BRIEF REVIEW OF Ph.D. PROJECT

SCIENTIFIC IMPORTANCE

Formate dehydrogenase (FDH, formate: NAD⁺ oxydoreductase E.C. 1.2.1.2) catalyzes NAD⁺-dependent oxidation of formate to carbon dioxide (1).

NAD⁺ + formate ==> NADH + CO₂ (1)

FDH is one of the best-studied NAD+-dependent dehydrogenases of 2-oxyacids. This group of enzymes has high-level homology of amino acid sequences and similar 3D-structures. In our laboratory, we crystallized and resolved at resolution (2Å) 3D-structures as apoenzyme (FDH) as well as holocomplexes: FDH-ADPR and FDH-NAD⁺-azide. We also identified amino acid residues crucial for the binding of coenzyme (NAD⁺) and azide anion, the closest structural analogue of the substrate and formulated the hypothetic mechanism of the enzyme action.

The active center of all dehydrogenases of 2-oxyacids consists of three conserved amino acid residues at the substrate-binding site (Figure 1): Arg, COO^- (Glu or Asp) and His. Back chain of Arg anchors the carboxyl group of the substrate. Carboxyl group and His form a proton transfer chain that facilitates the achievement of the transition state of the reaction (1).

There are two kinds of dehydrogenases of 2-oxyacids: ”d” and “l”. d-Dehydrogenases oxidize d-stereo isomers. In the active center they have Glu. In the active center of l-dehydrogenases, that convert l-stereo isomers of 2-oxyacids, -COO^- represented by Asp.

In the family of dehydrogenases of 2-oxyacids, FDH occupies a special place. First, the substrate, formate anion, is a C₁- compound. Moreover, the list of amino acid residues found at the substrate-binding site is not the same: Gln substituted the conserved Glu. This substitution disrupts the proton transfer chain that can’t function as it does in the active center of other dehydrogenases. Thus, in the active center of FDH, the conserved amino acid residues should play different role or have some extra functions.

In the active center of dehydrogenases of 2-oxyacids except FDH, carbonyl group of the substrate forms a fork-like non-covalent complex with guanidine group of conserved Arg (Figure 1). On the other hand, formation of the same complex in the active center of FDH, according to the results of X-ray analysis, will block the catalysis. Moreover, substitution of Glu by Gln (Gln313, Figure 2) should result in formation of hydrogen bond between its amide group and imidazol of conserved His (His332, Figure 2). Thus, His332, the structural analogue of conserved His of 2-oxydehydrogenases should be deprotonated and can’t participate in the transfer of proton.

Based on these facts, we could conclude that the molecular mechanism of formate dehydrogenase was not established yet and requires further concretization. Thus, the study of molecular mechanism of FDH remains actual and important as a scientific problem.

GOAL OF THE PROJECT

In the present study, we would improve our knowledge of molecular mechanism of action and catalysis of NAD⁺-dependent FDH. We studied the role of conserved Arg (Arg284) by site-directed mutagenesis (1). We also obtained, analyzed and interpreted the pH-dependences of different kinetic and spectral parameters of the enzyme (2). We studied the effect of thioformate as one of the closest structural analogues of the substrate on FDH, and the influence of modification with pyridoxal on kinetic properties of the enzyme (3) Finally, we also purified FDH from methylotrophic yeast Hansenula polymorpha and determined the type of its kinetic mechanism (4).

SCIENTIFIC INNOVATION

As a scientific hypothesis the molecular mechanism of action of FDH required further development. In the present study, we analyzed original and already published data of the enzyme kinetic as well as the structure of the active center of FDH. We studied the role of Arg284 in the enzyme molecular mechanism of action. We confirmed that this amino acid residue directly participated in the binding of the substrate. We also established how the conserved Arg could contribute in enzyme catalysis and maintenance of the native conformation of the active center. We obtained and interpreted pH- dependences of different kinetic and spectral parameters: identified amino acid residues responsible for the appearance pH-transitions in acidic pH-range and discussed possible reasons of the enzyme inactivation under basic pH values. Our results explained inhibition of FDH by pyridoxal and its closest structural analogues. Finally, we systemized the data about kinetic and structural properties of FDH purified from different sources.

PUBLICATIONS AND PRESENTATIONS

The present study summarizes 6 publications printed in international and national scientific journals. The data was presented at international (BIOCATALYSIS, 1995, Suzdal, Russia) and domestic (Autotrophic microorganisms, 1996, Moscow, Russia) conferences as well as an annual competition of the best scientific works in the A.N. Bakh Institute of Biochemistry (1995, Moscow, Russia).

