Expression of recombinant diaminopimelate epimerase in Escherichia coli. Isolation and inhibition with an irreversible inhibitor

Design and evaluation of substrate-product analog inhibitors for racemases and epimerases utilizing a 1,1-proton transfer mechanism

Racemases and epimerases catalyze the inversion of stereochemistry at asymmetric carbon atoms to generate stereoisomers that often play important roles in normal and pathological physiology. Consequently, there is interest in developing inhibitors of these enzymes for drug discovery. A strategy for the rational design of substrate-product analog (SPA) inhibitors of racemases and epimerases utilizing a direct 1,1-proton transfer mechanism is elaborated. This strategy assumes that two groups on the asymmetric carbon atom remain fixed at active-site binding determinants, while the hydrogen and third, motile group move during catalysis, with the latter potentially traveling between an R- and S-pocket at the active site. SPAs incorporate structural features of the substrate and product, often with geminal disubstitution on the asymmetric carbon atom to simultaneously present the motile group to both the R- and S-pockets. For racemases operating on substrates bearing three polar groups (glutamate, aspartate, and serine racemases) or with compact, hydrophobic binding pockets (proline racemase), substituent motion is limited and the design strategy furnishes inhibitors with poor or modest binding affinities. The approach is most successful when substrates have a large, motile hydrophobic group that binds at a plastic and/or capacious hydrophobic site. Potent inhibitors were developed for mandelate racemase, isoleucine epimerase, and α-methylacyl-CoA racemase using the SPA inhibitor design strategy, exhibiting binding affinities ranging from substrate-like to exceeding that of the substrate by 100-fold. This rational approach for designing inhibitors of racemases and epimerases having the appropriate active-site architectures is a useful strategy for furnishing compounds for drug development.

Kinetic and structural studies of the reaction of Escherichia coli dihydrodipicolinate synthase with (S)-2-bromopropionate

Dihydrodipicolinate synthase (DHDPS) catalyzes the first committed step in the lysine-biosynthetic pathway converting pyruvate and L-aspartate-β-semialdehyde to dihydrodipicolinate. Kinetic studies indicate that the pyruvate analog (S)-2-bromopropionate inactivates the enzyme in a pseudo-first-order process. An initial velocity pattern indicates that (S)-2-bromopropionate is a competitive inhibitor versus pyruvate, with an inhibition constant of about 8 mM. Crystals of DHDPS complexed with (S)-2-bromopropionate formed in a solution consisting of 50 mM HEPES pH 7.5, 18% polyethylene glycol 3350, 8 mM spermidine, 0.2 M sodium tartrate and 5.0 mg ml−1 DHDPS. The crystals diffracted to 2.15 Å resolution and belonged to space group P1. The crystal structure confirms the displacement of bromine and the formation of a covalent attachment between propionate and Lys161 at the active site of the enzyme. Lys161 is the active-site nucleophile that attacks the carbonyl C atom of pyruvate and subsequently generates an imine adduct in the first half-reaction of the ping-pong enzymatic reaction. A comparison of the crystal structures of DHDPS complexed with pyruvate or (S)-2-bromopropionate indicates the covalent adduct formed from (S)-2-bromopropionate leads to a rotation of about 180° of the β–δ C atoms of Lys61 that aligns the covalently bound propionate fairly closely with the imine adduct formed with pyruvate.

