Acyloin formation by benzoylformate decarboxylase from Pseudomonas putida.

Whole cells and cell extracts of Pseudomonas putida grown in a medium containing ammonium mandelate have the capacity to produce the acyloin compound 2-hydroxypropiophenone when incubated with benzoylformate and acetaldehyde. Benzaldehyde and benzyl alcohol were formed as reaction by-products. The enantiomeric excess of the 2-hydroxypropiophenone product was found to be 91 to 92%. The absolute configuration of the enzymatically prepared product at the carbinol carbon was found to be S. The thiamine PPi-linked enzyme benzoylformate decarboxylase, purified to give a single protein band on polyacrylamide gel electrophoresis, was shown to be responsible for the catalysis of this novel condensation reaction.     

Enantiospecific Production of S-(-)-2-Hydroxypropiophenone Mediated by Benzoylformate Decarboxylase from Acinetobacter Calcoaceticus

Whole cells of Acinetobacter calcoaceticus, grown in a medium containing mandelate, converted benzoylformate and acetaldehyde into the acyloin compound 2-hydroxypropiophenone. The optical purity of the product was found to be greater than 98%. The absolute configuration of the biotransformation product at the carbinol carbon was found to be (S). The enzyme responsible for this bioconversion was confirmed as benzoylformate decarboxylase by the demonstration that the purified homogeneous enzyme catalysed the condensation reaction.     

Optimisation of reaction conditions for production ofS-(−)-2-hydroxypropiophenone byAcinetobacter calcoaceticus

Summary Whole cells and cell-free extracts ofAcinetobacter calcoaceticus containing benzoylformate decarboxylase efficiently condensed benzoylformate and acetaldehyde to produce the acyloin compoundS-(–)-2-hydroxypropiophenone. Optimal concentrations of acetaldehyde cosubstrate for this reaction were found to be 1600 and 800 mM when whole cells and cell-free extracts were used respectively as biocatalysts. In both cases, optimal benzoylformate concentration was 100 mM. Temperature and pH optima for the biotransformation reaction were 30°C and 6.0 respectively. Under optimised conditions, maximum production of 2-hydroxypropiophenone, amounting to 8.4 g L^–1, occurred after a 2-h incubation. Product formation equivalent to 6.95 g in 1 h corresponded to a productivity of 267 mg acyloin per g dry cells per h.     

Factors affecting 2-hydroxypropiophenone formation by benzoylformate decarboxylase from Pseudomonas putida

Benzoylformate (100 mM) was quantitatively converted to the acyloin compound, 2-hydroxypropiophenone (61.76 mM) and benzaldehyde (38.2 mM) by an enzyme extract from Pseudomonas putida ATCC 12633 in the presence of 1.6M acetaldehyde. Biotransformations were carried out at pH 6.0 and 30 degrees C with an incubation time of 60 min. Activity of the acyloin forming enzyme, benzoylformate decarboxylase, was 1.23 units/mL in the biotransformation mixture. Acyloin formation increased dramatically with pH in the range 4-5 and had a broad activity plateau in the pH range 5-8. A broad temperature optimum for acyloin formation was also observed in the range 20-40 degrees C.     

Kinetics and mechanism of benzoylformate decarboxylase using 13C and solvent deuterium isotope effects on benzoylformate and benzoylformate analogues

