Redox properties and activity studieson a nickel-containing hydrogenase isolatedfrom a halophilic sulfate reducer Desulfovibrio salexigens
M. TEIXElRA*, 1. MOURA*, G. FAUQUE**, M. CZECHOWSKI**, Y. BERLIER**,P.A. LESPINAT**, J. LE GALL**, A.V. XAVIER* and J.J.G. MOURA*o
75
* Centro de Quimica Estrutural and UNL, Complexo I, 1ST, Av. Rovisco Pais, 1000 Lisbon, Portugal** ARBS, Equipe Commune d'Enzymologie CNRS-CEA, CEN Cadarache, 13108 Saint-Paul-lez-Durance,
France
(Received 12-7-1985. accepted 12-9-1985)
Summary - A soluble hydrogenase from the halophilic sulfate reducing bacterium Desulfovibriosalexigens, strain British Guiana (NCIB 8403) has been purified to apparent homogeneity with a finalspecific activity of 760 umoles H2 evolved/min/mg (an overall 180-fold purification with 20 % recoveryyield). The enzyme is composed of two non-identical subunits of molecular masses 62 and 36 kDa,respectively, and contains approximately 1 Ni, 12-15 Fe and 1 Se atoms/mole. The hydrogenase showsa visible absorption spectrum typical of an iron-sulfur containing protein (~oo/A280 = 0.275) and a molarabsorbance of 54 mM- 1 cm- l at 400 nm.
In the native state (as isolated, under aerobic conditions), the enzyme is almost EPR silent at 100 Kand below. However, upon reduction under H2 atmosphere a rhombic EPR signal develops at g-values2.22, 2.16 and around 2.0, which is optimally detected at 40 K. This EPR signal is reminiscent of thenickel signal C (g-values 2.19, 2.16 and 2.02) observed in intermediate redox states of the wellcharacterized D. gigas nickel containing hydrogenase and assigned to nickel by 61Ni isotopic substitution(J.J.G. Moura, M. Teixeira, I. Moura, A.V. Xavier and J. LeGal! (1984), J. Mol. Cat., 23, 305-314). Uponlonger incubation with H2 the "2.22" EPR signal decreases. During the course of a redox titration underH2 , this EPR signal attains a maximal intensity around - 380 mV. At redox states where this "2.22" signaldevelops (or at lower redox potentials), low temperature studies (below 10 K) reveals the presence ofother EPR species with g-values at 2.23, 2.21, 2.14 with broad components at higher fields. This newsignal (fast relaxing) exhibits a different microwave power dependence from that of the "2.22" signal,which readily saturates with microwave power (slow relaxing). Also at low temperature (8 K) typicalreduced iron-sulfur EPR signals are concomitantly observed with gmed - 1.94. The catalytic propertiesof the enzyme were also followed by substrate isotopic exchange DdH+ and H2 production measurements.
o To whom correspondence should be addressed.
Abbreviations .'EPR : electron paramagnetic resonance.
76 M. Teixeira and coil.
The general properties of the D. salexigens hydrogenase are compared with those of [NiFe]hydrogenases isolated from other sulfate reducers from the genus Desulfovibrio,
nickel I iron-sulfur clusters / hydrogenase / Desulfovlbrio sp. / oxidation-reduction I EPR
Resume - Une hydrogenase soluble a ete purifiee jusqu 'a homogeneite apparente a partir de la bacteriesulfato-reductrice Desulfovibrio salexigens souche British Guiana (NCIB 8403). L 'activite specifique finaleetait de 760 umoles Hz produttrmin/mg, soit une purification totale de 180 fois et un rendement de 20 %.L'enzyme est composee de deux sous-unites de masses moleculaires respectives 62 et 36 kDa et elle contientapproximativement 1 atome de Ni, 12 a 15 atomes de Fe et 1 atome de Se par mole. L'hydrogenase presenteun spectre d'absorption dans Ie visible typique d'une proteine a centres fer-soufre (A400/Az80 = 0.275) avecun coefficient d'extinction molaire de 54 mM- 1em:' a 400 nm.
A l'etat natif (sous conditions aerobies), l'enzyme ne presente pratiquement pas de signaux RPE atemperature egale ou inferieure a 100 K.
Cependant, apres reduction sous atmosphere d'hydrogene, un signal RPE rhombique avec des valeursg de 2.22, 2.16 et environ 2.0 peut etre detecte preferentiellement a 40 K. Ce signal rappe/Ie celui du nickelC (valeurs g de 2.19, 2.16 et 2.02) observe avec l'hydrogenase a nickel de D, gigas dans des etatsintermediaires d'oxydo-reduction. II a ete attribue a ce metal par substitution isotopique avec 61Ni(J.J.G. Moura, M. Teixeira, I. Moura, A. V. Xavier and J. Le Gall (1984), J. Mol. Cat., 23, 305-314).