Kinetic, spectral, and structural studies of the slow-binding inhibition of the Escherichia coli dihydrodipicolinate synthase by 2, 4-oxo-pentanoic acid

Dihydrodipicolinate synthase (DHDPS) catalyzes the first step in the biosynthetic pathway for production of l-lysine in bacteria and plants. The enzyme has received interest as a potential drug target owing to the absence of the enzyme in mammals. The DHDPS reaction is the rate limiting step in lysine biosynthesis and involves the condensation of l-aspartate-β-semialdehyde and pyruvate to form 2, 3-dihydrodipicolinate. 2, 4-oxo-pentanoic acid (acetopyruvate) is a slow-binding inhibitor of DHDPS that is competitive versus pyruvate with an initial K_i of about 20 μM and a final inhibition constant of about 1.4 μM. The enzyme:acetopyruvate complex displays an absorbance spectrum with a λ_max at 304 nm and a longer wavelength shoulder. The rate constant for formation of the complex is 86 M^-1 s^-1. The enzyme forms a covalent enamine complex with the first substrate pyruvate and can be observed spectrally with a λ_max at 271 nm. The spectra of the enzyme in the presence of pyruvate and acetopyruvate shows the initial formation of the pyruvate enamine intermediate followed by the slower appearance of the E:acetopyruvate spectra with a rate constant of about 0.013 s^-1. The spectral studies suggest the formation of a Schiff base between acetopyruvate and K161 on enzyme that subsequently deprotonates to form a resonance stabilized anion similar to the enamine intermediate formed with pyruvate. The crystal structure of the E:acetopyruvate complex confirms the formation of the Schiff base between acetopyruvate and K161.

RppH-dependent pyrophosphohydrolysis of mRNAs is regulated by direct interaction with DapF in Escherichia coli

Similar to decapping of eukaryotic mRNAs, the RppH-catalyzed conversion of 5′-terminal triphosphate to monophosphate has recently been identified as the rate-limiting step for the degradation of a subset of mRNAs in Escherichia coli. However, the regulation of RppH pyrophosphohydrolase activity is not well understood. Because the overexpression of RppH alone does not affect the decay rate of most target mRNAs, the existence of a mechanism regulating its activity has been suggested. In this study, we identified DapF, a diaminopimelate (DAP) epimerase catalyzing the stereoinversion of L,L-DAP to meso-DAP, as a regulator of RppH. DapF showed a high affinity interaction with RppH and increased its RNA pyrophosphohydrolase activity. The simultaneous overexpression of both DapF and RppH increased the decay rates of RppH target RNAs by about a factor of two. Together, our data suggest that the cellular level of DapF is a critical factor regulating the RppH-catalyzed pyrophosphate removal and the subsequent degradation of target mRNAs.

A new robust kinetic assay for DAP epimerase activity

DAP epimerase is the penultimate enzyme in the lysine biosynthesis pathway. The most versatile assay for DAP epimerase catalytic activity employs a coupled DAP epimerase-DAP dehydrogenase enzyme system with a commercial mixture of DAP isomers as substrate. DAP dehydrogenase converts meso-DAP to THDP with concomitant reduction of NADP(+) to NADPH. We show that at high concentrations, accumulation of NADPH results in inhibition of DAPDH, resulting in spurious kinetic data. A new assay has been developed employing DAP decarboxylase that allows the reliable characterisation of DAP epimerase enzyme kinetics.

The stereoselective synthesis of aziridine analogues of diaminopimelic acid (DAP) and their interaction with dap epimerase

Aziridine analogues of diaminopimelic acid (DAP) have been prepared stereoselectively for the first time and evaluated as inhibitors of DAP epimerase. (2R,3S,3'S)-3-(3'-Aminopropane)aziridine-2,3'-dicarboxylate was synthesised and shown to be a reversible inhibitor of DAP epimerase with an IC(50) value of 2.88 mM. (2S,4S)- and (2S,4R)-2-(4-Amino-4-carboxybutyl)aziridine-2-carboxylic acid (ll-azi-DAP and dl-azi-DAP ) were made as pure diastereomers, and both were shown to be irreversible inhibitors of DAP epimerase. ll-Azi-DAP selectively binds to Cys-73 of the enzyme active site whereas dl-azi-DAP binds to Cys-217 via attack of sulfhydryl on the methylene of the inhibitor aziridine ring. These observations are consistent with the two base mechanism proposed for the epimerization of ll-DAP and meso-DAP by DAP epimerase.

