We have cloned the gene encoding Desulfovibrio gigas ferredoxin using a photodigoxigenin-labelled probe synthesized with the polymerase chain reaction. The DNA sequence of the gene predicts a polypeptide of 58 residues after removal of the initial formyl methionine (polypeptide M(r) = 6,276). The ferredoxin gene was expressed in aerobically grown E. coli behind the lac promoter of pUC18 resulting in a high level of ferredoxin expression which comprises about 10% of the total cell protein. EPR analysis of recombinant ferredoxin revealed the presence of a [3Fe-4S] cluster which is characteristic of native D. gigas ferredoxin II.

Small electron-transfer proteins such as flavodoxin (16 kDa) and the tetraheme cytochrome c3 (13 kDa) have been used to mimic, in vitro, part of the complex electron-transfer chain operating between substrate electron donors and respiratory electron acceptors, in sulfate-reducing bacteria (Desulfovibrio species). The nature and properties of the complex formed between these proteins are revealed by 1H-NMR and molecular modeling approaches. Our previous study with the Desulfovibrio vulgaris proteins [Moura, I., Moura, J.J. G., Santos, M.H., & Xavier, A. V. (1980) Cienc. Biol. (Portugal) 5, 195-197; Stewart, D.E. LeGall, J., Moura, I., Moura, J. J. G., Peck, H.D. Jr., Xavier, A. V., Weiner, P. K., & Wampler, J.E. (1988) Biochemistry 27, 2444-2450] indicated that the complex between cytochrome c3 and flavodoxin could be monitored by changes in the NMR signals of the heme methyl groups of the cytochrome and that the electrostatic surface charge (Coulomb's law) on the two proteins favored interaction between one unique heme of the cytochrome with flavodoxin. If the interaction is indeed driven by the electrostatic complementarity between the acidic flavodoxin and a unique positive region of the cytochrome c3, other homologous proteins from these two families of proteins might be expected to interact similarly. In this study, three homologous Desulfovibrio cytochromes c3 were used, which show a remarkable variation in their individual isoelectric points (ranging from 5.5 to 9.5). On the basis of data obtained from protein-protein titrations followed at specific proton NMR signals (i.e., heme methyl resonances), a binding model for this complex has been developed with evaluation of stoichiometry and binding constants. This binding model involves one site on the cytochromes c3 and two sites on the flavodoxin, with formation of a ternary complex at saturation. In order to understand the potential chemical form of the binding model, a structural model for the hypothetical ternary complex, formed between one molecule of Desulfovibrio salexigens flavodoxin and two molecules of cytochrome c3, is proposed. These molecular models of the complexes were constructed on the basis of complementarity of Coulombic electrostatic surface potentials, using the available X-ray structures of the isolated proteins and, when required, model structures (D. salexigens flavodoxin and Desulfovibrio desulfuricans ATCC 27774 cytochrome c3) predicted by homology modeling.

Desulfovibrio gigas ferredoxin II (FdII) is a small protein (alpha 4 subunit structure as isolated; M(r) approximately 6400 per subunit; 6 cysteine residues) containing one Fe3S4 cluster per alpha-subunit. The x-ray structure of FdII has revealed a disulfide bridge formed by Cys-18 and Cys-42 approximately 13 A away from the center of the cluster; moreover, the x-ray structure indicates that Cys-11 forms a disulfide bridge with a methanethiol. In the oxidized state, FdIIoxm the 1H NMR spectra, exhibit four low-field contact-shifted resonances at 29, 24, 18, and 15.5 ppm whereas the reduced state, FdIIR (S = 2), yields two features at +18.5 and -11 ppm. In the course of studying the redox behavior of FdII, we have discovered a stable intermediate, FdIIint, that yields 1H resonances at 24, 21.5, 21, and 14 ppm. This intermediate appears in the potential range where the cluster (E'0 approximately -130 mV) is reduced from the [Fe3S4]1+ to the [Fe3S4]0 state. FdIIint is observed during reductive titrations with dithionite or hydrogen/hydrogenase or after partial oxidation of FdIIR by 2,6-dichlorophenolindophenol or air. Our studies show that a total of three electrons per alpha-subunit are transferred to FdII. Our experiments demonstrate the absence of a methanethiol-Cys-11 linkage in our preparations, and we propose that two of the three electrons are used for the reduction of the disulfide bridge. Mossbauer (and EPR) studies show that the Fe3S4 cluster of FdIIint is at the same oxidation level as FdIIox, but indicate some changes in the exchange couplings among the three ferric sites. Our data suggest that the differences in the NMR and Mossbauer spectra of FdIIox and FdIIint result from conformational changes attending the breaking or formation of the disulfide bridge. The present study suggests that experiments be undertaken to explore an in vivo redox function for the disulfide bridge.