Apres une plus longue incubation sous Hz, Ie signal "2.22" diminue. Par titration redox en presencede Hz, l'in tensile maximale est atteinte a environ - 380 mV. Pour un potentiel ega I au inferieur a cettederniere valeur mais atemperature inferieure a10 K, apparaissent d'autres signaux RPE (g 2.23, 2.21, 2.14)et des composantes larges aux champs magnetiques plus eleves. Ce nouveau signal (a relaxation rapide)presente une dependance vis-a-vis de la puissance microonde differente de celie du signal "2.22" qui, lui,sature rapidement avec celle-ci (relaxation lente). A basse temperature egalement (8 K) des signaux RPEtypiques des centres fer-soufre reduits sont observes avec g moyen de 1.94 environ. Les proprietes catalytiquesde l'enzyme ont ete quant a elles suivies par la reaction d'echange deuterium-proton et par la mesure de laproduction d'hydrogene.
Les proprietes generales de l'hydrogenase de D. salexigens sont companies avec celles d'autreshydrogenases a nickel isolees a partir de differentes souches de bacteries sulfato-reductrices du genreDesulfovibrio.
Several bacterial systems use the enzyme hydrogenase in order to metabolize the simplest'molecule, H2 • During the metabolic energy-yielding process, H2 is either oxidized or evolved asthe product of reduction of protons, the reactionbeing expressed as H2+±2e-+2H+. In the forward reaction, H2 serves as an electron donor andthe reaction initiates an energy yielding process.In the backward reaction, the proton serves asone of the terminal electron acceptors in anaerobic metabolism [1,2J.
Although the importance of its biologicalfunction has been recognized a long time ago,only recently have the structural features andphysico-chemical properties of its prostheticgroups begun to be understood, mainly through
the application of low temperature EPR studiescomplemented with Mossbauer spectroscopyusing metal isotopic substitutions ( 61Ni and s7Fe)[3,4,5].
Hydrogenases are generally recognized asiron-sulfur containing proteins, with four totwelve iron atoms in different cluster arrangements : [2Fe-2S], [3Fe-xS] and [4Fe-4S].
In the last few years, through physiological,chemical and spectroscopic studies (mainly EPR),nickel was found to be a constituent of severalhydrogenases.
The metabolism of molecular H2 has figuredcentrally in the development of our presentconcepts regarding the biochemistry and physiology of respiratory sulfate reduction carried out byDesulfovibrio sp. and the hydrogenase system hasbeen extensively studied [6].
D. salexigens hydrogenase 77
Generally, the hydrogenases isolated from thisbacterial group have been found to be confinedto the periplasmic space, but membrane boundand cytoplasm located enzymes have also beenreported.
The results so far obtained enabled, within theDesulfovibrio genus, two types of hydrogenases tobe distinguished : one type containing onlyiron-sulfur clusters (termed [Fe] hydrogenases)and the other the nickel-iron-sulfur containinghydrogenases (termed [NiFe] hydrogenases).
Within this bacterial group, the enzyme showsa large diversity with respect to structural features(e.g., the presence of subunit structure), catalyticproperties (Hz evolution versus Hz uptake andDz/H + exchange activities), activation step requirements in order to express full activity, as wellas sensitivity to thermal denaturation, unfoldingagents, and CO.
The [Fe] hydrogenase type has been mostextensively studied in D. vulgaris (Hildenborough) [7, 8] and the D. gigas enzyme has beenconsidered the prototype of the [NiFe] hydrogenases [9-14].
A tentative catalytic and activation scheme hasalready been proposed, showing the involvementof all the redox centers in the simple electrontransfer process (2H + +2e - +=t Hz) carried out bythis complex enzyme [13]. The improvement ofthis scheme and the full understanding of thebehaviour of this class of enzymes prompted usto characterize [NiFe] and [Fe] hydrogenases fromother Desulfovibrio strains.
In this paper, we describe the biochemicalcharacterization, redox properties (monitored byelectron paramagnetic spectroscopic measurements), and some relevant catalytic properties ofa [NiFe] hydrogenase isolated from Desulfovibriosalexigens strain British Guiana (NCIB 8403) .This desulfovibrione is the only well knownhalophilic strain within the genus. A few electrontransfer proteins have been previously isolatedfrom D. salexigens : a cytochrome C3 (M, 13000)[15], a fla vodoxin and a rubredoxin [16], as wellas desulfoviridin and a blue protein containingmolybdenum and iron-sulfur centers (our unpublished data).