Structural insight into gene duplication, gene fusion and domain swapping in the evolution of PLP-independent amino acid racemases

The X-ray crystal structure has revealed two similar alpha/beta domains of aspartate racemase (AspR) from Pyrococcus horikoshii OT3, and identified a pseudo mirror-symmetric distribution of the residues around its active site [Liu et al. (2002) J. Mol. Biol. 319, 479-489]. Structural homology and functional similarity between the two domains suggested that this enzyme evolved from an ancestral domain by gene duplication and gene fusion. We have expressed solely the C-terminal domain of this AspR and determined its three-dimensional structure by X-ray crystallography. The high structural stability of this domain supports the existence of the ancestral domain. In comparison with other amino acid racemases (AARs), we suggest that gene duplication and gene fusion are conventional ways in the evolution of pyridoxal 5'-phosphate-independent AARs.

Crystal structure of aspartate racemase from Pyrococcus horikoshii OT3 and its implications for molecular mechanism of PLP-independent racemization

There exists a d-enantiomer of aspartic acid in lactic acid bacteria and several hyperthermophilic archaea, which is biosynthesized from the l-enantiomer by aspartate racemase. Aspartate racemase is a representative pyridoxal 5'-phosphate (PLP)-independent amino acid racemase. The "two-base" catalytic mechanism has been proposed for this type of racemase, in which a pair of cysteine residues are utilized as the conjugated catalytic acid and base. We have determined the three-dimensional structure of aspartate racemase from the hyperthermophilic archaeum Pyrococcus horikoshii OT3 at 1.9 A resolution by X-ray crystallography and refined it to a crystallographic R factor of 19.4% (R(free) of 22.2%). This is the first structure reported for aspartate racemase, indeed for any amino acid racemase from archaea. The crystal structure revealed that this enzyme forms a stable dimeric structure with a strong three-layered inter-subunit interaction, and that its subunit consists of two structurally homologous alpha/beta domains, each containing a four-stranded parallel beta-sheet flanked by six alpha-helices. Two strictly conserved cysteine residues (Cys82 and Cys194), which have been shown biochemically to act as catalytic acid and base, are located on both sides of a cleft between the two domains. The spatial arrangement of these two cysteine residues supports the "two-base" mechanism but disproves the previous hypothesis that the active site of aspartate racemase is located at the dimeric interface. The structure revealed a unique pseudo mirror-symmetry in the spatial arrangement of the residues around the active site, which may explain the molecular recognition mechanism of the mirror-symmetric aspartate enantiomers by the non-mirror-symmetric aspartate racemase.

Active Site Residues of Glutamate Racemase†

Glutamate racemase, MurI, catalyzes the interconversion of glutamate enantiomers in a cofactor-independent fashion and provides bacteria with a source of D-Glu for use in peptidoglycan biosynthesis. The enzyme uses a "two-base" mechanism involving a deprotonation of the substrate at the alpha-position to form an anionic intermediate, followed by a reprotonation in the opposite stereochemical sense. In the Lactobacillus fermenti enzyme, Cys73 is responsible for the deprotonation of D-glutamate, and Cys184 is responsible for the deprotonation of L-glutamate; however, very little is known about the roles of other active site residues. This work describes the preparation of four mutants in which strictly conserved residues containing ionizable side chains were modified (D10N, D36N, E152Q, and H186N). During the course of this research, the structural analysis of a crystallized glutamate racemase indicated that three of these residues (D10, E152, and H186) are in the active site of the enzyme [Hwang, K. Y., Cho, C.-S., Kim, S. S., Sung, H.-C., Yu, Y. G., and Cho, Y. (1999) Nat. Struct. Biol. 6, 422-426]. Two of the mutants, D10N and H186N, displayed a marked decrease in the values of k(cat), but not K(M), and are therefore implicated as important catalytic residues. Further analysis of the primary kinetic isotope effects observed with alpha-deuterated substrates showed that a significant asymmetry was introduced into the free energy profile by these two mutations. This is interpreted as evidence that the mutated residues normally assist the catalytic thiols in acting as bases (D10 with C73 and H186 with C184). An alternate possibility is that the residues may serve to stabilize the carbanionic intermediate in the racemization reaction.

