Cloning and characterization of a phosphopantetheinyl transferase from Streptomyces verticillus ATCC15003, the producer of the hybrid peptide-polyketide antitumor drug bleomycin.

BACKGROUND: Phosphopantetheinyl transferases (PPTases) catalyze the posttranslational modification of carrier proteins by the covalent attachment of the 4'-phosphopantetheine (P-pant) moiety of coenzyme A to a conserved serine residue, a reaction absolutely required for the biosynthesis of natural products including fatty acids, polyketides, and nonribosomal peptides. PPTases have been classified according to their carrier protein specificity. In organisms containing multiple P-pant-requiring pathways, each pathway has been suggested to have its own PPTase activity. However, sequence analysis of the bleomycin biosynthetic gene cluster in Streptomyces verticillus ATCC15003 failed to reveal an associated PPTase gene. RESULTS: A general approach for cloning PPTase genes by PCR was developed and applied to the cloning of the svp gene from S. verticillus. The svp gene is mapped to an independent locus not clustered with any of the known NRPS or PKS clusters. The Svp protein was overproduced in Escherichia coli, purified to homogeneity, and shown to be a monomer in solution. Svp is a PPTase capable of modifying both type I and type II acyl carrier proteins (ACPs) and peptidyl carrier proteins (PCPs) from either S. verticillus or other Streptomyces species. As compared to Sfp, the only 'promiscuous' PPTase known previously, Svp displays a similar catalytic efficiency (k(cat)/K(m)) for the BlmI PCP but a 346-fold increase in catalytic efficiency for the TcmM ACP. CONCLUSIONS: PPTases have recently been re-classified on a structural basis into two subfamilies: ACPS-type and Sfp-type. The development of a PCR method for cloning Sfp-type PPTases from actinomycetes, the recognition of the Sfp-type PPTases to be associated with secondary metabolism with a relaxed carrier protein specificity, and the availability of Svp, in addition to Sfp, should facilitate future endeavors in engineered biosynthesis of peptide, polyketide, and, in particular, hybrid peptide-polyketide natural products.

Metal ion inhibition of nonenzymatic pyridoxal phosphate catalyzed decarboxylation and transamination.

Nonenzymatic pyridoxal phosphate (PLP) catalyzed decarboxylations and transaminations have been revisited experimentally. Metal ions are known to catalyze a variety of PLP-dependent reactions in solution, including transamination. It is demonstrated here that the rate accelerations previously observed are due solely to enhancement of Schiff base formation under subsaturating conditions. A variety of metal ions were tested for their effects on the reactivity of the 2-methyl-2-aminomalonate Schiff bases. All were found to have either no effect or a small inhibitory one. The effects of Al(3+) were studied in detail with the Schiff bases of 2-methyl-2-aminomalonate, 2-aminoisobutyrate, alanine, and ethylamine. The decarboxylation of 2-methyl-2-aminomalonate is unaffected by metalation with Al(3+), while the decarboxylation of 2-aminoisobutyrate is inhibited 125-fold. The transamination reaction of ethylamine is 75-fold slower than that of alanine. Ethylamine transamination is inhibited 4-fold by Al(3+) metalation, while alanine transamination is inhibited only 1.3-fold. Metal ion inhibition of Schiff base reactivity suggests a simple explanation for the lack of known PLP dependent enzymes that make direct mechanistic use of metal ions. A comparison of enzyme catalyzed, PLP catalyzed, and uncatalyzed reactions shows that PLP dependent decarboxylases are among the best known biological rate enhancers: decarboxylation occurs 10(18)-fold faster on the enzyme surface than it does free in solution. PLP itself provides the lion's share of the catalytic efficiency of the holoenzyme: at pH 8, free PLP catalyzes 2-aminoisobutyrate decarboxylation by approximately 10(10)-fold, with the enzyme contributing an additional approximately 10(8)-fold.

Computational studies on nonenzymatic and enzymatic pyridoxal phosphate catalyzed decarboxylations of 2-aminoisobutyrate.

