Activation of HydA Requires a Preformed [4Fe-4S] Cluster

The H-cluster is a complex bridged metal assembly at the active site of [FeFe]-hydrogenases that consists of a [4Fe-4S] subcluster bridged to a 2Fe-containing subcluster with unique nonprotein ligands, including carbon monoxide, cyanide, and a dithiolate ligand of unknown composition. Specific biosynthetic gene products (HydE, HydF, andHydG) responsible for the biosynthesis of the H-cluster and the maturation of active [FeFe]-hydrogenase have previously been identified and shown to be required for the heterologous expression of active [FeFe]-hydrogenase [Posewitz, M. C., et al. (2004) J. Biol. Chem. 279, 25711-25720]. The precise roles of the maturation proteins are unknown; the most likely possibility is that they are directed at the synthesis of the entire 6Fe-containing H-cluster, the 2Fe subcluster, or only the unique ligands of the 2Fe subcluster. The spectroscopic and biochemical characterization of HydA (the [FeFe]-hydrogenase structural protein expressed in the absence of the maturation machinery) reported here indicates that a [4Fe-4S] cluster is incorporated into the H-cluster site. The purified protein in a representative preparation contains Fe (3.1 ( 0.5 Fe atoms per HydA) and S (1.8 ( 0.5 S atoms per HydA) and exhibits UV-visible spectroscopic features characteristic of iron-sulfur clusters, including a bleaching of the visible chromophore upon addition of dithionite. The reduced protein gave rise to an axial S= /2 EPR signal (g= 2.04 and 1.91) characteristic of a reduced [4Fe-4S] cluster. M€ ossbauer spectroscopic characterization of Fe-enriched HydA provided further evidence of the presence of a redox active [4Fe-4S] cluster. Iron K-edge EXAFS data provided yet further support for the presence of a [4Fe-4S] cluster in HydA. These spectroscopic studies were combined with in vitro activation studies that demonstrate that HydA can be activated by the specific maturases only when a [4Fe-4S] cluster is present in the protein. In sum, this work supports a model in which the role of the maturation machinery is to synthesize and insert the 2Fe subcluster and/or its ligands and not the entire 6Fe-containing H-cluster bridged assembly. The [NiFe]and [FeFe]-hydrogenases are widely distributed in nature and efficiently catalyze the reversible oxidation of molecular hydrogen (H2 T 2H + + 2e). The [NiFe]-hydrogenases, present in archaea and bacteria, generally function to oxidize molecular H2 and provide reducing equivalents for metabolic processes, while the [FeFe]-hydrogenases, present in bacteria and eukarya, functionmore broadly to catalyze both proton reduction andH2 oxidation (2, 3).Recently, there has been a growing interest in these metalloenzymes because of their inherent applicability in the development of renewable H2-based energy technology. The active sites for both [NiFe]and [FeFe]-hydrogenases have been determined by X-ray crystallography and are united by the presence of π acceptor CO and CN ligands, which are not common in biology. These diatomic ligands stabilize low-spin and low-valent oxidation states of the metal centers at the active sites. For the [NiFe]-hydrogenase, the active sites from a variety of different sulfate-reducing bacterial sources have been determined to consist of a Ni atom coordinated to an Fe atom via two thiolate ligands and a bridging oxygen species (4). The Ni atom is further coordinated by two cysteine ligands from the protein, while the Fe atom is coordinated to two terminal CN ligands and one terminal CO ligand. In comparison, the [FeFe]-hydrogenase active site contains a 6Fe-containing complex cluster termed the H-cluster, as determined for This work was supported by AFOSR Multidisciplinary University Research Initiative Award FA9550-05-01-0365 (J.W.P.), NASA Astrobiology Institute FundedAstrobiologyBiogeocatalysisResearchCenter Grant NNA08C-N85A (J.W.P., J.B.B., and R.K.S.), National Institutes of Health Grant GM47295 (B.H.), and National Science Foundation Grant NSF0755676 (R.K.S.). *To whom correspondence should be addressed. Phone: (406) 9947211. Fax: (406) 994-7212. E-mail: [email protected]. edu. Abbreviations: CO, carbon monoxide; CN, cyanide; Fe-S, ironsulfur; HydA, HydA expressed in a genetic background devoid of HydE, HydF, and HydG; LB, Luria-Bertani; IPTG, isopropyl