Phenylhydrazine as probe for cofactor identification in amine oxidoreductases. Evidence for PQQ as the cofactor in methylamine dehydrogenase.

Homogeneous methylamine dehydrogenase (primary-amine:(acceptor) oxidoreductase (deaminating), EC 1.4.99.3, MADH) from the bacterium Thiobacillus versutus was treated with the inhibitor phenylhydrazine (PH). Derivatization of the cofactor in MADH took place in a fast reaction to give compound I. A different product, compound II, was formed in a slow reaction at high O2 concentrations. The compounds I and II could be removed from the protein by proteolysis with pronase and purified to homogeneity. Products showing identical absorption spectra and chromatographic behaviour were isolated from the reaction mixture after incubating pyrroloquinoline quinone (PQQ) with PH. Upon dissolving in dimethyl sulphoxide, both the enzyme-derived as well as the model-system-derived compounds I and II were nearly quantitatively transformed into PQQ. The conclusion is, therefore, that MADH from T. versutus contains covalently bound PQQ, removable from the protein with pronase, and not a structural analogue of this cofactor without the carboxylic acid groups, as was recently proposed for MADH from Bacterium W3A1 [(1986) Biochem. Biophys. Res. Commun. 141, 562-568]. The properties of compounds I and II suggest that they are the 'azo adduct' and the 'hydrazone adduct' of PH and PQQ at the C(5)-position, respectively. The finding that the reaction of a hydrazine with PQQ can lead to two different products, in enzymes as well as in a model system, has important implications for the interpretation of recent comparative studies aimed at detection of PQQ in amine oxidoreductases with Raman spectroscopy.

Dopamine beta-hydroxylase from bovine adrenal medulla contains covalently-bound pyrroloquinoline quinone.

Treatment of homogeneous dopamine beta-hydroxylase (DBH) preparations from bovine adrenals with the inhibitor phenylhydrazine (PH) changed the structureless absorption spectrum of DBH into spectra with a maximum at 350 nm. A product with this absorption spectrum could be detached with pronase, enabling its isolation. It appeared to be the C(5) hydrazone of pyrroloquinoline quinone (PQQ) and PH, as judged from its properties and the fact that it could be transformed into PQQ itself. From the yield obtained a ratio of 0.85 PQQ per enzyme subunit was calculated. In contrast to copper-quinoprotein amine oxidases (EC 1.4.3.6), hydrazone formation in DBH did not require saturation of the mixture with O2. DBH is the first copper-quinoprotein hydroxylase found so far. The implications of this finding for the current views on mechanism of action and inhibition by hydrazines are discussed. The success of the recently developed 'hydrazine method' [(1987) FEBS Lett. 221, 299-304] for all different types of amine oxidoreductases, suggest that the method could also be applied to other enzymes for which hydrazines are inhibitors and where the identity of the cofactors has not been established or the presence of PQQ is suspected.

Hydrazone formation of 2,4-dinitrophenylhydrazine with pyrroloquinoline quinone in porcine kidney diamine oxidase.

Homogeneous diamine oxidase (EC 1.4.3.6) from porcine kidney was treated with the inhibitor 2,4-dinitrophenylhydrazine (DNPH). The coloured compounds formed were detached with pronase and purified to homogeneity. When the reaction with DNPH was conducted under an O2 atmosphere, the product (obtained in a yield of 55%) was the C(5)-hydrazone of pyrroloquinoline quinone (PQQ) and DNPH, as revealed by its chromatographic behaviour, absorption spectrum and 1H-NMR spectrum. Only 6% of this hydrazone was formed under air, the main product isolated being an unidentified reaction product of DNPH with the enzyme. Porcine kidney diamine oxidase is the second mammalian enzyme shown to have PQQ as its prosthetic group. In view of the requirements for hydrazone formation with DNPH, it is incorrect to assume that inhibition of this type of enzymes with common hydrazines is simply due to blocking of the carbonyl group of its cofactor.

Bovine serum amine oxidase: a mammalian enzyme having covalently bound PQQ as prosthetic group.

