Hydrodynamic, structural and magnetic properties of Megasphaera elsdenii Fe hydrogenase reinvestigated

Megasphaera elsdenii hydrogenase has been purified to homogeneity using an FPLC procedure as the final step. The protein gives a single band in SDS/PAGE with an apparent molecular mass of 57-59 kDa. There is no second hydrogenase activity in the soluble fraction of M. elsdenii. The hydrodynamics of the enzyme have been compared to those of the two-subunit Fe hydrogenase from Desulfovibrio vulgaris (Hildenborough) in the analytical ultracentrifuge using the absorption of the intrinsic iron-sulfur clusters as the monitor. Sedimentation-velocity experiments indicate the M. elsdenii enzyme (s20,w = 4.95 S) to be essentially globular, while the D. vulgaris enzyme (s20,w = 4.1 S) has a less symmetric shape. From the sedimentation equilibrium measurements under a variety of conditions an average molecular mass is calculated of 58 kDa (M. elsdenii) and 54 kDa (D. vulgaris), respectively. Pure, maximally active M. elsdenii hydrogenase has A405/A280 = 0.36 and has a specific H2-production activity of 400 mumol H2.min-1.(mg protein)-1 at 30 degrees C and pH 8.0. The enzyme contains some 13-18 iron and acid-labile sulfur ions/58-kDa monomer. Eight of these Fe-S are present as two electron-transferring ferredoxin-like cubanes with Em approximately greater than -0.3 V, as indicated by pH-dependent EPR spectroscopy on the H2-reduced enzyme. In the (re)oxidized state the remainder iron gives rise to a single S = 1/2 rhombic EPR signal. Hydrogen-production activity, content of remainder iron and rhombic EPR signal intensity are mutually correlated. Purified hydrogenase appears to exist as a mixture of fully active holoenzyme and inactive protein still carrying the two cubanes but deficient in active-site iron.     

The pH in reversed micelles as imposed by the dihydrogen/proton redox couple and indicated by viologens and cytochrome c3 using hydrogenase as redox catalyst.

The pH values in reversed micelles were measured, making use of the hydrogenase enzyme as redox catalyst short-circuiting the viologen oxidized/semiquinone redox states. The hydrogenases from Desulfovibrio vulgaris (Hildenborough) and from Megasphaera elsdenii were applied. The observed pH values in reversed micelles were not dependent on the type of hydrogenase. Two cationic [cetyltrimethylammonium bromide and dodecylammonium propionate (DAP)] and two anionic sodiumdodecyl sulphate, sodium di(ethylhexyl)sulfosuccinate types of reversed micelles were used in combination with viologens having distinguishable valencies. It was observed that, in the cationic-reversed micelles, the dissociation constant for the semiquinone dimer had about the same value as compared to bulk water, while this value was significantly higher in the anionic-reversed micelles. Furthermore, the dissociation constant was independent of the concentration of viologen semiquinone in the reversed micelle, indicating that exchange kinetics are faster than the dimerisation process. With the exception of DAP, a linear relation exists, pH = a.pHrm + b, between the pH of the bulk water and the pH as measured in the reversed micelle (pHrm). In all these cases the value of a is smaller than unity, the value of b ranges between 1.6-2.7. For DAP the pHrm is always around 7. In DAP-reversed micelles, the counter-ion propionate probably serves as an internal buffer. Using cytochrome c3 as pH indicator in combination with N,N'-di(3-aminopropyl)-4,4'-bipyridinium)4+ to take care of electron transfer, in cetyltrimethylammonium-bromide-reversed micelles the pHrm is about the same as indicated by the viologen; in SDS-reversed micelles the pHrm is always lower than that indicated by N,N'-di(3-aminopropyl)4,4'-pyridinium4+. In contrast to cytochrome c3 from D. vulgaris, which in reversed micelles cannot become reduced directly by its D. vulgaris hydrogenase, the hydrogenase of M. elsdenii is able to reduce its ferredoxin directly.     

Further characterization of the [Fe]-hydrogenase from Desulfovibrio desulfuricans ATCC 7757.