β-D-1thiogalactopyranoside; PMSF, phenylmethanesulfonyl fluoride; DTT, dithiothreitol; DT, sodium dithionite; EPR, electron paramagnetic resonance; EXAFS, extended X-ray absorption fine structure analysis; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; EDTA, ethylenediaminetetraacetic acid; XAS, X-ray absorption spectroscopy; HydF*, HydF containing the ligand modified 2Fe subcluster. D ow nl oa de d by A IS T I on A ug us t 4 , 2 00 9 Pu bl is he d on M ay 1 2, 2 00 9 on h ttp :// pu bs .a cs .o rg | do i: 10 .1 02 1/ bi 90 00 56 3 Article Biochemistry, Vol. 48, No. 26, 2009 6241 Clostridium pasteurianum (CpI) (5, 6) and Desulfovibrio desulfuricans (7, 8). The H-cluster consists of a [4Fe-4S] subcluster coordinated to a 2Fe subcluster via a cysteine thiolate ligand. The two Fe centers in the 2Fe subcluster are bridged via a five-atom dithiolate ligand and a CO ligand. The chemical composition of the dithiolate ligand has not yet been determined unambiguously and has been proposed to be dithiomethylether (6), propane dithiolate (7), or dithiomethylamine (8). In addition, both Fe centers contain terminal CO and CN ligands. For the presumed oxidized state of CpI, a water molecule is present at the distal Fe center in the proximity of the [4Fe-4S] subcluster. Biosynthesis and maturation of the [NiFe]-hydrogenases have been thoroughly studied, including identification of at least six gene products involved in formation of active [NiFe]-hydrogenases and the interactions between gene products during maturation. The metabolic source of diatomic CN ligands has been identified to be carbamoyl phosphate (9), whereas the metabolic source for the CO ligand is still in question. In contrast, relatively little is known concerning the biosynthesis and maturation of [FeFe]-hydrogenases. By analysis of several mutant strains of Chlamydomonas reinhardtii that are unable to produce hydrogen, the genes hydEF and hydG were discovered to be required for maturation of [FeFe]-hydrogenases (1). Subsequent expression studies revealed that formation of an active [FeFe]-hydrogenase was achieved only whenHydAwas heterologously expressed in a background of coexpressed gene products HydEF and HydG in Escherichia coli (1). In most organisms, hydEF exists as two separate genes, hydE and hydF (1), and it has been shown that the coexpression in E. coli of HydE, HydF, and HydG from Clostridium acetobutylicum with the [FeFe]-hydrogenase structural gene product from various algal and bacterial sources is sufficient to effect expression of active [FeFe]-hydrogenase (10). Deduced amino acid sequence analysis of HydE, HydF, and HydGgene products revealsHydE andHydG tobe radical-SAM Fe-S enzymes as they both have the C-X3-C-X2-C radical-SAM signature motif and HydF to be a GTPase (1, 9). Also, preliminary biochemical characterization of HydE and HydG has revealed associated SAMcleavage activity (11). In addition, it has been shown that upon reconstitution HydF binds an Fe-S cluster and exhibits GTPase activity (12). The involvement of HydE, HydF, and HydG maturation enzymes in the biosynthesis of the H-cluster was further elaborated when it was shown that HydA, expressed in a genetic background devoid of HydE, HydF, and HydG (HydA), is a stable protein capable of being activated in vitro by the aforementioned proteins (13). It was also determined that HydF behaves as a scaffold protein in which an H-cluster precursor is assembled and can be subsequently transferred to HydA, resulting in the formation of an active [FeFe]-hydrogenase (14). Both in vitro activation studies imply that cluster biosynthesis does not take place on the structural protein (HydA) and that a chemical precursor to the H-cluster is synthesized in the absence ofHydA that upon transfer toHydA results in its activation. Although no chemical precursors or intermediates to the H-cluster have yet been characterized or even identified, it can be hypothesized that HydE, HydF, and HydG are directed toward the synthesis of (1) the entire 6Fe-containing H-cluster, (2) the 2Fe subcluster of the H-cluster, or (3) the biologically unique ligands of the 2Fe subcluster of the H-cluster. The characterization of HydA provides a critical link in our understanding of this fascinating process by defining the substrate for the Hyd maturation proteins. In this study, we present spectroscopic and biochemical characterizations