In this report, we describe the isolation of a 4020-bp genomic PstI fragment of Desulfovibrio gigas harboring the aldehyde oxido-reductase gene. The aldehyde oxido-reductase gene spans 2718 bp of genomic DNA and codes for a protein with 906 residues. The protein sequence shows an average 52% (+/- 1.5%) similarity to xanthine dehydrogenase from different organisms. The codon usage of the aldehyde oxidoreductase is almost identical to a calculated codon usage of the Desulfovibrio bacteria.

Flavodoxin was isolated and purified from Desulfovibrio desulfuricans ATCC 27774, a sulfate-reducing organism that can also utilize nitrate as an alternative electron acceptor. Mid-point oxidation-reduction potentials of this flavodoxin were determined by ultraviolet/visible and EPR methods coupled to potentiometric measurements and their pH dependence studied in detail. The redox potential E2, for the couple oxidized/semiquinone forms at pH 6.7 and 25 degrees C is -40 mV, while the value for the semiquinone/hydroquinone forms (E1), at the same pH, -387 mV. E2 varies linearly with pH, while E1 is independent of pH at high values. However, at low pH (< 7.0), this value is less negative, compatible with a redox-linked protonation of the flavodoxin hydroquinone. A comparative study is presented for Desulfovibrio salexigens NCIB 8403 flavodoxin [Moura, I., Moura, J.J.G., Bruschi, M. & LeGall, J. (1980) Biochim. Biophys. Acta 591, 1-8]. The complete primary amino acid sequence was obtained by automated Edman degradation from peptides obtained by chemical and enzymic procedures. The amino acid sequence was confirmed by FAB/MS. Using the previously determined tridimensional structure of Desulfovibrio vulgaris flavodoxin as a model [similarity, 48.6%; Watenpaugh, K.D., Sieker, L.C., Jensen, L.H., LeGall, J. & Dubourdieu M. (1972) Proc. Natl Acad. Sci. USA 69, 3185-3188], the tridimensional structure of D. desulfuricans ATCC 27774 flavodoxin was predicted using AMBER force-field calculations.

A [2Fe-2S] cluster has been detected in mammalian ferrochelatase, the terminal enzyme of the heme biosynthetic pathway. Natural ferrochelatase, purified from mouse livers, and recombinant ferrochelatase, purified from an overproducing strain of Escherichia coli, were investigated by electron paramagnetic resonance (EPR) and Mossbauer spectroscopy. In their reduced forms, both the natural and recombinant ferrochelatases exhibited an identical EPR signal with g values (g = 2.00, 1.93, and 1.90) and relaxation properties typical of [2Fe-2S]+ cluster. Mossbauer spectra of the recombinant ferrochelatase, purified from a strain of E. coli cells transformed with a plasmid encoding murine liver ferrochelatase and grown in 57Fe-enriched medium, demonstrated unambiguously that the cluster is a [2Fe-2S] cluster. No change in the cluster oxidation state was observed during catalysis. The putative protein binding site for the Fe-S cluster in mammalian ferrochelatases is absent from the sequences of the bacterial and yeast enzymes, suggesting a possible role of the [2Fe-2S] center in regulation of mammalian ferrochelatases.