A computational study of nonenzymatic and enzymatic pyridoxal phosphate-catalyzed decarboxylation of 2-aminoisobutyrate (AIB) is presented. Four prototropic isomers of a model aldimine between AIB and 5'-deoxypyridoxal, with acetate interacting with the pyridine nitrogen, were employed in calculations of both gas phase and water model (PM3 and PM3-SM3) decarboxylation reaction paths. Calculations employing the transition state structures obtained for the four isomers allow the demonstration of stereoelectronic effects in transition state stabilization as well as a separation of the contributions of the Schiff base and pyridine ring moieties to this stabilization. The unprotonated Schiff base contribution (approximately 16 kcal/mol) is larger than that of the pyridine ring even when it is protonated (approximately 10 kcal/mol), providing an explanation of the catalytic power of pyruvoyl-dependent amino acid decarboxylases. An active site model of dialkylglycine decarboxylase was constructed and validated, and enzymatic decarboxylation reaction paths were calculated. The reaction coordinate is shown to be complex, with proton transfer from Lys272 to the coenzyme C4' likely simultaneous with C alpha--CO(2)(-) bond cleavage. The proposed concerted decarboxylation/proton-transfer mechanism provides a simple explanation for the observed specificity of this enzyme toward oxidative decarboxylation.

Rapid kinetic and isotopic studies on dialkylglycine decarboxylase.

The two half-reactions of the pyridoxal 5'-phosphate (PLP)-dependent enzyme dialkylglycine decarboxylase (DGD) were studied individually by multiwavelength stopped-flow spectroscopy. Biphasic behavior was found for the reactions of DGD-PLP, consistent with two coexisting conformations observed in steady-state kinetics [Zhou, X., and Toney, M. D. (1998) Biochemistry 37, 5761--5769]. The half-reaction kinetic parameters depend on alkali metal ion size in a manner similar to that observed for steady-state kinetic parameters. The fast phase maximal rate constant for the 2-aminoisobutyrate (AIB) decarboxylation half-reaction with the potassium form of DGD-PLP is 25 s(-1), while that for the transamination half-reaction between DGD-PMP and pyruvate is 75 s(-1). The maximal rate constant for the transamination half-reaction of the potassium form of DGD-PLP with L-alanine is 24 s(-1). The spectral data indicate that external aldimine formation with either AIB or L-alanine and DGD-PLP is a rapid equilibrium process, as is ketimine formation from DGD-PMP and pyruvate. Absorption ascribable to the quinonoid intermediate is not observed in the AIB decarboxylation half-reaction, but is observed in the dead-time of the stopped-flow in the L-alanine transamination half-reaction. The [1-(13)C]AIB kinetic isotope effect (KIE) on k(cat) for the steady-state reaction is 1.043 +/- 0.003, while a value of 1.042 +/- 0.009 was measured for the AIB half-reaction. The secondary KIE measured for the AIB decarboxylation half-reaction with [C4'-(2)H]PLP is 0.92 +/- 0.02. The primary [2-(2)H]-L-alanine KIE on the transamination half-reaction is unity. Small but significant solvent KIEs are observed on k(cat) and k(cat)/K(M) for both substrates, and the proton inventories are linear in each case. NMR measurements of C2--H washout vs product formation give ratios of 105 and 14 with L-alanine and isopropylamine as substrates, respectively. These results support a rate-limiting, concerted C alpha-decarboxylation/C4'-protonation mechanism for the AIB decarboxylation reaction, and rapid equilibrium quinonoid formation followed by rate-limiting protonation to the ketimine intermediate for the L-alanine transamination half-reaction. Energy profiles for the two half-reactions are constructed.

Crystal structures of dialkylglycine decarboxylase inhibitor complexes.

