Analysis of the transcarbamoylation-dehydration reaction catalyzed by the hydrogenase maturation proteins HypF and HypE Melanie Blokesch, Athanasios Paschos,* Anette Bauer, Stefanie Reissmann, Nikola Drapal and August Bo¨ck Department Biologie I, Mikrobiologie, Ludwig-Maximilians-Universita ¨ tMu ¨ nchen, Mu ¨ nchen, Germany The h ydrogenase maturation proteins HypF and HypE catalyze the synthesis of the CN ligands of the active site iron of the NiFe-hydrogenases using carbamoylphosphate as a substrate. HypE protein from Escherichia coli was purified from a transformant overexpressing the hypE gene from a plasmid. Purified HypE in gel filtration experiments behaves predominantly as a monomer. It does not contain statisti- cally significant amounts o f metals or o f cofactors absorbing in the UV and visible light range. The protein displays low intrinsic ATPase activity with ADP and phosphate as the products, the apparent K m being 25 l M and the k cat 1.7 · 10 )3 s )1 . Removal of the C-terminal cysteine r esidue of HypE which accepts the carbamoyl mo iety from HypF affected the K m (47 l M ) but not significantly the k cat (2.1 · 10 )3 s )1 ). During the carbamoyltransfer reaction, HypE and HypF enter a complex which is rather tight at stoichiometric ratios of the two proteins. A mutant HypE variant was generated by amino acid replacements in the nucleoside triphosphate binding region, which showed no intrinsic ATPase activity. The variant was a ctive as an acceptor in the transcarbamoylation reaction but did not dehydrate the thiocarboxamide to the thiocyanate. The results obtained with the HypE variants and also with mutant HypF forms are integrated to explain the co mplex reaction pattern of protein HypF. Keywords: NiFe hydrogenase; maturation; CN ligand syn- thesis; hypE mutations; carbamoyl transfer. Escherichia coli possesses four hydrogenases (Hyd1–Hyd4) which are all members of the N iFe c lass [1,2]. In these enzymes, the bimetallic active centre is hooked to the protein via four cysteine thiolates whereby two of them act as ligands bridging the iron and the nickel (for a review, see [3]). The most intriguing feature, however, is that the iron carries three diatomic, nonprotein ligands which in the classical case consist of two cyanides and one carbon monoxide [4,5]. The NiFe metal centre i s positioned in the interior of the large subunit close to its interface with the small subunit. In addition to the operons coding for the structural proteins of the four hydrogenases there are six genes, designated hyp, whose products have a f unction in the maturation of the enzymes. Most o f them a ct pleiotropically in the synthesis of all four hydrogenases, in particular in the synthesis and insertion of the metal centre [6,7]. Two of the products of the hyp genes, namely HypA and HypB, are involved in nickel insertion [8,9]. From the other four hyp gene products, HypF and HypE have a function in the synthesis of the cyanide ligands. HypF functions as a carbamoyl transferase using carbamoylphosphate as a substrate and transferring the carboxamido moiety in an ATP-dependent reaction to the thiolate of the C-terminal cysteine of the HypE protein yielding a protein-S-carbox- amide [10–12]. Subsequent dehydration of the carboxamide residue via ATP dependent activation of the oxygen and dephosphorylation leads to HypE-thiocyanate. Chemical model reactions demonstrated that the cyano group can be nucleophilically transferred to an iron complex [12]. The origin of the carbonyl group of the Fe ligands in the E. coli hydrogenases is still unresolved. It is also open a s to w hether the ligandation of iron takes place at the hydrogenase large subunit apoprotein, which had accepted the iron before, or whether i t o ccurs at some scaffold protein from which the fully substituted m etal is then transferred to the target protein. Our present information supports the latter model. Arguments are that in cells deprived of carbamoylphos- phate a complex accumulates which consists of the matur- ation proteins HypC and HypD [13]. This complex is resolved in a time-dependent manner upon supply of a carbamoyl source delivered from citrulline. Moreover, the complex occurs in two electrophoretically different forms, a faster migr ating one from cells lack ing carbamoylphosphate and a slower one when carbamoylphosphate is present but the apoprotein of the large subunit is a bsent. On the basis of these results it was speculated that the site of Fe ligandation is at the HypC–HypD complex from where it is transferred to the large subunit apoprotein. HypC was supposed to be involved in this transfer as it also undergoes complex formation with the precursor of the large subunit [14,15]. Protein HypE thus possesses a key role in the process. It accepts the carbamoyl residue, dehydrates it to the cyano moiety and appears to transfer it to the iron. In this communication we describe the p rocedure for the purifica- tion of the HypE protein, its chemical properties and kinetic Correspondence to A. Bo ¨ ck, Department Biologie I, M ikrobiologie, Ludwig-Maximilians-Universita ¨ tMu ¨ nchen, Maria-Ward-Strasse 1a, D-80638 Mu ¨ nchen, Germany. Fax: + 49 89 218063857, Tel.: + 49 89 21806116, E-mail: august.boeck@lrz.uni-muenchen.de *Present a ddress: Department of Biology, McMaster University, 1280 Main St. West, Hamilton, Ontario, L8S 4K1, Canada. (Received 11 May 2004, revised 29 June 2004, accepted 7 July 2004) Eur. J. Biochem. 