of HydA from C. reinhardtii to provide insights into the [FeFe]-hydrogenase maturation and Hcluster biosynthesis. HydA from the eukaryotic green algae C. reinhardtii contains only the H-cluster binding domains and represents the simplest [FeFe]-hydrogenase known. Unlike [FeFe]-hydrogenases fromCl. pasteurianum andD. desulfuricans, the [FeFe]-hydrogenases from eukaryotic green algae do not contain additional accessory Fe-S clusters with plant-type ferredoxin domains that would complicate spectroscopic characterization of the Fe-S clusters present at the active site (15-17). Our characterization of HydA from C. reinhardtii indicates that a [4Fe-4S] cluster is present in HydA and is required for in vitro activation by the HydE, HydF, and HydG maturation enzymes. Accordingly, it follows that the aforementioned maturation enzymes are not directed at the synthesis of the entire 6Fe-containing H-cluster. EXPERIMENTAL PROCEDURES Cloning and Cell Growth Conditions. HydA from C. reinhardtii was cloned into a pET Duet vector as described previously (10) and modified for the presence of an N-terminal six-histidine tag. HydA from C. reinhardtii was expressed in E. coli BL21(DE3) cells and cultivated in either 2 L flasks with a 1 L medium volume or a 10 L benchtop fermentor (New Brunswick) containing modified MOPS minimal medium (18) supplemented with 5.5% glucose and 150 μg/mL ampicillin. Either 5 or 50 mL overnight cultures of BL21(DE3) cells were used to inoculate 1 or 10 L cultures, respectively. Later, cell expressions were cultivated in LB medium buffered with 50 mM phosphate (pH 7.6), supplemented with 5.5% glucose and 150 μg/mL ampicillin. The cells were grown at 37 C with vigorous shaking (flasks) and agitation and aeration (fermentor, 250 rpm, 3 L/min) to an optical density of 0.5 (measured at 600 nm with a visible spectrophotometer from Thermo Spectronic) and induced by addition of IPTG to a final concentration of 1 mM. (NH4)2Fe(SO4)2 3 6H2O (191 μM) was also added at induction. Induction was allowed to proceed aerobically for 2 h at 37 C followed by 16 h at 4 C under nitrogen purge with additional supplemented (NH4)2Fe(SO4)2 3 6H2O (191 μM). The cells were harvested anaerobically (8000 rpm and, 4 C), washed with buffer A [50 mM HEPES (pH 7.6), 150 mM NaCl, and 1 mM DTT] in an anaerobic Coy chamber (Coy Laboratories), and stored at -80 C. HydA Purification.Cell-free extracts were prepared by resuspending the cells (described above) in degassed anaerobic buffer B [5 mL of buffer/g of cells, 50 mM HEPES (pH 7.6), 50mMNaCl, 5% (w/v) glycerol, 1% (w/v) Triton X-100, 10mM MgCl2, 1 mM PMSF, 1 mM DTT, and trace quantities of lysozyme and DNAase] and placing the suspension in a pressure cell bomb (1000 psi nitrogen pressure, 1 h) followed by centrifugation of the lysate (36,000 G for 45 min). HydA was purified from the cell lysate by anion exchange chromatography on Q-Sepharose resin (GE Healthcare) followed by His-tag affinity chromatography to a Co resin (Talon resin; Clontech); all purification steps were done anaerobically using thoroughly degassed buffers under positive nitrogen pressure. Cell lysates were loaded onto a 100 mL Q-Sepharose column previously equilibrated in buffer C [50 mMHEPES (pH 7.6), 20% glycerol, and 1 mM DT]. The column was washed with buffer C supplemented with 100 mM NaCl, and HydA was eluted D ow nl oa de d by A IS T I on A ug us t 4 , 2 00 9 Pu bl is he d on M ay 1 2, 2 00 9 on h ttp :// pu bs .a cs .o rg | do i: 10 .1 02 1/ bi 90 00 56 3 6242 Biochemistry, Vol. 48, No. 26, 2009 Mulder et al. with buffer C in arrangement with a NaCl gradient up to a final concentration of 1 M while the absorbance at 405 nm was monitored.HydA protein-containing fractionswere goldenbrown in color, consistent with the presence of Fe-S clusters. These fractions were loaded onto a 25 mL Co affinity column. The column, equilibrated with buffer D [50 mM HEPES (pH 7.0), 300 mM NaCl, 20% glycerol, 0.5 mM DT, and 5 mM βME], was washed with buffer D supplemented with 10 mM imidazole, andHydA was eluted using an imidazole gradient up to 100 mM in buffer D while the absorbance at 405 nm was monitored. The golden-brown protein-containing fractions were concentrated using an Amicon concentration cell under positive argon pressure with a 30 kDa cutoff membrane filter. The proteinwas desalted in a Coy anaerobic