Tài liệu Báo cáo khoa học: Purification and characterization of a membrane-bound enzyme complex from the sulfate-reducing archaeon Archaeoglobus fulgidus related to heterodisulfide reductase from methanogenic archaea pdf

Purification and characterization of a membrane-bound enzyme complex from the sulfate-reducing archaeon Archaeoglobus fulgidus related to heterodisulfide reductase from methanogenic archaea Gerd J. Mander 1 , Evert C. Duin 1 , Dietmar Linder 2 , Karl O. Stetter 3 and Reiner Hedderich 1 1 Max-Planck-Institut fu ¨ r terrestrische Mikrobiologie, Marburg, Germany; 2 Biochemisches Institut, Fachbereich Humanmedizin, Justus-Liebig-Universita ¨ t Giessen, Germany; 3 Lehrstuhl fu ¨ r Mikrobiologie und Archaeenzentrum, Universita ¨ t Regensburg, Germany Heterodisulfide reductase (Hdr) is a unique disulfide reduc- tase that plays a key role in the energy metabolism of methanogenic archaea. The genome of the sulfate-reducing archaeon Archaeoglobus f ulgidus encodes s everal proteins of unknown function with high sequence similarity to the catalytic subunit of Hdr. Here we report on the purification of a multisubunit membrane-bound enzyme complex from A. fulgidus that contains a subunit related to the catalytic subunit of Hdr. The purified enzyme is a heme/iron-sulfur protein, as deduced by UV/Vis spectroscopy, EPR spec- troscopy, and the primary structure. It is composed of four different subunits encoded by a putative transcription unit (AF499, AF501–AF503). A fifth protein (AF500) encoded by this transcription unit could not be detected in the purified enzyme preparation. S ubunit A F502 is clo sely r elated to the catalytic subunit HdrD of Hdr from Methanosarcina bark- eri. AF501 encodes a membrane-integral cytochrome, and AF500 e ncodes a second integral m embrane protein. AF499 encodes an e xtracytoplasmic iron-sulfur protei n, and A F503 encodes an extracytoplasmic c- type cytochrome with th ree heme c-binding motifs. All of the subunits show high sequence similarity to proteins encoded by the dsr lo cu s of Allochromatium vinosum and to subunits of the Hmc complex from Desulfovibrio v ulgaris. The heme groups of the enzyme are rapidly reduced by redu ced 2,3-dimethyl-1,4- naphthoquinone (DMNH 2 ), which indicates that the enzyme functions as a menaquinol–acceptor oxidoreduc- tase. The physiological electron acceptor has not yet been identified. Redox t itrations monitored by EPR spectroscopy were carried out to characterize the iron-sulfur cl usters of the enzyme. In addition to EPR signals due to [4Fe-4S] + clus- ters, signals of an unusual p aramagnetic species with g values of 2.031, 1.994, and 1.951 were obtained. The paramagnetic species could be reduced in a one-electron transfer reaction, but could not be further oxidized, and shows EPR properties similar to t hose of a paramagnetic species r ecently identified in Hdr. In Hdr this paramagnetic species is specifically induced b y the substrates of the e nzyme a nd is thought to be an inte rmediate of the catalytic cycle. Hence, Hdr and the A. fulgidus enzyme not only share sequence similarity, but may also have a similar active site and a similar catalytic function. Keywords: Archaeoglobus f ulgidus; heterodisulfide reductase; Hmc complex; iron-sulfur proteins; sulfate-reducing bacteria. Heterodisulfide reductase (Hdr) is a key enzyme in the energy metabolism of methanogenic archaea. In the final step of methanogenesis, the mixed disulfide of the metha- nogenic thiol coenzymes coenzyme M and coenzyme B is generated in a reaction catalyzed by methyl-coenz yme M reductase [1]. This disulfide is reduced by a unique disulfide reductase, designated heterodisulfide reductase (Hdr). Two types of Hdr from phylo genetically distantly related meth- anogens have been identified a nd characterized. Neither type of enzyme belongs to t he family of pyridine nucleotide disulfide oxidoreductases [2]. Hdr from Methanothermobacter marburgensis is an iron-sulfur flavoprotein composed of the subunits HdrA, HdrB, and HdrC. The enzyme has been purified from the soluble fraction, and none of its subunits are predicted to form transmembran e helices. From sequence data, it has been deduced that HdrA contains an FAD-binding motif and four binding motifs for [4Fe-4S] clusters. HdrC contains two additional binding motifs for [4Fe-4S] clusters [2]. Hdr in t he two closely relate d Methanosarcina sp ecies M. barkeri and M. thermophila is tightly membrane- bound [3–5]. The enzyme is composed of two s ubunits, a membrane-bound b-type cytochrome (HdrE) and a hydrophilic subunit (HdrD) containing two binding motifs for [4Fe-4S] clusters. Subunit HdrD of the M. barkeri enzyme is a homologue of a hypothetical fusion protein of the M. marburgensis HdrCB subunits Correspondence to R. Hedderich, Max-Planck-Institut fu ¨ r terrestrische Mikrobiologie, Karl-von-Frisch-Strabe, D-35043 Mar burg/Germany. Fax: + 49 6421 178299, Tel.: + 49 6421 178230, E-mail: hedderic@mailer.uni-marburg.de Abbreviations: Hme, Hdr-like menaquinol-oxidizing enzyme; Hdr, heterodisulfide reductase; DMN, 2 ,3-dimethyl-1,4-naphthoquinone; H-S-CoM, coenzyme M or 2-mercaptoethanesulfonate; H-S-CoB, coenzyme B or 7-mercaptoheptanoylthreonine phosphate; CoM-S-S-CoB, heterodisulfide of H-S-CoM and H-S-CoB; Hmc, high-molecular-mass c-type cytochrome; Dsr, dissimilator y sulfite reductase. Enzyme: heterodisulfide reductase (EC 1.99.4 ). (Received 10 October 2001, revised 12 February 2002, accepted 15 February 2002) Eur. J. Biochem. 269, 1895–1904 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.02839.x [4]. A homologue of the M. marburge nsis HdrA subunit is lacking in H dr from Methanosarcina species. It has therefore been suggested that the conserved subunits HdrD and HdrCB must harbor the catalytic site for the reduction of the disulfide substrate. The active site of Hdr was recently shown to contain a [4Fe-4S] cluster that is directly involved in mediating heterodisulfide reduction [6,7]. This extra iron-sulfur cluster has been proposed to be co-ordinated by cysteine residues of the highly conserved sequence motif CX 31)38 CCX 33)34 CXXC found in subunits HdrD and HdrB. The nonconserved subunits HdrE and HdrA are thought to interact with the physio- logical electron donor, which differs in t he two types of Hdr. The physiological electron donor of Hdr from Methanosarcina species is thought to be the membrane- soluble electron c arrier methanophenazine [8]. Hdr from M. marburgensis forms a functional complex with the MvhAGD hydrogenase [9]. This complex catalyzes the reduction of CoM-S-S-CoB by H 2 . Hdr w as originally thought to be u nique to methanogenic archaea. However, in recent years, genes encoding pro- teins r elated to the catalytic subunit of Hdr have been identi- fied in a b road range of prokaryotes unable t o perform methanogenesis [2]. No function has so far been assigned to these Hdr-like proteins, and none has been purified and characterized. Archaeoglobus fulgidus is one of the organ- isms that en code the largest number of proteins related to Hdr [10]. This extremely thermophilic sulfate-reducing archaeon completely oxidizes organic substrates, such as lactate, t o CO 2 [11]. Acetate is oxidized to CO 2 by a modified acetyl- CoA pathway using typical methanogenic coenzymes [12,13]. Some of the reducing equivalents generated in the oxidative branch of the pathway are transferred to the deazaflavin coenzyme F 420 , which is reoxidized by the F 420 H 2 –menaquinone oxidoreductase. F 420 H 2 –menaqui- none oxidoreductase is an integral membrane protein that