271, 3428–3436 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04280.x characterization relative to the substrate ATP. The analysis of mutant variants of HypE demonstrates that the carb- amoylation of th e C-terminal thiol and the d ehydration of the S-carboxamide to the thiocyanate are independent reactions and that HypE p er se is a ble to catalyze the dehydration reaction. In addition, it is shown that during the carbamoyltransfer reaction HypE forms a complex w ith the HypF protein. The results reported for HypE and also for selected mutant variants of HypF are integrated into a model explaining the complex reaction pattern of the two proteins. Experimental procedures E. coli strains, plasmids and growth conditions E. coli strain MC4100 [16] was u sed a s wild-type a nd DH 5a [17] as host in transformations. DHP-E is a derivative of MC4100, which carries an in-frame deletion in the hypE gene [6]. The cultures were g rown at 37 °C in Luria broth [18] or in the buffered rich medium (TGYEP) described by Begg et al. [19]. Aerobic growth was achieved in rigorously shaken Erlenmeyer flasks, anaerobic g rowth in standing screw c ap flasks filled to the top. For the maintenance of t he plasmids, ampicillin was added at a concentration of 100 lgÆmL )1 . Plasmid pHypE was constructed by removing a 3006 bp large HpaI-StyI fragment, which contains most of the fhlA gene from plasmid pSA3 [20], and by religation after treatment w ith Klenow e nzyme. It had a size of 5156 bp and carried the hypE gene. pHypE was t hen employed to construct a variant lacking the codon for the C-terminal cysteine. It was achieved by inverse PCR [21] e mploying the overlapping primers cys336del-forward (5¢-C CGCGG AT ATGATAATAAAATTCTAAATCTCCTATAG-3¢)and cys336del-reverse (5¢-TCATATCCGCGG AAGCGGTT CGGCGTGTGGTAAATC-3¢) which harbour a SacII restriction site (bases in bold face letters) at their 5¢-ends. After the PCR reaction, DpnI was added to the mixture to digest the matrix plasmid pHypE. The PCR fragment was purified via p assage over a QIAquick Sp in Column (Qiagen GmbH, Hilden, Germany) and used directly to transfrom strain DH5a following the method of Ansaldi et al.[21]. The selection of accurate clones was accomplished by mini- preparation followed by Sac II digestion a nd its a uthenticity was confirmed by sequencing. For o verproduction of the wild-type HypE protein, a plasmid was constructed by excision of a 1.6 kb KpnI-MluI fragment f rom plasmid pSA3, treatment with Klenow enzyme to remove the protruding ends and by cloning into the SmaI r estricted vector pT7-7. I n this w ay the GTG start codon of the hypE gene was out of frame relative to an A TG of the vector. The plasmid was designated pTE-C2. I t represents a derivative of pT7-7 co ntaining a (ÔhypD-hypE- fhlAÕ) gene fragment from the E. coli chromosome. N-Terminal amino acid sequencing of the purified protein revealed the sequence MNNIQLAHG, which is in accord- ance with the observation made by Lutz et al.[22]and Jacobi et al. [6] that the translation of the hypE gene initiates at the GTG cod on 42 bases upstream of the previously assumed ATG codon. The GTG codon overlaps with the TGA termination codon of hypD [6]. Hence the hypE gene codes for a protein with 336 amino acids and a molecular mass of 35.1 kDa. The C-terminal a mino acid is therefore Cys336 rather than Cys322 as previously specified [12]. For overexpression of the hypED gene it was excised from plasmid pHypED by restriction with MluI, trea tment with Klenow enzyme and SpeI digestion and cloned into plasmid pTE-C2 via replacement of a HindIII-fragment that was treated with the Klenow enzyme and su bsequently cleaved with SpeI. The plasmid was designated pTEC-ED. Plasmid pTE-C2 was also used to construct hypE gene variants coding for products with an amino acid exchange in the nucleotide binding site of HypE. The variant containing the D83N replacement was obtained by inverse PCR employing primers t hat carried the desired mutation [23]. The plasmids harbouring the wild-type and mutant hypF genes have been described before [11]. For all constructions, the Expand High Fidelity PCR System from Roche Diagnostics GmbH (Mannheim, Germany) was employed. Amplified fragments generated by use of overlapping primers w ere purified by passage o ver a QIAquick Spin Column (Qiagen GmbH, Hilden, Ger- many) and used directly to transform strain DH5a.The authenticity of all constructs was verified by DNA sequen- cing using an ABI PRISM TM 310 sequencer (PE Applied Biosystems, Weiterstadt, Germany). Purification of HypE and HypF proteins Overproduction of HypE and of i ts derivatives took place in E. coli BL21(DE3) [24] transformed w ith the respective plasmid. The following proced ure was developed for purification of HypE from the wild-type and essentially the same could be adopted for the mutant variants. The transformants w ere grown aerobically in LB-medium in 2-L Erlenmeyer flasks a t 37 °C until the culture reached an A 600 of 1. The expression was initiated by the addition of 0.5 m M isopropyl thio-b- D -galactoside follo wed by a furthe r 3-h incubation period. The cells were collected by centrifugation at 3000 g 1,2 , washed in a buffer containing 50 m M Tris/HCl, 1,2 pH 7.4, centrifuged again and taken up in 1 : 10 of the volume of 50 m M Tris/HCl, 3 pH 7.4, 10 m M magnesium acetate, 50 m M sodium chloride, 0.1 m M dithiothreitol and 0.5 lgÆmL )1 each of leupeptin and p epstatin. A fter addition of 20 lgÆmL )1 each of phenylmethylsulfonyl fluoride and DNAse I, the cells were broken by a passage through a French Press cell at 118 Mpa. The crude extract was clarified by centrifugation (30 000 g for 30 min) and the supernatant was loaded on a 35 mL DEAE-Sepharose Fast Flow Column (Pharmacia, Freiburg, Germany), which had been equilibrated with 50 m M Tris/HCl, 4 pH 7.4, 10 m M magnesium acetate, 50 m M sodium chloride, and 0.1 m M dithiothreitol. Elution was performed with a linear gradient of sodium chloride reaching from 50 to 350 m M at a flow rate of 60 mLÆh )1 . The separation was followed via SDS/PAGE of each fraction. HypE-containing fractions were sampled, brought to an ammonium sulfate concentration of 30% saturation andslowlystirredat0°C for 0.5 h. The precipitate developed was collected by centrifugation at 15000 g for 30 min. The precipitate was dissolved in a minimum of a buffer containing 50 m M Tris/HCl 5 pH 7.4, 10 m M magnesium Ó FEBS 2004 Hydrogenase maturation protein HypE (Eur. J. Biochem. 