chamber using a Sephadex 2 mL G-25 column (GE Healthcare) equilibrated with buffer C and stored under liquid nitrogen. Overall, six different HydA protein preparations were used. For later preparations, it was determined that the quality of the spectroscopic data collected could be improved by selectively saving HydA fractions with the highest 405 nm to 280 nm absorbance ratios as measured after elution from the Q-Sepharose and Co His-tag affinity columns. Assays. Protein concentrations were determined using the Bradford assay (19) with bovine serum albumin (Sigma) as the standard. The specific hydrogenase activity of the purified HydA was measured using gas chromatography upon maximum in vitro activation with cell lysates containing maturationproteinsHydE,HydF, andHydG fromCl. acetobutylicum as described by McGlynn et al. (13). In the assays, dithionite (20 mM) was used as the reducing agent, methyl viologen (10 mM) was used as the electron carrier, and HydE, HydF, and HydG cell extracts from Cl. acetobutylicum were used to activate HydA. The iron content of HydA was determined using a procedure described by Fish (20), which uses ferrozine under reductive conditions after digestion of the protein in 4.5% (w/v) KMnO4 and 1.2 N HCl. Iron standards were prepared by dilution of a commercial Fe AA standard (Ricca Chemical Co.). Sulfide assays were conducted according to a procedure described by Beinert (21) and Broderick (22), and sulfide standards were prepared from Na2S 3 9H2O. Electronic Absorption Spectroscopy. For UV-visible spectroscopic experiments, samples were prepared inside an anaerobic Coy chamber and transferred to anaerobic 1 mL cuvettes (NSG Precision Cells, Inc.). UV-visible spectra were recorded at room temperature with a Cary 300 (Varian) spectrophotometer. Reduced samples were prepared by adding 2 mM DT. EPR Spectroscopy. Low-temperature X-band EPR spectra were recorded using a Bruker ESP300E spectrometer equipped with a liquid helium cryostat and temperature controller from Oxford Instruments. Typical EPR parameters were as follows: sample temperature, 12 K; microwave frequency, 9.36 GHz; attenuation, 20.4 dB; and microwave power, 1.85 mW. Sample concentrations varied between 128 and 512 μM. Reduced samples were prepared by adding 2 mM DT, and oxidized samples were prepared by titrating in increasing concentrations of ferricyanide (2-11 mM). The spin concentration was determined by double integration of the sample spectra using CuSO4 (0.20 mM) and EDTA (2.0 mM) as the standardmeasured under identical conditions. Basic analysis of the collected spectra was conducted using the computer software program SpinCount (M. Hendrich, Carnegie Mellon University, Pittsburgh, PA). M€ ossbauer Spectroscopy. Fe was purchased from Cambridge Isotope Laboratories, Inc., and dissolved in hot concentrated hydrochloric acid. The pH was adjusted with NaOH. HydA as-isolated Fe samples were prepared from 10 L cultures using the defined minimal medium (18) described above with Fe substituted at the same molar concentrations for Fe. Fe (191 μM) was also added at induction with IPTG. Protein samples (500-800 μM) were loaded into 450 μL cups and stored under liquid nitrogen. M€ ossbauer spectra were recorded on a M€ ossbauer spectrometer equipped with a Janis 8DT variabletemperature cryostat and operated at a constant acceleration mode in transmission geometry. The zero velocity refers to the centroid of a room-temperature spectrum of a metallic iron foil. Analysis of the spectra was performed with WMOSS (WEB Research). Fe K-Edge X-ray Absorption Spectroscopy. HydA samples (1.9 mM) were prepared from the HydA protein isolated as described above. The EXAFS cells (Delrin) sealed with thin Fe-free Kapton tape were loaded with ∼100 μL of sample. Fe K-edge X-ray absorption spectroscopic (XAS) measurements were conducted at beamline 7-3 (BL7-3) of the Stanford Synchrotron Radiation Lightsource (SSRL) under storage ring (SPEAR3) conditions with an energy of 3 GeV and a current of 100-80mA on two different occasions. BL7-3 is a 20-pole, 2 T Wiggler beamline equipped with a Si(220) downward reflecting, double-crystal monochromator. Data were collected in the energy range from 6785 eV to k= 17 Å above the Fe K-edge using an unfocused beam. The frozen solution samples were mounted under liquid nitrogen and measured in a liquid He