The properties of the periplasmic hydrogenase from Desulfovibrio desulfuricans ATCC 7757, previously reported to be a single-subunit protein [Glick, B. R., Martin, W. G., and Martin, S. M. (1980) Can. J. Microbiol. 26, 1214-1223] were reinvestigated. The pure enzyme exhibited a molecular mass of 53.5 kDa as measured by analytical ultracentrifugation and was found to comprise two different subunits of 42.5 kDa and 11 kDa, with serine and alanine as N-terminal residues, respectively. The N-terminal amino acid sequences of its large and small subunits, determined up to 25 residues, were identical to those of the Desulfovibrio vulgaris Hildenborough [Fe]-hydrogenase. D. desulfuricans ATCC 7757 hydrogenase was free of nickel and contained 14.0 atoms of iron and 14.4 atoms of acid-labile sulfur/molecule and had E400, 52.5 mM-1.cm-1. The purified hydrogenase showed a specific activity of 62 kU/mg of protein in the H2-uptake assay, and the H2-uptake activity was higher than H2-evolution activity. The enzyme isolated under aerobic conditions required incubation under reducing conditions to express its maximum activity both in the H2-uptake and 2H2/1H2 exchange reaction. The ratio of the activity of activated to as-isolated hydrogenase was approximately 3. EPR studies allowed the identification of two ferredoxin-type [4Fe-4S]1+ clusters in hydrogenase samples reduced by hydrogen. In addition, an atypical cluster exhibiting a rhombic signal (g values 2.10, 2.038, 1.994) assigned to the H2-activating site in other [Fe]-hydrogenases was detected in partially reduced samples. Molecular properties, EPR spectroscopy, catalytic activities with different substrates and sensitivity to hydrogenase inhibitors indicated that D. desulfuricans ATCC 7757 periplasmic hydrogenase is a [Fe]-hydrogenase, similar in most respects to the well characterized [Fe]-hydrogenase from D. vulgaris Hildenborough.     

The pH dependence of proton-deuterium exchange, hydrogen production and uptake catalyzed by hydrogenases from sulfate-reducing bacteria

Different patterns have been found in the pH dependence of hydrogenase activity with enzymes purified from different species of Desulfovibrio. With the cytoplasmic hydrogenase from Desulfovibrio baculatus strain 9974, the pH optima in H2 production and uptake were respectively 4.0 and 7.5 with a higher activity in production than in uptake. The highest D2-H+ exchange activity was found also at pH 4.0 but the optima differed for the HD and the H2 components. Both similarly rose when the pH decreased from 9.0 to 4.5, but the rate of H2 evolution slowed whereas the HD evolution continued rising till pH values around 3.0 were reached. The H2 to HD ratio at pH above 4.5 was higher than one. With the periplasmic hydrogenase from Desulfovibrio vulgaris Hildenborough, the highest exchange activity was near pH 5.5, the same value as in hydrogen production. The periplasmic hydrogenase from Desulfovibrio gigas had in contrast the same pH optimum in the exchange (7.5-8.0) as in the H2 uptake. The ratio of H2 to HD was below one for both enzymes. These different patterns may be related to functional and structural differences in the three hydrogenases so far studied, particularly in the composition of their catalytic centers.     

Electron paramagnetic resonance and other properties of hydrogenases isolated from Desulfovibrio vulgaris (strain Hildenborough) and Megasphaera elsdenii

The hydrogenases of Desulfovibrio vulgaris and Megasphaera elsdenii are compared with respect to some of their physical properties. In addition to Fe the only metal ions that are present in significant amounts are Ni and Cu. From cluster extrusion experiments it follows that the D. vulgaris enzyme contains three 4 Fe-4S clusters, while M. elsdenii hydrogenase only releases part of its Fe-S clusters. The resting D. vulgaris enzyme shows only a small 3 Fe-xS type of EPR signal (maximum 5% electron equivalent). This amount can be increased to approximately 25% by treatment with ferricyanide, with a concomitant large decrease in activity. The M. elsdenii enzyme shows in its oxidized state a normal Hipip (high-potential iron-sulphur protein) type of EPR spectrum. After a reduction/oxidation cycle the D. vulgaris enzyme also shows a weak Hipip type of EPR spectrum. In the reduced state both enzymes show complex spectra. By integration of those spectra it is shown that 1.5 electron equivalents are present. The complex spectra do not arise from nuclear hyperfine interactions but are partially due to electron spin interactions. It is proposed that the spectrum of reduced D. vulgaris hydrogenase consists of a sum of three different ferredoxin-like spectra.     