DESCRIPTION OF THE MANUSCRIPT

The manuscript consisted of the review (2 chapters) and experimental section. The experimental section subdivided at material and methods, results and discussion (4 chapters) conclusion and references. The manuscript contained original graphics and tables.

MATERIALS AND METHODS

MATERIALS

Microorganisms

Biomass of Pseudomonas sp. 101 was obtained from the Russian collection of microorganisms (VKM B-1545). Strains of methylotrophic yeast Candida boidini and Pichia pastoris were a kind gift of Prof. M.R. Kula (H. Heine Düsseldorf University, Germany) and Prof. I.I. Tolstorukov (VNII Genetic, Moscow, Russia), correspondently. Biomass of E. coli transformed with mutant FDH was obtained from Prof. V.I. Tishkov (Moscow State University, Moscow Russia).

Protein purification

Wild type bacterial FDH and mutant enzymes were purified according to a standard protocol from Pseudomonas sp. 101 and E. coli, correspondently. The mutant enzymes were detected by ELISA with monoclonal antibodies specific to the native bacterial enzyme. The yeast FDH was purified with original procedure by hydrophobic chromatography in descending gradient of NaCl followed by gel-filtration. Final protein preparations were homogeneous in SDS-polyacrylamide gel.

Methods

The spectrophotometer Hitachi 557 (Hitachi, Japan) was used for the spectral analysis, protein, and enzyme assay. Protein concentration was determined by biuret method or as absorbance at wavelength, l 280 nm (e=7.7 M^-1 x cm^-1). The FDH assay was based on the detection of NADH (e=6,220 M^-1 x cm^-1, l=340 nm). The fluorimeter Hitachi 4FM (Hitachi, Japan) was used for the fluorescence analysis. Calculation of K_b, (binding constant), K_i, (inhibition constant) as well as concentration of the enzyme active centers was done according to the protocols published earlier. Circular dichroism spectra recorded at the spectrophotometer JASCO J-40AS (Jasco, Japan). The purity of protein preparations was controlled by SDS-polyacrylamide electrophoresis.

RESULTS AND DISCUSSION

The role of Arg284 in the active center of FDH

To study the role of Arg284 in the active center of the enzyme we obtained two mutant FDH: Arg284>Gln and Arg284>Ala. For substitution, we chose amino acid residues that polarity; capability to form hydrogen bands and the length of back chain was different from Arg. Indeed, the back chain of Gln was semi-polar, 2 Å shorter and capable to formation of hydrogen bands. On the other hand, short methyl group of Ala had no charge was hydrophobic and not capable to formation of hydrogen bonds. According to the results of X-ray analysis, Arg284 should participate directly in the substrate binding by formation of a hydrogen band with one of two oxygen atoms of formate anion. We expected that substitution of Arg by Gln and Ala would abolish an interaction between the enzyme and substrate and affect on the enzyme activity.

We found (Table 1) that kinetic parameters of FDH were very sensitive to substitution of Arg284. Substitution of Arg284 with Gln resulted in sufficient decrease of V_max more than in 33 times (DG^@~8.8 kJ/M), K_i^azide and K_m^formate also dropped dramatically in 10 (DG~13.0 kJ/M) and 170 (DG~5.8 kJ/M) times, correspondently. Change of the free energy; DG^@ that caused the substitution of Arg by Gln (DG~5.8-8.8 kJ/M) was equivalent to disruption of 1 or 2 hydrogen bonds. More sufficient changes of DG for azide might direct that site-directed mutagenesis affected more on transition state of the reaction. Thus, some interactions inside the active center of native FDH that stabilized the transition state of the reaction were abolished or even terminated in the active center of mutant proteins (Figure 3). We speculated that binding of the substrate (substrate analogue) with flexible loop formed by Ile122 and Gly123 that also interacted directly with Arg284 could be one of these interactions. In mutant proteins, no hydrogen band could be formed between X284, Ile122 and Gly123. Thus, this important stabilizing interaction should be eliminated.

Substitution of Arg by Ala, which back chain (-CH₃) couldn’t form H-bands was fatal for the enzyme activity. The mutant protein did not bind azide anion.

We concluded that Arg284 as well as conserved Arg of the dehydrogenases of 2-oxyacids participated directly in the enzyme catalysis. It should form one or two hydrogen bands with one of two oxygen atoms of the substrate.