cryostat at ∼11 K. The beamline parameters were optimized at 8000 eV. The Fe KR fluorescence signal was collected using a 30-element Ge array detector and with a Soller slit andZ-1 (Mn) filter. The energywindowingof the detectorwas carefully done to minimize the fluorescence signal due to scattering and other non-Fe KR emission sources. ATHENA (23), a graphical-user interface to IFEFFIT (23), was used for averaging and background subtraction. The data were calibrated to the first rising-edge inflection point of the XAS spectra that was assigned to 7111.2 eV of an iron foil. The data are averages of at least five scans before normalization or background subtraction. ATHENA and AUTOBK were used to spline the postedge region and to obtain the EXAFS with an Rbkg of 1. ARTEMIS (23), ATOMS (23), and FEFF (24) were used to model and fit the data and calculate Fe 3 3 3Fe and Fe-S scattering paths, respectively. The structural models used in scattering calculations were derived from combination of average Fe 3 3 3Fe and Fe-S distances of reduced and oxidized [4Fe-4S] model compounds (25). Due to the presence of various Fe environments which lead to various reasonable fits [defined as R(fit) < 10], we used the composition of Fe-S clusters in the EXAFS fits as constraints from corresponding M€ ossbauer measurements. Reconstitution of Iron-Sulfur Clusters in HydA. HydA was subjected to reconstitution conditions following the general procedures described for biotin synthase (26). The protein (10 μM) was incubated with FeCl3 (100 μM), Na2S (100 μM), andDTT (1mM) in buffer E [50mMHEPES (pH7.0), 300 mM NaCl, and 20% glycerol] for 2-3 h with constant stirring in an anaerobic Coy chamber. Visually, the color of the diluted protein solutionwas observed to change to golden brown. All reagents were added sequentially, and following reconstitution, excess ions were removed using a G-25 Sephadex column. D ow nl oa de d by A IS T I on A ug us t 4 , 2 00 9 Pu bl is he d on M ay 1 2, 2 00 9 on h ttp :// pu bs .a cs .o rg | do i: 10 .1 02 1/ bi 90 00 56 3 Article Biochemistry, Vol. 48, No. 26, 2009 6243 The resulting reconstituted HydA was assayed for hydrogenase activity by in vitro activation with HydE, HydF, and HydG from Cl. acetobutylicum, and iron content was also analyzed as described earlier. Preparation ofApo-HydA andReconstitution.ApoHydA from C. reinhardtii was prepared by stripping out all Fe-S clusters using EDTA as an Fe chelator. HydA from C. reinhardtii was incubated aerobically with EDTA (100 mM) for 1 h in buffer E. Fe-S cluster chelation was monitored visually, and within 1 h, the golden-brown protein solution turned colorless. The resulting apoprotein was made anaerobic by degassing it under vacuum with sequential nitrogen purge and incubation with 5 mMDTT. Excess EDTA was removed with a G-25 Sephadex column, and verification that all Fe was stripped from the protein was achieved by Fe analysis. Apo-HydA from C. reinhardtii was reconstituted by incubating the protein (1 μM) with a large excess of FeCl3, Na2S, and DTT (1 mM) in buffer E for 2-3 h. Residual FeCl3 and Na2S were removed with a G-25 Sephadex column, and the protein was concentrated with a 30 kDa Centricon centrifugal filter device (Millipore). The colorimetric iron assay was used to monitor addition of Fe. The hydrogenase activities of the different HydA forms (HydA, apo-HydA, and reconstituted apo-HydA) were determined following activation with the HydE, HydF, and HydG maturation enzymes. RESULTS AND DISCUSSION HydA Binds a [4Fe-4S] Cluster. Anaerobic purification of heterologously expressed C. reinhardtii HydA [49 kDa (Figure 1A)] in E. coli gives a stable protein that is capable of being activated in vitro by Cl. acetobutylicum extracts containing HydE, HydF, and HydG. Specific activities of the six separateHydA protein purifications after in vitro activation with HydE, HydF, and HydG enzymes ranged from 10 to 38 μmol of H2 min -1 (mg of HydA). For in vivo coexpression of HydA from C. reinhardtii with the maturation genes fromCl. acetobutylicum inE. coli, H2 evolution activitywas previously reported to be 150 μmol of H2 min -1 mg (10). The difference in activity between in vitro and in vivo expression systems is likely attributed to the low occupancy of H-cluster activating precursors assembled by HydE, HydF, and HydG, as well as the heterogeneity of