In this report the first high-quality infrared spectra of [Fe]-hydrogenase are presented. Analyses of these spectra obtained under a variety of redox conditions strongly indicate that [Fe]-hydrogenases contain a low-spin Fe ion in the active site with one CN Ϫ group and one CO molecule as intrinsic, non-protein ligands. When in the ferric state, the presence of such an ion can explain the enigmatic EPR properties (the rhombic 2.10 signal) of the active, oxidised enzyme. To account for other, well-characterised properties of the active site, we propose that the active site of [Fe] A third class of hydrogenase, not containing any metals and only active in the presence of its cofactor, has been discovered in methanogenic Archaea by Thauer and coworkers [8,9].A few [NiFe]-hydrogenases have been extensively studied. The basic unit of these enzymes is formed by a large (47Ϫ 72 kDa) and a small (23Ϫ38 kDa) subunit. In the last 4 years FTIR studies on the enzymes from Chromatium vinosum [10Ϫ 13] and Desulfovibrio gigas [14,15], combined with the crystal structure of the D. gigas enzyme [14,16] have revealed that the active site is a bimetallic NiFe site with two CN Ϫ groups and one CO molecule bound to Fe. The whole site is bound to the large subunit via four thiols from strictly conserved Cys residues. In the D. gigas enzyme, the small subunit harbours one [11,14,15]. Also nitrile hydratase exhibits an FTIR band in this region : in its inactive state, this enzyme has an NO bound to Fe giving rise to an FTIR band around 1855 cm Ϫ1 [17].In this study we have examined the FTIR bands exhibited by [Fe]-hydrogenase from D. vulgaris, strain Hildenborough, in more detail. Also the behaviour of the bands under various conditions was studied and compared with that of [NiFe]-hydrogenase from C. vinosum studied under the same conditions. It is concluded that the active site in [Fe]-hydrogenase contains a low-spin Fe with one CN Ϫ and one CO as intrinsic, non-protein ligands. The Fe ion can be either Fe 3ϩ , giving rise to the rhombic 2.10 EPR signal, or Fe 2ϩ (EPR silent). It is proposed that this low-spin Fe is directly linked to a [4Fe-4S] cluster, presumably via two bridging thiols from conserved Cys residues. MATERIALS AND METHODSGrowth of D. vulgaris, subspecies Hildenborough (NCIB 8303), purification of its [Fe]-hydrogenase and determination of purity and activity were performed as described previously [7]. The purity index (A 400 /A 280 ) of the preparation used was 0.36 and the H 2-production activity was 2255 U/mg. Growth of C. vinosum (DSM 185), purification of its [NiFe]-hydrogenase and activity measurements were as previously de-

In Azotobucter vinelandii MoFe protein the oxidation of the P clusters to the S = 7/2 state is associated with a redox reaction with Em,7 = +90 2 10 mV (vs the normal hydrogen electrode), n = 1. A concomitant redox process is observed for a rhombic S = 1/2 EPR signal with g = 1.97, 1.88 and 1.68. This indicates that both S = 1/2 and S = 7/2 signals are associated with oxidized P clusters occurring as a physical mixture of spin states. The maximal intensity of the S = 1/2 and S = 7/2 signals in the mediated equilibrium redox titration is similar if not identical to that of solidthionine-treated samples. Summation of the spin concentration of the S = 1/2 spin state (0.25 2 0.03 spin/ad2) and the S = 7/2 spin state (1.3 * 0.2 spida2p2) confirms that the MoFe protein has absolutely no more than two P clusters. In spectra of enzyme fixed at potentials around -100 mV a very low-intensity g = 12 EPR signal was discovered. In parallel-mode EPR the signal sharpened and increased >lO-fold in intensity which allowed us to assign the g = 12 signal to a non-Kramers system (presumably S = 3). In contrast with the non-Kramers EPR signals of various metalloproteins and inorganic compounds, the sharp absorption-shaped g = 12 signal is not significantly broadened into zero field, implying that the zero field splitting of the non-Kramers doublet is smaller than the X-band microwave quantum. The temperature dependence of this g = 12 EPR signal indicates that it is from an excited state within the integer spin multiplet. A bell-shaped titration curve with Em,7 = -307 * 30 mV and + 81 2 30 mV midpoint potentials is found for the g = 12 EPR signal. We propose that this signal represents an intermediate redox state of the P clusters between the diamagnetic, dithionite-reduced and the fully oxidized S = 7/2 and S = 1/2 state. Redox transitions of two electrons (-307?30mV) and one electron (+90?10mV) link the sequence S = O*S = 3+(S = 7/2 and S = U2). We propose to name the latter paramagnetic oxidation states of the P clusters in nitrogenase POx1 and POx2, and to retain PN for the diamagnetic native redox state. The magnetic circular dichroism and Mossbauer data on thionine-oxidized MoFe protein have to be re-evaluated bearing in mind that the oxidized P clusters can exist in two redox-states. Finally, an account is given of the EPR spectroscopic properties of S = 9/2 and other systems obtained upon superoxidation of the MoFe protein.Nitrogenase is the biological catalyst for the activation of the dinitrogen molecule in aqueous solution. The enzyme complex consists of two dissociable metalloproteins, the =230-kDa ad2 tetrameric MoFe protein and the homodimeric = 62-kDa Fe protein. Substrate binding, activation and reduction takes place on the MoFe protein, presumably

No abstract

A conserved iron-binding site, the ferroxidase center, regulates the vital iron storage role of the ubiquitous protein ferritin in iron metabolism. It is commonly thought that two Fe(II) simultaneously bind the ferroxidase center and that the oxidized Fe(III)-O(H)-Fe(III) product spontaneously enters the cavity of ferritin as a unit. In contrast, in some bacterioferritins and in archaeal ferritins a persistent di-iron prosthetic group in this center is believed to mediate catalysis of core formation. Using a combination of binding experiments and isotopically labeled (57)Fe(II), we studied two systems in comparison: the ferritin from the hyperthermophilic archaeal anaerobe Pyrococcus furiosus (PfFtn) and the eukaryotic human H ferritin (HuHF). The results do not support either of the two paradigmatic models; instead they suggest a unifying mechanism in which the Fe(III)-O-Fe(III) unit resides in the ferroxidase center until it is sequentially displaced by Fe(II).

The crystal structure of the ferritin from the archaeon, hyperthermophile and anaerobe Pyrococcus furiosus (PfFtn) is presented. While many ferritin structures from bacteria to mammals have been reported, until now only one was available from archaea, the ferritin from Archaeoglobus fulgidus (AfFtn). The PfFtn 24-mer exhibits the 432 point-group symmetry that is characteristic of most ferritins, which suggests that the 23 symmetry found in the previously reported AfFtn is not a common feature of archaeal ferritins. Consequently, the four large pores that were found in AfFtn are not present in PfFtn. The structure has been solved by molecular replacement and refined at 2.75-Å resolution to R = 0.195 and R free = 0.247. The ferroxidase center of the aerobically crystallized ferritin contains one iron at site A and shows sites B and C only upon iron or zinc soaking. Electron paramagnetic resonance studies suggest this iron depletion of the native ferroxidase center to be a result of a complexation of iron by the crystallization salt. The extreme thermostability of PfFtn is compared with that of eight structurally similar ferritins and is proposed to originate mostly from the observed high number of intrasubunit hydrogen bonds. A preservation of the monomer fold, rather than the 24-mer assembly, appears to be the most important factor that protects the ferritin from inactivation by heat.