The crystal structures of four inhibitor complexes of dialkylglycine decarboxylase are reported. The enzyme does not undergo a domain closure, as does aspartate aminotransferase, upon inhibitor binding. Two active-site conformations have been observed in previous structures that differ in alkali metal ion content, and two active-site conformations have been shown to coexist in solution when a single type of metal ion is present. There is no indication of coexisting conformers in the structures reported here or in the previously reported structures, and the observed conformation is that expected based on the presence of potassium in the enzyme. Thus, although two active-site conformations coexist in solution, a single conformation, corresponding to the more active enzyme, predominates in the crystal. The structure of 1-aminocyclopropane-1-carboxylate bound in the active site shows the aldimine double bond to the pyridoxal phosphate cofactor to be fully out of the plane of the coenzyme ring, whereas the Calpha-CO2(-) bond lies close to it. This provides an explanation for the observed lack of decarboxylation reactivity with this amino acid. The carboxylate groups of both 1-aminocyclopropane-1-carboxylate and 5'-phosphopyridoxyl-2-methylalanine interact with Ser215 and Arg406 as previously proposed. This demonstrates structurally that alternative binding modes, which constitute substrate inhibition, occur in the decarboxylation half-reaction. The structures of d and l-cycloserine bound to the active-site show that the l-isomer is deprotonated at C(alpha), presumably by Lys272, while the d-isomer is not. This difference explains the approximately 3000-fold greater potency of the l versus the d-isomer as a competitive inhibitor of dialkylglycine decarboxylase.

Evidence for a two-base mechanism involving tyrosine-265 from arginine-219 mutants of alanine racemase.

A positively charged residue, R219, was found to interact with the pyridine nitrogen of pyridoxal phosphate in the structure of alanine racemase from Bacillus stearothermophilus [Shaw et al. (1997) Biochemistry 36, 1329-1342]. Three site-directed mutants, R219K, R219A, and R219E, have been characterized and compared to the wild type enzyme (WT) to investigate the role of R219 in catalysis. The R219K mutation is functionally conservative, retaining approximately 25% of the WT activity. The R219A and R219E mutations decrease enzyme activity by approximately 100- and 1000-fold, respectively. These results demonstrate that a positively charged residue at this position is required for efficient catalysis. R219 and Y265 are connected through H166 via hydrogen bonds. The R219 mutants exhibit similar kinetic isotope effect trends: increased primary isotope effects (1.5-2-fold) but unchanged solvent isotope effects in the L --> D direction and increased solvent isotope effects (1.5-2-fold) but unchanged primary isotope effects in the D --> L direction. These results support a two-base racemization mechanism involving Y265 and K39. They additionally suggest that Y265 is selectively perturbed by R219 mutations through the H166 hydrogen-bond network. pH profiles show a large pKa shift from 7.1-7.4 (WT and R219K) to 9. 5-10.4 (R219A and R219E) for kcat/KM, and from 7.3 to 9.9-10.4 for kcat. The group responsible for this ionization is likely to be the phenolic hydroxyl of Y265, whose pKa is electrostatically perturbed in the WT by the H166-mediated interaction with R219. Accumulation of an absorbance band at 510 nm, indicative of a quinonoid intermediate, only in the D --> L direction with R219E provides additional evidence for a two-base mechanism involving Y265.

pH studies on the mechanism of the pyridoxal phosphate-dependent dialkylglycine decarboxylase.

The pH dependence of the steady-state kinetic parameters for the dialkylglycine decarboxylase-catalyzed decarboxylation-dependent transamination between 2-aminoisobutyrate (AIB) and pyruvate is presented. The pH dependence of methylation and DTNB modification reactions, and spectroscopic properties, is used to augment the assignment of the kinetic pKa's to specific ionizations. The coincidence of pKa values (approximately 7.4) observed in kcat/KAIB, 1/KAIB, Kis for pyruvate, KPLP, and in absorbance and fluorescence titrations demonstrates that AIB is not a sticky substrate. It furthermore suggests that the decarboxylation step, or a conformational isomerization preceding it, limits the rate of the overall catalytic cycle. Coexisting, kinetically distinguishable conformers of DGD-PLP, originating from an alkali metal ion binding site, were previously demonstrated at pH 8.2 for DGD-PLP (Zhou, X., Toney, M. D. Biochemistry 37, 5761-5769). The pKa value of approximately 8.8 observed in kcat, kcat/KAIB, Kd for K+, spectrometric titrations, and the reaction of DGD-PLP with DTNB is tentatively assigned to the conformational change interconverting the two enzyme forms previously characterized. Three pKa's are observed in pH titrations of the DGD-PLP coenzyme absorbance. Individual spectra for the four ionization states are deconvoluted by fitting log-normal curves. All four ionization states have both ketoenamine and enolimine tautomers present. This and a review of spectral data in the literature lead to the conclusion that the pKa of approximately 7.4, which gives the largest spectral changes and controls kcat/KAIB, is not deprotonation of the aldimine nitrogen. Rather, it must be an active site residue whose ionization alters the ratio between ketoenamine and enolimine tautomers.

