Accessory proteins functioning selectively and pleiotropically in the biosynthesis of [NiFe] hydrogenases in Thiocapsa roseopersicina Gergely Maro ´ ti, Barna D. Fodor, Ga ´ bor Ra ´ khely, A ´ kos T. Kova ´ cs, Solmaz Arvani and Korne ´ l L. Kova ´ cs Institute of Biophysics, Biological Research Center, Hungarian Academy of Sciences, and Department of Biotechnology, University of Szeged, Hungary There are at least two membrane-bound (HynSL and HupSL) and one soluble (HoxEFUYH) [NiFe] hydrogen- ases in Thiocapsa roseopersicina BBS, a purple sulfur photosynthetic bacterium. Genes coding for accessory pro- teins that participate in the biosynthesis and maturation of hydrogenases seem to be scattered along the chromosome. Transposon-based mutagenesis was used to locate the hydrogenase accessory genes. Molecular analysis of strains showing mutant phenotypes led to the identification of hupK (hoxV ), hypC 1 , hypC 2 , hypD, hypE,andhynD genes. The roles of hynD, hupK and the two hypC genes were investigated in detail. The putative HynD was found to be a hydrogenase-specific endoprotease type protein, participa- ting in the maturation of the HynSL enzyme. HupK plays an important role in the formation of the functionally active membrane-bound [NiFe] hydrogenases, but not in the bio- synthesis of the soluble enzyme. In-frame deletion muta- genesis showed that HypC proteins were not specific for the maturation of either hydrogenase enzyme. The lack of either HypC protein drastically reduced the activity of every hydrogenase. Hence both HypCs might participate in the maturation of [NiFe] hydrogenases. Homologous comple- mentation with the appropriate genes substantiated the physiological roles of the corresponding gene products in the H 2 metabolism of T. roseopersicina. Keywords: hydrogenase; accessory genes; pleiotropic; metalloenzymes; [NiFe] center biosynthesis. Hydrogenases (EC class 1.12.1) [1] have the capability to reduce protons or oxidize molecular hydrogen. They are ancient metalloenzymes present in many archaea and bacteria, as well as occasionally in eukaryotes. Some microorganisms are known to contain several distinct hydrogenase enzymes [2] that vary in their cellular location. Two major groups of hydrogenases are distinguished according to their metal content, the Fe and the [NiFe] hydrogenases [1–3]. The [NiFe] hydrogenases are composed of at least two subunits. The small subunit transfers electrons via Fe–S clusters, while the large subunit contains the unique heterobinuclear [NiFe] metallocentre, which is the catalytic site. In the active centre two CN and one CO ligands are associated with the Fe atom [4]. The formation of an active hydrogenase requires a complex maturation process, including the incorporation of metal ions (Fe, Ni) and CO and CN ligands in the active centre, the orientation of the Fe–S clusters within the small subunit, and the proteolytic cleavage of the C-terminal end of the large subunit by an endoprotease [5,6]. Several steps in this maturation process have recently become understood. The HypFandHypEproteinswereproventoplayakeyrolein providing the CO and CN ligands from carbamoyl phosphate [7–9]. A complex of two other pleiotropic accessory gene products, the HypC and HypD proteins has been assumed to carry the iron atom during ligand formation and the assembled Fe-complex is somehow transferred to the C-terminal part of the hydrogenase large subunit as the HypC–HypD proteins dissociate [10–12]. There are additional accessory proteins, which are essential in the synthesis of mature [NiFe] hydrogenases although their particular role in the hydrogenase biosynthesis is less clear at this time. Some of these proteins are pleiotropic, as they participate in the biosynthesis of each [NiFe] hydro- genase present in the cell. Other accessory proteins are specific enzymes, that play a role only in the formation of a single hydrogenase [6]. In Ralstonia eutropha,thehypA, hypB and hypF genes are duplicated and any of the cognate gene products can mature thehydrogenasesinthisstrain[13].Itisanintriguing question, why two copies of the pleiotropic enzymes are needed, if one of them is sufficient to carry out the biological function? Remarkably, the chaperon-like entity, HypC, is a pleiotropic protein, although two variants of this protein have been identified in Escherichia coli. HypC is indispens- able for the maturation of the hydrogenase 3 in E. coli, although it can replace the function of the similar chaperon- type protein, HybG, in the maturation of hydrogenase 1 but not of hydrogenase 2 [14]. There are also two copies of the Correspondence to K. L. Kova ´ cs, Department of Biotechnology, University of Szeged, H-6726 Szeged, Temesva ´ ri krt. 62, Hungary. Fax: + 36 62 544 352, Tel.: + 36 62 544 351, E-mail: kornel@nucleus.szbk.u-szeged.hu Enzymes: Hydrogenases (EC 1.12.1). Note: Preliminary results were presented at the ÔBiohydrogen 2002Õ Conference, Ede-Wageningen, NL, April 21–24, 2002 and reviewed in Kova ´ cs,K.L.,Fodor,B.,Kova ´ cs, A ´ .T.,Csana ´ di, G., Maro ´ ti, G., Balogh, J., Arvani, S. & Ra ´ khely, G. (2002) Hydrogenases, accessory genes and the regulation of [NiFe] hydrogenase biosynthesis in Thiocapsa roseopersicina. Int. J. Hydrogen Energy 27, 1463–1469. (Received 29 January 2003, revised 12 March 2003, accepted 24 March 2003) Eur. J. Biochem. 