In work that is complementary to our investigation of the spectroscopic features of the cytochrome c peroxidase from Paracoccus denitrificans [Gilmour, Goodhew, Pettigrew, Prazeres, Moura and Moura (1993) Biochem. J. 294, 745-752], we have studied the kinetics of oxidation of cytochrome c by this enzyme. The enzyme, as isolated, is in the fully oxidized form and is relatively inactive. Reduction of the high-potential haem at pH 6 with ascorbate results in partial activation of the enzyme. Full activation is achieved by addition of 1 mM CaCl2. Enzyme activation is associated with formation of a high-spin state at the oxidized low-potential haem. EGTA treatment of the oxidized enzyme prevents activation after reduction with ascorbate, while treatment with EGTA of the reduced, partially activated, form abolishes the activity. We conclude that the active enzyme is a mixed-valence form with the low-potential haem in a high-spin state that is stabilized by Ca2+. Dilution of the enzyme results in a progressive loss of activity, the extent of which depends on the degree of dilution. Most of the activity lost upon dilution can be recovered after reconcentration. The M(r) of the enzyme on molecular-exclusion chromatography is concentration-dependent, with a shift to lower values at lower concentrations. Values of M(r) obtained are intermediate between those of a monomer (39,565) and a dimer. We propose that the active form of the enzyme is a dimer which dissociates at high dilution to give inactive monomers. From the activity of the enzyme at different dilutions, a KD of 0.8 microM can be calculated for the monomerdimer equilibrium. The cytochrome c peroxidase oxidizes horse ferrocytochrome c with first-order kinetics, even at high ferrocytochrome c concentrations. The maximal catalytic-centre activity ('turnover number') under the assay conditions used is 62,000 min-1, with a half-saturating ferrocytochrome c concentration of 3.3 microM. The corresponding values for the Paracoccus cytochrome c-550 (presumed to be the physiological substrate) are 85,000 min-1 and 13 microM. However, in this case, the kinetics deviate from first-order progress curves at all ferrocytochrome c concentrations. Consideration of the periplasmic environment in Paracoccus denitrificans leads us to propose that the enzyme will be present as the fully active dimer supplied with saturating ferrocytochrome c-550.

The same polypeptide chain (58 amino acids, 6 cysteines) is used to build up two ferredoxins in Desulfovibrio gigas a sulfate reducing organism. Ferredoxin II (FdII) contains a single [Fe3S4] core and ferredoxin I (FdI) mainly a [Fe4S4] core. The [Fe3S4] core can readily be interconverted into a [Fe4S4] complex (J.J.G. Moura, I. Moura, T.A. Kent, J.D. Lipscomb, B.H. Huynh, J. LeGall, A.V. Xavier, and E. Munck, J. Biol. Chem. 257, 6259 (1982)). This interconversion process suggested that the [Fe3S4] core could be used as a synthetic precursor for the formation of heterometal clusters. Co, Zn, Cd, and Ni derivatives were produced (I. Moura, J.J.G. Moura, E. Munck, V. Papaephthymiou, and J. LeGall, J. Am. Chem. Soc. 108, 349 (1986), K. Sureurs, E. Munck, I. Moura, J.J.G. Moura, and J. LeGall, J. Am. Chem. Soc. 109, 3805 (1986), and A.L. Macedo, I. Moura, J.J.G. Moura, K. Surerus, and E. Munck, unpublished results). The redox properties of a series of heterometal clusters (MFe3S4] are assessed using direct electrochemistry (square wave voltammetry--SWV) promoted by Mg(II) at a glassy carbon electrode (derivatives: Cd (-495 mV), Fe (-420 mV), Ni (-360 mV), and Co (-245 mV) vs normal hydrogen electrode (NHE)). In parallel, the electrochemical behavior (cyclic voltammetry--CV, differential pulse voltammetry--DPV and SWV) of FdI and FdII were investigated as well as the cluster interconversion process. In addition to the +1/0 (3Fe cluster) and +2/+1 (4Fe cluster) redox transitions, a very negative redox step, at -690 mV, was detected for the 3Fe core, reminiscent of a postulated further 2e- reduction step, as proposed for D. africanus ferredoxin III by F.A. Armstrong, S.J. George, R. Cammack, E.C. Hatchikian, and A.J. Thomson, Biochem. J. 264, 265 (1989). The electrochemical redox potential values are compared with those determined by independent methods (namely by electron paramagnetic resonance (EPR) and visible spectroscopy).