Material and Methods
All chemicals and reagents were of the highestpurity available.
Assays
Hydrogenase activity was assayed by the rate of H2evolution with sodium dithionite (IS mM) as electrondonor and methylviologen (I mM) as redox mediator[17], at 30°C and pH 7.6. Hydrogen evolved wasdetermined by gas chromatography using an AerographA-90 P3 chromatograph.
The D2/H"1- exchange reaction was performed aspreviously described [18] using a mass spectrometer(VGS-80 equipped with an Apple II data acquisitionsystem).
Total iron was determined by the 2,4,6-tripyridyl-S-I,3,5-triazine (TPTZ) method [19]. Metals werealso screened and quantified by plasma emissionspectroscopy using a Jarrell-Ash model 750Atomcomp.
Protein was determined by Lowry's method [201,using a bovine serum albumin standard solution purchased from Sigma.
The homogeneity of the preparations was checkedon 7 % polyacrylamide gel electrophoresis at pH 8.0[21]. The subunit structure was determined on 8DSpolyacrylamide gel electrophoresis [22], using the following molecular mass markers (Da) : phosphorylaseb (94000), bovine serum albumin (67000), ovalbumin(43000), carbonic anhydrase (30 000), chymotrypsinogen (25 000), soybean trypsin (20000)and lactoalbumin(14400).
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 Nicolett ll80 Computer, on which mathematical manipulations wereperformed. The visible/ultraviolet spectra were obtained on a Shimadzu model 260.
Oxidation-reduction potentiometric titrations
Oxidation-reduction titrations were carried out in anapparatus similar to that described by Dutton [23],equilibrating the enzyme under different partial pressures of hydrogen (using different proportions of argon+ hydrogen) at 30°C and pH 8.0 (100 mM Tris-H'Clbuffer), in the presence of the following oxidationreduction mediators at a final concentration of 50 11M :methylene blue (Eo= II mY); indigotetrasulphonate(Eo = - 46 mY); 2-hydroxy-I,4-naphthoquinone(Eo= -145 mY); anthraquinone-2-sulphonate (Eo=-225 mY) ; phenosafranine (Eo = -255 rriv); benzylviologen (Eo= - 345 mY); methylviologen (Eo= - 440mY); N.N-dimethyl-3-methyl-4,4-bipyridyl (Eo=-617 mY).
All redox potentials measured using a platinumlsaturated calomel electrode system are quoted relativeto the standard hydrogen electrode. The protein concentration in the titration vessel was 40 J.1M, as estimated by the molar absorbance coefficient. Typically,
78 M. Teixeira and coil.
the sample was first reduced under pure hydrogenatmosphere (I atm) and left to equilibrate. Samplereoxidation was accomplished varying the partialpressure of H2 gas, using the hydrogen and argonmixture. After equilibration at a fixed redox potential,a sample was transferred into an EPR tube under thetitration vessel pressure and immediately frozen at 77 Kfor further quantification.
Growth of the microorganisms and preparation ofcell crude extracts
D. sa/exigens strain British Guiana (NClB 8403) wasgrown at 37 DC on a standard lactate-sulfate medium[24] supplemented with 3 % sodium chloride. The cellswere then lysed and frozen until used. They wereslowly unfrozen and centrifuged at 20000 rpm for1.5 h. The supernatant from 250 g of cells (wet weight)was then centrifuged twice at 40 000 rpm for 2 h.
Purification of hydrogenase (Table I)
Scheme A
All purification procedures were carried out in airat 4 DC and the pH of the buffers (Tris-HCI andphosphate) was 7.6 (measured at 5°C). A summary ofthe purification steps is presented in Table 1.
The centrifuged extract was loaded onto a hydroxylapatite (Biorad) column (5 x 29 em) and the columnwashed with 500 ml of 0.2 M Tris-Hfll. A reversegradient of 500 ml of 0.2 M Tris-HCI to 500 ml of0.01 M Tris-HCI was applied. The column was thenwashed with 300 ml 0.01 M KPB and a phosphatelinear gradient of 0.01 up to 0.4 M phosphate buffer(1000 ml of each) was set up. No hydrogenase wasfound in the eluent. The column was further washedwith 500 ml 0.4 M phosphate buffer and the hydrogenase finally eluted from the column. About 80 % of thehydrogenase activity (in the H2 evolution) was recovered. The hydrogenase enzyme was then concentrated to8 ml in a Diaflow apparatus using a YM 30 membrane.It was then dialyzed against 4500 ml 0.01 M Tris-HC!.The dialysis resulted in the formation of a precipitate
which redissolved in 1 M Tris-H'Cl, The hydrogenaseactivity was then redetermined in the supernatant andfound to have decreased by 50 %. The resuspendedpellet was checked for activity and very little wasfound. A spectrum of the resuspended pellet showedonly cytochromes.