Benzoylformate decarboxylase (benzoylformate carboxy-lyase, BFD; EC 4.1.1.7) from Pseudomonas putida is a thiamine pyrophosphate (TPP) dependent enzyme which converts benzoylformate to benzaldehyde and carbon dioxide. The kinetics and mechanism of the benzoylformate decarboxylase reaction were studied by solvent deuterium and 13C kinetic isotope effects with benzoylformate and a series of substituted benzoylformates (pCH3O, pCH3, pCl, and mF). The reaction was found to have two partially rate-determining steps: initial tetrahedral adduct formation (D2O sensitive) and decarboxylation (13C sensitive). Solvent deuterium and 13C isotope effects indicate that electron-withdrawing substituents (pCl and mF) reduce the rate dependence upon decarboxylation such that decreased 13(V/K) effects are observed. Conversely, electron-donating substituents increase the rate dependence upon decarboxylation such that a larger 13(V/K) is seen while the D2O effects on V and V/K are not dramatically different from those for benzoylformate. All of the data are consistent with substituent stabilization or destabilization of the carbanionic intermediate (or carbanion-like transition state) formed during decarboxylation. Additional information regarding the mechanism of the enzymic reaction was obtained from pH studies on the reaction of benzoylformate and the binding of competitive inhibitors. These studies suggest that two enzymic bases are required to be in the correct protonation state (one protonated and one unprotonated) for optimal binding of substrate (or inhibitors).     

Immobilization and characterization of benzoylformate decarboxylase from Pseudomonas putida on spherical silica carrier

If an adequate biocatalyst is identified for a specific reaction, immobilization is one possibility to further improve its properties. The immobilization allows easy recycling, improves the enzyme performance, and it often enhances the stability of the enzyme. In this work, the immobilization of the benzoylformate decarboxylase (BFD) variant, BFD A460I-F464I, from Pseudomonas putida was accomplished on spherical silica. Silicagel is characterized by its high mechanical stability, which allows its application in different reactor types without restrictions. The covalently bound enzyme was characterized in terms of its activity, stability, and kinetics for the formation of chiral 2-hydroxypropiophenone (2-HPP) from benzaldehyde and acetaldehyde. Moreover, temperature as well as pressure dependency of immobilized BFD A460I-F464I activity and enantioselectivity were analyzed. The used wide-pore silicagel shows a good accessibility of the immobilized enzyme. The activity of the immobilized BFD A460I-F464I variant was determined to be 70% related to the activity of the free enzyme. Thereby, the enantioselectivity of the enzyme was not influenced by the immobilization. In addition, a pressure-induced change in stereoselectivity was found both for the free and for the immobilized enzyme. With increasing pressure, the enantiomeric excess (ee) of (R)-2-HPP can be increased from 44% (0.1 MPa) to 76% (200 MPa) for the free enzyme and from 43% (0.1 MPa) to 66% (200 MPa) for the immobilized enzyme.     

Studies of the reductive biotransformation of selected carbonyl compounds by whole cells and extracts of baker's yeast, Saccharomyces carevisiae

The progress of reductive biotransformations of a variety of earbonyl compounds by whole cells of baker's yeast was monitored with time. Biotransformations rates ranged from 0.11 to 112.12 mg product formed per g dry yeast per h. While rapid biotransformations of citronellal and ethyl benzoylformate were observed, complete conversion of substrate to product did not occur. Reductive conversions of ethyl- and methyl-acetoacetate went to completion in 6 and 12 h respectively. Ethyl mandelate was produced stereoselectively, favoring the (R)- stereoisomer and ethyl and methyl-3-hydroxybutyrate were produced with (S)-enantiospecificity. Yeast crude extract and resuspended presence of NAD(P)H. Ethyl benzoylformate and methyl-and ethyl-acetoacetate were preferentially reduced by yeast crude extract as compared to resuspended pellet and, in the case of the former two substrates, the reaction manifested a preference for NADPH over NADH.     

Lithium-stabilized nucleophilic addition of thiamin to a ketone provides an efficient route to mandelylthiamin,a critical pre-decarboxylation intermediate

Mandelylthiamin (MTh) is an accurate model of the covalent intermediate derived from the condensation of thiamin diphosphate and benzoylformate in benzoylformate decarboxylase. The properties and catalytic susceptibilities of mandelylthiamin are the subjects of considerable interest. However, the existing synthesis gives only trace amounts of the precursor to MTh as it is conducted under reversible conditions. An improved approach derives from the unique ability of lithium ions to drive to completion the otherwise unfavorable condensation of the conjugate base of thiamin and methyl benzoylformate. The unique efficiency of the condensation reaction in the presence of lithium ions is established in contrast to the effects of other Lewis acids. Interpretation of the pattern of the results indicates that the condensation of the ketone and thiamin is thermodynamically controlled. It is proposed that the addition of lithium ions displaces the equilibrium toward the product through formation of a stable lithium-alkoxide.     