270, 2218–2227 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03589.x HypC family members in Ralstonia eutropha, Rhodobacter capsulatus and Rhizobium leguminosarum [2]. Thiocapsa roseopersicina BBS is a mesophilic purple sulfur photosynthetic bacterium, containing at least two membrane-bound (HynSL, and HupSL) [15,16] and a soluble (HoxEFUYH) (G. Ra ´ khely, Gy. Csana ´ di, G. Maro ´ ti, B. D. Fodor & K. L. Kova ´ cs, unpublished observations) [NiFe] hydrogenase. No accessory genes could be identified in the vicinity of the hynSL genes [16] and the structural genes of the soluble hydrogenase. Downstream from the hupSL structural genes, accessory genes (hupDHI )andthehupR gene (corresponding to the regulator of a two component regulatory system) were found [15]. The lack of accessory genes in the vicinity of the structural genes is uncommon, as auxiliary genes tend to form gene clusters in most microorganisms harboring hydrogenase enzymes [1,2,6,17,18]. Our aim was to find and characterize the accessory genes needed for the maturation of functionally active hydro- genases in T. roseopersicina and to understand their physio- logical roles. The determination of the specificity of the accessory proteins is a challenging exercise in this micro- organism because of the presumed large number of hydro- genase-related genes. A transposon-based mutagenesis system and a reliable screening method has been established for T. roseopersicina [19]. The genetic approach was devel- oped further for producing in-frame deletion mutants in this strain. Here we show the molecular characterization of the T. roseopersicina mutant strains, where the hydrogenase biosynthesis is affected specifically and/or pleiotropically. Materials and methods Bacterial strains and plasmids Strains and plasmids are listed in Table 1. T. roseopersicina strains were grown photoautotrophically in Pfennig’s min- eral medium, under anaerobic conditions, in liquid cultures with continuous illumination at 27–30 °C for 4–5 days [20]. Plates were solidified with 7 gÆL )1 Phytagel (Sigma) [21] and supplemented with acetate (2 gÆL )1 ) when selecting for transconjugants. The plates were incubated in anaerobic jars using the AnaeroCult (Merck) system for two weeks. Escherichia coli strains were maintained on LB-agar plates. Antibiotics were used in the following concentrations (lgÆmL )1 ): E. coli: ampicillin (100), kanamycin (25), tetra- cyclin (20); for T. roseopersicina: kanamycin (25), strepto- mycin (5), gentamycin (5). Conjugation The conjugation was carried out as described in [19]. Transposon mutagenesis The mini transposon delivery plasmid pUT/mini-Tn5Km [23] was mobilized from E. coli S17-1(kpir) to T. roseo- persicina BBS. One hundred colonies were randomly selected after each mating and screened for a hydro- genase-deficient phenotype [19]. In this work, the M442, M1250, M4711, M646 and the M1343 mutants were chosen for detailed molecular analysis. DNA manipulations, PCR, sequencing, Southern blot and sequence analysis Preparation of genomic DNA, plasmids, cloning and Southern blots were done according to general practice [26], or the manufacturers’ instructions. PCR was carried out in a PTC-150 MiniCycler (MJ Research). Sequencing of both strands was done using an automatic Applied Biosys- tems 373 Stretch DNA sequencer. The searches in the NBRF, SwissProt, combined EMBL/GenBank and Prosite databases were carried out with the various BLAST programs (http://www3.ncbi.nlm.nih.gov/BLAST/). Mul- tiple alignments were performed with the CLUSTALW program ( DNASIS MAX v1.0, Hitachi Genetic System). Isolation of the hydrogenase-related genes Partial genomic libraries were prepared from the various mutants in pBluescript SK+ and ampicillin/kanamycin resistant clones were selected. A list of the positive clones is given in Table 1 (see also Fig. 1). The sequenced genes and regions has been deposited in the GenBank, under the accession numbers AY152822 and AY152823. Constructions for complementations Homologous complementations were performed using pBBR1MCS-5 based vectors [24]. On the pM4710 template the following primers were used to amplify the 949 bp PCR fragment carrying the hynD gene: HYDAZ04: 5¢-ATCGG GATACCGAGACACAT-3¢, HYDAZ05: 5¢-AATGGGT TGAACGAGAGTCG-3¢. First, this fragment was cloned into the HincII-digested pBluescribe plasmid (pHDS), then it was recloned into pBBR1MCS-5, as an SphI–SacI fragment (pBRHynD). pBRHupK was constructed by cloning a 2936 bp ApaI– ClaI fragment, containing the hupK gene with its regulatory region, into the ApaI–ClaI-digested pBBR1MCS-5 vector. pBRC1 was obtained by inserting the 1753 bp EcoRI– PstI fragment, containing the hypC 1 gene, from pM42-5 (pM42-5: the 6.3 kb NotI–BamHI fragment of the pM42-1 was cloned into the pBluescript SK+ NotI–BamHI sites) into EcoRI–PstI-digested pBBR1MCS-5. pBRC2 was pro- duced by insertion of the 552 bp RsaI fragment, containing the hypC 2 gene from pM47-13, into SmaI-digested pBBR1MCS-5. pBRCDE homologous complementation vector was constructed in three steps. The 1753 bp EcoRI– PstI fragment from pM42-5 was cloned into the EcoRI–PstI digested pBBR1MCS-5, which yielded the pBRC1 con- struct. The 293 bp PstI–BamHI fragment (part of the hypD gene), derived from pM1250, was ligated into the PstI– BamHI-digested pBRC1 (pBRCT2). The 1703 bp BamHI fragment (downstream region of the hypD gene and the entire hypE gene) from pM42-8 was transferred into the BamHI-digested pBRCT2, yielding pBRCDE. pBRKCDE homologous complementation vector was also constructed in three steps. The 2936 bp ApaI–ClaI fragment (harboring the hupK gene) from pM42-5 was inserted into ApaI–ClaI- digested pBBR1MCS-5, producing pBRHupK. pBRKT2 was obtained by cloning the 1069 bp ClaI–BamHI fragment (containing the hypC 1 and the 5¢ region of the hypD gene) from pM12-50 into the ClaI–BamHI-digested pBRHupK. Ó FEBS 2003 [NiFe] hydrogenase accessory proteins and assembly (Eur. J. Biochem. 270) 2219 pBRKT2 was digested with BamHI, and the 1703 bp BamHI fragment from pM42-8 was built into this vector (pBRKCDE). The homologous complementation con- structs were transformed into E. coli S17-1(kpir) strain, then conjugated into the appropriate T. roseopersicina strains. In-frame deletion mutagenesis The in-frame deletion vector constructs derived from the pK18mobsacB vector [25]. For deletion of the hupK gene, the 932 bp EcoRV–Eco47III fragment of pM42-5 (down- stream region of the hupK) was inserted into the SmaIsiteof pK18mobsacB (pDHuKA). The polished 878 bp BglI fragment from pM42-5 (the upstream homologous region) was ligated into the HindIII digested/blunted pDHuKA, resulting in pDHuK. For removal of hypC 1 and hypC 2 genes, the pDC1 and pDC2 in-frame deletion constructions were created as follows. The blunted 1423 bp SacI fragment (the downstream region of hypC 1 ) was cloned from pM42-5 into the SmaI-digested pK18mobsacB (pDC1A). The upstream region of hypC 1 was amplified