Coexisting kinetically distinguishable forms of dialkylglycine decarboxylase engendered by alkali metal ions.

The pyridoxal phosphate (PLP) dependent enzyme dialkylglycine decarboxylase (DGD) specifically binds alkali metal ions near the active site. Large ions (Rb+, K+) activate the enzyme while smaller ones (Na+, Li+) inhibit it. Crystallographic results have shown that DGD undergoes a metal ion size dependent structural switch [Hohenester, E., Keller, J. W., and Jansonius, J. N. (1994) Biochemistry 33, 13561], but no evidence for multiple conformations in crystalline DGD was obtained. Here, evidence is presented that DGD-K+ in solution exists in two conformations differing in catalytic competence. Initial rate traces for DGD-K+ exhibit a high degree of curvature due to decreasing activity over time. DGD remains tetrameric under the assay conditions as demonstrated by gel filtration experiments, arguing against the possibility of subunit dissociation as the source of activity loss. Likewise, the mass spectrum of DGD shows a single covalent form. A hysteretic model that assumes two slowly interconverting enzyme forms accounts well for the initial rate data when kinetic parameters from biphasic pre-steady-state kinetics are employed. The fit of the model to the data yields an estimate of 59 +/- 1% for the fast form. A cooperative model cannot account for the data. Double reciprocal plots for coenzyme binding to DGD exhibit two linear phases. Similarly, two kinetic phases are observed in PLP association kinetics. The substitution of Na+ or Rb+ for K+ alters the steady-state kinetic parameters of DGD. Preincubation of DGD-K+ with the competitive inhibitor 1-aminocyclopropane-1-carboxylate (ACC) lowers both kcat and KAIB apparently by drawing the enzyme toward the less active, tighter binding form observed in the pre-steady-state kinetics. These results suggest that the structure of the protein around the alkali metal ion determines the conformational distribution. The transamination reaction with l-alanine was coupled in the pre-steady-state to the LDH-catalyzed oxidation of NADH. This experiment yields an estimate of 68 +/- 4% for the fast form, in agreement with the hysteretic fit to the steady-state data. The reaction of DGD with dithiobis(nitrobenzoate) was used to probe the preexisting forms of DGD. Preincubation of DGD with ACC, like the exchange of Na+ for K+, shifts the conformational distribution, in agreement with the steady-state kinetics. These experiments clearly demonstrate that DGD is a hysteretic enzyme whose conformational distribution is controlled by the identity of the alkali metal ion bound near the active site, and that cooperativity does not play a role in catalysis or regulation.

Pre-steady-state kinetic analysis of the reactions of alternate substrates with dialkylglycine decarboxylase.