In addition to the metal ion, copper-containing amine oxidases possess an organic prosthetic group, the nature of which has long been controversial. We show here that in the case of bovine plasma amine oxidase, this second prosthetic group is covalently bound pyrroloquinoline quinone ( PQQ ). Until now the coenzyme PQQ has been found in several bacterial dehydrogenases. Thus the finding reported here is the first example of a quinoprotein oxidoreductase discovered in a eukaryotic organism.

The redox-cycling assay is not suited for the detection of pyrroloquinoline quinone in biological samples.

Based on the results of the so-called redox-cycling assay it has been claimed that various common foods and beverages as well as mammalian body fluids and tissues contain substantial quantities (microM) of free PQQ [M. Paz et al. (1989) in: PQQ and Quinoproteins (J.A. Jongejan and J.A. Duine, eds.) Kluwer Academic Publishers, Dordrecht, pp. 131-143 and J. Killgore et al. (1989) Science 245, 850-852]. However, by investigating samples from such sources with a biological assay of nM sensitivity, we could not confirm these claims. Analysis of the samples with procedures that proved adequate for the detection of PQQ adducts and conjugates gave equally negative results. To account for the positive response in the redox-cycling assay, as opposed to the negative results obtained by other methods, a search was made for those substances in these samples that caused the false-positive reactions. It was found that a number of commonly occurring biochemicals like ascorbic and dehydroascorbic acid, riboflavin and to a lesser extent pyridoxal phosphate, gave a positive response in the redox-cycling assay. The amounts of these interfering substances that were determined in the samples by independent methods could well explain the response. In separate experiments it was found that the effect of PQQ added to biological samples was obscured over an appreciable range of concentrations. For these reasons it must be concluded that the redox-cycling assay is not suited for the detection of PQQ in these samples. Any claims that are based on the results of this method should be disregarded.

Direct hydride transfer in the reaction mechanism of quinoprotein alcohol dehydrogenases: a quantum mechanical investigation.

Oxidation of alcohols by direct hydride transfer to the pyrroloquinoline quinone (PQQ) cofactor of quinoprotein alcohol dehydrogenases has been studied using ab initio quantum mechanical methods. Energies and geometries were calculated at the 6-31G(d,p) level of theory. Comparison of the results obtained for PQQ and several derivatives with available structural and spectroscopic data served to judge the feasibility of the calculations. The role of calcium in the enzymatic reaction mechanism has been investigated. Transition state searches have been conducted at the semiempirical and STO-3G(d) level of theory. It is concluded that hydride transfer from the Calpha-position of the substrate alcohol (or aldehyde) directly to the C(5) carbon of PQQ is energetically feasible. Copyright 2001 John Wiley & Sons, Inc. J Comput Chem 22: 1732-1749, 2001

Factors relevant to the production of (R)-(+)-glycidol (2,3-epoxy-1-propanol) from racemic glycidol by enantioselective oxidation with Acetobacter pasteurianus ATCC 12874.

Acetobacter pasteurianus oxidizes glycidol with high activity, comparable to the oxidation of ethanol. The organism has a preference for the S-enantiomer, and the kinetic resolution process obeys a simple relationship, indicating an enantiomeric ratio (E) of 19. The compound is converted into glycidic acid, although a transient accumulation of glycidaldehyde occurs initially. Determination of other parameters revealed a temperature optimum of 50 degrees C, long-term stability (cells in the resting state), and a pH optimum compatible with the chemical stability of glycidol. However, it was also noted that respiration rates decrease at concentrations of glycidol above 1 M. This is most likely caused by substrate inhibition of the glycidol-oxidizing enzyme, the quinohemoprotein ethanol dehydrogenase. Comparison with existing methods for enantiomerically pure glycidol production indicated a number of attractive points for the method described here, although definitive evaluation must await further studies on the long-term stability under process conditions, reusability of the cells, and the mechanism of glycidol inhibition.

On the relationship between affinity for molecular hydrogen and the physiological directionality of hydrogenases.