with the TRHC101 (5¢-GTTATCCTGAAGCGCGATCA-3¢) and TRHC102 Table 1. Strains and plasmids used in this study. Strain or plasmid Relevant genotype or phenotype Reference or source Thiocapsa roseopersicina BBS Wild type [22] DC1B hypC 1 D, wild type This work DC1G hypC 1 D, GB11 This work DC1H hypC 1 D, GB1121 This work DC12B hypC 1 D, hypC 2 D, wild type This work DC2B hypC 2 D, wild type This work DC2G hypC 2 D, GB11 This work DC2H hypC 2 D, GB1121 This work DHKG517 hupKD, GB11 This work DHKW426 hupKD, wild type This work GB11 hynSLD::Sm Unpublished observations a GB1121 hynSLD::Sm, hupSLD::Gm Unpublished observations a M1250 hypE::Km This work M1343 hypD::Km This work M442 hypD::Km This work M4711 hynD::Km This work M539 hypF::Km [19] M646 hynL::Km This work Escherichia coli S17-1(kpir) 294 (recA pro res mod) Tp r ,Sm r (pRP4-2-Tc::Mu-Km::Tn7), kpir [23] XL1-Blue MRF¢ D(mcrA)183, D(mcrCB-hsdSMR-mrr)173, endA1, supE44, thi-1, recA1, gyrA96, relA1 lac [F¢ proAB lacI q ZDM15 Tn10 (Tet r )] Stratagene Plasmids pBBR1MCS-5 Gm r , mob + [24] pBluescribe(+) Amp r , cloning vector, ColE1 Stratagene pBluescript SK(+) Amp r , cloning vector, ColE1 Stratagene pK18mobsacB Km r , mob + , sacB + , [25] pUTKm Amp r ; Tn5-based mini transposon delivery plasmid with Km r [23] pBRHynD Gm r , pBBR1MCS-5 carrying the hynD gene This work pBRC1 Gm r , pBBR1MCS-5 carrying the hypC 1 gene This work pBRC2 Gm r , pBBR1MCS-5 carrying the hypC 2 gene This work pBRCDE Gm r , pBBR1MCS-5 carrying the hypC 1 , hypD and hypE genes gene This work pBRHupK Gm r , pBBR1MCS-5 carrying the hupK gene This work pBRKT2 Gm r , pBBR1MCS-5 carrying the hupK, hypC 1 genes and the 5¢ region of the hypD gene This work pBRKCDE Gm r , pBBR1MCS-5 carrying the hupK, hypC 1 , hypD and hypE genes This work pDC1 Km r , in-frame up and downstream homologous regions of hypC 1 in pK18mobsacB This work pDC2 Km r , in-frame up and downstream homologous regions of hypC 2 in pK18mobsacB This work pDHuK Km r , in-frame up and downstream homologous regions of hupK in pK18mobsacB This work pHDS Amp r , pBS carrying the hynD gene This work pM12-50 4.3 kb SalI fragment harboring the transposon from M1250 in pBluescript SK(+) This work pM42-1 8.1 kb BamHI fragment harboring the transposon from M442 in pBluescript SK(+) This work pM42-8 3.5 kb PstI fragment harboring the transposon from M442 in pBluescript SK(+) This work pM47-10 7 kb SphI fragment containing the transposon from M4711 in pBluescribe(+) This work a G. Ra ´ khely, Gy. Csana ´ di, G. Maro ´ ti, B. D. Fodor & K. L. Kovacs. 2220 G. Maro ´ ti et al.(Eur. J. Biochem. 270) Ó FEBS 2003 (5¢-CTAGACACATGGACAAAAGA-3¢) primers and the 1441 bp PCR product was cloned into the HindIII- digested, Klenow filled pDC1A, resulting in pDC1. The upstream and downstream region of hypC 2 was amplified by PCR using Pwo polymerase. The following primers were used: HYDAZ04, HYDAZ05, TRHC201 (5¢-TGAGCA TGGTCGCAAACACG-3¢), TRHC202 (5¢-GGACGGC TCGAGGTTTGATC-3¢). pDC2A was obtained by cloning the HYDAZ04– HYDAZ05 PCR fragment covering the 949 bp upstream homologous region of hypC 2 into the polished SalIsiteof the pK18mobsacB vector. The 951 bp downstream homo- logous region was amplified with the TRHC201 and TRHC202 primers and cloned into the HindIII-digested, Klenow filled pDC2A (pDC2). The in-frame deletion constructs were transformed into E. coli S17-1(kpir) strain, then conjugated into T. roseo- persicina BBS, GB11 and GB1121 strains resulting the in-frame deletion mutants DHKW426 (DhupK BBS), DHKG517 (DhupK GB11), DC1B (DhypC 1 BBS), DC1G (DhypC 1 GB11), DC1H (DhypC 1 GB1121), DC2B (DhypC 2 BBS), DC2G (DhypC 2 GB11), DC2H (DhypC 2 GB1121) and DC12B (DhypC 1 DhypC 2 BBS) strains. Selection for the first recombination event was based on kanamycin resist- ance. The selection for the second recombination was based on the sacB positive selection system. In T. roseopersicina 3% sucrose was efficient to induce the sacB system [25]. The in-frame deletion mutant clones were verified using PCR, Southern analysis and sequencing. RNA isolation, reverse transcription (RT) and PCR RNA was isolated using the TRIzol TM reagent (Gibco BRL), following the manufacturer’s recommendation. Prior to RT-PCR, the RNA was DNase-treated at 37 °Cfor 60minin40m M of Tris/HCl (pH ¼ 7.5), 20 m M MgCl 2 , 20 m M CaCl 2 , 4 U RNase-free DNaseI. After phenol/ chloroform extraction and ethanol precipitation, the RNA was dissolved in 20 lLofH 2 O. RT-PCR experiments were carried out as described previously [19]. The TRHC102 primer (in hypC 1 , sequence see above) was used for the reverse transcription and PCR. The TRHD04 hypE A (8334 bps) hypD hypC 1 hupK ompR envZ 2000 4000 6000 8000 Pst I Sal I Not I Sac I Bam HI Pst I Eco RV Sac I Cla I Cla I Sal I Eco 47III Eco RI Bgl I Bgl I Apa I Bam HI (5130 bps) pntA orf hypC 2 hynD tnp B 1000 2000 3000 4000 5000 BamH I Rsa I Rsa I Bam HI Bam HI Fig. 1. Identified hydrogenase accessory genes in the M442, M1250 (A) and M4711 (B) transposon mutant strains. PntA is similar to transhydro- genases, orf is a putative conserved protein, tnp seems to encode a transposase. Black triangles show the positions where the transposon was inserted. The sequences have been deposited with GenBank, accession numbers AY152822 (A) and AY152823 (B). Ó FEBS 2003 [NiFe] hydrogenase accessory proteins and assembly (Eur. J. Biochem. 