Eur. J . Biochem. 142, 21-28 (1984) ( FEBS 1984

Purification, characterization and redox properties of hydrogenase from Methanosarcina barkeri (DSM 800)

Guy FAUQUE, Miguel TEIXEIRA, Isabel MOURA, Paul A. LESPINAT, Antbnio V. XAVIER, Daniel V. DER VARTANIAN, Harry D. PECK Jr., Jean LE GALL, and Jose G. MOURA Centro de Quimica Estrutural, Complexo I, Lisbon, Equipe Commune d’Enzymologie, Centre d’Etudes Nucleaires Cadarache; and Department of Biochemistry, University of Georgia, Athens

(Received October 24, 1983/March 1, 1984) - EJB 83 1148

A soluble hydrogenase from the methanogenic bacterium, Methanosarcina barkeri (DSM 800) has been purified to apparent electrophoretic homogeneity, with an overall 550-fold purification, a 45 % yield and a final specific activity of 270 pmol H2 evolved min-’ (mg protein)-’. The hydrogenase has a high molecular mass of z 800 kDa and subunits with molecular masses of z 60 kDa. The enzyme is stable to heating at 65 “C and to exposure to air at 4°C in the oxidized state for periods up to a week. The overall stability of this enzyme is compared with other hydrogenase isolated from strict anaerobic sulfate-reducing bacteria. Ms. barkeri hydrogenase shows an absorption spectrum typical of a non-heme iron protein with maxima at 275 nm, 380 nm and 405 nm. A flavin component, identified as FMN or riboflavin was extracted under acidic conditions and quantified to approximately one flavin molecule per subunit. In addition to this component, 8 - 10 iron atoms and 0.6-0.8 nickel atom were also detected per subunit.

The electron paramagnetic resonance (EPR) spectrum of the native enzyme shows a rhombic signal with g values at 2.24,2.20 and = 2.0. probably due to nickel which is optimally measured at 40 K but still detectable at 77 K. In the reduced state, using dithionite or molecular hydrogen as reductants, at least two types of g= 1.94 EPR signals, due to iron-sulfur centers, could be detected and differentiated on the basis of power and temperature dependence. Center I has g values at 2.04, 1.90 and 1.86, while center I1 has g values at 2.08, 1.93 and 1.85. When the hydrogenase is reduced by hydrogen or dithionite the rhombic EPR species disappears and is replaced by other EPR-active species with g values at 2.33, 2.23, 2.12, 2.09, 2.04 and 2.00. These complex signals may represent different nickel species and are only obser\’able at temperatures higher than 20 K. In the native preparation, at high temperatures (T > 35 K) or in partially red wed samples, a free radical due to the flavin moiety is observed. The EPR spectrum of reduced hydrogenase in 80 ”/, Me2S0 presents an axial type of spectrum only detectable below 30 K.

The methanogenic bacteria are strict anaerobic microor- ganisms that synthesize methane from a limited number of substrates (e. g. formate, acetate, methylamines, methanol, CO or C 0 2 and H2). They belong to a bacterial group designated as Archaebacteria which are only distantly related in the evol- utionary scale to eukaryotes and the strictly anaerobic bacteria such as Clostridia and sulfate-reducing organisms [ 11. Supporting this clasification, they contain unusual cofactors not found in other organisms such as: coenzyme M [2], F420 [3], F342 [4], and F,,, [5,6,52] which may be isolated in a protein-bound form. In contrast, the amino acid sequence of Methanosarcina barkeri ferredoxin is quite similar to the clostridial type [7], but the evolutionary significance of this observation is not understood.

Methanogens are basically chemoautotrophs in their me- tabolism and molecular hydrogen is often used as a sub- strate. The oxidation of this substrate is coupled with the reduction of C 0 2 to methane. Hydrogenase is expected to be found since it is needed for the activation of the hydrogen molecule and has been demonstrated in intact cells as well

Abbreviations. EPR, electron paramagnetic resonance; Me,SO, dimethylsulfoxide ; SDS, sodium dodecyl sulfate; CI,AcOH, tri- chloroacetic acid.

~~~

as in cell extracts of several methanogenic bacteria: Me- thanobacterium thermoautotrophicum [8], Methanobacterium strain M.0.H. [9], Methanobacterium G2R [8], Methano- bacterium ruminantium strain P.S. [lo], Methanosarcina har- keri (DSM 800) [8], Methanospirillum hungatii [8] and Me- thanobacterium bryantii [ 111. Purified fractions of hydro- genase have been obtained and studied from several strains of methanogens : Methanobacterium G2R [8] and M . thermo- autotrophicum [12 - 141. The properties of solubilized parti- culate hydrogenase from Methanobacterium strain G2R were reported [8]. The enzyme has a large molecular mass of 900 kDa with subunits of molecular mass 38.5 kDa, 50.70 kDa and 80 kDa, a pH optimum at 8.5 for hydrogen oxidation and is fully active at 52°C. Reports have been published on the purification and characterization of the hydrogenase from two strains of Methanobacterium thermo- autotrophicum [12, 131 (and unpublished results in [l]). Graf and Thauer [I21 demonstrated that the hydrogenase from M . thermoautotrophicum (strain Marburg DSM 2133) was a nickel-containing enzyme. The formation of active hydro- genase was dependent on the presence of nickel in the culture medium, and the enzyme ( M , 60000) contained approximately one nickel atom per molecule. The EPR properties of the enzyme were studied and rhombic EPR signals observed in the

22

native preparation (g values at 2.3, 2.2 and 2.0) were assigned to a Ni(II1) species by using 61Ni-enriched samples [14]. The introduction of an isotope with a nuclear spin of 312 showed nuclear hyperfine induced structure in the original EPR spectrum, unequivocally proving that these EPR signals were due to nickel. However, no EPR signals characteristic of iron- sulfur clusters were detected and the authors posed the question of whether iron-sulfur centers are present in this hydrogenase. The hydrogenase activity in M . thermoauto- trophicum (strain Marburg) was eluted in two peaks [I21 indicating the presence of two hydrogenases in this organism or two forms of hydrogenase. Only one of the forms was purified.