the maturation system (14). It was evident that HydA binds Fe-S clusters as observed by the dark brown color of the protein solution following purification as well as Fe (3.1 ( 0.5 Fe atoms/HydA) and S (1.8 ( 0.5 S atoms/HydA) analyses. It is important to note that the amount of Fe and S bound per HydA varied slightly between different purifications, suggesting that the Fe-S cluster present may be slightly labile during purification. The UVvisible absorbance spectrum shows a broad shoulder centered near 405 nm that upon reduction with DT decreases in intensity (Figure 1B). These absorbance features are consistent with the presence of Fe-S clusters in HydA. This was further confirmed by the low-temperatureX-bandEPR spectrum of reduced HydA which revealed an axial S= /2 signal (g= 2.04 and 1.91) characteristic of a reduced [4Fe-4S] cluster (Figure 2) (27). Also, a slight shoulder was observed at g = 2.04; however, the appearance of this feature varied between different protein preparations and could be minimized when the protein was further purified on the basis of collection of protein fractions with maximal 405 nm to 280 nm absorbance. The temperature and power dependence of the axial S = /2 signal showed a maximum intensity around temperatures of 10 K and diminished significantly at temperatures greater than 30 K at a power of 2 mW (Supporting Information). The signal was no longer FIGURE 1: (A) SDS-PAGE gel (10%) showing purification of HydA from C. reinhardtii following elution from the Co His-tag affinity column. The left lane shows the protein marker standards, and the right lanes show the pure HydA from C. reinhardtii at 49 kDa. (B) UV-visible spectra of HydA as isolated (;) and HydA reduced with 2 mM DT (--). The feature centered at 400 nm is characteristic of Fe-S cluster LMCT bonds. FIGURE 2: EPR spectrum of reduced HydA. Parameters for EPR measurement were as follows: sample temperature, 12 K; microwave frequency, 9.36 GHz; attenuation, 20.4 dB; microwave power, 1.85 mW; receiver gain, 2.0 10. The reduced sample was prepared by adding 2 mM DT to as-isolated HydA (235 μM). Upon oxidation of HydA with ferricyanide, the [4Fe-4S] cluster signal is no longer observed. D ow nl oa de d by A IS T I on A ug us t 4 , 2 00 9 Pu bl is he d on M ay 1 2, 2 00 9 on h ttp :// pu bs .a cs .o rg | do i: 10 .1 02 1/ bi 90 00 56 3 6244 Biochemistry, Vol. 48, No. 26, 2009 Mulder et al. observed above temperatures of 40 K and also did not saturate with an increase in microwave power. These temperature and power signal characteristics are typical for the presence of a [4Fe4S] cluster (28). Spin quantification of the signal, using CuSO4 (0.20 mM) with EDTA (2.0 mM) as a standard, gave 0.17 spin/Fe atom. Upon oxidation of HydA with ferricyanide, the [4Fe-4S] cluster signal disappeared and no EPR signal was observed (data not shown). M€ ossbauer Spectroscopic Characterization of HydA. Figure 3 shows the M€ ossbauer spectra of the Feenriched HydA in its as-purified (A) and dithionite-reduced (B and C) forms. The spectra were recorded at 4.2 K in a magnetic field of 50mTapplied parallel (A andB) and perpendicular (C) to the γ radiation. Analysis of the data indicates that all three spectra are composed of three subspectral components: a central quadrupole doublet (solid red lines in Figure 3), a magnetically split spectrum (blue lines), and an outer quadrupole doublet with broad and asymmetric absorption lines (green lines). The central quadrupole doublet is attributed to a [4Fe-4S] cluster and can be simulated as a superposition of two unresolved equal-intensity quadrupole doublets representing the two mixed-valence FeFe pairs within the cluster. The parameters obtained for the two unresolved doublets [ΔEQ(1) = 1.34 mm/s, δ(1) = 0.45 mm/s, and Γ(1) (line width)= 0.30 mm/s;ΔEQ(2)= 1.05 mm/s, δ(2)= 0.44 mm/s, and Γ(2) = 0.43 mm/s] are typical for [4Fe-4S] clusters (29, 30). The magnetically split spectral component exhibits an intensity pattern that depends on the direction of the applied field (blue lines in panels B and C), indicating that it originates from an EPR active Fe center. As reported above, reduced HydA displays an S = /2 EPR signal that can be assigned to a [4Fe-4S] cluster. The magnetic M€ ossbauer spectral component is therefore attributed to this S = /2 Fe cluster. Consistent with the [4Fe-4S] assignment, initial