shows high sequence similarity to energy-conserving NADH–quinone oxidoreductases [10,14]. It is assumed to function as a proton or sodium ion pu mp as well. In addition, the m embrane fraction o f A. fulgidus catalyzes the reduction of 2,3-dimethyl-1,4-naphthoquinone (DMN) by L -lactate, which indicates that lactate dehydrogenase direct- ly channels the reducing equivalents generated in the oxidation of lactate to pyruvate into the menaquinone pool [12]. A. fulgidus has been shown t o contain a modified menaquinone as a m embrane-soluble e lectron c arrier [15]. It is, however, not yet known how the reduced menaquinone pool is electrically connected to the enzymes of sulfate reduction, namely adenosine 5 ¢-phosphosulfate reductase and sulfite reductase. Here we report on the isolation and characterization of a heme-containing membrane protein from A. fulgidus related to Hdrfrom M. barkeri. Afunction ofthis enzymeas reduced menaquinone–acceptor oxidoreductase is discussed. MATERIALS AND METHODS Materials Redox dyes were obtained from Aldrich–Sigma. DMN w as from Sigma. Potassium trithionate was a gift from Peter M. H. Kroneck (Universita ¨ t K onstanz). All other chemicals were from Merck. The chromatographic materials were from Amersham Pharmacia Biotech. Growth of the organism A. fulgidus (VC16, DSMZ 304) was grown in a 300-L fermenter at 83 °C on lactate/sulfate medium as described previously [11]. Cells were harvested after shock cooling to 4 °C w ith a continuous flow centrifuge (Z61; Padberg Lahr, Germany) at 17 000 g; the pellet was frozen in liquid nitrogen and stored at )80 °Cbeforeuse. Enzyme purification All purification steps were carried out under strictly anoxic conditions under an atmosphere of N 2 /H 2 (95 : 5, v/v) at 18 °C. Cells were lysed by sonication and then centrifuged at 6400 g for 1 h. The supernatant was ultracentrifuged at 150 000 g for 2 h. The pellet was resuspended in 50 m M Mops (pH 7.0) using a Teflon homogenizer. Protein was solubilized from the membranes with 15 m M dodecyl-b- D - maltoside [ 2 mg dodecyl-b- D -maltosideÆ(mg protein) )1 ]at 4 °C for 12 h. The unsolubiliz ed proteins and the mem- branes were removed by u ltracentrifugation as described above. The supernatant was applied to a Q-Sepharose HighLoad column (2.6 · 10 cm) equilibrated with 50 m M Mops/KOH (pH 7.0) containing 2 m M dodecyl-b- D -malto- side (buffer A). Protein was eluted in a s tepwise NaCl gradient (80 mL each in buffer A): 0 m M , 300 m M , 400 m M ,500m M ,600m M ,and1 M . The majority of the heme-containing protein(s) were eluted at 600 m M NaCl. These f ractions were applied to a Superdex 200 gel-filtration column (2.6 · 60 cm) equilibrated w ith buffer A containing 100 m M NaCl. Protein was eluted using the same buffer. Heme-containing protein(s) were eluted after 120 mL (peak maximum). These fractions were applied to a Mono Q anion-exchange column (HR 10/10) equilibrated with buffer A. Protein was eluted using a linear NaCl gradient (0–1 M , 100 mL). Heme-containing protein(s) were eluted at 600 m M NaCl. The enzyme was concentrated by ultrafiltration (Molecular/Por ultrafiltration membranes; 100-kDa cut off; Spectrum, Houston, USA) and stored in buffer A at 4 °C und er N 2 . P rotein was judged to be >95% pure by SDS/PAGE. UV/Vis spectroscopy Spectra of samples in 1-mL quartz cuvettes in a n anaerobic chamber under N 2 /H 2 (95 : 5, v/v) were recorded using a Zeiss Specord S10 diode array spectrophotometer connected to a quartz photoconductor ( Hellma Mu ¨ llheim, Germany). The oxidation or reduction of the heme groups of the enzyme by DMN or DMNH 2 were followed spectropho- tometrically. DMN or DMNH 2 was added to the enzyme solution [1 mgproteinÆmL )1 in 50 m M Mops/KOH (pH 7.0)] to a final concentration of 150 l M , and spectra were recorded every 5 s. DMNH 2 was prepared a s described previously [16]. Analytical methods Non-heme iron was quantified colorimetrically with neo- cuproin (2,9-dimethyl-1,10-phenanthroline) and ferrozine 1896 G. J. Mander et al.(Eur. J. Biochem. 269) Ó FEBS 2002 [3-(2-pyridyl)-5,6-bis-(4-phenylsulfonate)-1,2,4-triazine] as described by Fish [17]. Acid-labile sulfur was analyzed as methyleneblue[18]. The protein concentration was routinely determined by the method of Bradford (Rotinanoquant; Roth Karlsruhe, Germany) using BSA as standard. Heme was extracted with acetone/HCl and the pyridine hemochrome derivate was formed as described. Reduced minus oxidize d differe nce spectra were recorded at room temperature [19]. The spectra obtained were compared with the pyridine hemochrome spectra obtained with heme extracted from hemoglobin. EPR spectroscopy measurements EPR spectra at X-band (9 GHz) were obtained with a Bruker EMX spectrometer. All spectra were recorded with a field modulation frequency of 100 kHz. The sample was cooled by an Oxford Instruments ESR 900 flow cryostat with an ITC4 temperature controller. Spin quantitations were carried out under nonsaturating conditions using 10 m M copper perchlorate as the standard (10 m M CuSO 4 , 2 M NaClO 4 ,10m M HCl). When the EPR signals over- lapped with o ther signals, e.g. radical signals from the redox dyes, the signals were simulated, and the simulations were double integrated t o obtain the spin intensity. Temperature- dependence studies were carried out under nonsaturating conditions where possible. For all signals, the peak ampli- tude was measured at different temperatures. These values were used to obtain Curie plots describing the temperature behavior of the respective signal. EPR signals were simu- lated using nonco mmercial programs based on formulas described previously [20]. Redox titrations Redox titrations were carried out at 18 °Cinananaerobic chamber under N 2 /H 2 (95 : 5, v/v). Potentials w ere a djusted with small amounts of freshly prepared sodium dithionite (20 m M stock solution) or freshly prepared potassium ferricyanide (20 m M stock solution). All redox potentials quoted here are relative t o the standard hydrogen electr ode. In these titrations, a selection of the following mediators (final concentration 20 l M ) were added individually to the enzyme solution: 1,2-naphthoquinone (E°¢ ¼ +134 mV), duroquinone (E ¢ ¼ +86 mV), 1,4-naphthoquinone (E ¢ ¼+69 mV), thionine (E ¢ ¼+64 mV), methylene blue (E ¢ ¼ +11 mV), indigodisulfonate (E ¢ ¼ )125 mV), 2-hydroxy-1,4-naphthoquinone (E ¢ ¼ )145 mV), anthra- quinone-1,4-disulfonate (E ¢ ¼ )170 mV), phenosafranin (E ¢ ¼ )252 mV), anthraquinone-2-sulfonate ( E°¢¼ )255 mV), safranin O (E ¢ ¼ )289 mV), and neutral red (E ¢ ¼ )325 mV). The final concentration of Hdr-like menaquinol-oxidizing enzyme (Hme) was 7 l M in 50 m M Mops/KOH (pH 7.0) containing 2 m M dodecyl-b- D -malto- side. After equilibration of the desired potential, a 0.3-mL aliquot was transferred to a calibrated EPR tube and immediately f rozen in liquid nitrogen. The redox potential wasmeasuredwithanAg/AgClredoxcombinationelec- trode (Mettler Toledo Giessen, Germany). To obtain potentials relative t o the standard hydrogen electrode, a value of 207 mV (corresponding to the potential of Ag/ AgCl at 25 °C) was added to t he measured redox potentials. Determination of amino-acid sequences For determination of N-terminal amino-acid sequences, polypeptides w ere separated by SDS/PAGE and b lotted on to poly(vinylidene difluoride) membranes (Applied Biosystems) as described previously [4]. Sequences were determined using an Applied Biosystems 4774 protein/ peptide sequencer and the protocol given by the m anu- facturer. Amino-acid sequence analysis For the