The dialyzed protein was diluted 1: 4 with 0.01 MTris-HCl and then applied to a DEAE-Biogel Acolumn (5 x 34 em). The column was washed with200 ml of 0.01 M Tris-HCI and a linear gradient wasthen constructed (1000 ml of 0.01 M Tris-HCI to1000 ml of 0.3 M Tris-HCI).
The hydrogenase was colIected at a concentrationof about 0.25 M Trls-HCl. About 85 % of the hydrogenase was recovered in this step of purification. The~OO/A2so ratio was 0.275 and the specific activity was602 umoles H2 produced/min/mg protein. The yield ofpurification after this step was 30 %.
After concentration in a YM 30 membrane, thishydrogenase fraction was introduced on a LKB HPLCgel filtration column (TSK G 3000 SW) equilibratedwith 0.5 M phosphate buffer at pH 7.4. By this procedure, a pure hydrogenase fraction was obtained (asjudged by polyacrylamide gel electrophoresis), containing 14.5 mg of hydrogenase, with an absorbance ratio~OO/A2so of 0.275 and a specific activity of 758 unitsof H1 evolved.
Scheme B
Another purification scheme was outlined in orderto decrease the number of chromatographic steps. Bothschemes A and B yield homogeneous preparations withthe same level of specific activity.
After disruption of the cells in a French Press at62 MPa, the crude extract was centrifuged at 8000 rpmfor 2 h and the supernatant dialyzed against distilledwater for 24 h. The dialyzed solution was then appliedonto a DEAE.Biogel A column (6 x 34 ern) equilibratedwith 0.01 M Tris-HC!. After elution with a lineargradient of 0.01-0.5 M Tris-HCI (1000 ml of each), thehydrogenase activity was found in a fraction eluted atabout 0.3 M Tris-H'Cl, This fraction was then applied
TABLE I
Purification (Scheme A) of hydrogenase from D. salexigens (British Guiana).
Fractions
Crude extractHydroxylapatite columnDialysis and centrifugationDEAE-Bio-GelHPLC
Protein(mg)
12690nd210
2714.5
Total activity(pmoles H2/min)
5400042500195001630011000
Specific activity(umoles Hs/min/rng)
4.3nd93
602758
Exchange activity(umoles HD + H2/min/mg)
2nd28
175378
Hydrogenase activity measured by the hydrogenase evolution assay or the D,/H+ exchange reaction (see Materials and Methods).nd : not determined.
D. salexigens hydrogenase 79
2 3 4 Time(min)
EPR spectroscopy
The EPR spectrum of the native ("as isolated")enzyme is shown in Figure 3-A. It shows a veryweak signal centered in the g= 2.0 region, observable only at low temperature. No EPR signalsare observed when the sample is examined at100 K.
The ratio between the initial H2 and HD evolutionis higher than 1 and the sum of HD + H2 evolved islower than the Hz production from dithionite reducedmethylviologen (Table I).
05
1.5
FIG. 2. - (D1/H +) exchange activity 0/ O. salexigenshydrogenase.
The curves refer to: (0) O2 uptake; (e) HD transientevolution then uptake; (0) H2 production following masspeaks 4, 3 and 2, respectively. Protein concentration 0,6 nM,gas-phase 20 % 0, in N2•
Enzyme activity (H2 evolution and D2/H+ exchange)
The aerobically isolated D. salexigens hydrogenasedoes not require a reductive activation step in order tocatalyze the methylviologen mediated H+ reduction.The Hz evolution rate is practically constant from timezero and no lag phase period was observed. Theenzyme preparation used has a specific activity of 760units of Hz evolved.
The hydrogenase activity was also followed atpH 7.6 by the Dz/H+ exchange reaction using 20 % Dzin N2• The D2/H+ exchange kinetics mediated by thepurified hydrogenase from D. salexigensis presented inFigure 2, with the three curves corresponding to D2
uptake, HD transient evolution and then uptake, andHz production (mass-peaks 4, 3 and 2, respectively).
enzyme obtained by the purification schemesdescribed is therefore I Ni, 12-15 Fe, I Se andtrace amounts of Zn.
Purity, metal content and optical absorptionspectra
The purification procedure is summarized inTable 1. The hydrogenase was purified 180-fold,showing a final specific activity of 760 units in H 2
evolution. The overall recovery yield was 30 %.The enzyme is composed of two subunits ofmolecular masses 62 and 36 kDa.
The native state of the enzyme (aerobicallyisolated) shows a typical U.V.lvisible spectrum ofa non-heme iron protein with broad bands at 400and 280 nm, and an absorbance ratio ~OO/A280 of0.275 (Fig. I). The molar absorbance at 400 mn is54 mM- 1 em:".