Isolation and characterization of a benzoylformate decarboxylase and a NAD+/NADP+-dependent benzaldehyde dehydrogenase involved in D-phenylglycine metabolism in Pseudomonas stutzeri ST-201

Following induction with D-phenylglycine both d-phenylglycine aminotransferase activity and benzoylformate decarboxylase activity were observed in cultures of Pseudomonas stutzeri ST-201. Induction with benzoylformate, on the other hand, induced only benzoylformate decarboxylase activity. Purification of the benzoylformate decarboxylase, followed by N-terminal sequencing, enabled the design of probes for hybridization with P. stutzeri ST-201 genomic DNA libraries. Sequencing of two overlapping genomic DNA restriction fragments revealed two open reading frames which were denoted dpgB and dpgC. Sequence alignments suggested that the genes encoded a thiamin-diphosphate-dependent decarboxylase and an aldehyde dehydrogenase, respectively. Both genes were isolated and expressed in Escherichia coli. The dpgB gene product was confirmed as a benzoylformate decarboxylase while the dpgC gene product was characterized as a NAD+/NADP+-dependent benzaldehyde dehydrogenase. In keeping with their high sequence identities (both greater than 85%) the kinetic properties of the two enzymes were similar to those of the homologous enzymes in the mandelate pathway of Pseudomonas putida ATCC 12633. However, Pseudomonas stutzeri ST-201 was unable to grow on either isomer of mandelate, and sequencing indicated that the dpgB gene did not form part of an operon. Thus it appears that the two enzymes form part of a d-phenylglycine, rather than mandelate, degrading pathway.     

Isolation and characterization of a benzoylformate decarboxylase and a NAD/NADP-dependent benzaldehyde dehydrogenase involved in d-phenylglycine metabolism in Pseudomonas stutzeri ST-201

Following induction with d-phenylglycine both d-phenylglycine aminotransferase activity and benzoylformate decarboxylase activity were observed in cultures of Pseudomonas stutzeri ST-201. Induction with benzoylformate, on the other hand, induced only benzoylformate decarboxylase activity. Purification of the benzoylformate decarboxylase, followed by N-terminal sequencing, enabled the design of probes for hybridization with P. stutzeri ST-201 genomic DNA libraries. Sequencing of two overlapping genomic DNA restriction fragments revealed two open reading frames which were denoted dpgB and dpgC. Sequence alignments suggested that the genes encoded a thiamin-diphosphate-dependent decarboxylase and an aldehyde dehydrogenase, respectively. Both genes were isolated and expressed in Escherichia coli. The dpgB gene product was confirmed as a benzoylformate decarboxylase while the dpgC gene product was characterized as a NAD⁺/NADP⁺-dependent benzaldehyde dehydrogenase. In keeping with their high sequence identities (both greater than 85%) the kinetic properties of the two enzymes were similar to those of the homologous enzymes in the mandelate pathway of Pseudomonas putida ATCC 12633. However, Pseudomonas stutzeri ST-201 was unable to grow on either isomer of mandelate, and sequencing indicated that the dpgB gene did not form part of an operon. Thus it appears that the two enzymes form part of a d-phenylglycine, rather than mandelate, degrading pathway.     