271) 3429 acetate, 100 m M sodium chloride, 0.1 m M dithiothreitol, dialyzed against the same buffer a nd subjected to gel filtration o ver a HiL oad TM 16/60 Superdex TM 75 pg column (1.6 · 60 cm) (Pharmacia, Freiburg, Germany) at a flow rate of 60 mLÆh )1 . F ractions containing apparently homo- genous HypE were sampled, dialyzed against the same buffer containing 50% glycerol (v/v) and stored at )20 °C. The purification of wild-type HypF protein has been described [11]. The purified HypF mutant proteins investi- gated were obtained employing an identical protocol. Electrophoretic separations Separation of proteins under denaturing conditions was conducted by SDS/PAGE employing gels made up of 10% or 12.5% polyacrylamide [25] and following the sample denaturation condition indicated. Electrophoresis took place at room temperature at a voltage of 150 V. For the immunological detection of HypE protein, the separated proteins were transferred onto a nitrocellulose membrane (BioTrace NT; P all Corp., Dreieich, Germany) 6 ,stainedwith amidoblack and the m embrane was subjected t o a standard immunoblotting procedure [14]. Polyclonal antibodies directed against HypE protein or H ypF protein were used in dilutions of 1 : 500 and 1 : 1000, respectively. D etection of the a ntibody–antigen complex on the membrane occurred by decoration with horseradish peroxidase cou- pled to Staphylococcus aureus protein A (dilution 1 : 3000) and by detection with the Lumi-Light Western Blotting Substrate (Roche Diagn ostics GmbH, Mannheim, G er- many) via exposure to WICORexÒ B+, Medical X-ray screen films. Separation of proteins under nondenaturing conditions was achieved with the procedure described previously [14]. Transcarbamoylation/dehydration assays The assay was performed by m ixing the HypE and HypF proteins at the i ndicated concentrations in buffer containing 50 m M Tris/HCl, 7 pH 7.5, 100 m M KCl, 5 m M magnes- ium acetate, 0.1 m M each of ATP and 14 C-labelled car- bamoylphosphate in the presence or absence 0.1 m M dithiothreitol. The reaction was performed at 25 °Cfor 10–30 min. The radioactivity transferred to H ypE was either determined by filtrating the samples through nitro- cellulose filters and subsequent scintillation counting of the label r etained by the filters as described previously [12] or by separating the m ixture in polyacrylamide gels. To detect the 14 C-labelled forms of HypE, non denaturing PAGE [14] and a Ômild -denaturingÕ SDS/PAGE was employed. In the latter case, the proteins of the reaction mixture were mixed with sample buffer containing dithiothreitol at a final concen- tration of 100 m M ,heatedfor10minto56°Cand subjected to SDS/PAGE afterwards. All separations were conducted in the cold room at a voltage of 110 V o r less. After the separation, the p roteins were transferred onto nitrocellulose membranes that were dried and exposed to Tritium Storage Phosphor Screen Cassettes (Amersham Biosciences, Freiburg, Germany). The radioactivity of the screens w as scanned with a Storm 840 PhosphoImager (Amersham Biosciences, F reiburg; Germany) and data were analyzed using IMAGEQUANT 5.2. Determination of the ATPase activity The hydrolysis of ATP by HypE was followed in assays containing 50 m M Tris/HCl 8 pH 7.4, 100 m M KCl, 5 m M MgCl 2 ,0.1m M dithiothreitol, 50 lgÆmL )1 of bovine serum albumin and HypE protein at the indicated concentration. The reaction took place at 25 °C in a final volume of 100 lL. Ten microliter samples were taken, mixed with 500 lL (5%) charcoal suspended in 50 m M KH 2 PO 4 for 30 s and clarified by centrifugation. The radioactivity of 100 lL samples of the supernatant was determined in a Liquid Scintillation Analyzer Tri-Carb 2100TR (Canberra Packard GmbH, Dreieich, Germany). Alternatively, ATP hydrolysis was followed in reaction mixtures of identical composition except containing [ 32 P]ATP[aP]. After incubation, 1-lL samples were spotted onto polyethyleneimine plates (Merck, D armstadt, Ger- many) which were developed with 0.5 M KH 2 PO 4 ,pH3.4. After drying, the plates were exposed to Storage Phosphor Screen Cassettes (Molecular Dynamics) and the radioactiv- ity of the screens was quantified in a Storm 840 Phospho- Imager (Amersham Biosciences, Freiburg, Germany). Enzymes and special chemicals Enzymes for restriction and modification of DNA were purchased from one of the following companies: MBI Fermentas (St. L eon-Rot, Germany), N ew England Biolabs (Frankfurt, Germany), Stratagene (Heidelberg, Germany) Roche Molecular Biochemicals (Penzberg, Germany) and Eurogentec (Ko ¨ ln, Germany). Oligonucleotides were syn- thesized by MWG (Ebersberg, Germany) o r Interactiva (Ulm, Germany). Carbamoylphosphate was obtained either from Sigma (Deisenhofen, Germany) or ICN Biomedical Inc. (Eschwege, Germany). It was provided in the form of the dilithium salt and had a purity between 90 and 95%. 