Hydrogenase activity in Thiocapsa roseopersicina according to the D2--H20 metabolic reaction]

Extracts of Thiocapsa roseopersicina cells show hydrogenase activity, measured by evolution of H2 from reduced methylviologene (MV) and by D2-H2O exchange reaction. According to these reactions the most part of hydrogenases is found to be in the soluble fraction. Hydrogenase activity measured in the exchange reaction is completely inhibited by p-chloromercurybenzoate (5-10- minus 3 M), iodacetate (1-10- minus 2 M) and 26% inhibited by KCN and o-phenanthroline (5-10- minus 3 M). Evolution of H2 from reduced MV was not inhibited by o-phenanthroline, KCN and iodacetate and was inhibited by 66% only with p-chloromercurybenzoate. Light and ATP stimulated hydrogenase activity of chromatophores did not affect on its activity in the soluble fraction. The results obtained show that there are certain differences in hydrogenase systems responsible for the exchange reaction and evolution of H2.     

Characterization of the hydrogen-deuterium exchange activities of the energy-transducing HupSL hydrogenase and H(2)-signaling HupUV hydrogenase in Rhodobacter capsulatus

Rhodobacter capsulatus synthesizes two homologous protein complexes capable of activating molecular H(2), a membrane-bound [NiFe] hydrogenase (HupSL) linked to the respiratory chain, and an H(2) sensor encoded by the hupUV genes. The activities of hydrogen-deuterium (H-D) exchange catalyzed by the hupSL-encoded and the hupUV-encoded enzymes in the presence of D(2) and H(2)O were studied comparatively. Whereas HupSL is in the membranes, HupUV activity was localized in the soluble cytoplasmic fraction. Since the hydrogenase gene cluster of R. capsulatus contains a gene homologous to hoxH, which encodes the large subunit of NAD-linked tetrameric soluble hydrogenases, the chromosomal hoxH gene was inactivated and hoxH mutants were used to demonstrate the H-D exchange activity of the cytoplasmic HupUV protein complex. The H-D exchange reaction catalyzed by HupSL hydrogenase was maximal at pH 4. 5 and inhibited by acetylene and oxygen, whereas the H-D exchange catalyzed by the HupUV protein complex was insensitive to acetylene and oxygen and did not vary significantly between pH 4 and pH 11. Based on these properties, the product of the accessory hypD gene was shown to be necessary for the synthesis of active HupUV enzyme. The kinetics of HD and H(2) formed in exchange with D(2) by HupUV point to a restricted access of protons and gasses to the active site. Measurement of concentration changes in D(2), HD, and H(2) by mass spectrometry showed that, besides the H-D exchange reaction, HupUV oxidized H(2) with benzyl viologen, produced H(2) with reduced methyl viologen, and demonstrated true hydrogenase activity. Therefore, not only with respect to its H(2) signaling function in the cell, but also to its catalytic properties, the HupUV enzyme represents a distinct class of hydrogenases.     

HupUV proteins of Rhodobacter capsulatus can bind H2: evidence from the H-D exchange reaction.

The H-D exchange reaction has been measured with the D2-H2O system, for Rhodobacter capsulatus JP91, which lacks the hupSL-encoded hydrogenase, and R. capsulatus BSE16, which lacks the HupUV proteins. The hupUV gene products, expressed from plasmid pAC206, are shown to catalyze an H-D exchange reaction distinguishable from the H-D exchange due to the membrane-bound, hupSL-encoded hydrogenase. In the presence of O2, the uptake hydrogenase of BSE16 cells catalyzed a rapid uptake and oxidation of H2, D2, and HD present in the system, and its activity (H-D exchange, H2 evolution in presence of reduced methyl viologen [MV+]) depended on the external pH, while the H-D exchange due to HupUV remained insensitive to external pH and O2. These data suggest that the HupSL dimer is periplasmically oriented, while the HupUV proteins are in the cytoplasmic compartment.     

Participation of hyf-encoded hydrogenase 4 in molecular hydrogen release coupled with proton-potassium exchange in Escherichia coli