Identification of Novel Benzoylformate Decarboxylases by Growth Selection

A growth selection system was established using Pseudomonas putida, which can grow on benzaldehyde as the sole carbon source. These bacteria presumably metabolize benzaldehyde via the β-ketoadipate pathway and were unable to grow in benzoylformate-containing selective medium, but the growth deficiency could be restored by expression in trans of genes encoding benzoylformate decarboxylases. The selection system was used to identify three novel benzoylformate decarboxylases, two of them originating from a chromosomal library of P. putida ATCC 12633 and the third from an environmental-DNA library. The novel P. putida enzymes BfdB and BfdC exhibited 83% homology to the benzoylformate decarboxylase from P. aeruginosa and 63% to the enzyme MdlC from P. putida ATCC 12633, whereas the metagenomic BfdM exhibited 72% homology to a putative benzoylformate decarboxylase from Polaromonas naphthalenivorans. BfdC was overexpressed in Escherichia coli, and the enzymatic activity was determined to be 22 U/ml using benzoylformate as the substrate. Our results clearly demonstrate that P. putida KT2440 is an appropriate selection host strain suitable to identify novel benzoylformate decarboxylase-encoding genes. In principle, this system is also applicable to identify a broad range of different industrially important enzymes, such as benzaldehyde lyases, benzoylformate decarboxylases, and hydroxynitrile lyases, which all catalyze the formation of benzaldehyde.     

Mechanism of benzaldehyde lyase studied via thiamin diphosphate-bound intermediates and kinetic isotope effects

Direct spectroscopic observation of thiamin diphosphate-bound intermediates was achieved on the enzyme benzaldehyde lyase, which carries out reversible and highly enantiospecific conversion of ( R)-benzoin to benzaldehyde. The key enamine intermediate could be observed at lambda max 393 nm in the benzoin breakdown direction and in the decarboxylase reaction starting with benzoylformate. With benzaldehyde as substrate, no intermediates could be detected, only formation of benzoin at 314 nm. To probe the rate-limiting step in the direction of ( R)-benzoin synthesis, the (1)H/ (2)H kinetic isotope effect was determined for benzaldehyde labeled at the aldehyde position and found to be small (1.14 +/- 0.03), indicating that ionization of the C2alphaH from C2alpha-hydroxybenzylthiamin diphosphate is not rate limiting. Use of the alternate substrates benzoylformic and phenylpyruvic acids (motivated by the observation that while a carboligase, benzaldehyde lyase could also catalyze the slow decarboxylation of 2-oxo acids) enabled the observation of the substrate-thiamin covalent intermediate via the 1',4'-iminopyrimidine tautomer, characteristic of all intermediates with a tetrahedral C2 substituent on ThDP. The reaction of benzaldehyde lyase with the chromophoric substrate analogue ( E)-2-oxo-4(pyridin-3-yl)-3-butenoic acid and its decarboxylated product ( E)-3-(pyridine-3-yl)acrylaldehyde enabled the detection of covalent adducts with both. Neither adduct underwent further reaction. An important finding of the studies is that all thiamin-related intermediates are in a chiral environment on benzaldehyde lyase as reflected by their circular dichroism signatures.     

Construction of an electro-enzymatic bioreactor for the production of (R)-mandelate from benzoylformate

Coupling both the electrocatalytic recycling of NADH and the enzymatic reduction of the substrate was used to produce (R)-mandelate from benzoylformate using benzoylformate reductase (BFR). The reduction of benzoylformate by BFR in combination with FAD-mediated electrolysis (at -0.5 V vs. Ag/AgCl) was complete in about 18 h and gave 47.5 mM (R)-mandelate from 50 mM substrate, while the process involving MV2+-mediated procedure (at -0.7 V vs. Ag/AgCl) produced 40 mM (R)-mandelate after 30 h. The overpotential for the NAD+ reduction could be decreased by about 0.2 V by substituting a toxic viologen derivative, MV2+, with a natural electron carrier, FAD. MV2+, however, decreased the productivity as BFR lost about 50% of its initial activity after 6 d in its presence.     

Asymmetric reduction of ethyl benzoylformate with chiral NADH model systems: Mechanistic and stereochemical consideration of the reactions based on the complexation properties of the model compounds

Asymmetric reductions of ethyl benzoylformate were conducted by use of NADH model compounds with C₁ or C₂ symmetry in the presence of magnesium perchlorate. It was found that NADH model compounds which form 2 : 1 chelation complexes with the magnesium ion showed the dependence of optical yield on the reaction conversion. The stereochemical behaviors of the model compounds were classified into three reaction types on the basis of the component ratio in the chelation complex between the reductants and the magnesium ion.     