270) 2221 (5¢-TTGCGGTTGTTGAGCCGCTG-3¢)servedasthe other primer in PCR. Using these primers a 524 bp fragment could be amplified. Preparation of membrane-associated and soluble protein fractions of T. roseopersicina T. roseopersicina culture (300 mL) was harvested in a Sorvall RC5C centrifuge at 7000 g. The cells were suspended in 3mLof20m M K-phosphate buffer (pH 7.0), and sonicated eight times for 10 s on ice. The broken cells were centrifuged at 10 000 g for 15 min. The debris (containing whole cells and sulfur crystals) was discarded and the supernatant was centrifuged twice at 100 000 g for 3 h [27]. The ultracentrif- ugation pellet was washed with 20 m M K-phosphate buffer (pH 7.0) and used as the membrane fraction. The super- natant was considered as the soluble fraction. Hydrogen uptake activity assay in vitro H 2 uptake, coupled to benzylviologen or methylviologen reduction, was assayed spectrophotometrically at 55 °C. The harvested cells, membrane or soluble fractions were suspended in 20 m M K-phosphate buffer (pH 7.0). Two millilitres of this mixture was placed into a cuvette, 18 lLof 20 m M benzylviologen was added, and the cuvettes were sealed with SubaSeal stoppers. The gas phase was flushed with N 2 for 5–10 min and then with H 2 for 5–10 min. Hydrogen evolution assay in vitro Sample (0.5 mL) was suspended in 1.2 mL of 20 m M K- phosphate buffer (pH ¼ 7.0) in Hypo-Vials (10 cm 3 volume, Pierce) and 1 mL of 1 m M methylviologen was added. In order to measure the activity of the Hyn enzyme selectively, cells were heat treated at 72 °C for 30 min prior to the assay. The gas phase was flushed with N 2 for 10 min, followed by the anaerobic addition of 0.5 mL of 0.1 gÆmL )1 dithionite. Samples were incubated at 40 °C for 30 min. Hydrogen production was measured by gas chromatograph [19]. Results Identification and characterization of the accessory genes Transposon-based mutagenesis was performed in order to create a mutant T. roseopersicina library and to find the hydrogenase accessory genes [19]. Six of 1600 mutant colonies showed a hydrogenase-deficient phenotype, five of which lost all hydrogenase activities and in one case (M646) the hydrogenase activity of the cells was dramatically reduced, but detectable. The M442 and M4711 strains were selected for detailed analysis. The hupK, hypC 1 , hypD and hypE genes An approximately 8.1 kb BamHI genomic fragment from the pleiotropic mutant M442 was isolated, subcloned and sequenced. The hypC 1 , hypD and hupK genes were identified in this clone (Fig. 1, Table 2). Upstream from the hupK gene, no hydrogenase-related gene could be identified, but two ORFs showed significant homology to the two-component regulatory system OmpR– EnvZ [28]. In T. roseopersicina,thehypD gene starts with GUG, and the Tn5 transposon was inserted at bp 792 of the 1146 bp-long ORF. As the BamHI fragment from M442 did not contain the whole hypD gene, an overlapping 3.5 kb PstI genomic fragment was cloned and sequenced. The hypE-type gene was found downstream from the hypD gene (Fig. 1, Table 2). In a separate hydrogenase-deficient mutant group (M1250), the transposon was inserted into the hypE gene. No additional accessory genes were found downstream from hypE (data not shown). The 8334 bp- long region was sequenced on both strands. The hynD and hypC 2 genes A 5130 bp-long chromosomal fragment surrounding the transposon in the M4711 nonpleiotropic mutant was sequenced on both strands. Two [NiFe] hydrogenase-related ORFs were found. The deduced amino acid sequence of the first ORF showed similarity to the HypC proteins (Fig. 1, Table 2) and the characteristic motif at the N-terminus of HypCs, namely M-C-(L/I/V)-(G/A)-(L/I/V)-P [10], could also be aligned. The second ORF (named hynD) encoded a putative protein, similar to the hydrogenase-specific endo- proteases of other microorganisms [2]. Multiple alignment indicated that the putative HynD was similar to the other [NiFe] hydrogenase-processing proteases, after a GTG codon (data not shown). The start codon of the hynD gene could not be identified. There was a long stretch (148 aa) upstream from this GTG without ATG in-frame, but the translated sequence was unrelated to any known protein. The codon usage of this upstream region is not character- istic of the known codon usage pattern of T. roseopersicina (among the 10 codons preceding the GTG, four are preferred at 1–10% frequency in this strain). If hynD starts at this codon, the putative HynD enzyme consists of 156 amino acids (16.6 kDa) and the transposon is inserted into Table 2. Identity between the accessory proteins of T. roseopersicina and the corresponding proteins from other organisms. Organism T. roseopersicina HupK (389 aa) HypC 1 (94 aa) HypD (381 aa) HypE (360 aa) HypC 2 (81 aa) HynD (156 aa) R. eutropha 30% (HoxV) 55% (HypC) 57% 76% 30% (HypC) 31% (HoxM) E. coli – 34% (HypC) 42% 46% 37% (HybG) 29% (HyaD) R. leguminosarum 27% 50% (HypC) 59% 61% 37% (HypC) 29% (HupD) Azotobacter sp. 26% 47% (HypC) 60% 80% 37% (HypC) 30% (HupM) 2222 G. Maro ´ ti et al.(Eur. J. Biochem. 270) Ó FEBS 2003 the hynD gene at bp 107 of the 471 bp-long gene. Thus, the hypC 2 and the hynD genes are separated by 120 bp, and they are in opposite orientation. It should be noted that the C-terminal end of HynD was slightly shorter than those of its counterparts from other microorganisms. HynD is a processing endopeptidase-like protein In the wild type T. roseopersicina, all hydrogenase activity, except that related to HynSL, could be eliminated by an appropriate heat treatment (see Materials and methods). Only heat labile hydrogenase activity could be detected in a DhynSL mutant strain (GB11). Likewise, in the mutant, in which the hynD gene was disrupted by the Tn5 insertion (M4711), no heat stable hydrogenase activity was observ- able. A series of hydrogenase activity measurements were performed using the wild type cells, the hynD::Km (M4711), the DhynSL (GB11) mutants and the complemented M4711 strain. Mutants lacking a functional hynD gene (M4711) or the heat stable [NiFe] hydrogenase, HynSL (GB11), showed the same behavior in the activity assays (Fig. 2). Comple- mentation of the hynD gene (pBRHynD) restored the heat stable HynSL hydrogenase activity to the level of the wild type control. As the in silico analysis of the putative HynD gene product clearly predicted a [NiFe] hydrogenase processing endopeptidase, it was concluded that HynD is a protease carrying out the post-translational modification of the C-terminus of the large subunit [29,30] during the maturation of the stable HynSL hydrogenase in T. roseo- persicina. Cotranscription of hupK and hypC 1 DE The hupK (hoxV) gene was separated from hypC 1 by 194 bp, the start codon of hypD was overlapping with the stop codon of hypC 1 ,andhypE started 94 bp downstream from the stop codon of hypD. The distances between the hupK, hypC 1 D and hypE genes are compatible with either an independent