According to the results of X-ray analysis, Arg284 could form five hydrogen bands with the substrate and following amino acid residues: Ile122, Gly123, Asp125 and Ala283. Thus, this amino acid residue could be most sufficient structural elements of the enzyme active center. The molecular modeling demonstrated that substitution of the Arg guanidine group by the Gln amide group should be accompanied by disruption or abolishment of some hydrogen bonds. That should make some local conformation changes inside the active center as well as proposed area of the catalytic process.

To make a conclusion about the structural role of Arg284, we studied spectral parameters that could be sensitive to conformational changes of the enzyme (CD, tryptophan fluorescence). Moreover, we also studied thermoinactivation of the enzyme. According to our results (Figure 4A, B), substitution of Arg284 should lead significant conformational changes in the enzyme molecule. The amplitude of CD-spectra as well as tryptophan fluorescence dramatically decreased in mutant proteins especially after the substitution of Arg284 with Ala. Moreover, the maximum of tryptophan fluorescence shifted at 10 nm to the short-wave side of the spectrum. Finally, substitution of Arg284 by Gln essentially improved the enzyme thermostability (Fig. 4C). We proposed that could be due to the removal of non-compensated positive Arg charge from the active center.

Thus, Arg284 should be important for maintaining of the catalytically active conformation of the enzyme active center.

Despite of significant conformational changes that happened in the active center of the enzyme as a result of site-directed mutagenesis, the binding of coenzyme (K_b^NAD) even became stronger (Table 1). In the mutant enzyme Arg284>Ala, it reached value 80 μM. This value was comparable to the binding constant of the reduced coenzyme (K_b^NADH) of the wild type FDH (100 μM) and was equivalent to the DG ~3.3 kJ/M.

According to the central dogma of catalysis, the enzyme should destabilize the reactants initial states and (or) stabilize the transition state of reaction. Indeed, if Arg284 participated in the catalysis it should destabilize the initial state of the coenzyme. Our results indirectly confirmed this hypothesis.

Thus, the positive charge of Arg284 promoted better binding of NADH to NAD and could contribute into the catalysis by destabilization of NAD as one of initial reactants.

pH-dependences of kinetic and spectral parameters

All known FDH had a wide optimum at neutral pH-values. The kinetic parameters of FDH purified from methylotrophic bacterium Pseudomonas sp. 101 remained constant at pH-range from 6 to 10. At pH-values lower than 6, the Michaelis constant for coenzyme (K_m^NAD) grew up and the maximal velocity of reaction (V_max) dropped down. In the alkaline buffer solutions (pH 10 or higher), binding of both reactants (NAD and formate) became worse. Despite of this, V_max remained constant (Fig. 5).

The proportional dependence of K_m^NAD on [H]² at low pH (Fig. 6) could be a result of two protons simultaneous binding to two functional groups with close or same pK-value located at the coenzyme-binding site (Scheme 1). According to our data, the pK-values for binary (FDH-NAD) and ternary (FDH-NAD-azide) complexes were equivalent to 6.0±0.1 and 5.4±0.1 correspondently. On the other hand, the ionization enthalpy of the process (DH^ion~-7.5±7.9 kJ/M) was close to zero. Thus, the protons should bind to the carboxyl groups. That could be –COO^-- groups of aspartatic or glutamic acid(s). Changes of V_max at acidic pH (pH 6 and lower) could be related to deionization of only one functional group (pK~5.35±0.05) of the enzyme ternary complex, FDH-NAD-formate. The close pK-values of deionized groups important for the substrate- (5.4±0.1) and coenzyme- (5.35±0.05) binding could direct that one of the functional groups that was important for the binding of substrate might be also important for the binding of coenzyme as well as for the catalysis.

According to the X-ray data, two carboxyl groups Asp221 and Asp308 were located at the coenzyme-binding site. Asp221 bound to 2’-OH- group of nicotinamide ribose of NAD. Asp308 also participated in the coenzyme binding. It bound to the amide group of NAD. This interaction promoted the nicotinamide ring correct orientation as well as redistributed the charges between the atoms in the coenzyme molecule. That improved the electrophilic properties of C4-atom of the coenzyme and simplified the transition of hydrogen ion from formate anion to NAD.

Thus changes of Vmax could be probably caused by deionization of Asp308. On the other hand, we changes of K_m^NAD would happen due to deionization of two groups, Asp221 and Asp308.