Identification of novel benzoylformate decarboxylases by growth selection

A growth selection system was established using Pseudomonas putida, which can grow on benzaldehyde as the sole carbon source. These bacteria presumably metabolize benzaldehyde via the beta-ketoadipate pathway and were unable to grow in benzoylformate-containing selective medium, but the growth deficiency could be restored by expression in trans of genes encoding benzoylformate decarboxylases. The selection system was used to identify three novel benzoylformate decarboxylases, two of them originating from a chromosomal library of P. putida ATCC 12633 and the third from an environmental-DNA library. The novel P. putida enzymes BfdB and BfdC exhibited 83% homology to the benzoylformate decarboxylase from P. aeruginosa and 63% to the enzyme MdlC from P. putida ATCC 12633, whereas the metagenomic BfdM exhibited 72% homology to a putative benzoylformate decarboxylase from Polaromonas naphthalenivorans. BfdC was overexpressed in Escherichia coli, and the enzymatic activity was determined to be 22 U/ml using benzoylformate as the substrate. Our results clearly demonstrate that P. putida KT2440 is an appropriate selection host strain suitable to identify novel benzoylformate decarboxylase-encoding genes. In principle, this system is also applicable to identify a broad range of different industrially important enzymes, such as benzaldehyde lyases, benzoylformate decarboxylases, and hydroxynitrile lyases, which all catalyze the formation of benzaldehyde.     

Controlled-release biocatalysis for the synthesis of D-phenylglycine

An isolated and immobilised aminotransferase cloned from Pseudomonas stutzeri ST-201 into Escherichia coli was used to synthesise d-phenylglycine. The reaction was characterised by an unfavourable equilibrium constant and substrate inhibition. The use of a controlled-release system via the use of Amberlite (IRA 400)-adsorbed benzoylformate proved a useful technique to circumvent these issues. This resulted in a four-fold improvement in product concentration achievable to yield a final d-phenylglycine concentration of 10.25 mg/ml.     

Exploring the active site of benzaldehyde lyase by modeling and mutagenesis

Benzaldehyde lyase (BAL) is a thiamin diphosphate-dependent enzyme, which catalyzes the breakdown of (R)-benzoin to benzaldehyde. In essence, this is the reverse of the carboligation reaction catalyzed by benzoylformate decarboxylase (BFD). Here, we describe the first steps towards understanding the factors influencing BFD to form a CC bond under conditions wherein BAL will cleave the same bond. What are the similarities and differences between these two enzymes that result in the different catalytic activities? The X-ray structures of BFD and pyruvate decarboxylase (PDC) were used as templates for modeling benzaldehyde lyase. The model shows that a glutamine residue, Gln113, replaces the active site histidines of BFD and PDC. Replacement of the Gln113 by alanine or histidine reduced the value of k(cat) for lyase activity by more than 200-fold. The residues in BFD interacting with the phenyl ring of benzoylformate have similarly positioned counterparts in BAL but Ser26, the residue known to interact with the carboxylate group of benzoylformate, has been replaced by an alanine (Ala28). The BAL A28S variant exhibited 7% of WT activity in the BAL assay but, in the most intriguing result, this variant was able to catalyze the decarboxylation of benzoylformate. Conversely, the BFD S26A variant was unable to cleave benzoin.     