The pre-steady-state kinetics of the half-reactions of several substrates with dialkylglycine decarboxylase are examined by multiwavelength kinetics and global analysis. The substrates examined fall into two groups: those that exhibit simple, monophasic kinetics and those that exhibit biphasic kinetics. The rate of the AIB half-reaction is likely limited by the decarboxylation step based on the simple kinetics and spectra obtained from global analysis. The spectra for the first species in the transamination half-reactions of L-alanine and L-aminobutyrate show long-wavelength absorption characteristic of a carbanionic quinonoid intermediate. This demonstrates that formation of the external aldimine intermediates and abstraction of the C alpha protons from them are rapid. The reactions of the slower substrates L-phenylglycine and 1-aminocyclohexane-1-carboxylate may have external aldimine formation as the rate-determining step. The biphasic reactions of 2-methyl-2-aminomalonate, 1-aminocyclopentane-1-carboxylate, isopropylamine, and glycine all have external aldimine formation as the rapid observable step, based on the spectral changes observed in absorption and circular dichroism measurements. 2-Methyl-2-aminomalonate reacts approximately 10(4)-fold slower than does AIB with dialkylglycine decarboxylase, compared to approximately 10(5)-fold faster with coenzyme in solution. It is proposed that this radical reactivity reversal is due to a slow protein conformational change that is a prerequisite to decarboxylation of MAM, which occurs rapidly thereafter. Circular dichroism measurements on active site bound coenzyme provide evidence supporting this proposal. The binding of the noncovalent inhibitors pyruvate or lactate or the covalently binding inhibitor 1-aminocyclopropane-1-carboxylate all induce a slow change in coenzyme circular dichroism that quantitatively parallels the slow decarboxylation of 2-methyl-2-aminomalonate. Fast circular dichroism changes are seen in the mixing time of these measurements for both 1-aminocyclopropane-1-carboxylate and 2-methyl-2-aminomalonate, indicating rapid external aldimine formation on this longer time scale.

Reactions of alternate substrates demonstrate stereoelectronic control of reactivity in dialkylglycine decarboxylase.

Kinetic and product analyses of the reactions of dialkylglycine decarboxylase with several alternative substrates are presented. Rate constants for the reactions of amino and keto acids of several substrates decrease logarithmically with increasing side-chain size. Conversely, kcat for L-amino acid decarboxylation increases with side-chain size. These and other data confirm a proposed model for three binding subsites in the active site. In this model, bond making and breaking in both the decarboxylation and transamination half-reactions occurs at the "A" subsite, which maintains the scissile bond aligned with the p orbitals of the conjugated aldimine and thus maximizes stereoelectronic effects. This strongly supports the proposal by Dunathan (Proc. Natl. Acad. Sci. U.S.A. 55, 712-716) that PLP-dependent enzymes can largely control reaction specificity by specific orientation about C alpha in the external aldimine intermediate. The "B" subsite can accept either an alkyl or a carboxylate group, while the "C" subsite accepts only small alkyl groups. This model predicts the existence of nonproductive binding modes for amino acids, which is proposed to be the ultimate origin of the kcat increase with side-chain size for L-amino acid decarboxylation. The specificity of the 2-aminoisobutyrate decarboxylation half-reaction toward oxidative decarboxylation is very high (< 1 in 10(5) turnovers yields nonoxidative decarboxylation). The origin of this specificity is explored with the reactions of amino- and methylaminomalonate. These substrates exhibit high yields of nonoxidative decarboxylation, providing support for a model in which the interaction between a carboxylate group in the B subsite and Arg406 is a prerequisite to proton donation to and removal from C alpha.

Active site model for gamma-aminobutyrate aminotransferase explains substrate specificity and inhibitor reactivities.

A homology model for the pig isozyme of the pyridoxal phosphate-dependent enzyme gamma-aminobutyrate (GABA) aminotransferase has been built based mainly on the structure of dialkylglycine decarboxylase and on a multiple sequence alignment of 28 evolutionarily related enzymes. The proposed active site structure is presented and analyzed. Hypothetical structures for external aldimine intermediates explain several characteristics of the enzyme. In the GABA external aldimine model, the pro-S proton at C4 of GABA, which abstracted in the 1,3-azaallylic rearrangement interconverting the aldimine and ketimine intermediates, is oriented perpendicular to the plane of the pyridoxal phosphate ring. Lys 329 is in close proximity and is probably the general base catalyst for the proton transfer reaction. The carboxylate group of GABA interacts with Arg 192 and Lys 203, which determine the specificity of the enzyme for monocarboxylic omega-amino acids such as GABA. In the proposed structure for the L-glutamate external aldimine, the alpha-carboxylate interacts with Arg 445. Glu 265 is proposed to interact with this same arginine in the GABA external aldimine, enabling the enzyme to act on omega-amino acids in one half-reaction and on alpha-amino acids in the other. The reactivities of inhibitors are well explained by the proposed active site structure. The R and S isomers of beta-substituted phenyl and p-chlorophenyl GABA would bind in very different modes due to differential steric interactions, with the reactive S isomer leaving the orientation of the GABA moiety relatively unperturbed compared to that of the natural substrate. In our model, only the reactive S isomer of the mechanism-based inhibitor vinyl-GABA, an effective anti-epileptic drug known clinically as Vigabatrin, would orient the scissile C4-H bond perpendicular to the coenzyme ring plane and present the proton to Lys 329, the proposed general base catalyst of the reaction. The R isomer would direct the vinyl group toward Lys 329 and the C4-H bond toward Arg 445. The active site model presented provides a basis for site-directed mutagenesis and drug design experiments.