The close values of pKs that we determined from the pH-dependences of K_m^NAD, K_m^formate and K_is of linear (CNO^-, azide and CNS^-) and planar (nitrate) anion inhibitors would propose of existence of an universal mechanism that could manage all these parameters (Fig. 5-7 and Table 3). We believe that this pH-transition could be caused by the change of the conformation.

The enzymes purified from methylotrophic yeast have same pH-transition at pH values 10 and higher (Fig. 8). We observed on pH-dependences of K_m^formate (FDH purified from Candida boidini and Hansenula polymorpha) and V_max (Candida boidini only). Moreover, the statistical analysis of pH-dependences demonstrated that the pH-transition had a cooperative character. Thus, this is another evidence the observed pH-transition was caused by the conformational changes

The data of X-ray analysis showed that only three residues (Asn146, Arg284 and His332) could participate in the substrate and coenzyme binding. The formation of hydrogen band should keep the Arg284 gunidine group deionized. Thus, the pK^Arg284 would be higher than standard Arg pK in water solutions (pK>11.6). Moreover, as we said above, His332 was already deionized at neutral pH. Thus, the pK^His332 should shift to acidic pH value (pH~5.5-6.0). Finally, the amide group (pK^Asn146) should be higher than 14.0.

Thus, we concluded that functional groups that caused pH-transition at alkaline pH-values were located outside the active center of the enzyme. The pH-transition was caused by the changes of conformation.

Effect of pyridoxal on the enzyme activity

According to the data of X-ray analysis demonstrated that the active center of FDH was loaded 15 Å below the surface of protein globule. There were to channels. One of them, substrate channel served for the substrate delivery. Another, so-called coenzyme binding channel could serve for delivery as the substrate as well as the coenzyme. The diameter hypothetic substrate channel (6-8 Å) allowed us to place inside it a molecule that size could be much larger than formate anion.

Pyridoxal was the only documented competitive inhibitor of the formate anion that could interact with both apo- and holoenzymes (binary complex FDH-NAD). Other competitive inhibitors of formate could interact only with holoenzyme.

Interaction of pyridoxal with FDH included two steps. At the first step, pyridoxal formed a non-covalent complex that appeared in competitive inhibition of formate. Then, a partial inactivation occurred due to a covalent complex formation. The inactivation became irreversible in the presence of sodium borhydride.

Inactivation accompanied by the changes in absorption and fluorescence spectra of the enzyme that were characteristic for a Shiff-base formation. Proportional dependence of remained enzyme activity on the concentration of pyridoxal (Fig. 9A) demonstrated that only one of lysine residues could be essential for the enzyme activity. X-ray analysis showed that it could be Lys286 located at the entry of substrate channel. Thus, the modification of Lys286 might cause partial inactivation of the enzyme. Other vise, FDH purified from methylotrophic yeast Candida boidini and Hansenula polymorpha was resistant to pyridoxal (Fig. 9B). Moreover, Ala substituted Lys in the protein sequence of yeast enzyme. Thus, that could be another argument for the identification of Lys286 as the amino acid residue responsible for the inactivation.

Computer modeling of the pyridoxal binding showed that chemical modification of Lys286 with pyridoxal could block the substrate channel. Moreover, in the substrate channel due to its close location to His332 pyridoxal also might affect on the substrate binding. Thus, modification of FDH with pyridoxal could have a competition with formate anion for the active center.

Thus, the mathematic modeling of pyridoxal binding to FDH as well as kinetic experiments confirmed the existence of the substrate channel. The channel served for the substrate delivery to the active center of the enzyme. Binding of pyridoxal to Lys286 located in the substrate channel could explain competition between pyridoxal and formate for the active center of FDH.

Different kinetic behavior of bacterial and yeast FDHs

FDH purified from different sources like plants, fungi, methylotrophic yeast and bacteria demonstrated the similarity of general properties (Mw, subunit composition, absence of prosthetic groups, close values of K_m^NAD and K_m^formate). We found that conserved amino acid residues of FDH purified from different sources had 100% similarity. After analysis of amino acid sequences, were able to separate the proteins at few groups. We found 80% similarity inside each group and 40-60% similarity between them. Fungi FDH as well as FDH of methylotrophic yeast have higher homology (60-65%). That made possible to combine these two groups of proteins into one.