prediction of transmembrane helices in proteins, noncommercial programs were u sed (http://www.sbc.su.se/ miklos/DAS/; http://www.cbs.dtu.dk/services/TMHMM- 2.0/). For sequence comparisons, multiple sequence align- ments were generated using the FASTA 3server(http:// www.ebi.ac.uk/fasta3/). RESULTS Purification of a heme-containing enzyme complex from the membrane fraction of A. fulgidus The genome of A. fulgidus encodes several membrane- bound oxidoreductases that share sequence similarity with subunits of Hdr from methanogenic archaea, in particular with the membrane-bound enzyme from M. barkeri [4,10], which is anchored in the cytoplasmic membrane via a b- type cytochrome [3]. We used this knowledge to identify and purify heme-containing membrane-bound enzymes from A. fulgidus cells cultivated on lactate/sulfate medium by following the characteristic absorption of heme proteins. The membrane fraction was isolated, and proteins were solubilized with the detergent dodecyl-b- D -maltoside. On anion-exchange chromatography on Q-Sepharose, the major heme-containing fraction was eluted at 600 m M NaCl. Approximately 70% of the heme presen t in solubi- lized membranes was found in this fraction. A further purification of the proteins in this heme-containing fraction by gel fi ltration on Superdex 200 resulted again in only one heme-containing fraction eluted after 120–150 mL. In the final purification step, the sample was chromatographed on a Mono Q anion-exchange column. The protein thus purified was subjected to SDS/PAGE (Fig. 1). Samples were either boiled for 5 min in SDS buffer or incubated in SDS buffer at room temperature for 1 h before electro- phoresis. The samples incubated at room temperature yielded four major polypeptide bands with apparent molecular masses of 53, 34, 31, and 16 kDa after SDS/ PAGE (Fig. 1, lane A1). In the boiled sample, the 34-kDa polypeptide was only detectable at lower intensities. This may be due to protein aggregation, which is typical of integral membrane proteins (Fig. 1, lane B2). From the results of SDS/PAGE, it can be deduced that the 16-kDa protein is only present in substoichiometric amounts. In some preparations, this protein was completely absent (Fig. 1B). As will be described below, the enzyme complex purified from A. fulgidus shows similarity to Hdr and has a menaquinol-oxidizing activity. The enzyme was therefore preliminarily designated Hdr-like menaquinol-oxidizing enzyme complex, abbreviated as Hme complex. Ó FEBS 2002 Hdr-like enzyme complex from A. fulgidus (Eur. J. Biochem. 269) 1897 Identification of the genes encoding the subunits of the Hme complex and sequence analysis The N-terminal sequences of the four polypeptides present in the purified enzyme preparation were determined by Edman degradation (Table 1). Using these sequences, the corresponding genes (AF499, AF501–503) were identified in thegenomeofA. fulgidus [10]. The noncoding regions between the different genes are short (less than 12 bp) or nonexistent (the genes overlap). The sequence region upstream o f A F499 is AT-rich and contains typical a rchaeal promoter elements. The sequence AAAGGTTAATATA was f ound 64 bp upstream of the start codon of AF499; this corresponds to the BRE element and the box A element of archaeal promoters [21,22]. The A F499–AF503 gene cluster can therefore be predicted to form a transcription unit (Fig. 2). This transcription unit contains one gene (AF500) for which no corresponding protein was found in the purified enzyme preparation. The results of the sequence analyses of the deduced proteins are given in Table 2. The protein encoded by AF502 has a calculated mole- cular mass of 64.4 kDa. The protein shows about 35% sequence identity with the proposed catalytic s ubunit H drD from M. barkeri. The closest relative of t he protein e ncoded by AF502 (40% sequence identity) is the dissimilatory sulfite reductase (Dsr)K protein from the sulfur-oxidizing phototrophic bacterium Allochromatium vinosum. The DsrK protein is encoded by the dsr locus, which also encodes the subunits of the siroheme sulfite reductase [23]. Another relative of the protein encoded b y AF502 is the high- molecula r-ma ss c-type cytochrome (Hmc)F protein of Desulfovibrio vulgaris (20% sequence identity) [24]. A characteristic of HdrD o f M. barkeri is the presence of two t ypical [4Fe-4S] cluster binding motifs in th e N -terminal part of the protein. HdrD contains 10 additional cysteine residues found in two CX 31)38 CCX 33)34 CXXC sequence motifs at the C-terminal part of the protein [4]. Multiple sequence alignments of HdrD, AF502, DsrK, and HmcF clearly identified the t wo typical CXXCXXCXXXCP binding motifs for [4Fe-4S] clusters in the N-terminal part of these proteins. AF502, DsrK, and HmcF also contain one of the two CX 31)38 CCX 33)34 CXXC motifs present in HdrD. Only in AF502 does an aspartate residue replace one of the five cysteines present in this motif [25]. The AF501 protein has a calculated molecular mass of 38 k Da. The molecular mass of this protein was estimated by SDS/PAGE to be 34 kDa. The protein shows the highest sequence similarity (30% identity) to the DsrM protein from A. vinosum, encoded b y the dsr locus,andtotheb-type cytochrome subunit of nitrate reductase from various organisms, such as NarI of nitrate reductase of Escherichia coli [26]. AF501 also has low sequence similarity to the b-type cytochrome (HdrE) of Hdr. A topological analysis suggests that AF501, like NarI, has five membrane-span- ning helices. In t he b-type c ytochromes of nitrate reductases, four histidine residues are conserved, two in helix b and two Fig. 1. SDS/PAGE of the purified Hme complex. Proteins were sepa- ratedina12%slabgel(8· 7 c m) which was subsequently stained with Coomassie Brilliant Blue R250. The polypeptide with an apparent molecular mass of 16 kDa, identified as a c-type cytochrome by N-terminal sequencing, was not found in all preparations. The pre- paration shown in (A) still contains the c-type cytochrome, while the preparation shown in (B) lacks this polypeptide. M, Low-molecular- mass markers ( Amersham Pharmacia Biotech). The m olecular mass es of the marker proteins are given o n the right side. Lane A1, 15 lgof the A. fulgidus Hme c omplex denatured fo r 30 min at room t emper- ature in SDS sample buffer; lane B1, 10 lg Hme complex denatured for 30 min at room temperatu re i n SDS sam ple buffer; lane B2, 10 lg Hme complex denatured for 5 min at 100 °C in SDS sample buffer. The polypeptide with an apparent molecular mass of 34 kDa, identi- fied as a b-type cytochro me-like protein by N-terminal sequencing, shows a lower intensity in the b oiled sample; it probably forms aggregates that do not run into the gel (lane B2). This behavior is typical of integral membrane proteins. The polypeptide with an apparent molecular mass of 53 kDa appears as a double band in unboiled samples (lanes A1 and B1). Table 1. N-Terminal sequences of the polypeptides of the purified enzyme. N-Terminal sequen ces were eit her obtained by Edman degradation (column 1) or d er ived from the genome sequence of A. fulgidus (column 2). The c orresp onding genes are given in column 3. Amino acids p re sent in both sequences are u nd erlined, a nd am ino acids th at co uld not be determined with ce rtainty i n the Ed man d egradation are g iven in parentheses. Sequence derived by Edman degradation Sequence derived from the A. fulgidus genome sequence Identified ORF ( M)(E)RMRE(I)IEIKAKFP MEEMPERIEIKQKFP AF502 MIGVIFGVIVFYIAV MIGVIFGVIVFYIAV AF501 ( K)TQFIESPEEV(V)EK MMSRRKFLLLTGAAAAGAILTPQISA KTQFIESPEEVREK AF499 MYNK-YVIPLILVFL MSEMYNKKYVIPLILVFL AF503 1898 G. J. Mander et al.