Integration of newly isolated biocatalyst and resin-based in situ product removal technique for the asymmetric synthesis of (R)-methyl mandelate

The enantioselective reduction of methyl benzoylformate to (R)-methyl mandelate, an important pharmaceutical intermediate and a versatile resolving agent, was investigated in this study. After minimizing the reaction-specific constraints (constraints dependent on the nature of the substrate and product) by preliminary selection of the reaction parameters, an effective whole cell biocatalyst (Saccharomyces cerevisiae AS2.1392) was obtained by simple screening procedures. Under further optimized conditions, a product concentration of 103 mmol L⁻¹ could be attained within 5 h with a yield of 85.8% and an enantiometric excess of 95.4%, indicating S. cerevisiae AS2.1392 an efficient biocatalyst for the asymmetric synthesis of (R)-methyl mandelate. Furthermore, resin-based in situ product removal (ISPR) technique was applied to alleviate the substrate and product inhibition or toxicity to the whole cells. The integration of newly isolated biocatalyst and proper ISPR technique provides a practical route for the preparation of optically active pharmaceutical intermediates.     

Structural and kinetic analysis of catalysis by a thiamin diphosphate-dependent enzyme,benzoylformate decarboxylase

Benzoylformate decarboxylase is a member of the family of enzymes that are dependent on the cofactor thiamin diphosphate. A structure of this enzyme binding (R)-mandelate, a competitive inhibitor, suggests that at least two hydrogen bonds are formed between the substrate, benzoylformate, and active site side chains. The first is between the carboxylate group of benzoylformate and the hydroxyl group of S26, and the second is between carbonyl group of the substrate and an imidazole nitrogen of H70. Steady-state kinetic studies indicate that the catalytic parameters are strongly affected in three active site mutants, S26A, H70A, and H281A. The K(m) of S26A was increased most dramatically, 25-fold more than that of the wild-type enzyme, while the K(i) of (R)-mandelate was increased 100-fold, suggesting that the serine hydroxyl is important for substrate binding. The k(cat) of H70A is reduced more than 3 orders of magnitude, strongly implicating this residue in catalysis, and H281 showed significant, but smaller magnitude, effects on both K(m) and k(cat). Stopped-flow experiments using an alternative substrate, p-nitrobenzoylformate, lead to kinetic resolution of the fate of key thiamin diphosphate-bound intermediates. Together, the experimental results suggest the following roles for residues in the active site. The residue H70 is important for the protonation of the 2-alpha-mandelyl-ThDP intermediate, thereby assisting in decarboxylation, and for the deprotonation of the 2-alpha-hydroxybenzyl-ThDP intermediate, aiding product release. H281 is involved in protonation of the enamine. Surprisingly, S26 appears to be involved not only in substrate binding but also in other steps of the reaction.     

Functional models of α-keto acid dependent nonheme iron oxygenases: synthesis and reactivity of biomimetic iron(II) benzoylformate complexes supported by a 2,9-dimethyl-1,10-phenanthroline ligand

Two biomimetic iron(II) benzoylformate complexes, [LFe^II(BF)₂] (2) and [LFe^II(NO₃)(BF)] (3) (L is 2,9-dimethyl-1,10-phenanthroline and BF is monoanionic benzoylformate), have been synthesized from an iron(II)–dichloro complex [LFe^IICl₂] (1). All the iron(II) complexes have been structurally and spectroscopically characterized. The iron(II) center in 2 is coordinated by a bidentate NN ligand (2,9-dimethyl-1,10-phenanthroline) and two monoanionic benzoylformates to form a distorted octahedral coordination geometry. One of the benzoylformates binds to the iron in 2 via both carboxylate oxygens but the other one binds in a chelating bidentate fashion via one carboxylate oxygen and the keto oxygen. On the other hand, the iron(II) center in 3 is ligated by one NN ligand, one bidentate nitrate, and one monoanionic chelating benzoylformate. Both iron(II) benzoylformate complexes exhibit the facial NNO donor environment in their solid-state structures. Complexes 2 and 3 are stable in noncoordinating solvents under an inert atmosphere, but react with dioxygen under ambient conditions to undergo oxidative decarboxylation of benzoylformate to benzoate in high yields. Evidence for the formation of an iron(IV)–oxo intermediate upon oxidative decarboxylation of benzoylformate was obtained by interception and labeling experiments. The iron(II) benzoylformate complexes represent the functional models of α-keto acid dependent oxygenases.