The kinetic properties of the electron-transfer process between reduced Desulfovibrio vulgaris cytochrome c3 and D. vulgaris flavodoxin have been studied by anaerobic stopped-flow techniques. Anaerobic titrations of reduced cytochrome c3 with oxidized flavodoxin show a stoichiometry of 4 mol of flavodoxin required to oxidize the tetraheme cytochrome. Flavodoxin neutral semiquinone and oxidized cytochrome c3 are the only observable products of the reaction. At pH 7.5, the four-electron-transfer reaction is biphasic. Both the rapid and the slow phases exhibit limiting rates as the flavodoxin concentration is increased with respective rates of 73.4 and 18.5 s-1 and respective Kd values of 65.9 +/- 9.4 microM and 54.5 +/- 13 microM. A biphasic electron-transfer rate is observed when the ionic strength is increased to 100 mM KCl; however, the observed rate is no longer saturable, and relative second-order rate constants of 5.3 x 10(5) and 8.5 x 10(4) M-1 s-1 are calculated. The magnitude of the rapid phase of electron transfer diminishes with the level of heme reduction when varying reduced levels of the cytochrome are mixed with oxidized flavodoxin. No rapid phase is observed when 0.66e(-)-reduced cytochrome c3 reacts with an approximately 25-fold molar excess of flavodoxin. At pH 6.0, the electron-transfer reaction is monophasic with a limiting rate of 42 +/- 1.4 s-1 and a Kd value of approximately 8 microM. Increasing the ionic strength of the pH 6.0 solution to 100 microM KCl results in a biphasic reaction with relative second-order rate constants of 5.3 x 10(5) and 1.1 x 10(4) M-1 s-1. Azotobacter vinelandii flavodoxin reacts with reduced D. vulgaris cytochrome c3 in a slow, monophasic manner with limiting rate of electron transfer of 1.2 +/- 0.06 s-1 and a Kd value of 80.9 +/- 10.7 microM. These results are discussed in terms of two equilibrium conformational states for the cytochrome which are dependent on the pH of the medium and the level of heme reduction [Catarino et al. (1991) Eur. J. Biochem. 207, 1107-1113].

The kinetic properties of the electron-transfer process between reduced Desulfovibrio vulgaris cytochrome c(3) and D. vulgaris flavodoxin have been studied by anaerobic stopped-flow techniques. Anaerobic titrations of reduced cytochrome c(3) with oxidized flavodoxin show a stoichiometry of 4 mol of flavodoxin required to oxidize the tetraheme cytochrome. Flavodoxin neutral semiquinone and oxidized cytochrome c(3) are the only observable products of the reaction. At pH 7.5, the four-electron-transfer reaction is biphasic. Both the rapid and the slow phases exhibit limiting rates as the flavodoxin concentration is increased with respective rates of 73.4 and 18.5 s(-1) and respective K-d values of 65.9 +/- 9.4 mu M and 54.5 +/- 13 CIM. A biphasic electron-transfer rate is observed when the ionic strength is increased to 100 mM KCl; however, the observed rate is no longer saturable, and relative second-order rate constants of 5.3 X 10(5) and 8.5 x 10(4) M(-1) s(-1) are calculated. The magnitude of the rapid phase of electron transfer diminishes with the level of heme reduction when varying reduced levels of the cytochrome are mixed with oxidized flavodoxin. No rapid phase is observed when 0.66e(-)-reduced cytochrome c(3) reacts with an similar to 25-fold molar excess of flavodoxin. At pH 6.0, the electron-transfer reaction is monophasic with a limiting rate of 42 +/- 1.4 s(-1) and a Kd value of similar to 8 mu M. Increasing the ionic strength of the pH 6.0 solution to 100 mu M KCl results in a biphasic reaction with relative second-order rate constants of 5.3 x 10(5) and 1.1 x 10(4) M(-1) s(-1) Azotobacter vinelandii flavodoxin reacts with reduced D. vulgaris cytochrome cs in a slow, monophasic manner with limiting rate of electron transfer of 1.2 +/- 0.06 s(-1) and a K-d value of 80.9 +/- 10.7 mu M. These results are discussed in terms of two equilibrium conformational states for the cytochrome which are dependent on the pH of the medium and the level of heme reduction [Catarino et al. (1991) Eur. J. Biochem. 207, 1107-1113].