The physiological significance of the generic reaction H(2)<-->2[H] is not always clear because hydrogenases may function in the breakdown of molecular hydrogen or in its synthesis or in both directions. Fe-hydrogenases have nevertheless been most often associated with proton reduction and NiFe-hydrogenases with hydrogen oxidation. A re-determination of the K(M) of H(2) oxidation by Pyrococcus furiosus NiFe-hydrogenase-I and by Desulfovibrio vulgaris Fe-hydrogenase suggests that affinity for hydrogen has been seriously underestimated and that the kinetics of hydrogen activation in relation to the directionality of hydrogenases should be re-evaluated.

Solvent effect on lipase enantioselectivity. Evidence for the presence of two thermodynamic states.

The enantioselectivity of lipase-catalyzed kinetic resolutions has been measured at various temperatures in binary mixtures of solvents. Varying the solvent composition and temperature had a profound effect on the enantiomeric ratio. The values for delta delta H(R-S)(#) and delta delta S(R-S)(#), calculated from the E values measured at various temperatures, were estimated as a function of the solvent composition. By plotting delta delta H(R-S)(#) versus delta delta S(R-S)(#) as a function of the solvent composition, an extreme was observed. The resulting "hairpin-type" enthalpy-entropy compensation plots can be described by assuming the presence of two thermodynamically distinct physical states, displaying different enantioselectivities, that are in equilibrium with one another. Changing the solvent composition results in a change in the equilibrium constant K(eq) for the two states. The intriguing bell-shaped curves of the enantioselectivity versus solvent composition observed for lipase-catalyzed kinetic resolutions can be described assuming a linear correlation for the logarithm of K(eq) and the solvent composition. Thus, a simulation of the two-state model adequately describes the solvent effects found for lipase-catalyzed kinetic resolutions in binary mixtures of solvents and possibly in series of homologous organic solvents.

Detection and determination of pyrroloquinoline quinone, the coenzyme of quinoproteins.

A convenient determination of pyrroloquinoline quinone (PQQ) in extracts of purified enzymes is possible with ion-pair chromatography on a HPLC reverse-phase column and with uv detection. However, when culture supernatants have to be analyzed, a fluorescence detection system is more appropriate. Proof for the presence of PQQ can be obtained by treating such a sample with butyraldehyde, which converts the coenzyme into a stable adduct having a suitable retention time in the system. The sensitivity and selectivity of the analysis can be further enhanced by reducing the sample with NaBH4, which produces the dihydrodiol form of the coenzyme (PQQH4) and oxidizing PQQH4 with NaIO4 to a strongly fluorescing compound. A procedure is described for the easy preparation of an apoenzyme from the quinoprotein glucose dehydrogenase of Pseudomonas aeruginosa strains. With this biological test system, very low amounts of PQQ can be detected. However, when PQQ is present in the form of adducts, the analysis has to be performed via reduction to PQQH4, oxidation with NaIO4, and HPLC.

Homology model of the quinohaemoprotein alcohol dehydrogenase from Comamonas testosteroni.

A molecular model of QH-ADH, the quinohaemoprotein alcohol dehydrogenase from Comamonas testosteroni, has been built by homology modelling. Sequence similarity of N-terminal residues 1-570 with the alpha-subunit of quinoprotein methanol dehydrogenases (MDHs) from Methylophilus methylotrophus W3A1 and Methylobacterium extorquens provided a basis for the design of the PQQ-binding domain of QH-ADH. Minimal sequence similarity with cytochrome c551 from Ectothiorhodospira halophila and cytochrome c5 from Azotobacter vinelandii has been used to model the C-terminal haem c-binding domain, residues 571-677, absent in MDHs. Distance constraints inferred from 19F-NMR relaxation studies of trifluoromethylphenylhydrazine-derivatized PQQ bound to QH-ADH apoenzyme as well as theoretical relations for optimal electron transfer have been employed to position the haem- and PQQ-binding domains relative to each other. The homology model obtained shows overall topological similarity with the crystal structure of cd1-nitrite reductase from Thiosphera pantotropha. The proposed model accounts for the following: (i) the site that is sensitive to in vivo proteolytic attack; (ii) the substrate specificity in comparison with MDHs; (iii) changes of the spectral properties of the haem c upon reconstitution of apo-enzyme with PQQ; (iv) electronic interaction between haem and PQQ; and (v) enantioselectivity in the conversion of a chiral sec alcohol.