A soluble hydrogenase has been purified from M . thermo- autotrophicum (strain AH) (Mortenson, L., Peck, H. D., Jr and LeGall, J., unpublished results in [l]). The enzyme has a molecular mass of % 185 kDa and is composed of four subunits with molecular masses of 64 and 40 kDa. This hydrogenase contains 30-40 nonheme iron atoms and its optimum activity is at pH 8.0.

Two hydrogenases from the methanogenic bacterium M . thermoautotrophicum (strain AH) have been purified recently [13]. Both enzymes contained tightly bound nickel as well as iron-sulfur centers in a ratio 15 -20 iron atoms per nickel atom. One hydrogenase (subunits of 40 kDa, 31 kDa and 26 kDa) reduces the coenzyme factor F420 whereas the other hydrogenase (subunits of 52 and 40 kDa) does not reduce F42, and has been purified as a methyl-viologen-reducing enzyme. Rhombic EPR signals detected at % 150 K were demonstrated to be associated with the nickel center by hyperfine coupling in "Ni-containing hydrogenase.

Nickel, in addition to iron-sulfur centers, has been shown to be a structural component of hydrogenases purified from bacteria other than the methanogens. Nickel-containing hy- drogenases have been purified from Desulfovibrio gigas [I5 - 171, Desulfovibrio desulfuricans (ATCC 27774) [18], D. desulfuricans (strain Norway 4) [ 191, Alcaligenes eutrophus [20], Chromatium zjinosum [21] and Vibrio succinogenes [22].

In this paper, we report on the purification, characteriza- tion and redox properties of the hydrogenase from Ms. barkeri (DSM 800). This is the first report on the purification of a hydrogenase from a methanogen capable of converting acetate to methane.

MATERIALS AND METHODS

available. All chemicals and reagents are of the highest purity

Assays

Hydrogenase activity was assayed by the rate of H2 evolution with sodium dithionite as electron donor and methyl viologen as mediator [23]. The assay mixture (3 ml) in 13-ml sealed vials contained 50 mM Tris/HCl buffer pH 7.6, bovine serum albumin (0.5 mg/ml) and methyl viologen (1 mM). After flushing with argon for 15 min, freshly prepared sodium dithionite was added (15 mM). The vials were shaken at 30 "C and the reaction started by the addition of hydrogenase. One unit of enzyme activity is defined as the amount of enzyme which catalyses the evolution of 1 bmol H,/min under these conditions. Hydrogen evolution was measured by means of gas chromatography using an Acrograph A-90 P3 chromato- graph. Total iron was determined by the 2,4,6-tripyridyl-s-

1,3,5-triazine method [24] and nickel by atomic absorption using a Perkin-Elmer model 403. Protein was determined by the Lowry [25] and biuret [26] methods, using a protein standard solution purchased from Sigma.

The molecular mass of the native preparation was esti- mated by gel filtration on a column (1 x 100 cm) of Sephadex G-200 (fine) [27]. The column was equilibrated and eluted with 20 mM phosphate buffer, pH 7.6, 1 M NaC1. Al- bumin (45 kDa), aldolase (158 kDa), catalase (240 kDa) and ferritin (450 kDa) were used as molecular mass markers. The standard proteins were individually passed through the columon. The homogeneity of the preparations was checked on 7 A polyacrylamide gels [28]. The subunit structure was determined on SDS/polyacrylamide gel electrophoresis [29] using the following molecular mass markers : trypsin (21 kDa), albumin (45 kDa), aldolase (158 kDa), catalase (240 kDa) and ferritin (450 kDa).

The flavin chromophore was extracted by denaturing the protein (1 mg/ml) with 8% C1,AcOH and the precipi- tate removed by centrifugation. The chromophore present in the supernatant was identified by thin-layer chromatog- raphy on cellulose and silica plates using FMN, FAD, and riboflavin as standards. Two running solvents were used: solvent A (5% NaH2P04, in water) and solvent B (n-butanollacetic acid/water, 4:3 :3) [30].

EPR samples were buffered in Tris/HCl, pH 7.6. An- aerobic atmospheres were created with purified argon gas. Reduction of anaerobic samples was accomplished by adding aliquots of 100 mM sodium dithionite solution prepared in 0.5 M Tris/HCl at pH 9.0 degassed with argon, using gas-tight Hamilton syringes or exposure to hydrogene atmosphere. Mixing was assured in order to obtain homogeneous samplts.

The EPR spectrum of dithionite-reduced protein in SOL Me2S0 was measured in a sample prepared as described in [31 J and [32].

The coupling effect of hydrogenase on the reduction of F420 cofactor and ferredoxin was assayed by a spectrophoto- metric method, following the reduction of the chromophore at 420 nm (F420) and 410 nm (ferredoxin) in the presence of hydrogenase under a hydrogen atmosphere (1 atm, 25 "C).

Spectroscopic instrumentation

Electron paramagnetic resonance spectroscopy (EPR) was carried out on a Bruker 200-tt spectrometer, equipped with an ESR-9 flow cryostat (Oxford Instruments Co., Oxford, UK) and a Nicolet 1180 computer with which mathematical manipulations were performed. The visible/ultraviolet spectra were obtained on a Beckman model 35. Fluorescence emission was measured on a fluorescence spectrometer Perkin-Elmer MPF 3. EPR signal intensities were determined by double integration with baseline corrections and were compared with a Cu-EDTA standard (1 mM) and an oxidized Desulfovibrio gigas ferredoxin I1 sample (0.174 mM).

