phosphate is removed during the reaction, and the product is primarily L-tryptophan. We conclude that pyridoxal phosphate enzymes may be able to catalyze some unnatural reactions involving bound reactants and bound coenzyme since the coenzyme itself has the intrinsic ability to promote a variety of reactions.     

31P nuclear-magnetic-resonance studies of pyridoxal and pyridoxamine phosphates. Interaction with cytoplasmic aspartate transaminase.

The 31P nuclear magnetic resonance (NMR) spectrum of the phosphate in free pyridoxal or pyridoxamine phosphate reveals a resonance signal that is coupled to the methylene protons of the 5-CH2 with JHP of 6.0 +/- 0.3 Hz. Proton noise decoupling results in a single signal with a pH-dependent chemical shift with deprotonation of the phosphate resulting in a shift of the 31P resonance to lower fields. A single 31P NMR signal at a frequency corresponding to fully ionized phosphate monoesters is observed in aspartate-transaminase-bound pyridoxal or pyridoxamine phosphate. The 31P resonance in the holotransaminase is pH-independent and is unaffected by saturating concentrations of substrates or inhibitors. Only denaturation with 6 M guanidine with HCl results in changes in the 31P of the holoenzyme. It appears that the phosphate group of pyridoxal phosphate is bound to a positive pocket in the holoenzyme and remains fully ionized in the pH range of 5.6 to 9.2. The phosphate-binding properties are present even in the apoenzyme which is able to bind inorganic phosphate which then can be displaced by pyridoxal or pyridoxamine phosphate in the process of holoenzyme formation.     

The separate effects of coenzyme components may not be additive. Roles of pyridoxal and inorganic phosphate in aspartate aminotransferase apoenzymes

Both cytosolic and mitochondrial aspartate transaminase can be resolved of pyridoxal phosphate. The resulting apoenzymes still bind individual structural components of the coenzyme. The separate contributions of coenzyme components to protein thermal stability have been independently assessed for phosphate ions (Pi) and for the pyridoxal or pyridoxamine components of the coenzyme. 31P NMR and differential scanning calorimetry reveal that the thermodynamic contributions of binding are not additive and are dissimilar for the two isozymes. High and low affinity sites for Pi binding are present in both apoenzymes with only the low affinity site being present in the holoenzyme forms. The contribution of both bound phosphates to increasing temperatures (Tm) and enthalpies (delta Hd) of denaturation differ between the isozymes and within sites. In either isozyme occupancy of the high affinity site by Pi produces only a 4- or 5- degree increase in the Tm value with respect to Pi-free apoenzyme. By contrast, in the mitochondrial apoenzyme, the presence of Pi at the second low affinity site increases the calorimetric parameters from Tm = 47 degrees C and delta Hd = 4.7 cal g-1 to Tm = 62 degrees C and delta Hd = 7 cal g-1. For cytosolic apoenzyme the respective changes are from 66 to 69.5 degrees C and 5.2 to 5.8 cal g-1. Addition of pyridoxal, but not pyridoxamine, displaces the high affinity Pi in both apoenzymes. This shows that the pyridine ring and Pi groups of pyridoxal-P bind exclusive of each other when they are not covalently linked as an ester, as in the coenzyme. The observation has been exploited as a method to prepare completely dephosphorylated mitochondrial apoenzyme. Electrostatic effects, structural differences in the phosphate binding pockets, and steric effects can be invoked to account for the Pi and pyridine binding behavior in the two proteins.     

Chemistry of the inactivation of 4-aminobutyrate aminotransferase by the antiepileptic drug vigabatrin

The chemical modification of pig liver 4-aminobutyrate aminotransferase by the antiepileptic drug 4-aminohex-5-enoate (Vigabatrin) has been studied. After inactivation by 14C-labeled Vigabatrin, the enzyme was digested with trypsin, and automated Edman degradation of the purified labeled peptide gave the sequence FWAHEHWGLDDPADVMTFSKK. Chymotryptic digestion of the tryptic peptide and sequencing of a resulting tripeptide identified the penultimate lysine residue of this peptide as the site of covalent modification. This lysine normally binds the coenzyme. Absorption spectroscopy demonstrated the absence of coenzyme from the tryptic peptide, and mass spectrometry showed its mass/charge ratio to be increased by 128. All of the bound coenzyme released after denaturation of the inactivated enzyme was as pyridoxamine phosphate. The structural nature of the modification is deduced, and mechanisms for its occurrence identified. Initially, 1 mol of radiolabeled inhibitor was bound per mol of monomer of the enzyme, although approximately half was released during denaturation and digestion, while the remainder was irreversibly bound. Coenzyme not released as pyridoxamine phosphate retained the absorbance characteristics of the aldimine, although the enzyme was completely inactive. Mass spectrometry of the sample of purified radiolabeled tryptic peptide revealed the presence of an approximately equal amount of a second fragment that contained no modification and from which the second lysine was absent, indicating that at the time of proteolysis the active site lysine was unaltered in 50% of the enzyme molecules.     