The classification that we proposed was also in good correspondence with kinetic properties and substrate specificity of yeast and bacterial FDH (Table 5). We purified FDH from methylotrophic yeast Hansenula polymorpha and studied its kinetic mechanism. Our results (Table 6) demonstrated that the enzyme that we purified had ordered kinetic mechanism (NAD as the first substrate and NADH as the second product). Moreover, yeast and bacterial FDH had different specificity to thioformate, which was the closest structural analogue of formate (Table 5). Yeast FDH purified from Candida boidini and Hansenula polymorpha were able to use thioformate as a substrate. Other vise, thioformate had a competition with formate for the active center of FDH of Pseudomonas sp. 101 and inhibited the enzyme.

The analysis of amino acid sequences allowed us separating FDH of different origin at few groups that corresponded to major systematic units (plants, fungi, yeast and bacteria). We also established the differences of the kinetic properties and substrate specificity between yeast and bacterial FDH.

Conclusion

1. The pH-dependences of FDH purified from Pseudomonas sp. 101 were studied. The deionization of two carboxyl groups with same or close constant of ionization values (pK~5.5-6.0) was required for the binding of NAD in the active center. According to the results of X-ray analysis, the amino acid residues responsible for coenzyme binding were identified as Asp221 and Asp308. One of them (Asp308) was also important for the catalysis.

Another pH-transition with pK~10.0-10.5 was present at pH-dependences of yeast and bacterial FDH, at pH-dependences of kinetic and spectral parameters. It was not related with ionization of any functional group in the active center but accompanied by sufficient changes of the enzyme conformation.

2. FDH with substituted Arg284 (Arg284>Gln and Arg284>Ala) was characterized.

a) The contribution Arg284 in the binding of the substrate was established.

b) It was shown that Arg284 could participate in the catalysis by destabilization of NAD.

c) It was found that Arg284 would be important for the maintenance of FDH in catalytically active state.

3.

Binding of pyridoxal to the bacterial FDH of Pseudomonas sp. 101 was studied. The molecular modeling of this process was also performed.

a) The presence of substrate channel in the FDH molecule was confirmed

b) It was shown that chemical modification of FDH with pyridoxal could block the enzyme substrate channel and cause the competition with formate for the active center.

4. Classification of FDH of different origin based on high homology of the amino acid sequences (~80%) was proposed. The differences in kinetic characteristics of yeast and bacterial FDH were established. The kinetic mechanism of yeast FDH purified from Hansenula polymorpha was established.

Publications

1. Mezentsev, A.V., Lamzin, V.S., Tishkov, V.I., Ustinnikova, T.B. and Popov, V.O. Effect of pH on kinetic parameters of NAD-dependent formate dehydrogenase (1997) Biochem. J. Vol. 321, P. 475-480

2. Mezentsev, A.V., Ustinnikova, T.B., Tikhonova, T.V. and Popov, V.O. Isolation and kinetic mechanism of action of NAD- dependent formate dehydrogenase from methylotrophic yeast Hansenula polymorpha. (1996) Applied Biochemistry and Microbiology (Moscow) Vol. 32, N 6, P. 850-854

3. Kutsenko, A.S., Mezentsev, A.V. and Popov, V.0. NAD-dependent formate dehydrogenase of methylotrophic bacteria Pseudomonas sp. 101: 3D-symmetry of domains (1996) Proceeding of the conference “Autotrophic microorganisms” (Moscow), P.86.

4. Mezentseva, N.V., Kuznetsov, D.A., Mezentsev A.V., Kutsenko A.S. and Popov, V.O. NAD-dependent formate dehydrogenase of methylotrophic bacteria Pseudomonas sp. 101: mapping of the active center with pyridoxal and its analogues (1996). Proceeding of the conference “Autotrophic microorganisms” (Moscow), P.86.

5. Tishkov, V.O. Matorin, A.D., Rojkova, A.M., Fedorchuk, V.V., Savitsky, P.A., Dementieva, L.A., Lamzin, V.S., Mesentzev A.V. and Popov V.O. Site-directed mutagenesis of the formate dehydrogenase active centre: role of the His332-Gln313 pair in enzyme catalysis. (1996) FEBS Lett. Vol. 390, P. 104-108

6. Tishkov V.I., Galkin, A.G., Mezentsev, A.V., Ustinnikova, T.B., Popov, V.O. and Shelukho, D.V. Molecular mechanism of NAD-dependent formate dehydrogenase. Site-directed mutagenesis of the enzyme active centre (1996) Proceeding of the conference BIOCATALYSIS 95 (Moscow).