(Eur. J. Biochem. 269) Ó FEBS 2002 in helix d. These histidine residues have been assigned as b-heme axial ligands for two heme groups that are located on different h alves o f t he membrane bilayer [26]. AF501 not only has the same topology as NarI, but also contains the two h istidine residues in h elix b and two in helix d. AF501 is therefore predicted to ligate two heme groups. The AF499 protein has a calculated molecular mass of 30.5 kDa. Sequence analysis revealed that the protein belongs to a group of iron-sulfur proteins with 16 conserved cysteine residues predicted to co-ordinate four [4Fe-4S] clusters. Members of this family include DsrO from A. vinosum, HmcB from D. vulgaris, and HybA, DmsB, and N rfC f rom E. coli. Some members, including AF499, have an N-terminal Ôtwin-arginineÕ signal sequence that is characteristic of cofactor-containing proteins translocated into the periplasm via the Tat translocase [27]. As deduced from the N-terminal sequence of AF499, the signal peptide is not present in the mature enzyme (Table 1). The AF503 protein has a calculated molecular mass of 16.7 kDa. The protein contains three CxxCH sequence motifs characteristic of proteins th at co-ordinate heme c. The protein is therefore predicted to co-ordinate three heme c molecules. AF503 shows the highest sequence similarity to a protein encoded by the dsr locus of A. vinosum, the DsrJ protein. The mature form of the AF503 protein contains an N-terminal hydrophobic s tretch predicted to form a transmembrane a helix, which may anchor the protein in the membrane. This stretch may function as a signal peptide of the Sec pathway [28]. The AF500 protein, which was not detected in the purified enzyme, has a calculated molecular mass of 43 kDa. This protein shows highest sequence identity to the DsrP protein from A. vinosum. It shows low sequence similarity to the HmcC protein of D. vulgaris. Topological analysis suggests that AF500, like DsrP and HmcC, has 10 membrane-spanning helices. These three proteins are also related to the DmsC protein of dimethylsulfoxide reductase [29]. The latter protein contains only eight predicted transmembrane helices. Catalytic properties of the Hme complex and characterization by UV/Vis spectroscopy To determine whether the cytochrome present in the Hme complex is reduced by menaquinone, in vitro assays were performed using the more hydrophilic analogue of men- aquinone, DMN. The enzyme purified u nder anoxic condi- tions generally contained the heme groups in the reduced state. Any enzyme molecules that contained oxidized h eme groups could be rapidly reduced by sodium dithionite. Addition of DMN to the reduced enzyme resulted in rapid oxidation of the heme present in the enzyme. The oxidized heme groups could be rapidly reduced using DMNH 2 as electron donor. The r ates of h eme reduction by DMNH 2 or oxidation by D MN were too rapid to be resolved. Figure 3 shows t he dithionite-reduced minus air-oxidized absorbance difference spectrum of an enzyme preparation containing only minor amounts of th e 16-kDa c-type cytochrome. The Fig. 2. Genomic o rganization of the genes e ncoding the subunits of the Hme complex from A. fulgidus. ThegenenamesannotatedbyTIGRaregiven above the arro w representing t he genes a n d their direction of transcription. The s ize in bp is given below each gene. Betwee n the genes A F498 and AF499 is a 385-bp-long noncoding region. The genes within the putative transcription unit from AF499 to AF503 have an intergenic region ranging from 1 t o 11 bp or even overlap (AF500 and AF501 overlap by 3 bp). The region 81–65 bp upstream of t he start c odon of AF499 was identified as an archaeal promoter element by seque nce analysis. The sequ ence AAAGGTTAATATA shows a high le vel of identity with th e consensus se quence ()35 to )23, AAANNN TTATATA) ; the sequence of the so-called BRE (transcription factor B recognition element) is in italics; the sequence of the so-called Box A is underlined. These eleme nts have been id entified as essentia l elements for archa eal transcription [21,22 ]. Table 2. Features of the subunits of the Hme complex from A. fulgidus. Data are eith er derived from the analysis of the sequence (calculated molecular mass, predicted transmembrane helices, cofactor binding sites, sequence identities) o r obtained experimentally (apparent molecular mass, cofactor content). Gene AF502 AF501 AF499 AF503 AF500 Apparent/calculated molecular mass 53/64.4 kDa 34/38 kDa 31/30.5 kDa 16/16.7 kDa – /43 kDa Transmembrane helices None 5 None 1 10 Cofactor binding sites 2[4Fe-4S], 4 highly conserved cysteine residues 2 heme groups 4[4Fe-4S] 3 heme c (CX 2 CH) – Highest sequence identity with DsrK DsrM DsrO DsrJ DsrP Further comments Related to the catalytic subunit of Hdr Cytochrome, integral membrane protein Extracytoplasmic iron-sulfur protein Extracytoplasmic c-type cytochrome Integral membrane protein Ó FEBS 2002 Hdr-like enzyme complex from A. fulgidus (Eur. J. Biochem. 269) 1899 absorption maxima at 420 nm (c band), 530 nm (b band) and 557 nm (a band) are characteristic of cytochrome b. Heme was extracted from the protein with acidic acetone and the pyridine hemochrome spectrum (reduced minus oxidized) was recorded. T he spectrum contained maxima of the a and b band at 553 and 521 nm, respectively. These maxima are blue-shifted by a bout 4 nm relative t o published values for protoheme IX [30]. In a control experiment the pyridine hemochrome spectrum of hemoglobin was deter- mined under identical conditions resulting in maxima identical with published values (525 nm for the b band and 557 nm for the a band). A similar blue shift was f ound in heme o of cytochrome bo [30]. As pyridine hemochrome spectra are very sensitive to substitutions on the porphyrin ring, the results indicate that th e e xtractable heme of Hme is not protoheme IX. Further studies are necessary to elucidate the nature of the extractable heme present in this enzyme. The oxidation of the heme groups of the enzyme by various compounds was tested to identify the physiological electron acceptor of the enzyme. The enzyme was not oxidized by the heterodisulfide of coenzyme M and coenzyme B, or by the homodisulfides of these two coenzymes. Also potassium trithionate and sodium tetra- thionate, w hich have been identified i n dissimilatory sulfate reducers [31], failed t o oxidize the enzyme. Characterization of the iron-sulfur clusters by EPR spectroscopy The enzyme preparation was shown to contain 90–110 nmol nonheme iron and 105–115 nmol acid-labile sulfur. I f one enzyme molecule has a mass of 150 kDa, this corresponds to  19–21 mol acid-labile sulfurÆ(mol enzyme) )1 and 16–20 mol nonheme ironÆ(mol enzyme) )1 . Analysis of the amino-acid sequence of the enzyme leads to the p rediction that the enzyme contains six [4Fe-4S] c lusters, four in AF499 and two in AF502. In addition to the conserved cysteine residues that co-ordinate these iron- sulfur clusters, the AF502 protein contains a cysteine cluster that is conserved in Hdr from various methanogens and in all Hdr-like proteins. Some of these cysteines in Hdr are thought to co-ordinate a special iron-sulfur cluster in the active site of the enzyme that is directly i nvolved in the reduction of disulfide substrate. The Hdr-like protein from A. fulgidus was t herefore studied by EPR spectroscopy. Redox titrations were monitored by EPR to characterize the different iron-sulfur clusters presen t in the enzyme. As expected, t