Pyrroloquinoline quinone as cofactor in galactose oxidase (EC 1.1.3.9).

Galactose oxidase from Dactylium dendroides was shown to contain one molecule of covalently bound pyrroloquinoline quinone (PQQ/enzyme molecule. From the spectroscopic characteristics reported for the enzyme forms, a mechanistic role for PQQ could be deduced. In analogy with other quinoproteins, the initial formation of a PQQ-substrate adduct is proposed. Following internal hydrogen transfer, leading to aldehyde product and reduced pyrroloquinoline quinone, reoxidation of the organic cofactor with molecular oxygen could be mediated by the PQQ-liganded copper ion with concomitant formation of hydrogen peroxide. With PQQ as an additional (two-electron) redox center the occurrence of a "superoxidized" enzyme form must be considered. Possible consequences of this view, in relation to a physiological function of the enzyme and interpretation of ESR data, are discussed.

Factors relevant in the reaction of pyrroloquinoline quinone with amino acids. Analytical and mechanistic implications.

In order to reveal the stability of pyrroloquinoline quinone (PQQ) in complex samples, its reaction on incubation with amino acids was followed spectrophotometrically by monitoring oxygen consumption, and with a biological assay. For several alpha-amino acids, the formation of a yellow coloured compound (lambda max = 420 nm) was accompanied by oxygen uptake and disappearance of biological activity from the reaction mixture. The yellow product appeared to be an oxazole of PQQ, the exact structure depending on the amino acid used. Oxazole formation also occurred under anaerobic conditions with concomitant formation of PQQH2, suggesting that PQQ is able to oxidize the presumed oxazoline to the oxazole. Besides the condensation reaction, there is also a catalytic cycle in which an aldimine adduct of PQQ and the amino acid is converted into the aminophenol form of the cofactor and an aldehyde resulting from oxidative decarboxylation of the amino acid. Addition of NH4+ salts, as well as that of certain divalent cations, greatly stimulated both the cyclic and the linear reaction. With basic amino acids, oxazole formation scarcely occurred. However, as oxygen consumption was observed (provided that certain divalent cations were present), conversion of these compounds took place. A reaction scheme is proposed accounting for the products formed and the effects observed. Since NH4+ ions activate several quinoproteins (PQQ-containing enzymes) and divalent cations (Ca2+, Fe2+, and Cu2+) are additional (co)factors in certain metallo quinoproteins, the effects of metal ions observed here could be related to the mechanistic features of these enzymes. Although all oxazoles were converted to PQQ by acid hydrolysis, PQQ was not detected when hydrolysis was carried out in the presence of tryptophan, a compound which appeared to have a deleterious effect on the cofactor under this condition. The results here described explain why analysis methods for free PQQ in complex samples fail in certain cases, or are not quantitative.

Inactivation of quinoprotein alcohol dehydrogenases with cyclopropane-derived suicide substrates. .

Quinoprotein alcohol dehydrogenases can be inactivated by cyclopropanol, cyclopropanone hydrate, and, depending on whether they can oxidize secondary alcohols, also by cyclopropanone ethyl hemiketal. Only enzyme molecules containing the oxidized coenzyme (PQQ), but not those with the coenzyme in the semiquinone form (PQQH), become inactivated with these compounds. The inactivation process proceeds without proton production or electron acceptor consumption and free radical is not observed in the inactivated enzyme. It could be demonstrated that a stoichiometric relationship exists between enzyme inactivation, PQQ converted, PQQ adduct formed, and cyclopropanol added. Thus the dimeric and monomeric enzyme become fully inactivated with two and one molecule of cyclopropanol, respectively, indicating that the dimeric enzyme contains two independently acting catalytic sites. Inactivation of the enzyme by cyclopropanol and cyclopropanone hydrate produces chromatographically different PQQ adducts. Since cyclopropanemethanol, cyclobutanol and cyclohexanol are not suicide substrates, the inactivation presumably proceeds via a ring opening such as proposed for the metal-ion-catalysed degradation of cyclopropane derivatives. The results are in accordance with our view on the reaction mechanism of these enzymes but not with that of others [Mincey et al. (1981) Biochemistry 20, 7502-7509]. The reasons why their model has to be refuted are discussed.