14 C-labelled carbamoylphosphate was purchased from American Radiolabelled Chemicals Inc. (St. Louis, MO, USA) at a specific radioactivity of 7 mCi Æmmol )1 .Itwas dissolved in water and distributed into small aliquots that were stored at )80 °C. Results Sequence characteristics of the HypE protein The in silico analysis of the amino acid sequence of the HypE protein revealed similarities with those of the ThiL protein (thiamin monophosphate kinase), the SelD protein (monoselenophosphate synthetase) and the PurM protein (aminoimidazole ribonucleotide synthetase) [26]. The simi- larity embraces several positions and sequence stretches that are assumed to be involved in the binding of ATP. The determination of the crystal structure of the PurM protein from E. coli in complex with its substrate then showed that these residues indeed are involved in liganding ATP, forming a novel nucleoside triphosphate binding site [26]. All members of the HypE family, in addition, possess the strongly conserved C-terminal tetrapeptide PR[I/V]C [12]. The sequences of HypE and of SelD could be modelled into the coordinates of the Pu rM 3D structure, which suggested that the three proteins might catalyze reactions 3430 M. Blokesch et al. (Eur. J. Biochem. 271) Ó FEBS 2004 mechanistically similar to that of PurM, namely an ATP- dependent dehydration. Indeed, HypE from E. coli dehy- drates the carboxamido residue linked t o the C-terminal cysteine of HypE to the protein thiocyanate [12]. Purification and properties of the HypE protein from E. coli HypE protein was overproduced in the transformant of E. coli BL21(DE3) harbouring plasmid pTE-C2 and puri- fied following the procedure outlined under Experimental procedures. In short, th is involved breakage of the cells by passage through a French Press cell, preparation of the 30 000 g supernatant, anion exchange chromatography over a DEAE column followed by ammonium sulfate precipitation and gel filtration. Figure 1 gives the path of purification as visualized by SDS/PAGE of the pooled fractions after each step and staining with Serva-Blue R-250. In a typical purification run about 14 mg apparently homogenous HypE protein were obtained from 4 g of cells (wet weight). Essentially the same purification protocol could be employed to purify the following two m utant variants of the HypE protein: (a) HypED which lack s the C-terminal cysteine residue and (b) HypE[D83N] in which the aspartate shown to participate in ATP binding [26] is replaced by asparagine. Inductively coupled atomic emission spectroscopy of the protein showed that the purified preparation does not contain metal in a significant amount. The UV and visible spectrum also did not indicate the existence of a cofactor absorbing in the wavelength range between 250 and 500 nm (results not shown). Upon gel filtration of the protein on a calibrated HiLoad TM 16/60 Superdex TM 75 pg column the majority migrated as an apparent monomer. However, there w as always HypE protein i n the elution position o f the apparent homodimer. This had already been observed in the electrophoretic separation in polyacrylamide gels under nondenaturing conditions [12]. To assess whether this putative dimer is the product of a chemical linkage via disulfide bridging or the result of a monomer d imer equilibrium, samples of the purified preparations of wild-type HypE, of HypE[D83N] and HypED were incubated in the presence of 50 m M dithio- threitol and subjected to SDS/PAGE under nondenaturing conditions (Fig. 2A). The preincubation converted the apparent homodimer present in the wild-type and HypE[D83N] preparations into the monomeric form. Because HypED, on the other hand, was devoid of the homodimeric form these results suggest that the homodimer is the result of disulfide bridging via the C-terminal cysteine residues. This conclusion is supported by the results of carbamoylation of the HypE forms by HypF protein in the presence of carbamoylphosphate and ATP (Fig. 2 B). Incu- bation with dithiothreitol grossly increased the capacity to accept the label from [ 14 C]carbamoylphosphate, which is in accord with the notion that the C-terminal cysteine is required for activity [12]. Although it is still open whether the disulfide-bridged dimer is o f biological significance, the results stress the necessity for the reductive activation of HypE in order to obtain maximal acceptor activity in vitro. HypE protein possesses intrinsic ATPase activity As suggested by the sequence signatures, HypE protein possesses ATPase activity delivering ADP and inorganic Fig. 1. Purification of HypE protein from an overexpressing strain as followed by SDS/PAGE (12.5%) of the pooled fractions of each step. Lane 1: molec ular mass ( kDa) s tandard ( b-galactosidase, bovine serum albumin, ovalbumin, lactate dehydrogenase, endonuclease Bsp98I, b-lactoglobulin, lysozyme); lane 2: cell lysate of BL21(DE3)/pTE-C2 before inductio n; lane 3: cell lysate of BL21(DE3)/pTE -C2 after induction with 1 m M isopropyl thio-b- D -galactoside; lane 4: S30 extract; lane 5: sediment of the 3 0 000 g centrifugation; lane 6: p ooled fractions after DEAE Sepharose chromatography and ammonium sulfate p recipitation up to 30% saturation; lane 7: HypE prote in after gel filtration (Superdex-75) and dialysis. The gel was stained for pro- teins with Serva Blue R-250. Fig. 2. Migration behaviour and activity of wild -type HypE and m uta nt variants in 10% nondenaturing SDS gels after preincubation with dithiothreitol. (A) Serva Blue R-250 stained SDS gel. Lane 1: molecular mass standard; lanes 2 and 3: wild -type HypE p rotein; lanes 4 an d 5: HypE[D83N]; lanes 6 and 7: HypED. In lanes 2, 4 and 6 proteins (12 l M ) were preincubated with 50 m M dithiothreitol fo r 1h on ice. The monomeric and dimeric formsofHypEanditsvariantsare indicated by arrows. T he chemical basis of the migration in two fo rms is unknown (lan es 3, 5 and 7). (B) Determination of 14 C-labelled HypE protein and its variants by binding to nitrocellulose filters [11]. Lanes 2–7 as in (A). Results are average s of three in depe ndent experiments ± standard deviation. Ó FEBS 2004 Hydrogenase maturation protein HypE (Eur. J. Biochem. 