he enzyme showed broad unresolved EPR signals at redox potentials u p to )100 mV that w ere only detectable at temperatures below 1 0 K. These signals are most probably due to the bulk of [4Fe-4S] + clusters present in the enzyme. At potentials higher than 0 mV, an unusual paramagnetic species was detected with g values at 2.031, 1.994, and 1.951. The resonance started to develop at potentials ‡ 0 mV and was stable at potentials up to +350 mV. The loss and formation of the resonance was associated with a one-electron redox process with a midpoint potential of +90 ± 10 mV (Fig. 4). The spin concentration of the signals in the different titrations was generally near 0.4 spinÆ(mol enzyme) )1 . Because of overlap with radical signals around g ¼ 2, the signal was simulated (Fig. 4) and double integrated to obtain the spin intensity. Temperature studies showed that the signal is readily power saturated at 4.5–15 K. At 15–35 K, the signal could be measured under nonsaturating conditions. At higher tem- peratures, the signal started to broaden a nd was broadened beyond detection at 60 K. The EPR signal observed has EPR characteristics very similar to a unique signal described for Hdr from metha- nogens. The two paramagnetic species have similar g values, show the same temperature behavior, and are only detect- able in the o xidized enzyme. The g value at 2.016 present in Hdr [6] is shifted in the A. fulgidus enzyme to 2.031. The midpoint potential of the paramagnetic species found in the A. fulgidus enzyme is shifted to higher redox potentials. In Hdr, this paramagnetic species is only observed in titrations carried out in the presence of one of the substrates of the enzyme. The physiological electron acceptor of the A. ful- gidus enzyme is still unknown. Therefore, no substrate could be added to the titration mixture. DISCUSSION A large number of protein sequences related to the catalytic subunit of Hdr from methanogenic archaea have been deposited in the databases. None of these putative pro teins has b een c haracterized and no f unction has been assigned to any of them [2]. In this study, we chose the sulfate-reducing archaeon A. fulgidus for the isolation of one of the Hdr-like proteins encoded by the genome of this organism. In cells cultivated on lactate/sulfate medium, the enzyme turned o ut to be a major membrane protein and contained m ost of the heme present in t he membran e fraction. Hdr from M. bark- eri is composed of only two subunits, a b-type cytochrome and the hydrophilic catalytic subunit [7]; the subunit structure of the Hme complex isolated from A. fulgidus is considerably more complex. The a nalysis o f the sequence o f the gene cluster encoding the enzyme predicts the presence of five subunits, but only four were detected in the purified enzyme preparation. The integral m embrane subunit A F500 Fig. 3. Room temperature reduced–oxidized difference spectrum of the purified Hme complex. Hme [1 mg proteinÆml )1 in 50 m M Tris/HCl (pH 7 .6)] was reduced with so dium dithionite and su bsequently oxi- dized by air. The oxidized spectrum was subt racted from the red uced spectrum. When the enzyme was oxidized by DMN, the same differ- ence spectrum was observed (not shown). The arrow indicates the absorption maximum of the aband at 557 nm. 1900 G. J. Mander et al.(Eur. J. Biochem. 269) Ó FEBS 2002 could not be detected after SDS/PAGE in gels stained with either Coomassie or silver (data not shown), which suggests that this subunit does no t copurify with the other subunits of the enzyme complex. From the primary structure, the b-type cytochrome-like protein AF501 and subunit AF500 are clearly predicted to be integral membrane proteins. The iron-sulfur protein AF499 contains a characteristic twin-arginine leader pep- tide. This strongly suggests t hat this p rotein is located at t he extracytoplasmic side of the membrane [27]. The c-type cytochrome AF503 contains a typical Sec-dependent hydro- phobic leader peptide [28] that is not cleaved off by a l eader peptidase as i t was still present in the purified protein. Therefore, this protein can also be predicted t o be located on the extracytoplasmic side of the membrane and to have an N-terminal membrane anchor. The AF502 protein, which is related to the catalytic subunit of Hdr, is a hydrophilic iron-sulfur protein. The protein does not contain a leader sequence and therefore may be attached to the integral membrane subunits on the cytoplasmic side. It cannot, however, be excluded that this protein binds to the AF499 protein in the cytoplasm and that this protein complex then is translocated across the cytoplasmic mem- brane via the TAT translocase. Such a mechanism has been found for periplasmic oxidoreductases [27]. The characterization of the A. fulgidus protein complex by EPR spectroscopy identified an unusual paramagnetic species with EPR characteristics and redox properties similar to those of the unusual paramagnetic species that has recently been described for Hdr from M. marburgensis and M. barkeri. In Hdr, this paramagnetic species, desig- nated CoM-Hdr, is formed on reaction o f the oxidized enzyme with coenzyme M (H-S-CoM) in the absence of coenzyme B (H-S-CoB). This paramagnetic species can be reduced in a one- electron step with a midpoint potential of )185 mV (M. marburgensis enzyme) or )142 mV (M. barkeri enzyme), but cannot be further oxidized. A broadening of the EPR sign al in the 57 Fe-enriched enzyme indicates that it is at least partially iron-based. The g values (g xyz ¼ 2.013, 1.991, 1.938 for the M. marburgensis enzyme and g xyz ¼ 2.011, 1.993, 1.944 for the M. barkeri enzyme) and the midpoint potential argue against a conventional [2Fe- 2S] + ,[3Fe-4S] + , [4Fe-4S] + ,or[4Fe-4S] 3+ cluster. CoM- Hdr reacts with H-S-CoB to produce an EPR-silent form. This indicates that only a half reaction is catalyzed when only H -S-CoM is pr esent and that a reaction intermediate of the catalytic cycle is trapped [6]. Variable-temperature magnetic circular dichroism spectroscop y studies of CoM- Hdr have provided compelling evidence f or the p resence of a novel type of [4Fe-4S] 3+ cluster at t he active site of Hdr [6,7]. When oxidized Hdr is incubated with H-S-CoB, an EPR signal with similar g values is obtained, but the midpoint potential is shifted t o higher values ()30 mV for Hdr from M. marburgensis and >0 mV for Hdr from M. barkeri). From these data it has been concluded that H-S-CoB also reacts with the active site of the enzyme. As this reaction is only observed at nonphysiological redox potentials, it has been proposed that this species could not be an intermediate of the catalytic cycle, but rather is the product of a side reaction that occurs at these high redox potentials. Similar results have been obtained with other thiols, such as dithiothreitol, which are not substrates of the enzyme [6]. In contrast with Hdr, the paramagnetic species in the enzyme co mple x f rom A. fulgidus co uld a lready be observed when the enzyme was poised at redox potentials higher than 0 mV. No substrate was added in these experiments. It cannot, however, b e excluded that the purified enzyme contains an unidentified tightly bound substrate. It also has to be considered that the formation of the paramagnetic species is an intrin sic property of the enzyme. In this case, the signal could, for example, b e g enerated by the co-ordination of a redox-active cysteine residue of the enzyme to a metal cluster. The midpoint-potential of the paramagnetic species in the A. fulgidus enzyme was determined to be +90 mV. Fig. 4. EPR-monitored r edox titration of the A. fulgidus Hme c omplex. Hme ( 7 l M )in50m M Mops/KOH (pH 7.0) was used. Titrations were carried out as described in Materials and methods. (A) Data points correspondtotheamplitudeofthetroughcenteredatg ¼ 1.951; as in the low po tential range, the radical signal of the dyes overlap in the g ¼ 2.0 region. The maximal spin concentration was 0.4 per enzyme molecule as determined by double integration of t he simulated EPR signal. (B) EPR spectrum obtained at +176 mV (solid line) and the EPR simulation (dashed line). EPR conditions: temperature, 20 K; microwave power, 2.007 mW; microwave frequency, 9458 MHz; modulation amplitude, 0.6 mT. Simu lation parameters: g 123 ¼ 2.031, 1.994, and 1.951; W 123 ¼ 1.25, 1.2, and 1.15 mT. Ó FEBS 2002 Hdr-like enzyme complex from A. fulgidus (Eur. J. Biochem. 