Structure/function studies on enzymes in the diaminopimelate pathway of bacterial cell wall biosynthesis

Within the past 18 months work has continued on the structure and mechanisms of enzymes involved in the diaminopimelic acid/lysine biosynthetic pathway. A novel structure has been determined for a PLP-independent epimerase, and structures with bound substrates have been solved for two other enzymes. Additionally, new studies have appeared describing the chemical mechanisms of three enzymes in the pathway.

Chemical mechanism of Haemophilus influenzae diaminopimelate epimerase

Seven unique enzymatic steps lead to the biosynthesis of L-lysine from L-aspartate semialdehyde and pyruvate in bacteria. The immediate precursor to L-lysine is D,L-diaminopimelate, a diamino acid which is incorporated into the pentapeptide of the Gram-negative peptidoglycan moiety. D,L-Diaminopimelate is generated from the corresponding L,L-isomer by the dapF-encoded epimerase. Diaminopimelate epimerase is a representative of the pyridoxal phosphate-independent amino acid racemases, for which substantial evidence exists supporting the role of two cysteine residues as the catalytic acid and base. The pH dependencies of the maximum velocities in the L,L --> D,L and D,L --> L,L direction depend on a single catalytic group exhibiting pK values of 7.0 and 6.1, respectively, which must be unprotonated for activity. The pH dependencies of the V/K values in both directions depend on the ionization of two groups, one exhibiting a pK value of 6.7 which must be unprotonated and one exhibiting a pK value of 8.5 which must be protonated. Primary kinetic isotope effects on V and V/K are unequal, with D(V/K) being larger than DV in both the forward and reverse directions. Solvent kinetic isotope effects in both directions are inverse on V/K, but normal on V. Both of these isotopic observations support a model in which proton isomerization after catalysis and substrate dissociation is kinetically significant. A single solvent "overshoot" is observed when L, L-diaminopimelate is incubated with enzyme in D2O; however, an unprecedented double overshoot is observed when D,L-diaminopimelate is incubated with enzyme in D2O. A model has been developed which allows these two overshoots to be simulated. A chemical mechanism is proposed invoking the function of two cysteine residues, Cys73 and Cys217, observed in the recently determined three-dimensional structure of this enzyme [Cirilli, M., et al. (1998) Biochemistry 37, 16452-16458], as the acid and base in the mechanism.

Catalytic acid/base residues of glutamate racemase

Glutamate racemase is a cofactor-independent enzyme that employs two active-site cysteine residues as acid/base catalysts during the interconversion of glutamate enantiomers. In a given reaction direction, a thiolate from one of the cysteines abstracts the alpha-proton, and the other cysteine thiol delivers a proton to the opposite face of the resulting carbanionic intermediate. This paper reports that the C73S and C184S mutants are still capable of racemizing glutamate with specificity constants about 10(3)-fold lower than those of the wild-type enzyme. A "one-base requiring" reaction, the elimination of water from N-hydroxyglutamate, has been used to deduce which thiol acts as the base for a given enantiomer. With D-N-hydroxyglutamate the C73S mutant is a much poorer catalyst than wild-type enzyme, whereas the C184S mutant is a somewhat better catalyst. This trend was reversed with L-N-hydroxyglutamate, suggesting that Cys73 is responsible for the deprotonation of D-glutamate and Cys184 is responsible for the deprotonation of L-glutamate. In addition, with C73S the Vmax/KM isotope effect on D-glutamate racemization was greater than that seen with wild-type enzyme, whereas the isotope effect with L-glutamate had decreased. The results were reversed with the C184S mutant. This is interpreted as being due to an asymmetry in the free energy profiles that is induced upon mutation, with the deprotonation step involving a serine becoming the more cleanly rate-determining of the two. These results support the above assignment and the notion that a carbanionic intermediate is formed during catalysis.