analysis of this spectral component indicates that it may be simulated as a superposition of two spectral components arising from the FeFe and FeFe pairs of a [4Fe-4S] cluster (see the Supporting Information) (29, 30). On the basis of the parameters obtained for the outer quadrupole doublet [ΔEQ = 2.77 mm/s, δ=1.30 mm/s, Γ(left) = 0.53 mm/s, and Γ(right) = 0.80 mm/s], this component is attributed to non-cysteine-coordinated extraneous Fe impurities, which may be generated via cluster degradation. Decomposition of the M€ ossbauer spectra into the abovedescribed spectral components shows that in the as-purified HydA sample (see Figure 3A) the majority (70%) of the Fe is present in the oxidized [4Fe-4S] form while a small percentage (15%) is in the reduced [4Fe-4S] form. The remaining Fe (15%) is present as Fe impurities. Upon reduction (see Figure 3B,C), a substantial amount of the [4Fe-4S] clusters are reduced to [4Fe-4S] clusters, resulting in a decrease in the M€ ossbauer absorption of the [4Fe-4S] cluster from 70 to 6% and an increase in the [4Fe-4S] absorption to 76%. That of the Fe impurities also increases slightly to 18%, suggesting that the [4Fe-4S] cluster inHydA is slightly unstable under dithionite reducing conditions. Thus, the M€ ossbauer data unambiguously show that the as-purified HydA contains predominantly [4Fe-4S] clusters. No other types of Fe-S clusters are detected. The observation that the [4Fe-4S] cluster in HydA can be reduced by dithionite to the [4Fe-4S] state establishes further that HydA contains a redox active [4Fe-4S] cluster. Taking into consideration the Fe and protein content (1.9 Fe atoms/HydA) determined for the as-purified Fe-enriched HydA and the total percent absorption of the [4Fe-4S] cluster detected by the M€ ossbauer measurement (85%), we find a stoichiometry of 0.4 [4Fe-4S] cluster/HydA for the as-purified HydA. Fe K-Edge X-ray Absorption Spectroscopic Characterization of HydA. The Fourier transforms of EXAFS (FTEXAFS) for the as-isolated HydA sample (black trace) and with a representative fit (red trace) are shown in Figure 4. The fitting parameters are summarized in Table 1. Among numerous reasonable fits with R(fit) values of 10, Table 1 and Figure 4 report the one that was obtained using theM€ ossbauer results (see above) as constraints for the amount and distribution of various Fe sites. The initial parameters of the fit were set to represent the 70% oxidized and 15% reduced protein-embedded [4Fe-4S] cluster content. The average Fe 3 3 3Fe, Fe-S , and Fe-S distances in synthetic [4Fe-4S] complexes are 2.74 ( 0.01, 2.25 ( 0.01, and 2.27 ( 0.03 Å, respectively (25). The protein-bound counterparts of Fe-S distances tend to be ∼0.02 Å longer due to dipole and hydrogen bonding interactions involving the sulfides and thiolate sulfurs, which reduces the nucleophilicity of the sulfur atoms and thus the covalency of the sulfur-iron bonds. These differences between the synthetic model and proteinbound Fe-S clusters can be observed by EXAFS as demonstrated by a series of FT-EXAFS analyses of Fe-S clusters (31-33). Using the initial distances with standard deviations as initial FIGURE 3: M€ ossbauer spectra of as-purified (A) and dithionite-reduced (B and C) HydA. The spectra (hatched marks) were recorded at 4.2K in amagnetic field of 50mTapplied parallel (A and B) and perpendicular (C) to the γ radiation. The solid lines plotted above the data are simulated spectra for the [4Fe-4S] cluster (red), [4Fe-4S] cluster (blue), and Fe impurities (green), normalized to the following percent absorptions: (A) 70% [4Fe-4S], 15% [4Fe4S], and 15% Fe and (B and C) 6% [4Fe-4S], 76% [4Fe-4S], and 18%Fe. The black lines overlaid with the experimental spectra are composite spectra. Parameters used for the simulations are given in the text (for the [4Fe-4S] and Fe impurities) and in the Supporting Information (for the [4Fe-4S] cluster). D ow nl oa de d by A IS T I on A ug us t 4 , 2 00 9 Pu bl is he d on M ay 1 2, 2 00 9 on h ttp :// pu bs .a cs .o rg | do i: 10 .1 02 1/ bi 90 00 56 3 Article Biochemistry, Vol. 48, No. 26, 2009 6245 Debye-Waller factors of 2.74( 0.005, 2.24( 0.003, and 2.28( 0.003 Å for Fe 3 3 3Fe, Fe-S , and Fe-S scattering paths, respectively, we obtained a reasonable fit that is shown in Figure 4. To prevent negative Debye-Waller factors (σ) and large deviations of