transcription of hupK, hypC 1 D and/or hypE,or all of these genes could be cotranscribed. In order to test this possibility, RT-PCR analysis was performed on total RNA isolated from T. roseopersicina. An mRNA species contain- ing both the hupK and hypC 1 genes was detected, which indicated the common transcriptional regulation of these genes. The transcript, however, appeared very weak (Fig. 3), and therefore, independent transcription had to be considered as well. The two possibilities were further examined in additional complementation experiments. Two constructs were made in order to complement the strain carrying a hypD::Km mutation (M442). The two constructs differed from each other in the hupK gene and its regulatory region. One of them contained the hupK-hypC 1 DE genes (pBRKCDE), and the other one contained only the hypC 1 DE genes (pBRCDE). The presence of the pleiotropic hypE gene in the constructs was necessary because of the possible polar effect of the transposon. A similar RT-PCR experiment as above showed that the hypD and hypE genes were cotranscribed (data not shown). Both constructions complemented the mutation in hypD::Km, but the comple- mentation was not complete in either case. It was signifi- cantly higher when the construct with hupK was used (18% without hupK and 43% with hupK, respectively, Table 3). These results again corroborate the presence of two sets of regulatory elements, one between hupK and hypC 1 , and one upstream from hupK. To some extent, it would explain the low complementation efficacy obtained in the hypC 1 complementation experiments, where hupK was omitted from the complementing construct (see above). Properties of the HupK protein The role of the HupK (HoxV) in the maturation process of the [NiFe] hydrogenases is unknown. Conserved regions could be recognized at the N- and C-termini, while the middle portion of the proteins appeared variable. The highest homology was found at the C-terminus and, H 2 evolution activity (arbitrary units) hynSL wt (BBS) hynD::K m complemented (M4711+ pBRHynD) hynD::K m (M4711) without heat treatment heat treated (GB11) 0 1 2 3 4 5 6 7 8 Fig. 2. Hydrogen evolution activity of the wild type and the HynD mutant T. r oseopersicina strains. The samples were or were not heat- treated before the measurements. (Strains given in Table 1.) It should be noted, that HynSL is a thermophilic enzyme, i.e. its activity increases with temperature (at least up to 80 °C) [31]. Therefore, heat treatment of the samples probably activates this hydrogenase, which explains the higher activity of the heat- treated samples. Ó FEBS 2003 [NiFe] hydrogenase accessory proteins and assembly (Eur. J. Biochem. 270) 2223 remarkably, this region showed significant identity to the HupL (hydrogenase large subunit) proteins as well, although half of the conserved cysteines were missing [32]. In-frame deletion mutagenesis was used to determine the specificity of the HupK protein. Thirty one amino acid residues in the truncated HupK originated from the N-terminus, 37 aa from the C-terminus of the protein and 13 aa came from the multiple cloning site of the pK18mob- sacB vector. The extensively shortened hupK derivative was cloned into the wild type and DhynSL (GB11) T. roseo- persicina strains. The physiological effects of the mutation on the hydrogenase enzyme activities were tested in H 2 uptake activity assays of each individual [NiFe] hydrogenase enzyme in T. roseopersicina. Approximately 90% of both HynSL and HupSL activity was lost in comparison to the wild type (Table 4). On the contrary, the soluble fraction retained almost all of its activity; around 75% of Hox activity was detectable in the HupK deleted strain, with respect to the wild type. Homologous complementation with the hupK gene (pBRHupK) fully restored the hydrog- enase activity of the cells (Table 3). This has further proven the selectivity of HupK, which is important for the formation of both functionally intact membrane-associated [NiFe] hydrogenases, but it is not involved in the maturation of the soluble Hox enzyme in this bacterium. The two HypC accessory proteins The role of the putative HypC proteins was studied by in-frame deletion mutagenesis in T. roseopersicina.Each hypC gene was deleted from the wild type, the GB11 (HynSL minus) and GB1121 (HynSL and HupSL minus) genomes individually. In addition, a double hypC mutant strain was also generated from the wild type T. roseopersicina BBS (Table 1). Hydrogenase activity assays, in uptake and evolution directions, were carried out both on membrane and soluble fractions of the various mutant strains. The absence of HypC 1 almost completely eliminated the activity of all [NiFe] hydrogenases: about 3–5% of the activities of both membrane-bound hydrogenases (Hup and Hyn), and 10% of the cytoplasmic (Hox) hydrogenase activity was detectable in the DhypC 1 mutant (Table 4). Homologous complementation with the hypC 1 gene (pBRC1), containing the hypC 1 upstream region, yielded incomplete restoration of activity: only 15% of the wild type activity was measurable (Table 3). The low complementation efficacy might be due either to the lack of the putative promoter preceding the hupK gene, or to the absence of the hypD gene in the complementing construct, i.e. an in-frame deletion of hypC 1 might also have a polar effect on the expression of hypD (M. Blokesch, Lehrstuhl fu ¨ r Mikrobiologie, Universi- ta ¨ tMu ¨ nchen, Germany). The mutation of the hypC 2 gene also affected all three hydrogenases, the HupSL and the HynSL activities decreased to 9–10% and the soluble Hox hydrogenase retained only 6% of its activity as compared to the wild type. Homologous complementation with the hypC 2 gene (pBRC2) was complete; the wild type Hup, Hyn and Hox activities of these [NiFe] hydrogenases were restored (Table 3). The results indicate that the two related putative proteins cannot replace one another in the matur- ation of the various hydrogenases. Discussion Thiocapsa roseopersicina harbors at least three hydrogenase enzymes, two of which are attached to the membrane and one that is located in the cytoplasm. Thus, it is intriguing and important to explore the functional relationship 250 500 750 1000 RT+ RT- gC bp hypC 1 hupK Fig. 3. RT-PCR analysis of the cotranscription of the hupK and hypC 1 genes. M, marker; bp, base pairs; RT+, reverse transcription was made before PCR reaction; RT–, reverse transcriptase was omitted; gC, control PCR made on genomic DNA. Table 3. H 2 uptake activities in homologous complementation experiments. The results are given as a percentage compared to the T. roseopersicina wild type strain. Complementing gene Plasmid hupKD, BBS (DHKW426) hypC 1 D, BBS (DC1B) hypC 2 D, BBS (DC2B) hypD::Km (M442) hupK pBRHupK 100 ± 8.1 – – – hypC 1 pBRC1 – 15 ± 4.3 – – hypC 2 pBRC2 – 0 100 ± 4.5 – hypC 1 DE pBRCDE – – – 18 ± 6.6 hupK, hypC 1 DE pBRKCDE – – – 43 ± 11.3 2224 G. Maro ´ ti et al.