Competitive lipase-catalyzed ester hydrolysis and ammoniolysis in organic solvents; equilibrium model of a solid-liquid-vapor system.

Enzymatic ester hydrolysis and ammoniolysis were performed as competitive reactions in methyl isobutyl ketone without a separate aqueous phase. The reaction system contained solid ammonium bicarbonate, which dissolved as water, ammonia, and carbon dioxide. During the reaction an organic liquid phase, a vapor phase, and at least one solid phase are present. The overall equilibrium composition of this multiphase system is a complex function of the reaction equilibria and several phase equilibria. To gain a quantitative understanding of this system a mathematical model was developed and evaluated. The model is based on the mass balances for a closed batch system and straightforward relations for the reaction equilibria and the solubility equilibria of ammonium bicarbonate, the fatty acid ammonium salt, water, ammonia, and carbon dioxide. For butyl butyrate as a model ester and Candida antarctica lipase B as the biocatalyst this equilibrium model describes the experiments satisfactorily. The model predicts that high equilibrium yields of butyric acid can be achieved only in the absence of ammoniolysis or in the presence of a separate water phase. However, high yields of butyramide should be possible if the water concentration is fixed at a low level and a more suited source of ammonia is applied.

Characterization of the Enantioselective Properties of the Quinohemoprotein Alcohol Dehydrogenase of Acetobacter pasteurianus LMG 1635. 1. Different Enantiomeric Ratios of Whole Cells and Purified Enzyme in the Kinetic Resolution of Racemic Glycidol.

Resting cells of Acetobacter pasteurianus LMG 1635 (ATCC 12874) show appreciable enantioselectivity (E=16-18) in the oxidative kinetic resolution of racemic 2,3-epoxy-1-propanol, glycidol. Distinctly lower values (E=7-9) are observed for the ferricyanide-coupled oxidation of glycidol by the isolated quinohemoprotein alcohol dehydrogenase, QH-ADH, which is responsible for the enantiospecific oxidation step in whole cells. The accuracy of E-values from conversion experiments could be verified using complementary methods for the measurement of enantiomeric ratios. Effects of pH, detergent, the use of artificial electron acceptors, and the presence of intermediate aldehydes, could be accounted for. Measurements of E-values at successive stages of the purification showed that the drop in enantioselectivity correlates with the separation of QH-ADH from the cytoplasmic membrane. It is argued that the native arrangement of QH-ADH in the membrane-associated complex favors the higher E-values. The consequences of these findings for the use of whole cells versus purified enzymes in biocatalytic kinetic resolutions of chiral alcohols are discussed.

Enzymes involved in the glycidaldehyde (2,3-epoxy-propanal) oxidation step in the kinetic resolution of racemic glycidol (2,3-epoxy-1-propanol) by Acetobacter pasteurianus.