Structural and mechanistic analysis of two refined crystal structures of the pyridoxal phosphate-dependent enzyme dialkylglycine decarboxylase.

Two refined structures, differing in alkali metal ion content, of the bifunctional, pyridoxal phosphate-dependent enzyme dialkylglycine decarboxylase (DGD) are presented in detail. The enzyme is an alpha 4 tetramer, built up as a dimer of dimers, with a subunit molecular mass of 46.5 kDa. The fold of DGD is similar to those of aspartate aminotransferase, omega-amino acid aminotransferase and tyrosine phenol-lyase. The structure has two binding sites for alkali metal ions. DGD with potassium in site 1 (near the active site) and sodium in site 2 (at the surface of the molecule) has been refined against 2.6A resolution data (R-factor = 17.6%), and DGD with sodium at both sites has been refined against 2.1 A resolution data (R-factor = 17.8%). The proximity of site 1 to the active site accounts for the dependence of enzyme activity on potassium ions, and the observed active site structural changes caused by ion exchange at this site explain the inhibition of activity by sodium. DGD catalyzes both the decarboxylation of dialkylglycine species and the transamination of L-amino acids in its normal catalytic cycle. The active site structure of DGD is moderately homologous to that of aspartate aminotransferase, which catalyzes only transamination; both the differences and similarities provide mechanistic guidelines for the DGD-catalyzed reactions. Models of the L-isovaline and L-alanine external aldimine intermediates suggest mechanisms by which the decarboxylation and transamination reactions could be accomplished within the single active site. Decarboxylation is proposed to be at least partially catalyzed by stereoelectronic activation of the C alpha-carboxylate bond achieved by orienting this bond perpendicular to the plane of the pyridinium ring in the dialkylglycine external aldimine intermediate. Transamination is proposed to be catalyzed by a similar effect on the C alpha-H bond of the L-amino acid external aldimine intermediate, combined with general base catalysis provided by Lys272, in analogy to the mechanism of aspartate aminotransferase.

Crystal structures of true enzymatic reaction intermediates: aspartate and glutamate ketimines in aspartate aminotransferase.

The crystal structures of the stable, closed complexes of chicken mitochondrial aspartate aminotransferase with the natural substrates L-aspartate and L-glutamate have been solved and refined at 2.4- and 2.3-A resolution, respectively. In both cases, clear electron density at the substrate-coenzyme binding site unequivocally indicates the presence of a covalent intermediate. The crystallographically identical environments of the two subunits of the alpha 2 dimer allow a simple, direct correlation of the coenzyme absorption spectra of the crystalline enzyme with the diffraction results. Deconvolution of the spectra of the crystalline complexes using lognormal curves indicates that the ketimine intermediates constitute 76% and 83% of the total enzyme populations with L-aspartate and L-glutamate, respectively. The electron density maps accommodate the ketimine structures best in agreement with the independent spectral data. Crystalline enzyme has a much higher affinity for keto acid substrates compared to enzyme in solution. The increased affinity is interpreted in terms of a perturbation of the open/closed conformational equilibrium by the crystal lattice, with the closed form having greater affinity for substrate. The crystal lattice contacts provide energy required for domain closure normally supplied by the excess binding energy of the substrate. In solution, enzyme saturated with amino/keto acid substrate pairs has a greater total fraction of intermediates in the aldehyde oxidation state compared to crystalline enzyme. Assuming the only difference between the solution and crystalline enzymes is in conformational freedom, this difference suggests that one or more substantially populated, aldehydic intermediates in solution exist in the open conformation. Quantitative analyses of the spectra indicate that the value of the equilibrium constant for the open-closed conformational transition of the liganded, aldehydic enzyme in solution is near 1. The C4' pro-S proton in the ketimine models is oriented nearly perpendicularly to the plane of the pyridine ring, suggesting that the enzyme facilitates its removal by maximizing sigma-pi orbital overlap. The absence of a localized water molecule near Lys258 dictates that ketimine hydrolysis occurs via a transiently bound water molecule or from an alternative, possibly more open, structure in which water is appropriately bound. A prominent mechanistic role for flexibility of the Lys258 side chain is suggested by the absence of hydrogen bonds to the amino group in the aspartate structure and the relatively high temperature factors for these atoms in both structures.