Reduction of the enzyme under H 2 gas decreases the absorbance in the visible region by approximately 15 % (Fig. I).
Analysis of D. salexigens hydrogenase by the.chemical (TPTZ) method gave a value of 12± Ig-atms of iron per minimal molecular mass of98 kDa. Plasma emission analysis detected thefollowing metals in relevant amounts:1.03 g-atms of nickel, 15.15 g-atms of iron, 1.08g-atms of selenium and 0.163 g-atms of zinc perminimal molecular mass. The metal content of the
FIG. 1. - Optical absorption spectra 0/ O. salexigenshydrogenase.
Protein concentration 3.75 I-tM, at pH 7.6, 50 mM Tris-Helbuffer....... native enzyme; - • - - - H, reduced (flushed underH, for 1 h).
l!.JUZ<l:CD0:::o(J)
CD<!
Results
to a hydroxylapatite column, equilibrated with 0.4 MTris-HC!. A discontinuous gradient of 0.4-0.01 MTris-HCI was performed and then a continuous gradient of 0.001-0.6 M phosphate (500 ml of each) was setup. The hydrogenase fractions eluted at 0.5 M phosphate were concentrated on a Diana Amicon with aYM 30 membrane.
80 M. Teixeira and coil.
I
2log P
o
A /~\\., I, I, Q
"," I,\• .' 0 \
0,'0,0\
t
, ~\ ,
......~ \
n ~
B0.0',,,' ~, \
I \
I,
I,.
~' ~
I,
}l.,11'
C ,7'I
,r{C
I
,e, ,
.a
-!i'.P'
-.c·
J .1 I.
Ul
:1::c::;:J
.Q....ell
)0r-i1iZwrZ
...J<tz<9ena:::Q.W
-100 -200 -300 -400 -500E(mV}
0.51-
FIG. 4. - EPR redox titration curves of the EPR activespecies detected in the hydrogen reduced states ofD. salexigenshydrogenase.
EPR signal intensities (arbitrary units) of the EPR signalsdetected upon poising the enzyme under different H, partialpressures at 25°C, pH 8.5, in the presence of redox mediators,as described in Materials and Methods, at the followingg-values and temperatures: A) g=2.22 (0) and 2.16 (e), at20 K. B) g=2.23 at 4 K (o), C) g= 1.87 at 4 K (0).
FIG. 5. - Microwave power sail/ration curves oj D. salexigens hydrogenase EPR detectable signals under H2 atmosphere.
(ct) g=2.22, 4 K. (e) g=2.22. 20 K. (0) g=2.23, 4 K.
FIG. 3. - EPR spectra 0/ D. salexigens hydrogenase.
A) Native state ("as isolated") at II K; gain 3.2 x io',B) Intermediates H, reduced state at 25 K. The sample waspoised at -380 mV under H, atmosphere (see redox titrationconditions under Materials and Methods); gain 2 x lOs.C) Same as B, at 4 K; gain 2 x lOs. Other experimentalconditions: microwave power 2 mW, modulation amplitudeI mT, microwave frequency 9.45 GHz.
Upon exposure to different partial pressures ofhydrogen gas, in the presence of redox mediators,a rhombic EPR signal develops with g-values at2.22, 2.16 and around 2.0 (Fig. 3-B). During thecourse of the redox titration, using H 2 gas aselectron donor, the "2.22" EPR signal reaches amaximal intensity (Fig.4-A). This signal is optimally detected at 40 K and is readily saturated bymicrowave power at low temperature, beingcharacteristic of a slow relaxing species.
Long incubation of the enzyme under H2atmosphere decreases the intensity of this signal.When the enzyme is poised at redox potentialswhere the "2.22" signal attains maximal intensity(- - 380 mV, Fig. 4-A), studies at low temperature reveal the presence of other EPR activespecies. Below 10 K the "s» 2.22" signal startssaturating and a new set of signals at 2.23, 2.21,2.14, and broad features at higher field develop(Fig. 3-C). These latter signals exhibit a differentpower dependence from that of the previous EPRsignal ("g= 2.22"), as shown in Figure 5. Thecomplex set of signals only observable below10 K shows fast electronic relaxation properties.At redox potentials below -300 mY, the lowtemperature (8 K) EPR spectra also reveal thepresence of reduced iron-sulfur centers with g med
at 1.94. Temperature and microwave power dependence of this spectral region indicate that atleast two types of iron-sulfur centers are present(Fe/S center I (g min at 1.87) has faster relaxationproperties than Fe/S center II (g min at 1.90)). Thelow intensity of these signals prevents a detailed
D. salexigens hydrogenase 81
Discussion
electrons. Thus, the system has considerableadvantages for the study of the bioenergetics andthe physiology of H2 metabolism.