Modification of pyruvate, phosphate dikinase with pyridoxal 5'-phosphate: evidence for a catalytically critical lysine residue

Pyruvate, phosphate dikinase from Bacteroides symbiosus is strongly inhibited by low concentrations of pyridoxal 5'-phosphate. The inactivation follows pseudo-first-order kinetics over an inhibitor concentration range of 0.1-2 mM. The inactivation is highly specific since pyridoxine and pyridoxamine 5'-phosphate, analogues of pyridoxal 5'-phosphate, which lack an aldehyde group, caused little or no inhibition even at high concentrations. The unreduced dikinase-pyridoxal 5'-phosphate complex displays an absorption maxima near 420 nm, typical for Schiff base formation. Following reduction of the Schiff base with sodium borohydride, N6-pyridoxyllysine was identified in the acid hydrolysate. When the enzyme was incubated in the presence of pyridoxal 5'-phosphate and reducing agent, the ATP/AMP, Pi/PPi, and pyruvate/phosphoenolpyruvate isotopic exchange reactions were inhibited to approximately the same extent, suggesting that the modification of the lysyl moiety causes changes in the enzyme that affect the reactivity of the pivotal histidyl residue. Phosphorylation of the histidyl group appears to prevent the inhibitor from attacking the lysine residue. On the other hand, addition of pyridoxal 5'-phosphate to the pyrophosphorylated enzyme promotes release of the pyrophosphate and yields the free enzyme which is subject to inhibition.     

Neoglycoproteins: preparation of noncovalent glycoproteins through high-affinity protein- (glycosyl) ligand complexes

This work was undertaken as part of a search for well-characterized glycoprotein models in which both the oligosaccharide structure, the number of oligosaccharide chains, and the precise location of these chains in the protein are known. On the basis of the fact that high-affinity ligand binding sites have been defined precisely for several proteins in terms of both number and relative location, the hypothesis to be tested was that if oligosaccharide chains were covalently attached to such high-affinity ligands, they would be specifically bound in the ligand sites of the appropriate protein, thus permitting the preparation of neoglycoproteins of precise predetermined oligosaccharide valency and topography. To test this hypothesis, pyridoxal 5'-phosphate was reductively (NaB3H4) aminated with the alpha-amino group of the asparagine oligosaccharide Man6-GlcNAc2-Asn from ovalbumin. When the resulting phosphopyridoxylated oligosaccharide (PG) was added to the apo form of aspartate aminotransferase (AAT; EC 2.6.1.1, the cytosolic enzyme from pig heart, consisting of two subunits and containing two coenzyme binding sites), a 2:1 (PG-AAT) complex was formed which could be characterized on the basis of tritium content, the absorbance and fluorescence of the pyridoxamine phosphate moiety of PG, and the concanavalin A binding properties acquired by AAT through the incorporation of the oligosaccharide. As expected from the established properties of the holoenzyme, the AAT-PG complex is stable in the absence of phosphate or vitamin B6 derivatives and can be dialyzed for 24 h without any significant loss of PG. According to the three-dimensional model of AAT, the oligosaccharide chain of PG should be partially masked in the coenzyme binding pocket.(ABSTRACT TRUNCATED AT 250 WORDS)     

Observations on the pyridoxal 5'-phosphate inhibition of DNA polymerases.

Pyridoxal 5'-phosphate at concentrations greater than 0.5 mM inhibits polymerization of deoxynucleoside triphosphate catalyzed by a variety of DNA polymerases. The requirement for a phosphate as well as aldehyde moiety of pyridoxal phosphate for inhibition to occur is clearly shown by the fact that neither pyridoxal nor pyridoxamine phosphate are effective inhibitors. Since the addition of nonenzyme protein or increasing the amount of template primer exerted no protective effect, there appears to be specific affinity between pyridoxal phosphate and polymerase protein. The deoxynucleoside triphosphates, however, could reverse the inhibition. The binding of pyridoxal 5'-phosphate to enzyme appears to be mediated through classical Schiff base formation between the pyridoxal phosphate and the free amino group(s) present at the active site of the polymerase protein. Kinetic studies indicate that inhibition by pyridoxal phosphate is competitive with respect to substrate deoxynucleoside triphosphate(s).     