Dialkylglycine decarboxylase structure: bifunctional active site and alkali metal sites.

The structure of the bifunctional, pyridoxal phosphate-dependent enzyme dialkylglycine decarboxylase was determined to 2.1-angstrom resolution. Model building suggests that a single cleavage site catalyzes both decarboxylation and transamination by maximizing stereoelectronic advantages and providing electrostatic and general base catalysis. The enzyme contains two binding sites for alkali metal ions. One is located near the active site and accounts for the dependence of activity on potassium ions. The other is located at the carboxyl terminus of an alpha helix. These sites help show how proteins can specifically bind alkali metals and how these ions can exert functional effects.

Lysine 258 in aspartate aminotransferase: enforcer of the Circe effect for amino acid substrates and general-base catalyst for the 1,3-prototropic shift.

The replacement of Lys258 by alanine (K258A) in aspartate aminotransferase reduces the rate constant for the central, 1,3-prototropic shift by 10(6)-10(8)-fold, confirming the role of Lys258 as the general-base catalyst for this step. The rate constant for the 1,3-prototropic shift interconverting K258A aldimine and ketimine intermediates is pH-independent like that of the wild-type enzyme (WT-AATase). K258A binds amino acid substrates in external aldimine intermediates 10(5)-fold more tightly than does WT-AATase. The excess amino acid binding energy observed in the mutant is sacrificed by the WT-AATase in order to increase the value of kcat. The net result is that the kcat/KM values for amino acid substrates are reduced only 3-100-fold by the mutation. This provides a clear example of the Circe effect propounded by Jencks [Jencks, W. P. (1975) Adv. Enzymol. Rel. Areas Mol. Biol. 43, 219]. Part of the increase in kcat due to the inclusion of Lys258 is accomplished by a 10(4)-10(5)-fold acceleration of external aldimine formation and hydrolysis. This step is partially rate-determining for K258A, but not for WT-AATase. A significant consequence of the utilization of amino acid binding energy for catalysis is the raising of the dissociation constants for these substrates to levels near the physiological concentrations of amino acids. The major product of the reaction of K258A with oxalacetate is pyruvate due to decarboxylation of the beta-imine formed in the ketimine intermediate.

Brønsted analysis of aspartate aminotransferase via exogenous catalysis of reactions of an inactive mutant.

Primary amines functionally replace lysine 258 by catalyzing both the 1,3-prototropic shift and external aldimine hydrolysis reactions with the inactive aspartate aminotransferase mutant K258A. This finding allows classical Brønsted analyses of proton transfer reactions to be applied to enzyme-catalyzed reactions. An earlier study of the reaction of K258A with cysteine sulfinate (Toney, M.D. & Kirsch, J.F., 1989, Science 243, 1485) provided a beta value of 0.4 for the 1,3-prototropic shift. The beta value reported here for the transamination of oxalacetate to aspartate is 0.6. The catalytic efficacy of primary amines is largely determined by basicity and molecular volume. The dependence of the rate constants for the reactions of K258A and K258M on amine molecular volume is nearly identical. This observation argues that the alkyl groups of the added amines do not occupy the position of the lysine 258 side chain in the wild type enzyme. Large primary C alpha and insignificant solvent deuterium kinetic isotope effects with amino acid substrates demonstrate that the amine nitrogen of the exogenous catalysts directly abstracts the labile proton in the rate-determining step.