Growth of the microorganism and preparation of the crude extract

Methanosarcina barkeri (DSM 800) was grown at 37 "C in a medium as previously described [33]. The cells were har- vested by centrifugation at the end of the growth phase and stored at -25°C. The purity of the cultures used for hydrogenase purification was checked by microscopic exa- mination with ultraviolet light. In a typical preparation of the crude extract, 1.9 kg of cells (wet weight) grown in methanol,

23

Table 1. PuriJication scheme for Methanosarcina barkeri (DSM 800) hydrogenase 1 unit is the amount catalyzing the evolutioa of 1 pmol H,/min

Purification step Volume Total Specific Yield activity activity

ml units units/mg protein % Crude extract 2400 87024 0.49 100 After DEAE- cellulose and DEAE-Bio-Gel 991 57167 4.12 65.7 After hydroxylapatite 131 51 103 16.6 58.7 After centrifugation 4.7 39240 270 45.1

were suspended in 800 ml of 10 mM Tris/HCl buffer pH 7.6. DNase was added and the extract was ruptured once in a French press at 62 MPa under N2 atmosphere. The extract was centrifuged at 12000 rpm for 30 min and a total volume of 2400 ml of crude cell extract was obtained.

Purgication of hydrogenase

Unless otherwise stated, the entire purification procedure was carried out at 0 -4 "C under argon atmosphere and all buffers were flushed with argon before use. Tris/HCl or phosphate buffers (pH 7.6) at appropriate molarities were used.

The crude cell extract was separated in two equal parts: one was adsorbed on a DEAE-cellulose (DE 52) column (6 x 36 cm) and the other one on a DEAE-BioGel A column (6x34cm). After elution with a continuous gradient of 10 - 500 mM Tris/HCl buffer, the hydrogenase activity was found to be distributed in a membrane fraction not absorbed, also containing cytochrome b, as well as in a soluble fraction eluted at low ionic strength ( < 200 mM). The soluble fraction was collected from both columns (total volume 991 ml) and adsorbed on an hydroxylapatite column (4.5 x 16 cm) equi- librated with 200 mM Tris/HC1 buffer. The hydrogenase fraction was not retained in this step, and was concentrated in a ultrafilter Diaflo Amicon with a YM-30 membrane. The concentrated fraction (131 ml) presented a high turbidity and was centrifuged for 40 min at 8500 rpm. The brownish pre- cipitate was collected and dissolved gently under argon in 300 mM Tris/HCl buffer pH 7.6, and centrifuged again for 20 min at 8500 rpm. The supernatant (4.7 ml) contained an active hydrogenase with an absorption ratio A400/A275 = 0.29. By this procedure, 145 mg of pure hydrogenase were obtained.

RESULTS The hydrogenase activity in the crude cell extract of

Methanosurcina barkeri (DSM 800) distributes between the particulate and soluble fractions. Only the hydrogenase pre- sent in the soluble fraction collected after the first DEAE- cellulose and DEAE-BioGel purification steps was purified; the procedure is summarized in Table 1 . The hydrogenase was purified 550-fold with a final specific activity of z 270 pmol H, evolved min-' (mg protein)-' (at pM = 7.5, showing a p- crease in activity at higher pH). The overall yield was 45 A.

% 1.6 c

8 s 1.2

3 0.8

2

L

0

al 0

0

1 0.L

0 300 100 500 600

Wavelength (nm) Fig. 1 . Electronic absorption spectrum of' oxidized Ms. barkeri (DSM 800) hydrogenase at 25°C. (--) Absorbance of isolated hydrogenase (left scale) ; (-.-.-) spectrum of the chromophore released by acid treatment (right scale). The cuvette was sealed under argon and the protein concentration was 0.62 mg/ml in 50 mM Tris/HCl buffer, pH 7.6

In the presence of hydrogen the purified hydrogenase can reduce the F420 cofactor (either free or protein bound) [l, 31 as well as the ferredoxin isolated from Ms. barkeri [7]. In the same conditions, it can also reduce methyl and benzyl viologen and it is active in the ferredoxin-dependent reduction of sulfite to sulfide [l].

Molecular mass

The molecular mass was estimated after gel filtration on a Sephadex G-200 column. The elution volume of the hydro- genase fraction was very close to the exclusion limit. Using standard molecular mass proteins, an extrapolated value of z 800 kDa was determined. This value is compatible with the elfctrophoretic data since the hydrogenase is unable to enter 7 4 polyacrylamide gels. SDS/polyacrylamide gel electro- phoresis showed bands at z 50 - 70 kDa, 120 - 140 kDa and z 200 kDa.

Spectral data and active-center composition

The Ms. barkeri hydrogenase has a golden-brown colour and the absorption spectrum shown in Fig. 1 is typical of a non-heme iron protein. Prominent bands are observed at 275 nm, 380 nm and 405 nm (shoulder). The ratio A400/A275

was 0.29. On reduction under hyodrogen a bleaching was observed which corresponded to 4 0 4 of the total absorption at 400 nm. When the apoprotein was precipitated with 8% C1,AcOH a fluorescent chromophore was released from the enzyme. Comparison of the emission spectra of the native enzyme and the treated supernatant revealed a large increase in fluorescence at 520 nm when the chromophore was released (Fig. 2). On reduction with dithionite the fluorescence of the free chromophore disappears. The visible spectrum of the free chromophore is also shown in Fig. I, with bands at 370 nm and 435 nm. The chromophore was identified as a flavin (FMN or riboflavin), by thin-layer chromatography, as indicated in Table 2. The amount of flavin was estimated using a prepara- tion of flavodoxin from Desulfovibrio gigas as standard [34].