In a previous work (Trchounian et al., Biol. Membrany 16:416-428 (1999) (in Russian)) we reported the interrelations between production of H2 and H+-K+ exchange in fermenting Escherichia coli grown under anaerobic conditions at pH 7.5. The ion fluxes had stable stoichiometry 2H+/K+ and were N,N'-dicyclohexylcarbodiimide (DCC)-inhibitable at different external pH and K+ activity. In the present study, the H2 production was further studied in fermenting bacteria grown at pH 7.5 or 6.5. The H2 production was inhibited by DCC and did not occur if bacteria were grown at pH 7.5 in a medium containing formate or upon hypoosmotic stress. The H2 production was not sensitive to osmotic stress when bacteria were grown at pH 6.5. Formation of H2 and 2H+/K+ exchange were not observed in mutants with deletions of the hyfoperon genes, encoding membrane-associated hydrogenase 4. K+ influx in these mutants was not sensitive to valinomycin, in contrast to the K+ influx in the parental strain. If grown at pH 6.5, the mutants produced H2 and carried out 2H+/K+ exchange, when subjected to the hyperosmotic stress. The results suggest a participation of hydrogenase 4 in the production of H2 and proton-potassium exchange in fermenting E. coli grown at pH 7.5. In bacteria grown at pH 6.5 or in a medium containing formate, another membrane-bound hydrogenase, namely hydrogenase 3, may be responsible for the H2 production.     

Characterization of the oxygen tolerance of a hydrogenase linked to a carbon monoxide oxidation pathway in Rubrivivax gelatinosus

A hydrogenase linked to the carbon monoxide oxidation pathway in Rubrivivax gelatinosus displays tolerance to O2. When either whole-cell or membrane-free partially purified hydrogenase was stirred in full air (21% O2, 79% N2), its H2 evolution activity exhibited a half-life of 20 or 6 h, respectively, as determined by an anaerobic assay using reduced methyl viologen. When the partially purified hydrogenase was stirred in an atmosphere containing either 3.3 or 13% O2 for 15 min and evaluated by a hydrogen-deuterium (H-D) exchange assay, nearly 80 or 60% of its isotopic exchange rate was retained, respectively. When this enzyme suspension was subsequently returned to an anaerobic atmosphere, more than 90% of the H-D exchange activity was recovered, reflecting the reversibility of this hydrogenase toward O2 inactivation. Like most hydrogenases, the CO-linked hydrogenase was extremely sensitive to CO, with 50% inhibition occurring at 3.9 microM dissolved CO. Hydrogen production from the CO-linked hydrogenase was detected when ferredoxins of a prokaryotic source were the immediate electron mediator, provided they were photoreduced by spinach thylakoid membranes containing active water-splitting activity. Based on its appreciable tolerance to O2, potential applications of this hydrogenase are discussed.     

Aerobically purified hydrogenase from Azotobacter vinelandii: activity, activation, and spectral properties

The hydrogenase from Azotobacter vinelandii is typically purified under anaerobic conditions. In this work, the hydrogenase was purified aerobically. The yields were low (about 2%) relative to those of the anaerobic purification (about 20%). The rate of enzyme activity depended upon the history of the enzyme. The enzyme preparations were active as isolated in H2 oxidation, and isotope exchange. The activity increased during the assay to a new maximal level (turnover activation). Treatment with reductants (e.g., H2, dithionite, dithiothreitol, indigo carmine) resulted in greater activation (reductant activation). Activation of the hydrogenase was accompanied by decrease in visible light absorption (300-600 nm) with maximal decreases at 450 and 345 nm which indicated the reduction of iron-sulfur clusters. The aerobically purified hydrogenase was susceptible to irreversible inactivation by cyanide. Pretreatment with acetylene did not influence activation of the hydrogenase. Once activated, the aerobically purified hydrogenase was indistinguishable from the anaerobically purified hydrogenase with respect to the catalytic properties tested.     

Human 5-lipoxygenase contains an essential iron

The iron content of human 5-lipoxygenase has been determined by a colorimetric assay using the chromogenic ligand FerroZine. The highly active enzyme was obtained from a baculovirus expression system and purified using an ATP-agarose chromatography column (Denis, D., Falgueyret, J.-P., Riendeau, D., and Abramovitz, M. (1991) J. Biol. Chem. 266, 5072-5079). A linear correlation was observed between the enzyme's specific activity and iron content in six different preparations. Enzyme with the highest specific activity (24 mumol of 5-hydroperoxyeicosatetraenoic acid/mg of protein) contained 1.1 mol of iron/mol of enzyme, whereas inactive enzyme contained no detectable iron. The iron is tightly bound to the enzyme and could only be released after inactivation of the enzyme by exposure to oxygen.     