The precursor of mitochondrial aspartate aminotransferase is translocated into mitochondria as apoprotein

Mitochondrial aspartate aminotransferase is synthesized on free polysomes as a higher molecular weight precursor (Sonderegger, P., Jaussi, R., Christen, P., and Gehring, H. (1982) J. Biol. Chem. 257, 3339-3345). The present study examines whether the coenzyme pyridoxal phosphate or pyridoxamine phosphate is required for the uptake of the precursor into mitochondria. Chicken embryo fibroblasts were cultured in medium prepared with and without pyridoxal. In cells grown in the presence of pyridoxal only holoform of aspartate aminotransferase and no apoenzyme was detected. Cells cultured under pyridoxal deficiency contained about 30% of apoenzyme in secondary cultures. All of this apoform was identified as mitochondrial isoenzyme. In order to differentiate whether this apoenzyme corresponded to newly synthesized protein or originated from pre-existing holoenzyme, double isotope-labeling experiments were performed. Secondary cultures of chicken embryo fibroblasts grown under pyridoxal depletion were labeled with [3H]methionine, and then pulsed with [35S]methionine. In another series of experiments, the 3H-labeled cells were pulsed with [35S]methionine in the presence of the protonophore carbonyl cyanide m-chlorophenylhydrazone in order to accumulate the precursor. Subsequently, the accumulated precursor was chased into the mitochondria by addition of the carbonyl cyanide m-chlorophenylhydrazone antagonist cysteamine. The holo- and apoenzyme from the ultrasonic extract of the double-labeled cells were separated by affinity chromatography on a phosphopyridoxyl-AH-Sepharose column, immunoprecipitated, and analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and fluorography. Under both experimental conditions, the 3H/35S ratio of the apoenzyme was less than half of that of the holoenzyme. Therefore, the apoenzyme and not the holoenzyme is the first product of the precursor in the mitochondria. Apparently, the precursor of mitochondrial aspartate aminotransferase is transported into mitochondria as apoprotein and is processed there independently of the coenzyme.     

Spectroscopic characterization of true enzyme-substrate intermediates of aspartate aminotransferase trapped at subzero temperatures

Absorption and circular dichroism spectra of stable enzyme-substrate intermediates of aspartate aminotransferase were recorded at subzero temperatures (down to -65 degrees C) in the cryosolvent water/methanol. The intermediates were formed either between the pyridoxal form of the enzyme and its amino acid substrates, or between the pyridoxamine form and its oxo acid substrates. Kd values determined by spectroscopic titration were very close to the Km values reported for the different substrates. The adsorption complex of the pyridoxal form was probably obtained on addition of cysteine sulfinate. This complex is characterized by an increased absorption at 430 nm together with a positive Cotton effect, as also observed in the case of the complex with the competitive inhibitor maleate indicating protonation of the internal aldimine. Addition of the substrates aspartate or glutamate to the pyridoxal form seemed to result in the direct accumulation of the external aldimine which showed a slight decrease in both the absorbance and the Cotton effect at 360 nm. Additionally, a bathochromic shift of 5 nm was observed in the case of glutamate. At 430 nm, only a minor increase in absorbance, but not in circular dichroism, was observed with aspartate, and no changes were found with glutamate and the substrate analog 2-methylaspartate, indicating a deprotonated external aldimine. Presumably, the ketimine intermediate was obtained on addition of the oxo acids 2-oxoglutarate or oxalacetate to the pyridoxamine form. The intermediate showed a slight bathochromic shift (2 nm) of the absorption band and decreased circular dichroism. On formation of the ketimine, a tyrosine residue, probably active-site Tyr225, becomes partly ionized. The finding that the external aldimine can probably be accumulated in the conversion of the pyridoxal to the pyridoxamine form with the natural substrates would confirm the proton abstraction at C alpha to be the rate-limiting step in the tautomerization, although with cysteine sulfinate, the formation of the external aldimine might contribute to the rate limitation. Accumulation of the ketimine in the reverse direction would indicate that the proton abstraction at C4' is rate-limiting in this half-reaction. The results demonstrate the feasibility of further structural investigations of true enzyme-substrate intermediates.