Hydrogenases isolated from Desulfovibrio sp.have recently been extensively studied. Table IIsummarizes the data on the localization, activity,metal center composition and relevant physicochemical data on hydrogenases isolated from thisgroup of bacteria. A common feature emerges:all the enzymes contain nickel as a relevantconstituent. So far, only D. vulgaris (Hildenborough) hydrogenase does not contain this transition metal [7].
D. salexigens hydrogenase is isolated from theonly well characterized halophilic sulfate-reducer.Its properties have many common features withthe group of the nickel containing hydrogenasesisolated from sulfate-reducing bacteria of thegenus Desulfovibrio. However, important differences are found within the group. D. salexigenshydrogenase is practically EPR silent, when isolated. The same is observed with the solubleD. desulfuricans (Norway 4) hydrogenase [25] andD. baculatus strain 9974 enzyme [26].
In contrast, D. desulfuricans (ATCC 27774) [4],membrane bound D. desulfuricans (Norway 4)[27], D. multispirans n.sp. [28] and D. gigas [10]hydrogenases exhibit a rhombic EPR signal withg-values around 2.3, 2.2 and 2.0. This signal hasbeen termed nickel signal A. These rhombic EPRsignals observed for the oxidized state of bacterialhydrogenases were assigned to Ni(lII) based onNi model compounds, relaxation properties (thesignal is observable at 100 K), EPR g-values, and61Ni isotopic substitutions (performed for D. gigas[3] and D. desulfuricans (ATCC 27774) [4]).Besides the rhombic signal, this group of hydrogenases also exhibits a strong isotropic signal atg "'" 2.02, observable below 30 K. This signal isassigned to a [3Fe-xS] center, based on Mossbauerspectroscopic studies using unenriched naturallyabundant and 57Fe enriched D. gigas and D. desulfuricans (ATCC 27774) hydrogenases [4,10].
The rhombic EPR nickel signal accounts for50-100 % of the chemically detectable nickel,depending on the preparation and source. The"2.02" isotropic signal integrates from 0.2 spins(D. multispirans hydrogenase) [28] up to 0.9 spins(D. gigas hydrogenase) [10]. Redox titrations haveonly been performed for the D. gigas hydrogenase[10,11]. The "2.02" signal titrates at -70 mY (atlO K) and the nickel signal A disappearence isassociated with a redox process which titrates at
(
Sulfate reducing bacteria carry out the reduction of sulfur containing compounds, an important biological set of reactions with relevance tothe biocycle of this element. Also Desulfovibrionesis the only bacterial group which clearly participates in inter-species hydrogen transfer, workingeither as Hy-producing or as H2-utilizing microorganisms [6].
The bidirectional hydrogenase functions as an"energy valve", supplementing or disposing of
FIG. 6. - EPR spectra of D, salexigens hydrogenase poisedat -450 mV under H, atmosphere (see redox titration conditions under Materials and Methods).
A) Temperature 4 K. B) Temperature II K. C) Temperature 22 K. Gain loj. Other experimental conditions as incaption of Figure 3. The top spectrum (insert) represents anidentical spectrum of D. gigas hydrogenase, under the sameexperimental conditions as D. salexigens hydrogenase spectrum A.
analysis of their spectral features. Figure 6 showsa temperature dependence of a D. salexigenshydrogenase sample poised at approximately- 450 mY, a redox stage where the "2.22" EPRsignal has decreased in intensity (Fig. 4-A). Lowtemperature reveals the presence of the "2.23"signal features (Fig. 6). The different relaxationproperties for the two EPR signals detected inintermediate redox stages of the enzyme enablea complete description of the intensity profiles tobe obtained by measuring the spectra at 20 and4 K, respectively, as shown in Figure 4. Thisfigure also indicates the development of the EPRfeatures associated with the FelS clusters.
c
82 M. Teixeira and coli.
TABLE II
Comparison of physico-chemical properties of Desulfovibrio sp. hydrogenases.