269) 1901 This value is more positive than the standard redox potentials of the APS/sulfite couple ()60 mV) and the sulfite/sulfide couple ()116 mV), which are thought to be the final electron a cceptors (see b elow). It therefore has to be considered that the signal in the A. fulgidus enzyme is generated nonspecifically at high redox potentials. The reaction of the e nzyme with its physiological substrate may result in a shift of the midpoint potential of this species to lower values, as has been observed w ith Hdr [6]. The sequence analysis of the A. fulgidus enzyme clearly shows that t he AF502 protein is related to the catalytic subunit HdrD of H dr from M. barkeri. I n particular, AF502 and HdrD share a common cysteine motif that in Hdr is thought to co-ordinate the special [4Fe-4S] cluster in the active site. I n the four Hdr sequences currently available, this motif (CX 31)38 CCX 33)34 CXXC) is present in two copies in each sequence. The Hdr-like p roteins contain either one or two copies of this sequence motif. The three closely related proteins AF502, DsrK, and HmcF contain only one copy, and this may be sufficient for metal-cluster binding. Only in the A F502 protein does an aspartate residue replace one of the five cysteine residues. Aspartate can in principle also function as a ligand of an iron-sulfur cluster [25]. Enzymes related to the Hme complex from A. fulgidus described in this work are also encoded by the genomes of the sulfate-reducing bacterium D. vulgaris and the phototrophic sulfur bacterium A. vinosum. Anaerobic sul- fate-reducing bacteria such as D. vulgaris contain a high- molecular-m ass cytochrome c with 16 covalently bound hemes [32]. This multiheme cytochrome has been purified and extensively characterized. In D. vulgaris, this protein is encoded by a large operon, called the hmc operon [24]. The operon consists of eight genes: two encode regulatory proteins and six encode the s tructural proteins o f the enzyme complex (hmcA to hmcF). hm cA encodes a high-molecular- mass c-type cytochrome, hmcB encodes a periplasmic i ron- sulfur protein, hmcE encodes a b-type cytochrome, hmcD encodes a small hydrophilic protein with a single hydro- phobic, potentially membrane-spanning sequence, hmcC encodes an integral membrane p rotein, a nd hmcF encodes an iron-sulfur protein related to the catalytic subunit of Hdr. A comparison of the Hmc complex from D. vulgaris with the Hme complex from A. fulgidus shows that the two complexes h ave a set o f s equence-related subunits, with only two major differences: a homologue of the H mcD p rotein is not encoded by the operon from A. fulgidus, and the high- molecular-mass c ytochrome c of D. vulgaris is replaced b y a low-molecular-mass cytochrome c with only three heme - binding motifs in A. fulgidus. The D. vulgaris Hmc complex has not yet been purified, but expression of the hmc operon has been monitored in an immunoassay using HmcA-specific or HmcF-specific anti- sera. T he le vel of expression is highest in cells cultivated on H 2 /sulfate medium, and expression is about fourfold lower in cells cultivated on lactate/sulfate or pyruvate/sulfate medium [33]. In addition, a mutant strain in which most of the hmc operon is deleted has been constructed. This deletion mutant grows normally when lactate or pyruvate serve as e lectron donors for sulfate r eductio n. T he mutant is still able to grow on H 2 /sulfate, although at a growth rate lower t han that of the wild-type. The mutant is a lso d eficien t in low-redox-potential niche establishment [34]. From these various observations, it has been concluded that the Hmc complex is involved in t he electron transfer from H 2 ,which is activated by a periplasmic hydrogenase, to an electron acceptor on the cytoplasmic side o f the membrane. As growth of the hmc deletion mutant on H 2 /sulfate is not completely abolished, the organism may be able to synthe- size an alternative enzyme complex with a function similar to that of Hmc. Proteins with the highest sequence similarity to the five subunits of the A. fulgidus enzyme complex were found to be encoded by the dsr locus of A. vinosum [23]. dsrA and dsrB encode the a and b subunit of the dissimilatory sulfite reductase of this organism. These two g enes are o rganized in a cluster with genes encoding proteins highly related to the AF499–AF503 proteins (Table 2) [23]. Polar insertion mutations immediately downstream of dsrA,andindsrB, dsrH,anddsrM, lead to an inability of the cells to oxidize intracellular sulfur to sulfite under photolithoautotrophic conditions. The ability of the mutant cells to oxidize sulfide to sulfur, thiosulfate to tetrathionate, or sulfite to sulfate under photolithoautotrophic conditions is unaltered. Two models suggesting a function of the dsr gene products in the oxidation of sulfur to sulfite have been presented [23]. In these m odels, t he pr oducts of the dsrO to dsrN genes a re not yet included. These genes have o nly recently b een identified (EMBL accession number U84760). A. fulgidus strain VC16 completely oxidizes organic substrates, s uch a s l actate, to C O 2 . T he reducing equivalents thus generated are transferred to the menaquinone pool by different oxidoreductases of the oxidative branch of t he pathway. It is, however, not yet k nown how the reduced menaquinone pool is electrically connected to th e enzymes of sulfate reduction, namely APS reductase and sulfite reductase. APS reductase from A. fulgidus is an iron-sulfur flavoprotein composed of two subunits [35]. The enzyme has been isolated from the soluble fraction, and t he primary structure does not indicate any transmembrane helices. The enzyme is closely related to APS reductase from sulfate- reducing b acteria a nd from chemotrophic and phototrophic sulfur-oxidizing bacteria [36]. Sulfite reductase from A. ful- gidus has also been characterized [37]. I n c ommon with other Dsrs, the enzyme h as an a 2 b 2 structure and contains siroheme, nonheme iron, and acid-labile sulfur [36]. An additional protein with an apparent molecular mass of 11 k Da is associated with sulfite reductase from D. vulgaris [38] and De sulfovibrio desulfuricans [39,40]. The function of this so calle d c subunit is not yet known. In most of the organisms t hat h ave been studied, t he enzyme has been isolated from the soluble fraction. Sulfite reductase from D. desulfuricans was found to be partially membrane- associated after gentle disruption of t he cells [39,40]. On the basis of our results and comparisons with published results for other organisms, we propose that the Hme complex of A. fulgidus functions as a menaquinol oxidoreductase. The sequence analysis of the enzyme indicates that it is composed of two modules that may have distinct functions. The first module is related to Hdr from M. barkeri [3]. It is composed of the b-type cytochrome-like protein AF501 and the AF502 protein, which has sequence similarity to the catalytic subunit of Hdr. We propose that this module of the enzyme complex mediates the electron transfer from menaquinol to an unidentified electron acceptor on the cytoplasmic side o f the membrane. This is supported by the findin g that the heme g roups of the 1902 G. J. Mander et al.