It is already known that kinetic resolution of racemic glycidol (2,3-epoxy-1-propanol) takes place when Acetobacter pasteurianus oxidizes the compound to glycidic acid (2,3-epoxy-propionic acid) with glycidaldehyde (2,3-epoxy-propanal) proposed to be the transient seen in this conversion. Since inhibition affects the feasibility of a process based on this conversion in a negative sense, and the chemical reactivity of glycidaldehyde predicts that it could be the cause for the phenomena observed, it is important to know which enzyme(s) oxidise(s) this compound. To study this, rac.- as well as (R)-glycidaldehyde were prepared by chemical synthesis and analytical methods developed for their determination. It appears that purified quinohemoprotein alcohol dehydrogenase (QH-ADH type II), the enzyme responsible for the kinetic resolution of rac.-glycidol, also catalyses the oxidation of glycidaldehyde. In addition, a preparation exhibiting dye-linked aldehyde dehydrogenase activity for acetaldehyde, most probably originating from molybdohemoprotein aldehyde dehydrogenase (ALDH), which has been described for other Acetic acid bacteria, oxidised glycidaldehyde as well with a preference for the (R)-enantiomer, the selectivity quantified by an enantiomeric ratio (E) value of 7. From a comparison of the apparent kinetic parameter values of QH-ADH and ALDH, it is concluded that ALDH is mainly responsible for the removal of glycidaldehyde in conversions of glycidol catalysed by A. pasteurianus cells. It is shown that the transient observed in rac.-glycidol conversion by whole cells, is indeed (R)-glycidaldehyde. Since both QH-ADH and ALDH are responsible for vinegar production from ethanol by Acetobacters, growth and induction conditions optimal for this process seem also suited to yield cells with high catalytic performance with respect to kinetic resolution of glycidol and prevention of formation of inhibitory concentrations glycidaldehyde.

Do organic solvents affect the catalytic properties of lipase? Intrinsic kinetic parameters of lipases in ester hydrolysis and formation in various organic solvents.

When it is assumed that organic solvents do not interfere with the binding process nor with the catalytic mechanism, the contribution of substrate-solvent interactions to enzyme kinetics can be accounted for by just replacing substrate concentrations in the equations by thermodynamic activities. It appears from the transformation that only the affinity parameters (K(m), K(sp)) are affected by this. Thus, in theory, the values of these corrected, intrinsic parameters (K(m) (int), k(sp) (int)) and the maximal rate (V(1)) should be equal for all media. This was tested for hydrolysis, transesterification, and esterification reactions catalyzed by pig pancreas lipase and Pseudomonas cepacia lipase in various organic solvents. Correction was carried out via experimentally determined activity coefficients for the substrates in these solvents or, if not feasible, from values in data bases. However, although the kinetic performances of each enzyme in the solvents became much more similar after correction, differences still remained. Analysis of the enzyme suspensions revealed massive particles, which explains the low activity of enzymes in organic solvents. However, no correlation was found between estimates of the amount of catalytically available enzyme (present at the surface of suspended particles or immobilized on beads) and the maximal rates observed. Moreover, the solvents had similar effects on the intrinsic parameters of suspended and immobilized enzyme. The possible causes for the effects of the solvents on the catalytic performance of the enzymes, remaining after correction for solvent-substrate interactions and the amount of participating enzyme, are discussed with respect to the premises on which the correction method is based. (c) 1995 John Wiley & Sons, Inc.

Description of the kinetic mechanism and the enantioselectivity of quinohaemoprotein ethanol dehydrogenase from Comamonas testosteroni in the oxidation of alcohols and aldehydes.

Initial rate studies were performed on the oxidation of (racemic) alcohols as well as aldehydes by quinohaemoprotein ethanol dehydrogenase, type 1, from Comamonas testosteroni with potassium ferricyanide as electron acceptor. The data could be fitted with an equation derived for a mechanism (hexa-uni ping-pong) in which alcohols are oxidized to the corresponding carboxylic acids and the intermediate aldehyde becomes released from the enzyme. However, for some substrates it was necessary to assume that they exert uncompetitive inhibition. The same model was used to fit the data of conversion processes. Reversible inactivation of the enzyme takes place during the conversion, the extent being inversely proportional to the concentration of ferricyanide present at the start. From the values of the kinetic parameters obtained for (R)- and (S)-solketal [2,2-dimethyl-4-(hydroxymethyl)-1,3-dioxolane] and their corresponding aldehydes, it appeared that the second step in (S)-solketal conversion is much faster than the first one and that opposite enantiomeric preferences exist for the alcohol and the aldehyde substrates. Since the initial rate measurements as well as the progress curve analysis gave similar kinetic parameter values and product analysis revealed intermediates in the amounts predicted, it is concluded that the kinetic and enantioselective behaviour of the enzyme is adequately described by the model presented here. Finally, the results indicate that both kinetic approaches should be used in conversions with consecutive reactions since they provide complementary information.