Crystallization and preliminary X-ray diffraction studies of dialkylglycine decarboxylase, a decarboxylating transaminase.

The pyridoxal phosphate-dependent enzyme dialkylglycine decarboxylase (E.C. 4.1.1.64) has been crystallized by vapor diffusion from a 15% polyethyleneglycol solution with sodium pyruvate as coprecipitant. The space group of the crystals is either P6(2)22 or the enantiomorph, P6(4)22, with one subunit of 46,500 Da per asymmetric unit. The unit cell has dimensions a = b = 152.7 A, c = 86.6 A, alpha = beta = 90 degrees, gamma = 120 degrees, and a solvent content of approximately 61%. diffraction extends to 2.3 A resolution.

The K258R mutant of aspartate aminotransferase stabilizes the quinonoid intermediate.

Lys-258 of aspartate aminotransferase forms a Schiff base with pyridoxal phosphate and is responsible for catalysis of the 1,3-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 alpha-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 unreactive toward keto acid substrates, and the coenzyme bound to this species is not dissociable from the protein.

Tyrosine 70 fine-tunes the catalytic efficiency of aspartate aminotransferase.

The aspartate aminotransferase mutant Y70F exhibits kcat = 8% and kcat/KM = 2% of the wild type values for the transamination of aspartate and alpha-ketoglutarate. The affinity of the enzyme for the noncovalently bound inhibitor maleate is reduced 17-fold by the mutation, while only a 2.5-fold reduction is observed for alpha-methylaspartate, which forms a stable, covalent external aldimine. The high population of the quinonoid intermediate formed in the reaction of the wild type with beta-hydroxyaspartate is more than 75% diminished by the mutation. The values of the Y70F C alpha-H kinetic isotope effects for the aspartate reaction are larger than those of wild type (DV = 2.4 vs 1.52; D(V/K) = 2.5 vs 1.7). Conversely, the Y70F value of D(V/K) for the glutamate reaction is decreased compared to wild type (1.75 vs 2.5). These results, combined with previous studies of Lys258 mutants, eliminate Tyr70 as an essential component of the catalytic apparatus, with the caveat that the functionally of the deleted hydroxyl group is possibly replaced by a water molecule.

Kinetics and equilibria for the reactions of coenzymes with wild type and the Y70F mutant of Escherichia coli aspartate aminotransferase.

The Y70F mutant of aspartate aminotransferase has reduced affinity for coenzymes compared to the wild type. The equilibrium dissociation constants for pyridoxamine phosphate (PMP) holoenzymes, KPMPdiss, were determined from the association and dissociation rate constants to be 1.3 nM and 30 nM for the wild type and mutant, respectively. This increase in KPMPdiss for Y70F is due to a 27-fold increase in the dissociation rate constant. Pyridoxal phosphate (PLP) association kinetics are complex, with three kinetic processes detectable for wild type and two for Y70F. A directly determined, accurate value of KPLPdiss for wild type enzyme has been difficult to obtain because of the low value of this constant. The values of KPLPdiss for the holoenzymes were determined indirectly through the measured values for KPMPdiss, glutamate-alpha-ketoglutarate half-reaction equilibrium constants, and the equilibrium constant for the transamination of PLP by glutamate catalyzed by Y70F. The values of KPLPdiss obtained by this procedure are 0.4 pM for wild type and 40 pM for Y70F. The increases in KPMPdiss and KPLPdiss for Y70F correspond to delta delta G values of 1.9 and 2.7 kcal/mol, respectively, and are directly attributed to the loss of the hydrogen bond from the phenolic hydroxyl group of Tyr70 to the coenzyme phosphate. The delta G for association of PLP with wild type enzyme is 4.7 kcal/mol more favorable than that for PMP.