(Eur. J. Biochem. 270) Ó FEBS 2003 between the biosynthesis and maturation of the various hydrogenases. Mini Tn5 transposon mutagenesis was used to identify the hydrogenase accessory genes required for the maturation of the [NiFe] hydrogenase enzymes in this particular strain. Six independent mutant strains were isolated from a library of 1600 colonies [19]. Besides the previously identified hypF gene [19], detailed molecular investigation of the mutant strains resulted in the identifi- cation of one locus containing the hupK-hypC 1 DE accessory genes and another one, where the hypC 2 and hynD genes were found. The organization of the accessory genes in this bacterium is unusual, as the corresponding genes are frequently organized into large gene clusters in other organisms [2,6,17]. In order to examine the specificity of the auxiliary proteins, hydrogenase deletion mutant strains were generated (G. Ra ´ khely, Gy. Csana ´ di, G. Maro ´ ti, B. D. Fodor & K. L. Kova ´ cs, unpublished observations), and the effect of the accessory genes was studied through hydro- genase activity assay measurements. In three mutants the transposon was inserted into the hypD or the hypE gene abolishing all hydrogenase activities in the cells. The corresponding gene products have obviously fundamental roles in the formation of any [NiFe] hydrogenase. The physiological functions of the HynD, HupK and HypC 1 and HypC 2 proteins were investigated in detail. The hynD gene of T. roseopersicina showed a high level of homology to the ORFs encoding the specific endoproteases of the [NiFe] hydrogenases of other bacteria. These proteases have a function in one of the last steps of hydrogenase maturation, when the C-terminal end of the precursor large subunit polypeptide is cleaved, as soon as the [NiFe] heterobinuclear center with its diatomic ligands [2,6,29,30] has been successfully assembled and inserted into the active site of the enzyme. Downstream from the hupSLC genes, the hupD gene was identified, which also encodes a related putative protein, likely to be involved in the processing of the HupL subunit [15]. It is plausible to assume that HynD is involved in the maturation of the HynL protein. Indeed, in the strain harboring the Tn5 transposon-inactivated hynD gene no HynSL enzyme activity could be detected. HynSL activity was completely restored by hynD complementation. The location of the hupK gene, upstream from hypC 1 DE, is somewhat surprising because this gene has been found in the hup operon of other organisms [2]. The distance between hupK and hypC 1 raised the question of whether hupK- hypC 1 DE constituted a single operon or whether the transcription of hupK was regulated separately from hypC 1 DE. Homologous complementation experiments clearly indicated that the hypC 1 DE genes had their own regulatory element, independent from that of the hupK, but they could also be transcribed from the promoter of the hupK gene. RT-PCR analysis between the hupK and hypC 1 corroborated these conclusions. The role of HupK is ambiguous in the strains studied so far. In R. eutropha, deletion of hoxV (hupK) reduced the activity of the membrane-bound hydrogenase to 30% compared to the wild type [33]. On the contrary, inactivation of hupK led to the accumulation of the immature form of the inactive hydrogenase subunits in R. leguminosarum [34]. In T. roseo- persicina the activities of both membrane-associated [NiFe] hydrogenases (HynSL and HupSL) decreased dramatically in the absence of the HupK protein, whereas the soluble HoxEFUYH enzyme remained apparently unaffected. Remarkably, this protein does not occur in all microbes containing [NiFe] hydrogenase, hence the role of the HupK protein is still uncertain. It resembles the large subunit of the [NiFe] hydrogenases, therefore HupK has been suggested to function as a scaffolding protein during metal cofactor assembly [32]. Although our study did not uncover the precise function of HupK, this was the first demonstration that it made a selection among the various [NiFe] hydro- genases in the cell, and participated in the biosynthesis of the membrane-bound ones. HypC is a small, chaperon-like protein that participates in two protein complexes, and thus a dual function has been assigned to it. HypC interacted with the large subunit of the hydrogenase 3 (HycE) in E. coli [10] and it was recently shown to form a complex with the HypD protein [12]. In the model based on the observations in E. coli,firsttheHypC– HypD complex is formed, where the Fe gets liganded by CO and two CN with the involvement of HypF and HypE [9]. Then HypC, equipped with the Fe-CO-(CN) 2 complex, is transferred to the HycE subunit with the concomitant dissociation of HypD [12]. HypC selectively interacts with hydrogenase 3 and it can take over the functions of the homologous HybG in processing the hydrogenase 1 to some extent in E. coli [14]. The molecular phenotype of HypC mutations is strikingly different in T. roseopersicina. In our case, both HypC proteins are important for the maturation of all three hydrogenases, i.e. both of them have a task in every stage, even if they can partially substitute each other. Consequently, both HypCs are truly pleiotropic accessory proteins in T. roseopersicina. The findings in the two bacteria can be assembled into a generalized [NiFe] hydrogenase maturation scheme if we assume that two