271) 3431 phosphate as products (not shown). The hydrolysis rate is linear with time and is not influenced by the presence of carbamoylphosphate (data not presented). The D83N exchange leads to the abolition of activity (not sho wn), whereas the HypED variant displays about half of the activity of the wild-type protein under the assay conditions employed. The following kinetic constants of the ATP cleavage reaction were determined for wild-type HypE in five independent determinations: K m: 25 ± 1.8 l M ; K cat 1.7 ± 0.16 · 10 )3 s )1 .TheHypED mutant protein, on the other hand, showed considerable variations in the kinetic assays indicating stability problems. The average values obtained in six independent determinations were K m : 47 ± 13.3 l M ; K cat 2.1 ± 2.6 · 10 )3 s )1 . Analysis of the transcarbamoylation reaction catalyzed by HypF The formation of the HypE-thiocyanate involves first the carbamoylation of the C -terminal c ysteine of HypE b y interaction with HypF, then the release of HypF and the subsequent dehydration of the protein th iocarboxamide to protein t hiocyanate [12]. T o r esolve these partial reactions it was necessary to develop a method via which the carbamo- ylated form of HypE could be differentiated from the cyanated protein. Use w as made of the previous observation that HypE-CN is labile in the presence of thiols. When incubated in the presence of 1 m M dithiothreitol for 15 min at 40 °C the yield of HypE-CN had been much lower in comparison to samples incubated i n the absence of dithio- threitol [12]. Alteration of the incubation conditions to 10 min at 56 °C in the presence of 100 m M dithiothreitol (Ômild-denaturing Õ conditions) lead to the complete disap- pearance of the cyanated form after SDS/PAGE (Fig. 3B, lane 1) whereas it was still distinctly resolved upon electrophoresis under nondenaturing conditions in the absence of dithiothreitol and omission o f heating of the mixture (Fig. 3A, lan e 1). On the other hand, samples from assays that were blocked in the dehydration reaction because of the inclusion of ADP-CH 2 -P instead of ATP in the reaction [ 12] exhibited the presence of the HypE- thiocarboxamide after SDS gel electrophoresis (Fig. 3B, lane 2). It is striking that nondenaturing PAGE does not resolve the presence of HypE-thiocarboxamide. A possible reason could be that HypE-thiocarboxamide and HypE- thiocyanate might possess differential stabilities under the conditions of electrophoresis, in particular at the alkaline pH of the gels. To follow t his assumption the HypE protein was carbamoylated by HypF in the presence of ADP-CH 2 - Pand[ 14 C]carbamoylphosphate and the substrates were removed by filtration. Parallel samples were incubated at different pH values and the retention of the radioactivity bound to HypE was a ssessed ( Fig. 3C) by Ômild-denaturing Õ SDS/PAGE as described above. It is evident that alkaline pH leads to the loss of the thio carboxamide moiety; intriguingly, the apparent hydrolysis of the thiocarboxamide requires native H ypE prote in as it is fully stable when the samples are denatured in SDS sample buffer containing 100 m M dithiothreitol and s eparated in SDS g els possessing the same pH. The radioactive material migrating on the top of the nondenaturing gel (Fig. 3A) coincides with a signal in immunoblots detected bo th with anti-HypE and anti-HypF antibodies (not shown). It therefore may denote a complex between the HypE and HypF proteins that might constitute an inte rmediate in the transcarbamoylation reaction. To follow this a ssumption, transcarbamoylation/dehydration reactions were carried out at different ratios between HypF and HypE proteins and the products were separated by nondenaturing PAGE (Fig. 4). (An incubation time was chosen in which a 10-fo ld lower mount of HypF protein still was able to convert all radioactivity on HypE i nto the thiocyanate form, not shown). At close to stoichiometric ratios, the major amount of radioactivity migrated in a position i ndicated b y i mmunoblots to contain both proteins (not shown). Decrease of HypF in the a ssay gradually shifted the migration position of HypE, which refle cts a Fig. 3. Differential detection of HypE-thiocarboxamide and HypE- thiocyanate via separation by nondenaturing PAGE (A) and SDS/ PAGE (B) and instability of the HypE-thiocarboxamide (C). HypE protein (2 l M ) was mixed with HypF protein (0.5 l M ), 14 C-labelled carbamoylphosphate ([ 14 C]CP; 100 l M ) and eith er ATP (100 l M ;lane 1) or ADP-CH 2 -P (100 l M ;lane2)andincubatedat25°Cfor10min. The sample for the nondenaturing PAGE was mixed prior to appli- cation with a s ample buffer co ntaining 50 m M Tris/HCl (p H 6.8), 5% glycerol and 0.025% bromophenol blue (final concentrations). The sample for the SDS/PAGE was mixed prior to application with a sample buffer containing 50 m M Tris/HCl (pH 6.8), 2% SDS, 5% glycerol, 100 m M dithiothreitol, 0.025% bromophenol blue and heated to 56 °C for 10 min. (C) Instability of HypE-thiocarboxamide in 9 dependence o n the pH. HypE (6 l M ) and HypF (1 l M )proteinswere mixedwith[ 14 C]CP (100 l M )andADP-CH 2 -P (100 l M ) and incuba- ted for 15 min at 25 °C. Substrates and buffer were removed by filtration and extensive washing (nanosep MWCO 10 kDa). Th e protein f ractio n was further incubated for 10 min at 25 °Cin100m M Tris/HCl pH 7.5 (lane 1), pH 8.0 (lane 2), pH 8.5 ( lane 3), pH 8.8 (lane 4) and pH 9.2 (lane 5) followed by mixing with s ample buffer and separation in a 10% SD S g el as ind icated f or (B). 3432 M. Blokesch et al. (Eur. J. Biochem. 