D. salexigens
PropertyD. I'II/garis (strain British D. gigas D. deslI!jllricalls D. bacuianu D. deslI!jur!calls D. deslIlfllricalls D. multispirans
Localiza tion periplasm peri plasm periplasm membrane" periplasm" periplasm NR* cytoplasmMolecul arweight 49000 98000 89500 58000 100000 52000 77 600 82500Subunits I 2 2 1 1 I 2 2Nickel 0 1 I +(EPR) I NR 1 ISelenium NR 1 0 I I NR 0 0Non-heme iron 12 12-15 11 6 12 12 11[Fe1S]] 0 NR 1 "g=2.02" NR NR I I[Fe,S,] 3 +(probably 2) 2 NR +(probably 2) NR 2 +(probably 2)Specific activity(umolesHs/min/mg)
Evolution 4600 760 420 70 527 9000 152 790Consum ption 50000 NR 1200 200 NR NR NR 590References [7,8] This work [9,10,12] [27] [261 1341 141 1281
* non reported.(a) A soluble form was also purified [25].(b) A cytoplasmic and a membrane-bound form were also purified [26].
-220 mV, and was shown to be pH-dependent(60 mV/pH unit) [11). After the first sequence ofreductive events, an EPR silent state is attained.Evidence was previously accumulated [13) suggesting that in this EPR silent state one [Fe4S4)+1cluster (S = 1/2) is present and coupled to theNi(III) center. This proposal implies that the- 220 mV redox transition represents the midpoint redox potential of the iron-sulfur center. Inthis context, D. salexigens, D. baculatus strain9974 and soluble D. desulfuricans (Norway 4)hydrogenases would have been isolated in thisspin coupled state. The low intensity of the NiEPR signals in the native preparations or even itsabsence [26,28], could be due to spin couplingbetween the Ni center and the Fe/S cluster.
The pattern observed for the reductive eventsfollowing the EPR silent state are now commonto all Desulfovibrio [NiFe) hydrogenases studied.The appearence of a transient rhombic signal(termed nickel signal C) detected in D. gigashydrogenase with g-values 2.19,2.16 and 2.0 [13)is also observed in D. salexigens, D. baculatusstrain 9974 [26), D. desulfuricans (ATCC 27774) [4)and D. multispirans [28] enzymes. In all thesecases, EPR studies reveal the presence of otherEPR active species at redox stages where the
transient is observed. Due to its relaxation properties (fast relaxing) this last signal is onlyobservable at low temperature and with highlevels of microwave power [13).
The redox titration data obtained for D. salexigens hydrogenase, under H2 atmosphere, show avery similar behaviour to that of the D. gigasenzyme [13). The transient nickel signal developsto maximal intensity at - - 380 mV. Also, thestudy of the development of this signal followedat two temperatures (20 K and 4 K) clearly showsthat the two species are not directly correlated.
Below - 450 mV, the slow relaxing speciesdisappears (not observable at 20 K) but thecomplex fast relaxing species is still detected. Asimilar study recently conducted in the D. gigashydrogenase fully supports this data analysis (ourunpublished results).
An important point to consider in the reactional mechanism, is that some of the so-called"oxygen stable" [NiFe) hydrogenases are not fullyactive when isolated. This state is EPR active(rhombic nickel signal A and isotropic g = "2.02"signal). A main conclusion is that these EPRactive species are not relevant for the mechanism.The enzyme must go through an activation process that represents a complex phenomenon:
O. salexigens hydrogenase 83
removal of oxygen (lag phase) followed by areductive step [18, 19]. It is important to note thatthe enzymes which are EPR silent as isolated, donot show a lag phase for activation. The D. salexigens and D. baculatus hydrogenases have aconstant rate of H2 evolution, but for the D. gigasand D. multispirans enzymes an activation step(lag phase) is required in order to express fullactivity.
Upon reduction (either by long exposure to Hzatmosphere or by chemical reduction with excessdithionite) gmed - 1.94 EPR signals have beenobserved for D. gigas [13], D. desulfuricans (ATCC27774) [4J, soluble D. desulfuricans (Norway 4)[25], and D. baculatus strain 9974 hydrogenases[26], as well as for the D. salexigens enzyme.
All these pieces of information have beendiscussed in general terms as a basis for a"working hypothesis" which represents a usefulframework for discussing the mechanistic involvement of these redox species [13].
The analysis of the structural and physicochemical properties of the redox centers of thesehomologous hydrogenases will enable a wealth ofinformation to be built up, useful to delineate ageneral approach to the mechanism of enzymeaction.
Another important piece of information thathas not been fully explored in the study ofbacterial hydrogenases is the mechanism of activation of the hydrogen molecule by the enzymeactive centers, which can be directly probed byD2/H+ exchange experiments. Activity measurements are generally only related with the measurement of the overall evolution or consumptionof Hz. However, when these data are correlatedwith the exchange activities (Dz/H+) of thereactional center it is possible to probe theoperating mechanisms, i.e. homolytic or heterolytic cleavage [30,31].