Based on a minimum molecular mass of 60 kDa and on the protein concentration determined as described in Ma-

24

1 x

In c m C

m U c m U m

c .-

c

.-

: 3

LL -

- _ - - _ I I I I I I I I

160 500 5LO 580 I I I I I I I I

160 500 5LO 580 Wavelength lnrn)

Fig. 2. Fluorescence emission spectra (-.-.) native Ms. barkeri hy- drogenase and ( ~ - --) extracted chromophore. Experimental con- ditions: excitation wavelength, 400 nm; concentration 1.2 mg/ml in 8 7; trichloroacetic acid

Table 2. Identification of the jluvin chromophore of Ms. barkeri hydrogenuse The chromophore was released by precipitation of the hydrogenase with 8% C1,AcOH (see Materials and Methods)

Substance R, value on plates of

cellulose in solvent silica in solvent

A B B

FAD 45.9 4.9 7.3 FMN 53.5 36.2 16.6 Riboflavin 50.9 16.9 16.3 Ms. burkeri chromophore 57.3 18 17.2

terials and Methods, 1 FMN or riboflavin molecule, 8 - 20 iron atoms and 0.6 -0.8 nickel atom per 60 kDa were found. The peak at 380nm of the optical spectrum (Fig. 1) may represent the flavin contribution to the total absorption.

EPR data

The following redox states of the enzyme were analyzed using electron paramagnetic resonance spectroscopy : the native, partially reduced, and fully reduced states, using hydrogen or dithionite as electron donors.

In the native state (as isolated) the purified hydrogenase does not exhibit an ‘isotropic’ type of EPR spectra in the g = 2 region. when observed at low temperature (< 20 K). How- ever, a rhombic signal, that is highly saturated with power at low temperature, is observed with g values at 2.24, 2.20 and 22.0. The signal is optimally detected at 40 K but is still observed at 77 K (Fig. 3). A minor EPR component barely discernible with features at 2.30 and 2.12 is also observable. Minor components assigned to nickel and dependent on protein preparation were also reported in D. gigas hydro-

>

In C

c

a, >

0 a, [L

c .-

- .-

.- - -

2.33 2.23

I I

H 10mT

-__ Magnetic f ield s t r e n g t h

Fig. 3. Nickel EPR signals in Ms. barkeri ( D S M 800) hydrogennse. (A) Native preparation 45.4 K, gain 5 x lo5, microwave power 3.2 mW. (B) Enzymereducedwithalargedithioniteexcess, 60 K, gain3.2 x lo5, microwave power 20 mW. Other experimental conditions : protein concentration, 32 mg/ml in 50 mM Tris/HCl buffer, pH 7.6, modu- lation amplitude ImT

genase [16]. The integration of the nickel signal in the native state yields a low value coompared with the chemically deter- mined nickel (less than 10/0). Although the integrated values reported so far for nickel EPR signals in nttive preparations of hydrogenase only account for about 50/0 of the chemically detectable nickel [13,15-171 this is a quite unusually low value. Probably the protein is isolated in a partially reduced state, as indicated by the presence of a semiquinone flavin signal observable in the native preparation (see below).

The partially and fully reduced states of the enzyme were analyzed in detail. When the enzyme is incubated under hydrogen atmosphere for a short time, the rhombic signal previously observed disappears. A complex spectrum develops at low temperature ( < 20 K), with features at g values of 2 1.94 (Fig. 4A). Combining power and temperature dependence studies of this signal (Fig. 4B shows the temperature effect), two different sets of g values can be resolved describing two types of iron-sulfur centers : center I, with g values at 2.04, 1.90 and 1.86, and center I1 with g values at 2.08, 1.93, and 1.85. After a long exposure to hydrogen atmosphere (Fig. 4C) the relative intensities of center I and center I1 change with an increase in intensity of the features associated with the g = 2.04 signal. Hydrogenase samples reduced with an equivalent amount of dithionite gave Fe-S EPR signals identical to those of the hydrogen-reduced hydrogenase (Fig. 4C and D). In the presence of a twofold molar excess of dithionite the signal at g = 1.90 disappears and new signals at g values 2.09, 2.04, 1.93, 1.86 are observed (Fig. 4E). The double integration of the iron- sulfur EPR signals of a dithionite-reduced sample (Fig. 4D) yields 1.7-2.0 spins per molecule (Mr 60000).

When the fully reduced samples are examined at tempe- ratures higher than 30 K a very complex signal with g values at 2.33, 2.23, 2.12, 2.09, 2.04 and 2.00 is observed (Fig. 3B; this may represent multi-nickel species and may account for at least two types of rhombic signals. These signals are already observed in partially H,-reduced samples.

The flavin contribution is suggested in Fig. 3 and 4, in the g = 2 region of the spectrum. A radical signal can be detected in

25

2.08 2.01 1.93 1.90 ' ' 1.86

Table 3. Thermal stability of diflerent hydrogenasespurified from strict anaerobic bacteria The values are presented as a percentage of the maximum activity obtained. A dash (-) indicates the value was not determined

Hydrogenase Incubation source time

Activity at temperature of

30°C 50°C 65°C 72°C 80°C 95°C

min %

Ms. barkeri 20 60

120 D. vulgaris 20

60 120

D. gigus 10 20 60

120

97 91 75 14 1 100 88 70 5 0 100 86 55 2 0

100 93 20 0.1 0 100 94 12 0 0

95 100 93 78 55 98 96 97 74 49 98 97 97 71 47

98 99 28 0.25 0.1

- - - - - 2.6 0.8 0 0

Magnetic field strength

Fig. 4. Iron-sulfur EPR signuls in reduced Ms. barkeri (DSM 800) hydrogenase. (A) Enzyme partially reduced under hydrogen atmo- sphere, 8.4 K, gain 3.2 x 1 04, microwave power 0.63 mW. (B) Enzyme partially reduced under hydrogen atmosphere (as A) 15.7 K, gain 5 x lo4, microwave power 2 mW. (C) Enzyme 'fully' reduced under hydrogen atmosphere, 9.8 K, gain 2 x lo4, microwave power 2 mW. (D) Enzyme dithionite-reduced, 8.3 K, gain 1.2 x lo4, microwave power 2 mW. (E) Enzyme reduced withexcess of dithionite 8.7 K, gain 5.0 x lo4, microwave power 2 mW. (F) Dithionite-reduced enzyme in 80% Me,SO, 9.8 K, gain 0.3 x lo4, microwave power 2 mW. Other experimental conditions: protein concentration 32 mg/ml in 50 mM Tris/HCI buffer, pH 7.6, modulation amplitude 1 mT

the native preparations at high temperature (77 K) and has a large intensity in the partially reduced samples (Fig. 4A). However, when attempts are made to reduce the enzyme fully, the signal decreases in intensity (Fig. 4C and D). With a large excess of dithionite (Fig. 4E) the signal is not detectable.