Desulfoferrodoxin, a non-heme iron protein, was purified previously from extracts of Desulfovibrio desulfuricans (ATCC 27774) (Moura, I., Tavares, P., Moura, J. J. G., Ravi, N., Huynh, B. H., Liu, M.-Y., and LeGall, J. (1990) J. Biol. Chem. 265, 21596-21602). The as-isolated protein displays a pink color (pink form) and contains two mononuclear iron sites in different oxidation states: a ferric site (center I) with a distorted tetrahedral sulfur coordination similar to that found in desulforedoxin from Desulfovibrio gigas and a ferrous site (center II) octahedrally coordinated with predominantly nitrogen/oxygen-containing ligands. A new form of desulfoferrodoxin which displays a gray color (gray form) has now been purified. Optical, electron paramagnetic resonance (EPR), and Mossbauer data of the gray desulfoferrodoxin indicate that both iron centers are in the high-spin ferric states. In addition to the EPR signals originating from center I at g = 7.7, 5.7, 4.1, and 1.8, the gray form of desulfoferrodoxin exhibits a signal at g = 4.3 and a shoulder at g = 9.6, indicating a high-spin ferric state with E/D approximately 1/3 for the oxidized center II. Redox titrations of the gray form of the protein monitored by optical spectroscopy indicate midpoint potentials of +4 +/- 10 and +240 +/- 10 mV for centers I and II, respectively. Mossbauer spectra of the gray form of the protein are consistent with the EPR finding that both centers are high-spin ferric and can be analyzed in terms of the EPR-determined spin Hamiltonian parameters. The Mossbauer parameters for both the ferric and ferrous forms of center II are indicative of a mononuclear high spin iron site with octahedral coordination and predominantly nitrogen/oxygen-containing ligands. Resonance Raman studies confirm the structural similarity of center I and the distorted tetrahedral FeS4 center in desulforedoxin and provide evidence for one or two cysteinyl-S ligands for center II. On the basis of the resonance Raman results, the 635 nm absorption band that is responsible for the gray color of the oxidized protein is assigned to a cysteinyl-S-->Fe(III) charge transfer transition localized on center II. The novel properties and possible function of center II are discussed in relation to those of mononuclear iron centers in other enzymes.

A c-type monohemic ferricytochrome c552 (11 kDa) was isolated from the soluble extract of a marine denitrifier, Pseudomonas nautica strain 617, grown under anaerobic conditions with nitrate as final electron acceptor. The NH2-terminal sequence and the amino acid composition of the cytochrome were determined. The heme iron of the cytochrome c552 has histidine-methionine as axial ligands, and a pH-dependent mid-point redox potential, equal to 250 mV at pH 7.6. The presence of methionine was demonstrated by visible, EPR and NMR spectroscopies. The assignment of most of the hemic protons was performed applying two-dimensional NOE spectroscopy (NOESY), and the aromatic region was assigned through two-dimensional correlated spectroscopy (COSY) experiments. The EPR spectrum of the oxidised form of the cytochrome c552 is typical of a low-spin ferric heme.

The cytochrome c peroxidase of Paracoccus denitrificans is similar to the well-studied enzyme from Pseudomonas aeruginosa. Like the Pseudomonas enzyme, the Paracoccus peroxidase contains two haem c groups, one high potential and one low potential. The high-potential haem acts as a source of the second electron for H2O2 reduction, and the low-potential haem acts as a peroxidatic centre. Reduction with ascorbate of the high-potential haem of the Paracoccus enzyme results in a switch of the low-potential haem to a high-spin state, as shown by visible and n.m.r. spectroscopy. This high-spin haem of the mixed-valence enzyme is accessible to ligands and binds CN- with a KD of 5 microM. The Paracoccus enzyme is significantly different from that from Pseudomonas in the time course of high-spin formation after reduction of the high-potential haem, and in the requirement for bivalent cations. Reduction with 1 mM ascorbate at pH 6 is complete within 2 min, and this is followed by a slow appearance of the high-spin state with a half-time of 10 min. Thus the process of reduction and spin state change can be easily separated in time and the intermediate form obtained. This separation is also evident in e.p.r. spectra, although the slow change involves an alteration in the low-spin ligation at this temperature rather than a change in spin state. The separation is even more striking at pH 7.5, where no high-spin form is obtained until 1 mM Ca2+ is added to the mixed-valence enzyme. The spin-state switch of the low-potential haem shifts the midpoint redox potential of the high-potential haem by 50 mV, a further indication of haem-haem interaction.