HypC proteins are needed in the ÔHypC cycleÕ [12]. In our working hypothesis one HypC interacts with HypD, while the other one holds the unprocessed large subunit protein in an open confor- mation. Iron binding and ligation occurs on the HypC– HypD complex then this metal complex (possibly without the HypC protein) is transferred to the HypC–unprocessed Table 4. Hydrogenase activities of the wild type and in-frame deletion mutant T. roseopersicina strains. H 2 uptake activities were measured on the membrane and soluble fractions, respectively. The results are given in percentage activity compared to the wild type strain (100%). Experimental error was within 10%. For the description of the strains, see Table 1. Inactivated genes Strain Activity Hyn Hup Hox None (wild type) BBS 100 100 100 hupK DHKW426 7 12 76 hupK, hynSL DHKG517 0 12 73 hypC 1 DC1B 3 5 10 hypC 1 , hynSL DC1G 0 5 12 hypC 1 , hynSL, hupSL DC1H 0 0 14 hypC 2 DC2B 9 11 6 hypC 2 , hynSL DC2G 0 11 8 hypC 2 , hynSL, hupSL DC2H 0 0 6 hypC 1 , hypC 2 DC12B 0 0 0 Ó FEBS 2003 [NiFe] hydrogenase accessory proteins and assembly (Eur. J. Biochem. 270) 2225 large subunit complex formed independently. The HypCs involved in the two separate steps can be the same proteins or homologous counterparts, which may have dissimilar affinities to the HypD and to the unprocessed large subunit of the [NiFe] hydrogenases. The difference in the affinity may determine the specificity of the various HypC chaperons. There are at least two considerations, which are compat- ible with a ÔHypC cycleÕ involving two (iso)enzymes. On the one hand, all known HypC type proteins share the N-terminal highly conserved region M-C-(L/I/V)-(G/A)- (L/I/V)-P [10], which is the sequence element essential for the interaction with both target proteins [12]. In our model, this interaction is made possible without competition for the same binding site between the HypD and the unprocessed large subunit as only the iron complex is transferred from the HypC–HypD complex to the HypC–unprocessed large subunit assembly. On the other hand, it should be noted that there are two copies of the small chaperon-like protein in every [NiFe] hydrogenase-containing microorganism stud- ied in detail, e.g. in E. coli HypC and HybG [14], in R. eutropha HypC and HoxL [33,35], in R. leguminosarum [36], R. capsulatus and Bradyrhizobium japonicum [2] HypC and HupF, and in T. roseopersicina HypC 1 and HypC 2 .Our model offers a function for both chaperons. Experimental evidence that supports the cooperativity-based model are as follows. First, in T. roseopersicina both HypC proteins are required for the biosynthesis of each hydrogenase. A similar situation was observed in R. eutropha [33,35] where a mutation in either the hypC or in the homologous hoxL resulted in the dramatic reduction but not the complete loss of membrane-bound hydrogenase activity. Second, it was shown in E. coli that the HypC–preHycE complex exists on HypD – background [12]. This demonstrated the independent formation of the HypC–HypD and the HypC–preHycE complexes in E. coli also. Third, the distinct affinity of the two chaperon-like proteins, HypC and HybG, to the target protein was demonstrated in E. coli, when both HybG and HypC proteins were expressed in HybG – background and only the HybG–HypD complex was detectable, although this experiment was not evaluated quantitatively [12]. It should be noted that this is only a working hypothesis, which can interpret the data obtained in various microbes, but further validation of the universal nature of the model is necessary. Experiments to test this model and to identify the intermediates in the various T. roseopersicina mutants are in progress. In summary, HupK is selectively involved in the biosyn- thesis of the various [NiFe] hydrogenases. In contrast, both HypCs are truly pleiotropic proteins, which are very important for the maturation of all [NiFe] hydrogenases. We propose that the two HypCs might have distinct functions in the maturation process, and they can replace each other to some extent. Acknowledgements This research is supported by EU 5th Framework Programme projects (QLK5-1999-01267, QLK3-2000-01528, QLK3-2001-01676, ICA1-CT- 2000-70026) and by domestic sources (OTKA, FKFP, OMFB, OM KFHA ´ T, NKFP). International collaboration through the EU network COST Action 841 is greatly appreciated. References 1. Vignais, P.M., Billoud, B. & Meyer, J. (2001) Classification and phylogeny of hydrogenases. FEMS Microbiol. Rev. 25, 455–501. 2. Cammack, R., Frey, M. & Robson, R., eds. (2001) Hydrogen as a fuel. Learning from Nature. Taylor & Francis, London. 3. Cammack, R., Fernandez, V.M. & Hatchikian, E.C. (1994) Nickel iron hydrogenases. Methods Enzymol. 243, 43–68. 4. Volbeda, A., Charon, M H., Piras, C., Hatchikian, E.C., Frey, M. & Fontecilla-Camps, J.C. (1995) Crystal structure of the nickel- iron hydrogenase from Desulfovibrio gigas. Nature 373, 580–587. 5. Maier, T. & Bo ¨ ck, A. (1996) Nickel incorporation into hydro- genases. In Advances in Inorganic Biochemistry (Hausinger, R., Eichlorn, G.L. & Marzilli, L.G., eds), pp. 173–192.VHC Publishers Inc., New York. 6. Casalot, L. & Rousset, M. (2001) Maturation of the [NiFe] hydrogenases. Trends Microbiol. 9, 228–237. 7. Paschos, A., Glass, R.S. & Bo ¨ ck, A. (2001) Carbamoylphosphate requirement for synthesis of the active center of [NiFe]-hydro- genases. FEBS Lett. 488, 9–12. 8. Paschos, A., Bauer, A., Zimmermann, A., Zehelein, E. & Bo ¨ ck, A. (2002) HypF, a carbamoyl phosphate-converting enzyme involved in [NiFe] hydrogenase maturation. J. Biol. Chem. 277, 49945– 49951. 9. Reismann, S., Hochletitner, E., Wang, H., Pachos, A., Lottspeich, F., Glass, R.S. & Bo ¨ ck, A. (2003) Taming of a poison: biosynthesis of the NiFe-hydrogenase cyanide ligands. Science 299, 1067–1070. 10. Magalon, A. & Bo ¨ ck, A. (2000) Analysis of the HypC-HycE complex, a key itermediate in the assembly of the metal center of the Escherichia coli hydrogenase 3. J. Biol. Chem. 275, 21114– 21120. 11. Magalon, A., Blokesch, M., Zehelein, E. & Bo ¨ ck, A. (2001) Fidelity of metal insertion into hydrogenases. FEBS Lett. 499, 73–76. 12. Blokesch,M.