The dithionite-reduced protein was unfolded in 80% Me2S0 and a nearly axial type of spectrum was obtained with g values at 2.08, and 1.94 only detected below 30 K (Fig. 4F).

Thermal stability of the enzyme

As shown in Table 3 Ms. barkeri hydrogenase is rather stable at high temperatures under anaerobic conditions. After 2 h of incubation at 65 "C the enzyme still retains 55% of its initial activity. Above this temperature there is a rapid loss in activity. A comparison of the thermal stabilities of different hydrogenases purified from strict anaerobic bacteria shows the following stability order (Table 3 and [19]): Desulfovibrio desulfuricans (Norway strain) N D. gigas > Ms. barkeri (DSM 800) p Desulfovibrio vulgaris (Hildenborough).

The native Ms. barkeri hydrogenase did not lose activity after exposure to air at 4 "C for 8 days. The hydrogenases from several species of Desulfovibrio have also been found to be stable in air (D. desulfuricans) (Norway starin) [19], D . vulgaris (Hildenborough) [35] and D. gigas (G. Fauque and J. LeGall, unpublished results).

DISCUSSION

The information available on the hydrogenases from methanogens is very limited. Our interest in the purification of Methanosarcina barkeri hydrogenase stems from the need to characterize this enzyme in the group of methanogens that catalyze the formation of methane from acetate.

Hydrogenases catalyze the simplest oxidation-reduction process: H,e2Hf +2e, and they have been purified from many different groups of bacteria and algae. Hydrogenases have been generally recognized as iron-sulfur-containing proteins, usually with four to twelve iron atoms in different cluster arrangements : [2Fe-2S], [3Fe-xS] and [4Fe-4S]. In addition, some bacterial species contain hydrogenases with flavin prosthetic groups [I 3,301. Recently nickel was shown to be a constituent of some bacterial hydrogenases and the number of nickel-containing hydrogenases has been increas- ing : Desulfovibrio gigas [ 15 - 171, Desulfovibrio desulfuricans (ATCC 27774) [18], and D. desuljuricans (Norway strain 4) [19], Methanobacterium thermoautotrophicum [I 3,141, Alcali- genes eutrophus [20], Chromatrum vinosum [21], and Vibrio succinogenes [22].

The soluble hydrogenase of Ms. barkeri is shown in the present report to be a flavin-nickel-iron-sulfur-containing protein. The isolated enzyme has a very large molecular mass (z 800 kDa) and is composed of subunits of molecular mass 60 kDa. The high molecular mass is consistent with the observation of an R, = 0 in 4 % polyacrylamine gels for Ms. barkeri hydrogenase activity, reported by Mckellar and Sprott [8]. Also the hydrogenase isolated from Methanobacterium strain G2R [ 8 ] has a high molecular mass ( ~ 9 0 0 kDa).

Ms. barkeri hydrogenase was shown to mediate electron transfer from molecular hydrogen to either free or protein- bound F,,, (unpublished results) ; however, the specificity of hydrogenases from methanogens towards this cofactor is not very clear. In general, crude preparations and high-molecular- mass forms of hydrogenase reduce free F420, but low- molecular-mass forms do not reduce this unusual cofactor. Fuchs et al. [36] reported that hydrogenase (MI 60000) from M . therrnoautotrophicum (strain Marburg), reduces F,,,. A soluble hydrogenase purified from M . thermoautotrophicum (Mortenson, Peck, LeGall, unpublished results in [I]),

26

M , 185000 (subunits of M , 40000 and 64000) did not reduce free F420. Recently, two hydrogenases were purified from M . thernzoautotrophicum [ 131 ; the higher-molecular-mass species contains FAD, iron-sulfur centers and Ni and reduces free F420

under H, , while the lower-molecular-mass species contains only iron-sulfur centers plus Ni, and did not reduce free F,,, under H,. A clear distinction should be made between free and protein-bound cofactors such as F420 since the reactivity of hydrogenase may change accordingly. As an example, pure hydrogenase from D. gigas can reduce free FMN but not the same cofactor bound to flavodoxin [37]. Flavin groups have also been found in the nickel-containing hydrogenase from A . eutrophus [30] which contain two FMN groups.

The g values of the rhombic EPR signals observed in Ms. barkeri hydrogenase are within the range found for nickel-

Table 4. The g values of the nickel EPR signals of'hydrogenasepurij2ed j i o m d!ferent bacterial sources and model compounds

Source of hydrogenase gmax g m e d gmin Ref.