the edge positions (E0), we obtained the final fit by linking all the σ andE0 values while allowing the individual path lengths and their amplitudes to vary. The parameters in Table 1 fit acceptably well with the bulk of the experimental data (black trace), but a poor fit was found at the shorter distances. The residual FT-EXAFS intensity at these distances is consistent with the presence of ∼15% free iron that is likely partially solvated and/or coordinated with low-Z atoms, such as O and N from protein residues as suggested by M€ ossbauer results. The peak at around 1.9( 0.1 Å can be reasonably well fitted with an Fe center surrounded by six low-Z (O and N) scatterers in 15% abundance. The relatively short Fe-O distance of 1.9 Å is indicative of the presence of negatively charged O or N ligands for the Fe impurities (see above). The fit components and parameters for the latter are given as Supporting Information. For both FT-EXAFS fits, the resulting Fe 3 3 3Fe (2.71-2.72 Å) and Fe-S (2.27 Å) distances are highly similar to those observed for the oxidized and reduced [4Fe-4S] cluster in Av2 (2.72 and 2.73 Å, and 2.27 and 2.29 Å, respectively) (33) and thus further support the presence of the [4Fe-4S] cluster in HydA. Reconstitution of the [4Fe-4S] Cluster in HydA. It was previously reported that HydA can be activated by cell extracts containing the maturation enzymes HydE, HydF, and HydG (13). This observation, coupled with our evidence described above for a [4Fe-4S] cluster in HydA, suggests that the [4Fe-4S] form of HydA is the substrate for assembly of the H-cluster by HydE, HydF, and HydG and further that the [4Fe-4S] cluster present in HydA becomes part of the H-cluster upon activation. To further explore the requirement for a preformed [4Fe-4S] cluster in HydA during activation with HydE, HydF, and HydG, HydA was chemically reconstituted with FeCl3 and Na2S. Fe analysis of the reconstituted HydA gave 4.0 ( 0.1 Fe/HydA, and the EPR spectrum of reconstituted reduced HydA indicates the presence of a [4Fe-4S] cluster, with an axial S = /2 EPR signal (g= 2.04 and 1.91) essentially identical to that of the as-isolated enzyme. In vitro activation of the reconstituted HydA with HydE, HydF, and HydG gave hydrogenase activity [31.5 ( 0.5 μmol of H2 min -1 (mg of HydA)] approximately 2-fold greater than that of the as-isolated HydA activated with HydE, HydF, and HydG [17.1 ( 2.4 μmol of H2 min -1 (mg of HydA)]. These observations provide additional evidence that the [4Fe-4S] form of HydA is the substrate for H-cluster assembly byHydE,HydF, andHydG, as increasing the [4Fe-4S] content improves the ability to activate HydA. In addition, the reconstitution results are consistent with our previous suggestion that the [4Fe-4S] cluster in HydA is labile during purification and is subsequently repopulated or repaired during reconstitution. In contrast, if the [4Fe-4S] cluster present in HydA was not required for activation by HydE, HydF, and HydG, it is likely that reconstitution would inhibit in vitro activation by generating a cluster at the H-cluster site that would prevent activation by the maturation enzymes. Metal Chelation and Reconstitution. To further explore the hypothesis that the [4Fe-4S] cluster present in HydA is required for in vitro activation with HydE, HydF, and HydG, apo-HydA was prepared and then subsequently reconstituted with FeCl3 and Na2S; both apo and reconstituted samples were subjected to in vitro activation by HydE, HydF, and HydG (Figure 5). The apo sample was found to contain zero Fe atoms per HydA, and upon in vitro activation with HydE, HydF, and HydG maturation enzymes, the apo-HydA protein exhibited no hydrogenase activity. Following chemical reconstitution of apoHydA with FeCl3 and Na2S, the protein contained 4.0 ( 0.1 Fe atoms/HydA and had an EPR spectrum essentially identical to that of the as-isolated enzyme. Subsequent in vitro activation of the reconstituted enzyme with HydE, HydF, and HydG produced hydrogenase activity FIGURE 4: FT-EXAFS plot (A) and individual EXAFS contributions (B) for the as-isolated HydA sample with a representative fit containing Fe 3 3 3Fe, Fe-S(sulfide), and Fe-S(thiolate) scattering paths. Table 1: Representative Fitting Parameters for the As-Isolated HydA Sample Using the Fe-S Composition from M€ ossbauer Measurements scatterer parameter fitted value