Pedro Tavares

A new type of non-heme iron protein was purified to homogeneity from extracts of Desulfovibrio desulfuricans (ATCC 27774) and Desulfovibrio vulgaris (strain Hildenborough). This protein is a monomer of 16-kDa containing two iron atoms per molecule. The visible spectrum has maxima at 495, 368, and 279 nm and the EPR spectrum of the native form shows resonances at g = 7.7, 5.7, 4.1 and 1.8 characteristic of a high-spin ferric ion (S = 5/2) with E/D = 0.08. Mossbauer data indicates the presence of two types of iron: an FeS4 site very similar to that found in desulforedoxin from Desulfovibrio gigas and an octahedral coordinated high-spin ferrous site most probably with nitrogen/oxygen-containing ligands. Due to this rather unusual combination of active centers, this novel protein is named desulfoferrodoxin. Based on NH2-terminal amino acid sequence determined so far, the desulfoferrodoxin isolated from D. desulfuricans (ATCC 27774) appears to be a close analogue to a recently discovered gene product from D. vulgaris (Brumlik, M.J., and Voordouw, G. (1989) J. Bacteriol. 171, 49996-50004), which was suggested to be a rubredoxin oxidoreductase. However, reduced pyridine nucleotides failed to reduce the desulforedoxin-like center of this new protein.

The structure of desulfoferrodoxin (DFX), a protein containing two mononuclear non-heme iron centres, has been solved by the MAD method using phases determined at 2.8 Angstrom resolution. The iron atoms in the native protein were used as the anomalous scatterers. The model was built from an electron density map obtained after density modification and refined against data collected at 1.9 Angstrom. Desulfoferrodoxin is a homodimer which can be described in terms of two domains, each with two crystallographically equivalent non-heme mononuclear iron centres. Domain I is similar to desulforedoxin with distorted rubredoxin-type centres, and domain II has iron centres with square pyramidal coordination to four nitrogens from histidines as the equatorial ligands and one sulfur from a cysteine as the axial ligand. Domain I in DFX shows a remarkable structural fit with the DX homodimer. Furthermore, three beta-sheets extending from one monomer to another in DFX, two in domain I and one in domain II, strongly support the assumption of DFX as a functional dimer. A calcium ion, indispensable in the crystallisation process, was assumed at the dimer interface and appears to contribute to dimer stabilisation. The C-terminal domain in the monomer has a topology fold similar to that of fibronectin III.