A coupled enzyme assay for isopenicillin N synthetase

The development of a coupled enzyme assay for the determination of isopenicillin N synthetase activity in purified extracts from Cephalosporium acremonium was described. Isopenicillin N formed from its precursor, delta-(L-alpha-aminoadipyl)-L-cysteinyl-D-valine (ACV), by the synthetase was hydrolyzed by beta-lactamase I to the corresponding penicilloic acid. Automatic titration of the acid with standard sodium hydroxide delivered by a pH-stat gave a continuous plot of product formed vs time. This assay has been used in kinetic studies and to determine the effects of pH, ionic strength, and temperature on the enzyme's activity.     

Characterization of the Oxygen Tolerance of a Hydrogenase Linked to a Carbon Monoxide Oxidation Pathway in Rubrivivax gelatinosus

A hydrogenase linked to the carbon monoxide oxidation pathway in Rubrivivax gelatinosus displays tolerance to O₂. When either whole-cell or membrane-free partially purified hydrogenase was stirred in full air (21% O₂, 79% N₂), its H₂ evolution activity exhibited a half-life of 20 or 6 h, respectively, as determined by an anaerobic assay using reduced methyl viologen. When the partially purified hydrogenase was stirred in an atmosphere containing either 3.3 or 13% O₂ for 15 min and evaluated by a hydrogen-deuterium (H-D) exchange assay, nearly 80 or 60% of its isotopic exchange rate was retained, respectively. When this enzyme suspension was subsequently returned to an anaerobic atmosphere, more than 90% of the H-D exchange activity was recovered, reflecting the reversibility of this hydrogenase toward O₂ inactivation. Like most hydrogenases, the CO-linked hydrogenase was extremely sensitive to CO, with 50% inhibition occurring at 3.9 μM dissolved CO. Hydrogen production from the CO-linked hydrogenase was detected when ferredoxins of a prokaryotic source were the immediate electron mediator, provided they were photoreduced by spinach thylakoid membranes containing active water-splitting activity. Based on its appreciable tolerance to O₂, potential applications of this hydrogenase are discussed.     

Amino acid replacements at the H2-activating site of the NAD-reducing hydrogenase from Alcaligenes eutrophus

The role of amino acid residues in the H(2)-activating subunit (HoxH) of the NAD-reducing hydrogenase (SH) from Alcaligenes eutrophus has been investigated by site-directed mutagenesis. Conserved residues in the N-terminal L1 (RGxE) and L2 (RxCGxCx(3)H) and the C-terminal L5 (DPCx(2)Cx(2)H/R) motifs of the active site-harboring subunit were chosen as targets. Crystal structure analysis of the [NiFe] hydrogenase from Desulfovibrio gigas uncovered two pairs of cysteines (motifs L2 and L5) as coordinating ligands of Ni and Fe. Glutamate (L1) and histidine residues (L2 and L5) were proposed as being involved in proton transfer [Volbeda, A., Charon, M.-H., Piras, C., Hatchikian, E. C., Frey, M., and Fontecilla Camps, J. C. (1995) Nature 373, 580-587]. The A. eutrophus mutant proteins fell into three classes. (i) Replacement of the putative four metal-binding cysteines with serine led to the loss of H(2) reactivity and blocked the assembly of the holoenzyme. Exchange of Cys62, Cys65, or Cys458 was accompanied by the failure of the HoxH subunit to incorporate nickel, supporting the essential function of these residues in the formation of the active site. Although the fourth mutant of this class (HoxH[C461S]) exhibited nickel binding, the modified protein was catalytically inactive and unable to oligomerize. (ii) Mutations in residues possibly involved in proton transfer (HoxH[E43V], HoxH[H69L], and HoxH[H464L]) yielded Ni-containing proteins with residual low levels of hydrogenase activity. (iii) The most promising mutant protein (HoxH[R40L]), which was identified as a metal-containing tetrametric enzyme, was completely devoid of H(2)-dependent oxidoreductase activity but exhibited a remarkably high level of D(2)-H(+) exchange activity. These characteristics are compatible with the interpretation of a functional proton transfer uncoupled from the flow of electrons.     

Bidirectional measurement of hydrogenase activity by the pH-stat method

The protons produced by the catalytic activity of hydrogenase in H2 evolution from dithionite-reduced methyl viologen or through benzyl viologen reduction by H2 gas are automatically titrated by a pH-stat device. This approach allows the measurement of hydrogenase activity and ensures the constancy of pH during the reaction in absence of buffers. Kinetic assays and pH and temperature-dependence experiments with Desulfovibrio gigas hydrogenase performed by this method basically confirm the results obtained with customary manometric assay.     

Regulation of H2 oxidation activity and hydrogenase protein levels by H2, O2, and carbon substrates in Alcaligenes latus.