271) Ó FEBS 2004 rapid equilibrium between HypF (81.9 kDa molecular mass) and HypE (35.1 kDa molecular mass). When HypE was in excess, HypE-CN monomer carried all the radio- activity. The analysis of the carbamoyltransferase reaction cata- lyzed by HypF is dependent on the cleavage of ATP into AMP and pyrophosphate [11]. Accordingly, ADP-CH 2 -P serves as a substrate but not AMP-CH 2 -PP [ 12]. The easy discrimination between HypE-thiocarboxamide and HypE- thiocyanate a llowed the analysis whether AMP-CH 2 -PP is a substrate in the dehydration reaction. To this end, trans- carbamoylation/dehydration assays were carried out in the presence of ATP, of AMP-CH 2 -PP and of different ratios between ATP and its a nalogue and the p roducts were analyzed (Fig. 5). It was found that the analogue is unable to support HypF-catalyzed carbamoyltransfer to HypE as the p rotein does not carr y any radioactivity (lane 2). Surprisingly, it also inhibits conversion of the HypE- thiocarboxamide into HypE-thiocyanate when offered together with ATP: Formation of HypE-CN is gradually decreased (Fig. 5A) and HypE-CONH 2 appears (Fig. 5B). Mutant variants of protein HypE It has been speculated previously that carbamoylphos- phate, besides being the educt for synthesis of the CN ligands, may also give rise to the formation of the carbonyl ligand [12]. The thiocarboxamide of the HypE protein thus could provide the carbamoyl moiety for the reductive deamination t o deliver the carbonyl group either at the HypE protein itself or after donating it to the iron of the metal centre. A possibility to test this assumption may be offered by the construction of a mutant protein, which can accept the carbamoyl residue but is unable to dehydrate it because of the lack of the ATP-dependent phosphorylation activity. Position D83 was an attractive candidate as this residue was shown to be involved in the binding of ATP in the PurM protein [26]. Moreover, D83 is strictly conserved in all proteins belonging to the PurM family of dehydratases [26]. A D83N exchange was therefore introduced into HypE and the protein was overproduced and purified. Figure 6 gives the activity of the protein species in the transcarbamoy- lation/dehydration reaction in comparison to that of wild- type HypE and o f the variant lacking the C-terminal cysteine (HypED). As expected, HypED does not act as an acceptor in the carbamoylation reaction (lanes 3 and 4). The radioactive label in the D83N (lanes 5 and 6) and D83N A76V variants (lanes 7 and 8) is present in the thiocarboxamide form, irrespective of whether ATP or ADP-CH 2 -P was present as substrate. These variants will be analyzed in future whether they can transfer the carboxamido moiety to the HypC · HypD complex. Analysis of the transcarbamoylation reaction catalyzed by the HypF protein When the activity of the HypF protein was tested in the absence of HypE (which acts as the natural substrate accepting the carbamoyl group) it was shown to display the following three activities: (a) carbamoylphosphate Fig. 4. Autoradiograph of a nondenaturing PAG 10 in which the products of transcarbamoylation/dehydration assays were separated. Reaction mixtures co ntained HypE at 2 l M in the ratio to HypF indicated a bove each lane (2 l M down to 0.2 l M ). [ 14 C]Carbamoylphosphate and ATP were present at 100 l M each; i ncubation time was 30 min at 25 °C. The identity of the material in the labelled bands was shown by immuno - blotting (not sho wn). Fig. 5. Inhibition of the ATP-dependent dehydration r eaction by AMP- CH 2 -PP. HypE (2 l M ) was incubated with HypF ( 0.2 l M )and [ 14 C]carbamoylphosphate (100 l M ) in the presence o f ATP ( 100 l M where indicated) or/and AMP-CH 2 -PP at the concentrations indicated on top of the gel. The samples were separated by nondenaturing PAGE (A) or SDS/PAGE (B) and the proteins were transferred t o nitrocellulose memb ranes t hat were a utorad iographed. Fig. 6. Activity of wild-type and mutant HypE proteins in the trans- carbamoylation/dehydration reaction. Four micromoles of wild-type HypE protein (lanes 1 and 2), of HypED (lanes 3 and 4), HypE[D83N] (lanes 5 and 6) and HypE[D83N A76V] ( lanes 7 and 8) were incubated with 0.5 l M HypF, 100 l M [ 14 C]carbamoylphosphate and 100 l M ATP(lanes1,3,5and7)or100l M ADP-CH 2 -P (lan es 2, 4, 6 and 8) for 30 min at 25 °C. Th e reaction pro ducts were separated in non- denaturing gels (A) and SDS gels (B), transferred to a nitrocellulose membrane which was autoradiographed. Ó FEBS 2004 Hydrogenase maturation protein HypE (Eur. J. Biochem. 271) 3433 phosphatase activity in the absence of ATP; (b) a car- bamoylphosphate-dependent cleavage of ATP into AMP and pyrophosphate and (c) a carbamoylphosphate-depend- ent ATP-pyrophosphate exchange reaction [11]. The latter activity, however, levelled off far before equilibrium was reached. Knowing that HypF transfers the carbamoyl moiety to a free protein thiol group, it was first tested whether nonprotein thiols can replace HypE as acceptor substrate. To this end, determination of the carbamoylphosphate- dependent ATP cleavage reaction in the presence or absence of a thiol compound was tested. However, presence of dithiothreitol in concentrations between 0.5 and 100 l M did not influence the kinetics of ATP hydrolysis (data not shown). Analysis of HypF mutant proteins The results described thus far support the contention that the various activities of HypF reflect the particular experimental condition, namely that carbamoylphosphate phosphatase activity and carbamoylphosphate-dependent ATP hydrolysis to AMP and pyrophosphate might simply be side reactions followed in the absence of the natural substrate HypE. To provide further proof fo r this assumption, mutant variants of the HypF protein were purified and analyzed. Two of the variants chosen, HypF[R23Q] and HypF[R23E] have amino acid replace- ments in the acylphosphatase motif; in particular, R23 is part of an anion cradle of HypF which interacts with the phosphate of carbamoylphosphate i n a crystal o f the acylphosphatase domain complexed with