&Bo ¨ ck, A. (2002) Maturation of [NiFe] hydro- genases in Escherichia coli: the HypC cycle. J. Mol. Biol. 324, 287–296. 13. Wolf, I., Buhrke, T., Dernedde, J., Pohlmann, A. & Friedrich, B. (1998) Duplication of hyp genes involved in maturation of [NiFe] hydrogenases in Alcaligenes eutrophus H16. Arch. Microbiol. 170, 415–419. 14. Blokesch, M., Magalon, A. & Bo ¨ ck, A. (2001) Interplay between the specific chaperone-like proteins HybG and HypC in matura- tion of hydrogenases 1, 2, and 3 from Escherichia coli. J. Bacteriol. 183, 2817–2822. 15. Colbeau, A., Kova ´ cs, K.L., Chabert, J. & Vignais, P.M. (1994) Cloning and sequencing of the structural (hupSLC)andaccessory (hupDHI) genes for hydrogenase biosynthesis in Thiocapsa roseopersicina. Gene 140, 25–31. 16. Ra ´ khely, G., Colbeau, A., Garin, J., Vignais, P.M. & Kova ´ cs, K.L. (1998) Unusual organization of the genes coding for HydSL, the stable (NiFe) hydrogenase in the photosynthetic bacterium Thiocapsa roseopersicina BBS. J. Bacteriol. 180, 1460–1465. 17. Friedrich, B. & Schwartz, E. (1993) Molecular biology of hydro- gen utilization in aerobic chemolithotrophs. Annu. Rev. Microbiol. 47, 351–383. 18. Tamagnini, P., Axelsson, R., Lindberg, P., Oxelfelt, F., Wu ¨ ns- chiers, R. & Lindblad, P. (2002) Hydrogenases and hydrogen metabolism of cyanobacteria. Microbiol. Mol. Biol. Rev. 66, 1–20. 19. Fodor, B., Ra ´ khely, G., Kova ´ cs, A ´ .T. & Kova ´ cs, K.L. (2001) Transposon mutagenesis in purple sulfur photosynthetic bacteria: Identification of hypF, encoding a protein capable to process [NiFe] hydrogenases in a, b and c subdivision of proteobacteria. Appl. Environ. Microbiol. 67, 2476–2483. 20. Kova ´ cs, K.L., Bagyinka, C., Bodrossy, L., Csa ´ ki, R., Fodor, B., Gyo ¨ rfi, K., Hancza ´ r, T., Ka ´ lma ´ n, M., O ¨ sz, J., Perei, K., Polya ´ k, 2226 G. Maro ´ ti et al.(Eur. J. Biochem. 270) Ó FEBS 2003 B., Ra ´ khely, G., Taka ´ cs, M., To ´ th, A. & Tusz, J. (2000) Recent advances in biohydrogen research. Pflugers Arch. 439, R81–R83. 21. Ra ´ khely, G. & Kova ´ cs, K.L. (1996) Plating hyperthermophilic archea on solid surface. Anal. Biochem. 243, 181–183. 22. Bogorov, L.V. (1974) The properties of Thiocapsa roseopersicina BBS, isolated from an estuary of the White Sea. Mikrobiologija 43, 326–332. 23. Herrero,M.,Lorenzo,V.&Timmis,K.N.(1990)Transposon vectors containing non-antibiotic resistance selection markers for cloning and stable chromosomal insertion of foreign genes in gram-negative bacteria. J. Bacteriol. 172, 6557–6567. 24. Kovach, M.E., Elzer, P.H., Hill, D.S., Robertson, G.T., Farris, M.A., Roop, R.M.I.I. & Peterson, K.M. (1995) Four new derivatives of the broad-host-range cloning vector pBBR1MCS, carrying different antibiotic-resistance cassettes. Gene 166, 175–176. 25. Scha ¨ fer,A.,Tauch,A.,Jager,W.,Kalinowski,J.,Thierbach,G.& Pu ¨ hler, A. (1994) Small mobilizable multi-purpose cloning vectors derived from the Escherichia coli plasmids pK18 and pK19: selection of defined deletions in the chromosome of Corynebacterium glutamicum. Gene 145, 69–73. 26.Sambrook,J.,Maniatis,T.&Fritsch,E.F.(1989)Molecular Cloning: a Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. 27. Hancza ´ r, T., Csa ´ ki, R., Bodrossy, L., Murrell, J.C. & Kova ´ cs, K.L. (2002) Detection and localization of two hydrogenases in Methylococcus capsulatus (Bath) and their potential role in methane metabolism. Arch. Microbiol. 177, 167–172. 28. Cai, S.J. & Inouye, M. (2002) EnvZ–OmpR interaction and osmoregulation in Escherichia coli. J. Biol. Chem. 277, 24155–24161. 29. Theodoratou, E., Paschos, A., Mintz-Weber, S. & Bo ¨ ck, A. (2000) Analysis of the cleavage site specificity of the endopeptidase involved in the maturation of the large subunit of hydrogenase 3 from Escherichia coli. Arch. Microbiol. 173, 110–116. 30. Theodoratou, E., Paschos, A., Magalon, A., Fritsche, E., Huber, R. & Bo ¨ ck, A. (2000) Nickel serves as a substrate recognition motif for the endopeptidase involved in hydrogenase maturation. Eur. J. Biochem. 267, 1995–1999. 31. Gogotov, I.N., Zorin, N.A., Serebiakova, L.T. & Kondratieva, E.N. (1978) The properties of hydrogenase from Thiocapsa roseopersicina. Biochim. Biophys. Acta 523, 335–343. 32. Imperial, J., Rey, L., Palacios, J.M. & Ruiz-Argu ¨ eso, T. (1993) HupK, a hydrogenase-ancillary protein from Rhizobium leguminosarum, shares structural motifs with the large subunit of [NiFe] hydrogenases and could be a scaffolding protein for hydrogenase metal cofactor assembly. Mol. Microbiol. 9, 1305–1306. 33. Bernhard,M.,Schwartz,E.,Rietdorf,J.&Friedrich,B.(1996) The Alcaligenes eutrophus membrane-bound hydrogenase gene locus encodes functions involved in maturation and electron transport coupling. J. Bacteriol. 178, 4522–4529. 34. Brito, B., Palacios, J.M., Hidalgo, E., Imperial, J. & Ruiz- Argu ¨ eso, T. (1994) Nickel availability to pea (Pisum sativum L.) plants limits hydrogenase activity of Rhizobium leguminosarum bv. viciae bacteroids by affecting the processing of the hydrogenase structural subunits. J. Bacteriol. 176, 5297–5303. 35. Dernedde, J., Eitinger, T., Patenge, N. & Friedrich, B. (1996) hyp gene products in Alcaligenes eutrophus are part of a hydrogenase- maturation system. Eur. J. Biochem. 235, 351–358. 36. Rey, L., Murillo, J., Hernando, Y., Hidalgo, E., Cabera, E., Imperial, J. & Ruiz-Argu ¨ eso, T. (1993) Molecular analysis of a microaerobically induced operon required for hydrogenase synthesis in Rhizobium leguminosarum biovar viciae. Mol. Microbiol. 8, 471–481. Ó FEBS 2003 [NiFe] hydrogenase accessory proteins and assembly (Eur. J. Biochem. 270) 2227 . Accessory proteins functioning selectively and pleiotropically in the biosynthesis of [NiFe] hydrogenases in Thiocapsa roseopersicina Gergely Maro ´ ti, Barna D protein, likely to be involved in the processing of the HupL subunit [15]. It is plausible to assume that HynD is involved in the maturation of the HynL protein. Indeed, in the strain harboring the. T. roseopersicina and to understand their physio- logical roles. The determination of the specificity of the accessory proteins is a challenging exercise in this micro- organism because of the