(Eur. J. Biochem. 269) Ó FEBS 2002 purified A. fulgidus enzyme were rapidly reduced by DMNH 2 or were rapidly oxidized by DMN. Furthermore, it has b een shown that the membrane fr action of A. fulgidus catalyzes the reduction of the heme groups present in the membrane fraction by F 420 H 2 at high ra tes [14 ]. The Hme com plex contains th ree a dditional s ubunits: th e integral membrane protein AF500, the extracytoplasmic iron-sulfur protein AF499, and the extracytoplasmic c-type cytochrome AF503. AF500 shows low sequence similarity to the subunit DmsC of dimethylsulfoxide reductase, which functions as a menaquinol oxidase [41]. Likewise, AF500 may function as a second menaquinol-oxidizing site of the Hme complex, and, together with the iron-sulfur protein AF499 and the c-type cytochrome AF503, may f orm a second module of the enzyme complex. We propose that this modu le catalyzes the electron transfer from menaquinol to the c-type cytochrome. The c-type cytochrome may function as the electron donor of alternative electron acceptors or oxidoreductases. One possible candidate is oxygen. It has been shown that sulfate-reducing bacteria first consume oxygen i n their environment and then start to reduce sulfate. I n Desulfovibrio species, the highest oxygen- uptake activity is found in the periplasmic fraction, with H 2 as electron donor. C ytochrome c 3 was found to play a major role in oxygen reduction [42,43]. It has been proposed that the Hmc complex from D. vulgaris mediates the electron t ransfer from a periplasmic hydrogenase to t he cytoplasmic side where reduction of sulfate occurs [24,44]. A. fulgidus is also able to grow with H 2 as electron donor for sulfate reduction [11]. The genome of A. fulgidus, however, does not encode a s oluble periplasmic hydrogenase as found in sulfate-reducing bacteria. Instead, thegenomeofA. fulgidus encodes an extracytoplasmic hydrogenase (AF1379 to AF1381), which is predicted to contain a b-type cytochrome subunit (AF1379) as a membrane anchor [10]. This hydrogenase is therefore predicted to catalyze the hydrogen-dependent reduction of menaquinone, as do oth er hydrogenases of this type [45]. Methanogenic archaea belonging to the family Meth- anosarcinales contain two different energy-conserving elec- tron-transport chains t hat catalyze the reduction of the heterodisulfide. W hen the organism grows on methanol, reduced coenzyme F 420 is generated during methanol oxidation to CO 2 [46]. The organism contains a mem- brane-bound electron-transport chain which mediates the reduction of the heterodisulfide by F 420 H 2 . I t is c omposed of F 420 H 2 –methanophenazine oxidoreductase, methanophen- azine, and Hdr [47]. When the organism g rows on H 2 /CO 2 , H 2 serves as the electron donor for the reduction of th e heterodisulfide. In this case, the electron-transport chain is composed of a m ethanophenazine-reducing hydrogenase, methanophenazine, and Hdr [47]. The reduced thiols thus generated then function as the electron donor for the reduction of the methyl group to methane in a nonenergy- conserving reaction [1]. On the basis of the above, we propose that, in A. fulgidus, two similar electron-transport chains operate in which menaquinone and not methanophenazine is the membrane- soluble electron carrier. Menaquinone is either reduced by the F 420 H 2 –MQ o xidoreductase or by the h ydrogenase. T he Hme complex described in this work then reoxidizes menaquinol and transfers the e lectrons either to an electron acceptor on the extracytoplasmic side of the membrane or to an acceptor in the cytoplasm. The latter electron acceptor, which is still unknown, is thought to function in its reduced form as electr on donor of the e nzymes of sulfate reduction. ACKNOWLEDGEMENTS This work was supported by the Max-Planck-Gesellschaft,theDeutsch e Forschungsgemeinsc haft,theFonds der Chemischen Industrie, and by a fellowship f rom the Humboldt Stiftung to E. D. We thank P eter M. H. Kroneck for the gift of potassium trithionate. We thank Karen A. Brune for editing the manuscript. REFERENCES 1. T hauer, R.K. (1998) Biochemistry of methanogenesis: a tribute to Marjory Stephenson. Microbiology 144, 2377–2406. 2. Hedderich, R., Klimmek, O., Kro ¨ ger, A., Dirmeier, R., Keller, M. & Stetter, K.O. (1998) Anaerobic respiration w ith elemental sulfur and with disulfides. FEMS Microbiol. Rev. 22, 3 53–381. 3. Heiden, S., Hedderich, R., Setzke, E. & Thauer, R .K. (1994) Purification of a two-subunit cytochrome-b-containing het- erodisulfide reductase from methanol-grown Methanosarcina barkeri. Eur. J. Biochem. 221, 855–861. 4. K u ¨ nkel, A ., Vaupel, M., Heim, S., Thauer, R.K. & Hedderich, R. (1997) Heterodisulfide reductase from methanol-grown cells of Methanosarcina barkeri is not a flavoenzyme. Eur. J. Biochem. 244, 226–234. 5. Simianu, M., Murakami, E., Brewer, J.M. & Ragsdale, S.W. (1998) Purification and properties of the heme- and iron-sulfur- containing heterodisulfide reductase from Methanosarcina ther- mophila. Biochemistry 37, 10027–10039. 6. Madadi-Kahkesh, S., Duin, E.C., Heim, S., Albracht, S.P.J., Johnson, M.K. & Hedderich, R. (2001) A paramagnetic species with unique EPR characteristics in the active s ite of heterodisulfide reductase from methano genic archaea. Eur. J. Biochem. 268, 2566–2577. 7. Duin, E.C., Madadi-Kakesh, S., Hedderich, R., Clay, M.D. & Johnson, M.K. (2002) H eterodisulfide reductase from Metha- nothermobacter marburgensis contains an active-site [4Fe-4S] cluster that is directly involved in mediating heterodisulfi de reduction. FE BS Lett. 512, 263–268. 8. Abken,H.J.,Tietze,M.,Brodersen,J.,Baumer,S.,Beifuss,U.& Deppenmeier, U. (1998) Isolation and characterization of methanophenazine and function of phenazines in membran e- bound electron transport of Methanosarcina mazei Go ¨ 1. J. Bac- teriol. 180, 2027–2032. 9. Setzke, E., Hedderich, R., Heiden, S. & Thauer, R.K. (1994) H 2 :heterodisulfide oxidoreductase complex from Methanobacter- ium thermoautotrophicum: composition and properties. Eur. J. Biochem. 220, 139–148. 10. K lenk, H.P., Clayton, R.A., T omb, J.F., White, O., Nel son, K.E., Ketchum, K.A., Dodson, R.J., Gwinn, M., Hickey, E.K., Peterson, J.D., et al. (1997) The complete genome s eq uence of t he hyperthermophilic, sulphate-reducing archaeon Archaeoglobus fulgidus. Nature 390, 364–370. 11. Stetter, K.O., Lauerer, G., Thomm, M. & Neuner, A. (1987) Isolation of extremely thermophilic sul fate reducers: evidence for a novel branch of archaebacteria. Science 236, 822–824. 12. Mo ¨ ller-Zinkhan, D., Bo ¨ rner,G.&Thauer,R.K.(1989)Function of methanofuran, tetrahydromet hanopterin, and co enzyme F 420 in Archaeoglobus fulgidus. Arch. Microbiol. 152, 362–368. 13. Mo ¨ ller-Zinkhan, D. & Thauer, R.K. (1990) Anaerobic lactate oxidation to 3 CO 2 by Archaeoglobus fulgidus via the carbon monoxide dehydrogenase pathway: demonstration of the acetyl- CoA carbon-carbon cleavage reaction in cell extracts. Arch. Microbiol. 153, 215–218. Ó FEBS 2002 Hdr-like enzyme complex from A. fulgidus (Eur. J. Biochem. 