the substrate [27]. Another variant, HypF[H476A] has a replacement in the histidine-rich motif close to the C-terminus, which is a characteristic of O-carbamoyltransferases [11]. Previous results had demonstrated that the replacement R23E leads to a gene product inactive in vivo and devoid of acylphosphatase activity in crude extracts. In contrast, the exchange R23Q had only diminished these in vivo and in vitro activities. The phenotype of the mutant harbour- ing the gene for HypF[H476A] was indistinguishable from that of the wild-type [11]. When the activity of the purified HypF variant proteins in comparison to that of the wild-type HypF protein were determined in the carbamoylphosphate-dependent ATP hydrolysis reaction it was found that HypF[H476A] possesses less than 10% of the activity of wild-type HypF (results not shown). However, this activity appeared to suffice for the generation of a wild-type-like phenotype, especially when the gen e was expressed from a plasmid [11]. From the two proteins with an exchange in the acylphosphatase domain HypF[R23E] displays no detectable activity whereas HypF[R23Q] has some minute activity, ranging between 0.1 and 0.3% of wild-type HypF (data not shown). Discussion The results presented above and reported in previous communications [11,12] suggest the sequence of reactions catalyzed by the HypF a nd HypE proteins, which are depicted in Fig. 7. HypF catalyses the formation of a protein-bound putative carbamoyl-adenylate with the con- comitant liberation of pyrophosphate. The identity of the adenylate has not been shown yet. In the absence of HypE the a denylate is avidly h ydrolyzed into AMP and possibly carbamate which is unstable. When HypE is present in the reaction mixture the carbamoyl moiety is transferred to the thiol of the C-terminal cysteine of HypE followed by its ATP-dependent dehyd ration to the thiocyanate [12]. In the course of the reaction HypF has to dock to the HypE protein and this complex has been experimentally demon- strated now (Fig. 4; and data not shown). It is intriguing that at stoichiometric ratios of the two proteins most of the substituted HypE is caught in the complex. This implicates the existence of some mechanism to displace the product, either via replacement by free HypE, some conformational switch conferred to the HypE protein during the reaction or the t ransfer of the HypE protein from Hyp F to the HypC · HypD complex (our unpublished results). An app arent ATP-pyrophosphate exchange reaction, which i s d ependent o n carbamoylphosphate, has been described p reviously to be catalyzed by HypF [11]. The scheme of Fig. 7 now offers an explanation why this exchange did not reach equilibrium. Whereas the observed formation of r adioactive ATP from l abelled pyrophosphate can be readily explained by the reversion of the reaction, attainment of the equilibrium would also necessitate the formation o f carbamoylphosphate from the postulated enzyme-bound carbamoyladenylate at the expense of inor- ganic phosphate. The situatio n is thus different from a classical A TP-PP i exchange reaction like that catalyzed by aminoacyl-tRNA synthetases in which the substrate (the amino acid) drives both the forward a nd the reverse reaction. An issue that is still open, however, i s why carbamoylphosphate-dependent cleavage of ATP reaches a Fig. 7. Scheme of the postulated reaction pattern of HypE (E) and HypF (F) proteins. I indicates the carbamoylphosphate phosphatase activity in the a bsence of ATP, II the carbamoylphosphate-dependent cleavage of ATP into AMP and pyrophosphate, III the carbamoyl- phosphate-depe ndent ATP-pyrophosphate exch ange re action, and IV the ATP-dependent dehydration catalyzed by HypE. 3434 M. Blokesch et al. (Eur. J. Biochem. 271) Ó FEBS 2004 plateau at product concentrations well below equilibrium in the absence of HypE. Inactivation of HypF during the reaction might be one of several possible reasons. The putative HypF-bound carbamoyl-adenylate is extremely prone to hydrolysis, which removes it from the reaction even in the p resence of HypE. This is also in distin ct contrast to the p roperties of aminoacyl-adenylates, which are shielded from water in the active site of aminoacyl-tRNA synthetases and therefore protected from hydrolysis. A reason for the instability o f the c arbamoyl-adenylate might exist in th e observation that HypE can be retrieved from cells as a complex with two other hydrogenase maturation proteins, namely HypC and HypD (M. Blokesch and A. Bo ¨ ck, unpublished results). This HypE in the triple complex i s fully active and it might represent the actual state of the protein within the cell. Until now it was open a s to w hether HypF acts solely as a carbamoyltransferase or whether it a lso participates in the dehydration reaction. The property of mutant proteins w ith amino acid exchanges in the nucleotide binding site of HypE shows that the dehydration is catalyzed by HypE per se,as the mutant proteins accept the carbamoyl-residue but are unable to convert it into the thiocyanate. However, it is still an open question whether dehydration is the result of an intramolecular reaction or whether intermolecular HypE– HypE interactions are in volved. The sites of carbamoyl- binding and ATP-dependent d ehydrations definitely display rather weak interdependence as HypE can act as acceptor without possessing phosphorylation activity and exhibits only marginally affected intrinsic ATPase activity in the absence of the C-te rminal thiol. Availability of a HypE variant that carries the c arboxamide m oiety but is unable to convert it into the thiocyanate will facilitate the analysis whether the carbonyl ligand also arises from carbamoyl- phosphate. Acknowledgements We are greatly indebted to R. Thauer for discussion and helpful suggestions. We t hank E. Ze helein for expert purification of HypF and HypE, H. H artl for the ICP spectroscopy of the proteins an d F. Lottspeich for determination of the N -terminal amino acid sequence of HypE. This work was supported by the Deutsche Forschungsge- meinschaft and the Fonds der C hemischen Industrie. References 1. Vignais, P.M., Billoud, B. & Meyer, J. (2001) Classification and phylogeny of hydrogenases. FEMS Microbiol. Rev. 25, 455–501. 