M . thermoautotrophicum 2.3 2.2 (Marburg) M . tliermoautotrophicum (AH) 2.31 2.23 D. desuljuricans (Norway strain) 2.32 2.23 D . desulfuricans (ATCC 27774) 2.32 2.16 D. gi&s native 2.31 2.23

under H, 2.19 2.16 Ms. barkeri native 2.24 2.20

under Hz 2.33 2.12

2.23 2.09

2.0 [14]

2.01 [13] 2.014 [19] 2.01 [I81 2.02 [15] 2.02 [40] 2.12 2.01 this

2.04 work

Ni(II1) model compounds

Macrocyclic complexes - 2.205-2.012 - 2.144-2.016 [48] (81) ( S i i )

Bis(mallonitrilodithio1ato) 2.2055 -2.081 2.062 1491 nickel trianion

Oligopeptide complexes 2.227 2.011 P O I (H- 433NH3) (gl) (g//)

Tripeptide complexes 2.30 2.24 2.01 [511 ( Ala-GI y -Gly)

containing hydrogenases. Table4 shows the g values for nickel-containing hydrogenases and model compounds. With the exception of Vibrio succinogenes, A . eutrophus and Ch. vinosum enzymes, a careful analysis of the nickel EPR signals shows that they are redox-sensitive. The detection of nickel signals in biological systems was first reported by Lancaster [38] in membrane preparations of Methanobacterium bryantii; the tentative assignment was based on data available for nickel model compounds and on the fact that the g values, line-shape and relaxation properties did not fit with other known biologically relevant metals. The rhombic signal observed (g values at 2.3, 2.2 and 2.0) was unequivocally identified from the 61Ni-enriched preparations of M . bryantii [39] and hydro- genases from M . thermoautotrophicum [13,14] and D. gigas [40], since the nuclear spin of the 61Ni induced hyperfine structure in the EPR lines.

The rhombic EPR signal detected in the native form of hydrogenases was attributed to nickel in the 3 + state since one electron was added upon reduction [15,16] and the EPR signals disappear when the enzyme is reduced under anaerobic conditions with dithionite or molecular hydrogen, presumably due to the formation of diamagnetic nickel(I1). However, in the case of D. gigas hydrogenase, during an extended exposure to hydrogen atmosphere, another rhombic EPR signal devel- ops with g values at 2.19, 2.16 and 2.02. This new signal goes through a maximum intensity and after longer incubation under H,, an EPR-silent state is obtained at 77 K [41]. Because of its low intensity and appearance under hydrogen atmo- sphere, this signal was proposed to represent a transient nickel species trapped in the enzyme turnover [40]. A similar 2.19- type of rhombic signal was also recently observed in prepara- tions of M . thermoautotrophicum (strain AH) hydrogenase [13]. Other results [I31 also support the presence of a paramagnetic nickel signal with a transient nature. When hydrogen was replaced by argon in fully reduced samples of M. thermoautotrophicum hydrogenase (which was an EPR-silent state at high temperature) the appearance of the 2.19-type signal was observed, which was interpreted as an oxidative conversion of nickel(I1) in the fully reduced state to nickel(II1). Small but detectable hyperfine differences in the 2.19-type signal are observed when 61Ni-enriched and natural abun- dance enzyme are compared [41], confirming the origin of the paramagnetic signal. The different g values for the transient

Table 5. Redox center composition of nickel-containing hydrogenases Flavin indicates FMN or riboflavin; n.r. =not reported; + =detected; - =not detected

Hydrogendse source Ni Fe/S Other centers Isotope Ref. substitution performed

this work Ms. barkeri (DSM 800) 0.5 2 x [4Fe-4S] flavin n.r. D. gigas 1 [3Fe-xS] - 61Ni [I 6,401

M . thermoautotrophicum (Marburg) 1 [Fe-S] -

n.r. t211 Chr. rinosum + [3Fe-xS] -

n.r. [221 V. succinogenes + [Fe-S] - D. desulfuricans (ATCC 27774) 0.7 [3Fe-xS] - 61Ni, 57Fe [I81

2 x [4Fe-4S] 61Ni 1141

M. thermoautotrophicum (AH) + n x [Fe-S] FAD 61Ni ~ 3 1

[4Fe-4S]

14Fe-4S1 D. desulfuricans (Norway 4) + i4Fe-4S j -

I3Fe-xS1 n.r. 1191

A. eutrophus 1-2 2 x [2Fe-2S] 120,431 2 x [4Fe-4S] 2 x FMN n.r.

27

species as well as the difference in hyperfine coupling constants may reflect a replacement of ligands in the nickel or in the coordination number, upon hydrogen activation.

The study of the evolution of the rhombic EPR features associated with the nickel center of the Ms. barkeri hydro- genase upon exposure to molecular hydrogen also indicates that a two-step process takes place: first there is the disap- pearance of the EPR signal with g values at 2.24, 2.20 and 2.0 (Fig. 3); secondly, after a long exposure to hydrogen another complex group of EPR signals develop with g values at 2.33, 2.23, 2.12, 2.09, 2.04, ~ 2 . 0 (Fig. 3B).

The data known so far indicate that if nickel is involved in the primary phase of activation of the hydrogen molecule it is plausible to consider a hydride species, but the hypothesis of a further reduced state of the nickel center, Ni(I), should not be disregarded at the present time [15,16,40]. Furthermore, EPR studies on D. gigas hydrogenase partially reduced by dithionite or hydrogen suggested that the nickel redox center is the catalytic site of activation of the hydrogen molecule [15]. The mid-point redox potential of the nickel center is known for the D. gigas hydrogenase (Eo z -220 mV) [15,16] and was shown to be pH-dependent [I 61. Although such a value is much higher than the hydrogen potential (EA = -440 mV) it must be emphasized that the activation of molecular H2 by a metal catalyst does not necessarily require a redox change of the metallic center which may act simply as a hydride acceptor. Such a model is known and has been described in detail using a palladium' + catalyst [42]. Thus, the observed redox change of the nickel center of hydrogenase could represent the necessary step for the activation of the enzyme allowing the binding of hydride ions.