269) 1903 14. Kun ow, J., Linder, D., S tetter, K.O. & Thauer, R.K. (1994) F 420 H 2 :quinone oxidoreductase from Archaeoglobus fulgidus. Characterization of a membrane-bound multisubu nit complex containing FAD and iron-sulfur clusters. Eur. J. Biochem. 223, 503–511. 15. Tindall, B.J., Stetter, K.O. & Collins, M.D. (1989) A novel fully saturated menaquinone from the thermophilic sulfate-reducing archaebacterium Archaeoglobus fulgidus. J. Gen. Microbiol. 13 5, 693–696. 16. Unden, G. & Kro ¨ ger, A. (1986) Reconstitution of a functional electron-transfer chain from purified formate dehydrogenase and fumarate reductase complexes. Methods Enzymol. 126, 387–399. 17. Fish, W.W. (1988) Rapid colorimetric micromethod for the quantitation of comp lexed iron in bio logical samples. Methods in Enzymology (Riordan,J.F.&Vallee,B.L.,eds),pp.357–364. Academic Press Inc, New York. 18. Cline, J.D . (196 9) Spectrophotometric determination o f hydrogen sulfide in natural waters. Limnol. Oceanogr. 14, 454–458. 19. Berry, E.A. & Trumpower, B.L. (1987) Simultaneous determina- tion of h emes a, b,andc from pyridine he mochrome spectra. Anal. Biochem. 161, 1–15. 20. Beinert, H. & A lbracht, S.P. (198 2) New insights, ideas and unanswered questions concerning iron-sulfur clusters in mito- chondria. Biochim. Biophys. Acta 683, 245–277. 21. Soppa, J. (1999) Normalized nucleotide frequencies allow the definition of archaeal promoter elements for different archaeal groups and reveal base-specific TFB contacts upstream of the TATA box. Mol. Microbiol. 31, 1589–1601. 22. Brown, J.W., Daniels, C.J. & Reeve, J.N. (1989) Ge ne structure, organization, and expression in archaebacteria. Crit.Rev.Micro- biol. 16, 287–338. 23. Pott, A.S. & D ahl, C. (1998) Sirohaem s ulfite reductase and o ther proteins encoded by g e nes at th e dsr locus of Chrom atium vinosum are involved in the oxidation of intracellular sulfur. M icrobiology 144, 1881–1894. 24. Rossi, M., Pollock, W.B., Reij, M.W., Keon, R.G., Fu, R. & Voordouw, G. (1993) The hmc operon of Desulfovibrio vulgaris subsp. vulgaris Hildenboro ugh encodes a potential transm em- brane redox protein compl ex. J. Bacteriol. 175, 4699–4711. 25. Gorst, C.M., Zhou, Z.H., Ma, K ., Teng, Q., Howard, J .B., Adams, M.W. & La Mar, G .N. ( 1995) Participation of the disulfide bridge in the redox cycle of the ferredoxin from the hyperthermophile Pyrococcus furiosus: 1 H nuclear magnetic resonance time resolution of t he four redox states at amb ient temperature. Biochemistry 34, 8788–8795. 26. Berks, B.C., P age, M.D., Richardson, D .J., Reilly, A., Cavill, A., Outen, F. & F erguson, S.J. (1995) Sequence ana lysis of subunits of the m embrane-bou nd nitrate reductase from a denitrifying bac- terium: the integral membrane subunit provides a prototype for the dihaem e lectron-carrying arm of a redox loop. Mol. Microbiol. 15, 319–331. 27. Berks, B.C., Sargent, F. & Palmer, T. (2000) The Tat protein export pathwa y. Mol. Microbiol. 35, 260–274. 28. Fekkes, P. & Driessen A rnold, J.M. (1999) Protein t argeting to the bacterial cytoplasmic membrane. Microbiol. Mol. Biol. Rev. 63, 161–173. 29. Bilous, P.T., Cole, S.T., Anderson, W.F. & Weiner, J.H. (1988) Nucleotide sequence of the dmsABC operon enco ding the a nae- robic d imethylsulph oxide reductase of E scherich ia coli. Mol. Microbiol. 2, 785–795. 30. Puustinen, A. & Wikstro ¨ m, M. (1991) The heme groups of cyto- chrome o from Escherichia coli. Proc. Natl Acad. Sci. USA 88, 6122–6126. 31. Sass, H., Steuber, J., Kroder, M., Kroneck, P.M.H. & Cypionka, H. (1992) Formation of thionates by freshwater and marine strains of sulfate-reducing bacteria. Arch. M icrobiol. 158 , 418–421. 32. Pollock, W.B., Loutfi, M., Bruschi, M., Rapp-Giles, B.J., Wall, J.D. & Voordouw, G. ( 1991) Cloning, seque ncing, and expression of the gene e ncoding the high-molec ular-weight cytochrome c from Desulfovibrio v ulgaris Hildenborough. J . Bacteriol. 173,220– 228. 33. Keon, R.G. & Voordouw, G. (1996) Identification of the Hmcf and Topology of th e Hmcb Subunit of the Hmc Complex of Desulfovibrio vulgaris. Anaerobe 2, 231–238. 34. Dolla, A., P ohorelic, B.K., Voordouw, J.K. & Voordouw, G. (2000) Deletion of the hmc op eron o f Desulfovibrio v ulgaris su bsp. vulgaris Hildenborough hampers hydrogen metabolism and low- redox-potential niche establishm ent. Arch. Microbiol. 174 ,143– 151. 35. Speich, N., Dahl, C., Heisig, P., Klein, A., Lottspeich, F., Stetter, K.O. & Tru ¨ per, H.G. (1994) Adenylylsulphate r eductase from the sulphate-reducing archaeon Archaeoglobus f ulgi dus: cloning and characterization of the genes and c omparison of t he enzyme with other i ron-sulphu r flavoproteins. Microbiology 140, 1273–1284. 36. Hipp,W.M.,Pott,A.S.,Thumschmitz,N.,Faath,I.,Dahl,C.& Tru ¨ per, H.G. (1997) Towards the phylogeny of APS reductases and sirohaem sulfite reductases in sulfate-reducing and sulfur- oxidizing prokaryotes. Microbiology 143, 2891–2902. 37. Dahl,C.,Kredich,M.K.,Deutzmann,R.&Tru ¨ per, H.G. (1 993 ) Dissimilatory sulphite reductase from Archaeoglobus fulgidus: physico-chemical properties of the enzyme and cloning, sequ en- cing and a nalysis of the reductase genes. J. Gen. Microbiol. 139, 1817–1828. 38. Pierik, A.J., Duyvis, M.G., Van Helvoort, J.M.L.M., Wolbert, R.B.G. & Hagen, W.R. ( 1992) The third subunit of desulfoviridin- type dissimilatory sulfite reductases. Eur. J. Bi oche m. 205,111– 115. 39. Steuber, J ., C ypionka, H . & Kroneck, P.M.H. (1994) Mech anism of dissimilatory sulfite reduction by Desulfovibrio desulfuricans: purification of a membrane-bound sulfit e reductase and coupling with cytochro me c-3 and hydrogenase. Arch. M icrobiol. 16 2,255– 260. 40. Steuber, J., Arendsen, A.F., Hagen, W.R. & Kroneck, P.M. (1995) Molecular properties of the dissimilatory sulfite reductase from Desulfovibrio desulfuricans (Essex)andcomparisonwiththe enzyme from Desulfovibrio vulgaris (Hildenborough). Eur. J. Biochem. 233, 873–879. 41. Sambasivarao, D. & We iner, J .H. ( 1991) Dim ethyl sulfoxide reductase of Escherichia coli: an investigation of function and assembly by use of in vivo complementation . J. Bacteriol. 17 3, 5935–5943. 42.Baumgarten,A.,Redenius,I.,Kranzoch,J.&Cypionka,H. (2001) Periplasmic o xygen r eduction by Desulfovibrio species. Arch. Microbiol. 176, 306–309. 43. Cypionka, H. (2000) Oxygen respiration by Desulfovibrio species. Annu. Rev. Microbiol. 54, 827–848. 44. Fauque,G.,Peck,H.D.Jr,Moura,J.J.,Huynh,B.H.,Berlier,Y., DerVartanian, D.V., Teixeira, M ., Przybyla, A.E., Lespinat, P.A. & Moura, I. (1988) The three cl asses of hydrogenases from sulfate- reducing bacteria of the genus Desulfovibrio. FEMS Microbiol. Rev. 54, 299–344. 45. Gross,R.,Simon,J.,Theis,F.&Kro ¨ ger, A. (199 8) Tw o mem- brane anchors of Wolinella succinogenes hydrogenase and their function in fumarate and p olysulfide respiration. Ar ch. M icrobiol. 170, 50–58. 46. Keltjens, J.T. & Vogels, G.D. ( 1993) Conversion of methanol a nd methylamines to methane and carbon dioxide. M ethanogenesis (Ferry, J.G., ed.), pp. 253–303. Chapman & Hall, New York & London. 47. De ppenmeier, U., Lienard, T. & Gottsc halk, G. (1999) Novel reactions involved in energy conservation by methanogenic ar- chaea. FEBS Lett. 457 , 291–297. 1904 G. J. Mander et al.(Eur. J. Biochem. 269) Ó FEBS 2002 . Purification and characterization of a membrane-bound enzyme complex from the sulfate-reducing archaeon Archaeoglobus fulgidus related to heterodisulfide. ¢-phosphosulfate reductase and sulfite reductase. Here we report on the isolation and characterization of a heme-containing membrane protein from A. fulgidus related to