Regulation of H2 oxidation activity and hydrogenase protein levels in the free-living hydrogen bacterium Alcaligenes latus was investigated. Hydrogenase activity was induced when heterotrophically grown cells were transferred to chemolithoautotrophic conditions, i.e., in the presence of H2 and absence of carbon sources, with NH4Cl as the N source. Under these conditions, H2 oxidation activity was detectable after 30 min of incubation and reached near-maximal levels by 12 h. The levels of hydrogenase protein, as measured by a Western blot (immunoblot) assay of the hydrogenase large subunit, increased in parallel with activity. This increase suggested that the increased H2 oxidation activity was due to de novo synthesis of hydrogenase protein. H2 oxidation activity was controlled over a surprisingly wide range of H2 concentrations, between 0.001 and 30% in the gas phase. H2 oxidation activity was induced to high levels between 2 and 12.5% O2, and above 12.5% O2, H2 oxidation activity was inhibited. Almost all organic carbon sources studied inhibited the expression of hydrogenase, although none repressed hydrogenase synthesis completely. In all cases examined, hydrogenase protein, as detected by Western blot, paralleled the level of H2 oxidation activity, suggesting that control of hydrogenase activity was mediated through changes in hydrogenase protein levels.     

Comparison of the membrane-bound and detergent-solubilised hydrogenase from paracoccus denitrificans. Isolation of the hydrogenase.

The hydrogenase from Paracoccus denitrificans is an integral membrane protein and has been solubilised by Triton X-100. The membrane-bound and detergent-solubilised forms of the enzyme have been compared. Both forms of the enzyme show a pH optimum for reduction of benzyl viologen at pH 8.5--9.0 and are both inhibited by concentrations of NaCl greater than 30 mM. An Arrhenius plot of the activity of hydrogenase in the membrane shows no 'break'. The form of the Arrhenius plot and the activation energy are not significantly changed on solubilisation of the enzyme. The Km and V values for benzyl viologen, methyl viologen and H2 are unaltered when the enzyme is extracted from the membrane. Therefore, solubilisation of hydrogenase from the membrane by Triton X-400 is unlikely to disrupt the native conformation of the enzyme. The detergent-solubilised hydrogenase has subsequently been purified using ammonium sulphate precipitation, sucrose density gradient centrifugation and chromatography on hydroxyapatite. The overall yield of activity is 23%, with a final purification of over 100-fold.     

Partial purification and characterization of the first hydrogenase isolated from a thermophilic sulfate-reducing bacterium.

A soluble [NiFe] hydrogenase has been partially purified from the obligate thermophilic sulfate-reducing bacterium Thermodesulfobacterium mobile. A 17% purification yield was obtained after four chromatographic steps and the hydrogenase presents a purity index (A398 nm/A277 nm) equal to 0.21. This protein appears to be 75% pure on SDS-gel electrophoresis showing two major bands of molecular mass around 55 and 15 kDa. This hydrogenase contains 0.6-0.7 nickel atom and 7-8 iron atoms per mole of enzyme and has a specific activity of 783 in the hydrogen uptake reaction, of 231 in the hydrogen production assay and of 84 in the deuterium-proton exchange reaction. The H2/HD ratio is lower than one in the D2-H+ exchange reaction. The enzyme is very sensitive to NO, relatively little inhibited by CO but unaffected by NO2-. The EPR spectrum of the native hydrogenase shows the presence of a [3Fe-4S] oxidized cluster and of a Ni(III) species.     

Secondary and tertiary structure characteristics of Megasphaera elsdenii flavodoxin in the reduced state as determined by two-dimensional 1H NMR

The secondary structure of two-electron-reduced Megasphaera elsdenii flavodoxin has been determined by visual, qualitative inspection of the sequential connectivities involving C alpha H, C beta H and NH protons observed in NOESY (two-dimensional nuclear Overhauser enhancement spectroscopy) spectra. Results from an amide proton exchange experiment were used to confirm the secondary structure assignment and to demonstrate the compactness and stability of the protein. After the secondary structure elements were established, the global fold of the protein and the flavin binding site have been determined using nonsequential interresidual NOE connectivities as primary source of information. The secondary structure and the global fold of M. elsdenii and Clostridium MP flavodoxin appeared to be very similar, differences are observed however. M. elsdenii flavodoxin consists of a central parallel beta-sheet including five strands surrounded on both sides by a pair of alpha-helices.