The present knowledge of the iron-sulfur core arrangement in nickel-containing hydrogenases is limited. A [3Fe-xSI3+ cluster, associated with an intense 'isotropic' EPR signal centered at g= 2.02 and two EPR-silent [4Fe-4SI2+ clus- ters were demonstrated to be present in D. gigas [15] and D. desulfuricans (ATCC 27774) [17] by Mossbauer and EPR spectroscopy. The redox mid-point potential of the [3Fe-xSI3+"+ cluster is -70 mV [15,16]. EPR studies of D. desulfuricans (Norway strain) hydrogenase in the oxidized state indicates the presence of a [4Fe-4SI3+ or [3Fe-xSI3+ [IS]. A . eutrophus hydrogenase seems to contain 2 x [2Fe-2S] and 2 x [4Fe-4S]2+(2+9'+) centers [43]. The EPR spectrum of oxidized Chr. vinosum hydrogenase has been considered indicative for a [4Fe-3S]3+(3t,2+) [21]. The M . thermoauto- trophicum (strain Marburg) enzyme did not show iron-sulfur EPR signals either in the native or in the hydrogen-reduced or ferricyanide-oxidized states [I 31. Data on the redox center composition of several nickel-containing hydrogenases is presented in Table 5 , indicating a wide variation of active- center composition.

Iron-sulfur EPR signals could be detected at low temper- ature in the reduced state of the Ms. barkeri enzyme. Two types of iron-sulfur centers were discernible based on power and temperature dependence. No evidence was found for a [3Fe-xSI3+ center in the native preparations.

When the reduced protein is unfolded in SO% Me,SO a nearly axial type of EPR spectrum was obtained, with g values at 2.08 and 1.94, which is only detectable below 30 K. This signal is rather similar to those of reduced [4Fe-4S]If- containing ferredoxins and analogous compounds [44]. Ferre- doxins containing [2Fe-2S] centers show in the same con- ditions an isotropic EPR signal that can readily be detected at 77 K [45]. This difference has been suggested as a method for distinguishing between the two types of clusters by examining

the line shape and temperature dependence of the reduced protein in 80% Me,SO [31, 32, 451. The finding of [ ~ F - x S ] ' ~ cores has introduced an additional difficulty since easy interconversion was shown to occur boetween 3Fe and 4Fe cores [46]. The unfolding process in S O L Me,SO may cause rearrangements of the cores and therefore experiments must be supported unequivocally by independent spectroscopic me- thods. However, the detection of the axial type of signals in Me,SO and the lack of an isotropic signal in the native preparations may be indicative that [4Fe-4SI2 +(' +,' +) centers are present in the Ms. barkeri hydrogenase.

The thermal stability of different hydrogenases as reported in Table 3 shows that the resistance to high temperatures of a particular enzyme is not necessarily related to the physiology of the organism from which it has been isolated. The hydrogenases from M . barkeri, D. vulgaris and D. gigas all show high resistance to elevated temperatures and in the case of the enzyme from the latter organism, it is particularly striking since full activity is retained after an incubation at 65 "C for 2 h; however the optimal temperature for the growth of the organism is only 35 "C [47].

This research was supported by grants from Znstituto Nacional de InvestigaCcio Cientifica and Junta Nacional de Investigaccio Cientifica e Tecnologica, Portugal, and a NATO grant (J.J.G.M.), National ScienceFoundationgrant PCM 8111325(J.L.G., D.V.D. andH.D.P.) and SERI-Solar Energy Research Institute sub-contract XD-1-1155-1 (J.L.G.). We are indebted to Mrs I. Carvalho for her skilful technical help and to Dr M. Scandellari and Mr R. Bourelli for growing the bacteria which have been used for the present study.

REFERENCES

1. LeGall, J., Moura, J. J. G., Peck, H. D. Jr &Xavier, A. V. (1982) in Iron-Suljiur Proteins (T. G. Spiro, ed.) pp. 177 -248, John Wiley and Sons, New York.

2. McBride, B. C. & Wolfe, R. S. (1971) Biochemistry 10, 12317. 3. Cheeseman, P., Tomswood, A. & Wolfe, R. S. (1972) J. Bacteriol.

4. Gunsalus, R. P. & Wolfe, R. S. (1978) FEBS Lett. 3, 191. 5. Whitman, W. B. & Wolfe, R. S. (1980) Biochem. Biophys. Res.

Commun. 92, 1196. 6. Diekert, G., Kler, B. & Thauer, R. K. (1980) Arch. Microbiol. 128,

256. 7. Hausinger, R. P., Moura, I. Moura J. J. G., Xavier, A. V., Santos,

H., LeGall, J. & Howard, J. B. (1982) J. Biol. Chem. 257,

8. McKellar, R. C. & Sprott, G. D. (1979) J. Bacteriol. 139,

9. Eirich, L. D., Vogels, G. D. & Wolfe, R. S. (1979) J. Bacteriol.

10. Tzeng, F. S., Wolfe, R. S. &Bryant, M. P. (1975) J. Bacteriol. 121,

11. Pol, A,, van der Drift, C., Vogels, G. D., Guppen, T. J. H. M. & Laarhoven, W. H. (1980) Biochem. Biophys. Res. Commun. 92,

112, 527.

14192-14194.

231 -238.

140, 20 -27.

184 - 191.

255 -260. 12. Graf, E. & Thauer, R. K. (1981) FEBS Lett. 136, 165 -169. 13. Kogima, N., Fox, J., Hausinger, R., Daniels, L., Orme-Johnson,

W. H. & Walsh, C. (1983) Proc. Natl Acad. Sci. USA 80,

14. Albracht, S . P., Graft, E. G. & Thauer, R. K. (1982) FEBS Lett.

15. LeGall, J., Ljungdahl, P., Moura, I., Peck, H. D. Jr., Xavier, A. V., Moura, J. J. G., Teixeira, M., Huynh, B. H. & DerVartanian, D. V. (1982) Biochem. Biophys. Res. Commun.

16. Teixeira, M., Moura, I., Xavier, A. V., DerVartanian, D. V., LeGall, J. & Moura, J. J. G. (1983) Eur. J. Biochem. 130,

378 - 382.

140, 311 -313.

106, 610-616.

481 -484.