2. Bo ¨ ck, A. & Sawers, G. (1996) Fermentation. In Escherichia coli and Salmonella (Neidhardt, F.C., Curtiss,R., III, Ingraham, J.L., Lin, E.C.C., Low, K.B., Magasanik, B., Reznikoff, W.S., Riley, M., Schaechter, M. & Umbarger, H.E., eds), pp. 262–282, American Society for Microbiology, Washington, DC. 3. Frey, M., Fontecilla-Camps, J.C. & Volbeda, A. (2001) Nickel- iron hydrogenases. In Handbook of Metalloproteins (Messer- schmidt,A.,Huber,R.,Poulos,T.&Wieghardt,K.,eds),pp.880– 896. John Wiley & Sons, Chichester, UK. 4. Happ e, R.P., Roseboom, W., P ierik, A.J., Albracht, S. P. & Bag- ley, K.A. (1997) Biological activation o f hydrogen. Nature 385, 126. 5. Pierik, A.J., Roseboom, W., Happe, R.P., Bagley, K.A. & Albracht, S .P.J. (1999) C arbon monoxide and cyanide as intrinsic ligands to iron in the active site of [NiFe]-hydrogenases. J. Biol. Chem. 274, 3331–3337. 6. Jacobi, A., Rossmann, R. & Bo ¨ ck, A. (1992) The hyp operon gene products are requir ed for the m aturatio n of catal ytically ac tive hydrogenase isoenzymes in Escherichia coli. Arch. Microbiol. 158, 444–451. 7. Blokesch, M., Paschos, A., Theodoratou, E., Bauer, A., Hube, M., Huth,S.&Bo ¨ ck, A. (2002) Metal insertion into NiFe-hydro- genases. Biochem. Soc. Trans. 30, 674–680. 8. Mehta, N., O lson, J.W. & Maier, R.J. (2003) Characterization of Helicobacter pylori nickel metabolism accessory proteins needed for m aturation of both urease and hydrogenase. J . Bacteriol. 185, 726–734. 9. Blok esch, M., Rohrmoser, M., Rode, S. & Bo ¨ ck, A. (2004) HybF, a zinc containing protein involved in NiFe hydrogenase matura- tion. J. Bacteriol. 186, 2603–2611. 10. 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. 11. 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. 12. Reissmann, S., Hochleitner, E., Wang, H., Paschos, A., Lottspe- ich, F., Glass, R.S. & Bo ¨ ck, A. (2003) Taming of a poison: bio- synthesis of the NiFe-hydrogenase cyanide ligands. Science 299, 1067–1070. 13. Blokesc h, M. & Bo ¨ ck, A. (2002) Maturation of [NiFe]-hydro- genases in Escherichia coli: the HypC cycle. J. Mol. Biol. 324, 287– 296. 14. Drapal, N. & Bo ¨ ck, A. (1998) Interaction of the hydrogenase accessory protein HypC with HycE, the large subunit of Escher- ichia coli hydrogenase 3 du rin g enzyme m aturation. B i oche mistr y 37, 2941–2948. 15. Magalon, A. & Bo ¨ ck, A. (2000) Analysis of the HypC-HycE complex, a key int ermediate in the assembly of t he metal center of the Escherichia coli hydrogenase 3. J. Biol. Chem. 275, 21114– 21220. 16. Casadaban, M.J. & Cohen, S.N. (1979) Lactose genes fused to exogenous promoters in one step using a Mu-lac bacteriophage: in vivo probe for transcriptional control sequences. Proc. Natl Acad. Sci. USA 76, 4530–4533. 17. Yanisch-Perron, C., Vieira, J.&Messing,J.(1985)ImprovedM13 phage cloning vec tors a nd ho st strains: nucleotide sequences o f the M13mp18 and pUC19 v ectors. Gene 33, 103–119. 18. Miller, J.H. ( 1992) A s hort course in bacterial ge netics. In AShort Course in Bacterial Genetics. Cold Springer Harbor Laboratory Press, Col Spring Harbor, New Y ork. 19. Begg, Y.A., Whyte, J.N. & Haddock, B.A. (1977) The identifica- tion of mutants of Escherichia coli deficient in formate dehy- drogenase and nitrate reductase activities using dye indicator plates. FEMS Microbiol. Lett. 2, 47–50. 20. Schlensog, V. & Bo ¨ ck, A. (1990) Identification and sequence analysis of the gene encoding the transcriptional activator of the formate hydrogenlyase system of Escherichia c oli. Mol. Microbiol. 4, 1319–1327. 21. Ansaldi, M., Lepelletier, M. & Mejean, V. (1996) Site-specific mutagenesis by using an accurate recombinant polymerase chain reaction method. Anal. B iochem. 23 4, 110–111. 22. Lutz,S.,Jacobi,A.,Schlensog,V.,Bo ¨ hm,R.,Sawers,G.&Bo ¨ ck, A. (1991) Molecular characterization of an operon (hyp) necessary for the activity of the three hydrogenase i soenzymes in Escherichia coli. Mol. Microbiol. 5, 123–135. 23. Ochman, H., Gerber, A.S. & Hartl, D.L. (1988) Genetic applica- tions of an inverse polymerase chain reaction. Genetics 120, 621–623. Ó FEBS 2004 Hydrogenase maturation protein HypE (Eur. J. Biochem. 271) 3435 24. Studier, F.W. & Moffatt, B.A. (1986) Use of bacteriophage T7 RNA polymerase to direct selective high-level expression of cloned genes. J. Mol. Biol. 189, 113–130. 25. Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the he ad of b acteriophage T4. Na tu re 227, 680–685. 26. Li, C., Kappock, T.J., S tubbe, J., Weaver, T.M. & Ealick, S.E. (1999) X-Ray crystal structure of aminoimidazole ribonucleotide synthetase (PurM), from the Escherichia coli purine biosynthetic pathway at 2.5 A ˚ resolution. Structure 7, 1155–1166. 27. Rosano,C.,Zuccotti,S.,Bucciantini, M., Stefani, M., Ramponi, G. & Bolognesi, M . (2002) Crystal s tructure and anion binding i n the prokaryotic hydrogenase maturation factor H ypF acylphos- phatase-lik e do main. J. Mol. Biol . 321, 785–796. 3436 M. Blokesch et al. (Eur. J. Biochem. 271) Ó FEBS 2004 . complex. Analysis of the transcarbamoylation reaction catalyzed by the HypF protein When the activity of the HypF protein was tested in the absence of HypE. proteins HypF and HypE catalyze the synthesis of the CN ligands of the active site iron of the NiFe-hydrogenases using carbamoylphosphate as a substrate. HypE