According to the ratio found for the initial HDand Hz evolution in Dz/H+ exchange reactions,the hydrogenases isolated from the Desulfovibriogenus can be divided into two classes. i) One classis represented by the hydrogenases from D. salexigens and D. baculatus strain 9974 (our unpublished results) which have a H2/HD ratio. higherthan 1 (ratio H2/(HD + Hz) around 0.6). ii) Another class is represented by the hydrogenases from D. gigas and D. multispirans (our unpublished data) which have Hz/HD ratios lower than1 (ratio Hz/(HD + H2) around 0.3). The solublehydrogenase from Methanosarcina barkeri alsohas a Hz/HD ratio lower than I [32].
These H2/HD ratios are generally used todifferentiate between a heterolytic versus a homolytic cleavage of the hydrogen molecule. In similar experiments with metal salts, ratios of 0.95with platinum oxide and of 0.30 with rutheniumchloride were obtained (Y. Berlier, 0. Fauque,P.A. Lespinat and J. LeGall, unpublished results).These salts could serve as analogs for the homolytic and heterolytic activations of the hydrogenmolecule.
Another possible explanation for these differences is the kinetics of the Hz binding site of theenzymes with respect to the exchange with water.The general mechanism proposed for hydrogenase activity [30,31] involves the heterolyticactivation of hydrogen, with the formation ofa hydride:
Enzyme + Hz +± E-H- + B-H+
where E represents the hydride binding site andB the proton accepting site. The mechanism,based on the primary formation of HD ratherthan Oz in the Hz/O+ exchange reaction, indicates that only one of the bound atoms of Hz canfreely exchange with the protons of the medium.However, this may represent only a limitingsituation and it is possible that both sites exchange protons with water but with differentvelocities. Also, the lability of the H+ site may bemodulated by the pI<,. value of the proton acceptor site. In this context an extreme situationwithin the framework of the heterolytic mechanism would correspond to very different exchange .velocities of both sites, resulting in aHz/HD ratio lower than 1.
The different exchange kinetics of hydrogenbinding sites could reflect differences on theactive center of the hydrogenases, either at theproton or at the hydride binding site. Assumingthat nickel is the hydride binding center, as hasbeen proposed for several nickel containinghydrogenases [13], the observed difference couldreflect different ligation to this metal. In thisrespect, it is noteworthy that in the D. salexigensand the D. baculatus hydrogenases, the presenceof selenium was detected in a I: I 'ratio withnickel. These are the hydrogenases where theHz/HD ratio was found to be higher than 1.Selenium has also been found in Methanococcusvannielii [33J, soluble D. desulfuricans (Norway 4)[25] (and D. baculatus strain 9974 hydrogenases).It is not yet known whether selenium is presentat a new catalytic site. If this is so, it could berelated with the observed differences.
84 M. Teixeira and coli.
This explanation is also in better agreementwith the EPR studies of [NiFe] hydrogenases. Infact, not only do the hydrogen reduced states ofthe D. gigas, D. bacu/atus and D. salexigens hydrogenases show very similar EPR spectra, butthe redox patterns of the EPR spectra uponincubation under hydrogen are also identical,suggesting a common mechanism for the activation and reduction/oxidation of the hydrogenmolecule.
Acknowledgments
We thank M. Scandellari and R. Bourelli for growingthe bacteria. We are indebted to N. Galliano, I. Carvalho,B. Dimon and P. Carrier for their skillful technical helpand to M. Martinez and I. Ribeiro for carefully typing thismanuscript. This research was supported by grants fromInstituto Nacional de Investigaciio Cientifica and JuntaNacional de Investigacdo Cientifica e Tecnologica (Portugal), NATO 0341/83 and AID 936-5542 G-SS-4003-00(J.J.G. Moura), and National Science Foundation GrantDMB-84I5632 (J. LeGall). Part of this work has beenmade possible thanks to a special collaborative agreementbetween the University ofGeorgia (USA), Centre Nationalde la Recherche Scientifique and Commissariat Iil'Energie Atomique (France).
2. Schlegel, H.G., & Schneider, K. (1978) in : Hydrogenases : their catalytic activity, structure andfunction. (Schlegel, H.G. and Schneider, K., eds)pp. 15-44, E. Goltze, Gottingen,
3. Moura, J.J.G., Moura, I., Huyhn, B.H., Kruger,H.I., Teixeira, M., DuVarney, R.C., DerVartanian, D.V., Xavier, A.V., Peck, RD., Jr., &LeGall, J. (1982) Biochem. Biophys.Res. Commun., 108, 1388-1393.
4. Kruger, H.J., Huyhn, B.H., Ljungdahl, P.O., Xavier, AV., DerVartanian, D.V., Moura, I., Peck.R.D., Jr., Teixeira, M., Moura, J.J.G. & LeGall,J. (1982) J. Bioi. Chem., 257, 14620-14622.