28

17. Cammack, R., Patil, D., Aguirre, R. & Hatchikian, E. C. (1982)

18. Kriiger, H.-J., Huynh, B. H., Ljungdahl, P. O., Xavier, A. V., DerVartanian, D. V., Moura, I., Peck, H. D. Jr., Teixeira, M., Moura, J. J. G. & LeGall, J. (1982) J. Bid. Chem. 257,

19. Lalla-Maharajh, W. V., Hall, D. O., Cammack, R., Rao, K. K. &

20. Friedrich, C. G., Schneider, K. & Friedrich, B. (1982) J . Bacteriol.

21. Albracht, S. P. J., Albrecht, Elmer, K. Y., Schmedding, D. J. M. &

22. Unden, G.,BOcher, R.,Knecht, J. &KrOger,A.(1982)FEBSLett.

23. Peck, H. D. Jr & Gest, H. (1956) J . Bacteriol. 71, 70-80. 24. Fischer, D. S. & Price, D. C. (1964) Clin. Chem. 10, 21 -25. 25. Lowry, 0. H., Rosebrough, N. J., Fan, A. L. & Randall, R. J.

26. Levin, R. & Brauer, R. W. (1951) J . Lub. Clin. Med. 38, 474-

27. Whitaker, J. R. (1963) Anal. Chem. 35, 1950-1953. 28. Davis, B. J. (1964) Ann. N . Y. Acad. Sci. 121, 404-427. 29. Weber, K. & Osborn, M. (1969) J . Biol. Chem. 244,4406 -4412. 30. Schneider, K. & Schlegel, A. (1978) Biochem. Biophys. Res.

31. Cammack, R. (1975) Biochem. SOC. Trans. 3, 482-488. 32. Cammack, R. & Evans, M. C. W. (1975) Biochem. Biophys. Res.

33. Blaylock, B. A. & Stadtman, T. C. (1966) Arch. Biochem. Biophys.

34. Dubourdieu, M. & LeGall, J. (1970) Biochem. Biophys. Rex

35. van der Westen, H. M., Mayhew, S. G. & Veeger, C. (1978) FEBS

FEBS Lett. 142, 289 -292.

14620 -14623.

LeGall, J. (1983) Biochem. J . 209,445 -454.

152, 42 -48.

Slater, E. C. (1982) Biochim. Biophys. Acta 681, 330 -334.

145, 230 -234.

(1951) J . Bid. Chem. 193, 265-275.

479.

Commun. 84, 564-571.

Commun. 67, 544 - 549.

116, 138-158.

Commun. 38, 965 -972.

Lett. 86. 122-126.

36. Fuchs, G., Woll, J., Scherer, P. & Thauer, R. H. (1976) in Hydrogenuses : Their Catalytic Structure and Function (H. G. Schlegel, ed.) pp. 83 -92, Goltur, Gottingen.

37. Bell, G. R., Lee, J. P., Peck, H. D. Jr & LeGall, J. (1978) Biochimie

38. Lancaster, J. R. (1980) FEBS Lett. 115, 285-288. 39. Lancaster, J. R. (1982) Science (Wush. DC) 216, 1324-1325. 40. Moura, J. J. G., Moura, I., Huynh, B. H., Kruger, H.-J., Teixeira,

M., DuVarney, R. C., DerVartanian, D. V., Xavier, A. V., Peck, H. D. Jr & LeGall, J. (2982) Biochem. Biophys. Res. Commun.

41. Moura, J. J. G., Teixeira, M., LeGall, J., Moura, I. &Xavier, A. V.

42. Olive, H. & Olive, W. S. (1975/76) J. Mol. Catal. 6, 307-310. 43. Schneider, K., Cammack, R., Schlegel, H. G. &Hall, D. 0. (1974)

44. Cammack, R. (1975) Biochem. Soc. Trans. 3, 482. 45. Cammack, R., Dickson, D. P. E. & Johnson, C. E. (1977) in Zron-

Sulfur Proteins (W. Lovenberg, ed.) pp. 283 -330, Academic Press, New York.

46. Moura, J. J. G., Moura, I., Kent, T. A,, Lipscomb, J. D., Huynh, B. H., LeGall, J., Xavier, A. V. & Munck, E. (1982) J . Biol. Chem. 257, 6259 -6267.

(Pari.y) 60, 315 -320.

108, 1388-1393.

(1983) J . Mol. Catal., in the press.

Biochim. Biophys. Actu 578, 445 -461.

47. LeGall, J. (1963) J. Bucteviol. 86, 1120. 48. Lovechio, F., Gere, E. S. & Busch, D. H. (1974) J. Am. Chem. SOC.

49. Mines, T. & Geiger, W. (1973) Inorg. Chem. 12, 1189-1191. 50. Lappin, A. G., Murray, C. K. & Margerum, D. W. (1978) Inorg.

51. Sakurai, T., Hongo, J. & Nakahara, A. (1980) Inorg. Chim. Acta

52. Moura, I., Moura, J. J. G., Santos, H., Xavier, A. V., Burch, G., Peck, H. D. Jr & LeGall, J. (1982) Biochim. Biophys. Acta 742,

96, 3109-3118.

Chem. 17, 1630-1634.

46, 205 -210.

184-190.

M. Teixeira, I. Moura, A. V. Xavier, and J. J. G. Moura, Centro de Quimica Estrutural, Complexo Interdisciplinar, Instituto Supkrior Ttcnico, Avenida Rovisco Pais, P-1000 Lisboa, Portugal G. Fauque and P. A. Lespinat, Laboratoire Commun d’Enzymologie du Centre National de la Recherche Scientifique-Commissariat B 1’Energie Atomique, Dkpartement de Biologie, Centre $Etudes Nucleaires, Cadarrache, Boite postale 1, F-13115 Saint Paul Lez Durance, France

D. V. dervartanian, H. D. Peck Jr. and J. LeGall, Department of Biochemistry, University of Georgia, Boyd Graduate Studies Research Center, Athens, Georgia, USA 30602