Structural symmetry: the three-dimensional structure of Haemophilus influenzae diaminopimelate epimerase

The Haemophilus influenzae diaminopimelate epimerase was cloned, expressed, purified, and crystallized in the C2221 space group (a = 102.1 A, b = 115.4 A, c = 66.3 A, alpha = beta = gamma = 90 degrees). The three-dimensional structure was solved to 2.7 A using a single Pt derivative and the Se-Met-substituted enzyme to a conventional R factor of 19.0% (Rfree = 24.2%). The 274 amino acid enzyme consists of two structurally homologous domains, each containing eight beta-strands and two alpha-helices. Diaminopimelate epimerase is a representative of the PLP-independent amino acid racemases, for which no structure has yet been determined and substantial evidence exists supporting the role of two cysteine residues as the catalytic acid and base. Cys73 of the amino terminal domain is found in disulfide linkage, at the domain interface, with Cys217 of the carboxy terminal domain, and we suggest that these two cysteine residues in the reduced, active enzyme function as the acid and base in the mechanism.

Enzymology of bacterial lysine biosynthesis

Bacteria have evolved three strategies for the synthesis of lysine from aspartate via formation of the intermediate diaminopimelate (DAP), a metabolite that is also involved in peptidoglycan formation. The objectives of this chapter are descriptions of mechanistic studies on the reactions catalyzed by dihydrodipicolinate synthase, dihydrodopicolinate reductase, tetrahydrodipicolinate N-succinyl-transferase, N-succinyl-L,L-DAP aminotransferase, N-succinyl-L,L-DAP desuccinylase, L,L-DAP epimerase, L,L-DAP decarboxylase, and DAP dehydrogenase. These enzymes are discussed in terms of kinetic, isotopic, and X-ray crystallographic data that allow one to infer the nature of interactions of each of these enzymes with its substrate(s), coenzymes, and inhibitors.

Bacterial glutamate racemase has high sequence similarity with myoglobins and forms an equimolar inactive complex with hemin

Glutamate racemase (EC 5.1.1.3), an enzyme of microbial origin, shows significant sequence homology with mammalian myoglobins, in particular in the regions corresponding to the E and F helices, which constitute the heme binding pocket of myoglobins. Glutamate racemase binds tightly an equimolar amount of hemin, leading to loss of racemase activity. Although this enzyme shows homology with aspartate racemase, the latter does not bind hemin. The glutamate racemase gene of Pediococcus pentosaceus has a 795-nt open reading frame and encodes 265-amino acid residues, which form a monomeric protein (M(r) 29,000). Neither racemase has cofactors, but they contain essential cysteine residues [Yohda, M., Okada, H. & Kumagai, H. (1991) Biochim. Biophys. Acta 1089, 234-240].

Active-site directed irreversible inhibition of diamine oxidase by a homologous series of aziridinylalkylamines

Three electrophilic homologous aminoalkylaziridine analogues of putrescine, cadaverine, and 1,3-diaminopropane were synthesized and found to represent a mechanistically distinct class of irreversible inhibitors of diamine oxidase. The putrescine analogue, N-(4-aminobutyl)aziridine gave the lowest calculated IC50 value, whereas N-(3-aminopropyl)aziridine, an analogue of the poorest substrate of the series, showed the highest IC50. The findings suggest that the aziridinylalkylamines tested are site-directed agents that form irreversible complexes at the active site of diamine oxidase. Affinity of the inhibitors for the active site appeared to be dependent on alkyl chain length, suggesting that binding promotes the reactivity of the aziridinyl group.