Redox intermediates of D. desulfuricans ATCC 27774 [NiFe] hydrogenase were generated under dihydrogen. Detailed redox titrations, coupled to EPR measurements, give access to the mid-point redox potentials of the iron-sulfur centers and of the Nickel-B signal that represents the ready form of the enzyme. The interaction between the dihydrogen molecule and the nickel centre was probed by the observation of an isotopic effect on the EPR signals detected in turnover conditions, by comparison of the H2O/H2 and D2O/D2-reacted samples.

The molybdenum [iron-sulfur] protein, first isolated from Desulfovibrio gigas by Moura et al. [Moura, J. J. G., Xavier, A. V., Bruschi, M., Le Gall, J., Hall, D. O., & Cammack, R. (1976) Biochem. Biophys. Res. Commun. 72, 782-789], was later shown to mediate the electronic flow from salicylaldehyde to a suitable electron acceptor, 2,6-dichlorophenolindophenol (DCPIP) [Turner, N., Barata, B., Bray, R. C., Deistung, J., LeGall, J., & Moura, J. J. G. (1987) Biochem. J. 243, 755-761]. The DCPIP-dependent aldehyde oxidoreductase activity was studied in detail using a wide range of aldehydes and analogues. Steady-state kinetic analysis (KM and Vmax) was performed for acetaldehyde, propionaldehyde, benzaldehyde, and salicylaldehyde in excess DCPIP concentration, and a simple Michaelis-Menten model was shown to be applicable as a first kinetic approach. Xanthine, purine, allopurinol, and N1-methylnicotinamide (NMN) could not be utilized as enzyme substrates. DCPIP and ferricyanide were shown to be capable of cycling the electronic flow, whereas other cation and anion dyes [O2 and NAD(P)+] were not active in this process. The enzyme showed an optimal pH activity profile around 7.8. This molybdenum hydroxylase was shown to be part of an electron-transfer chain comprising four different soluble proteins from D. gigas, with a total of 11 discrete redox centers, which is capable of linking the oxidation of aldehydes to the reduction of protons.

The kinetics of electron transfer between the Desulfovibrio gigas hydrogenase and several electron-transfer proteins from Desulfovibrio species were investigated by cyclic voltammetry, square-wave voltammetry and chronoamperometry. The cytochrome c3 from Desulfovibrio vulgaris (Hildenborough), Desulfovibrio desulfuricans (Norway 4), Desulfovibrio desulfuricans (American Type Culture Collection 27774) and D. gigas (NCIB 9332) were used as redox carriers. They differ in their redox potentials and isoelectric point. Depending on the pH, all the reduced forms of these cytochromes were effective in electron exchange with hydrogenase. Other small electron-transfer proteins such as ferredoxin I, ferredoxin II and rubredoxin from D. gigas were tentatively used as redox carriers. Only ferredoxin II was effective in mediating electron exchange between hydrogenase and the working electrode. The second-order rate constants k for the reaction between reduced proteins and hydrogenase were calculated based on the theory of the simplest electrocatalytic mechanism [Moreno, C., Costa, C., Moura, I., Le Gall, J., Liu, M. Y., Payne, W. J., van Dijk, C. & Moura, J. J. G. (1993) Eur. J. Biochem. 212, 79-86] and the results obtained by cyclic voltammetry were compared with those obtained by chronoamperometry. Values for k of 10(5)-10(6) M-1 s-1 (cytochrome c3 as electron carrier) and 10(4) M-1 s-1 (ferredoxin II as the electron carrier) were determined. The rate-constant values are discussed in terms of the existence of an electrostatic interaction between the electrode surface and the redox carrier and between the redox carrier and a positively charged part of the enzyme.