Nitrous oxide (N2O) is a greenhouse gas, the third most significant contributor to global warming [1]. As a key process for N2O elimination from the biosphere, nitrous oxide reductase (N2OR) catalyses the two-electron reduction of N2O to N2. This periplasmic enzyme is found in denitrifying bacteria which obtain metabolic energy by using nitrogen-oxidised compounds, instead of oxygen, as terminal electron acceptors in anaerobic respiration. Nitrous oxide reductase is a 1160 residue, homodimer, containing 2 copper centres called CuA and CuZ. The CuA centre is the electron entry site, and is structurally homologous to the CuA centre of cytochrome c oxidase [2]. The structure of the catalytic CuZ centre, however, is unknown. In order to solve the structure of N2OR and thereby elucidate the structure of the CuZ centre, a MAD data collection was performed on BM14, exploiting the anomalous absorption of the copper K-edge. The coordinates of the twelve copper clusters were identified from low resolution anomalous Patterson syntheses. Phases were calculated from these twelve clusters and improved by solvent flattening. After iterative rounds of model building and six-fold, non-crystallographic symmetry averaging, the improved map enabled the complete polypeptide chain to be built. The copper centres were then constructed by refining the preliminary model against a 2.4 Å data set collected from ID14-4.

Each N2OR monomer is composed of two distinct domains. The N-terminal domain (residues 10-443) adopts a 7-bladed ß-propeller fold, with the CuZ centre located at one end of the propeller on the pseudo 7-fold axis. The C-terminal domain (residues 478-581) forms an anti-parallel ß-sandwich in the Greek key motif, and adopts a cupredoxin fold already seen in bovine cytochrome c oxidase, with the CuA centre located in a loop region between strands ß8 and ß9. In the dimer, the C-terminal domain of one monomer faces the N-terminal domain of the second monomer reminiscent of the phenomenon of 'domain-exchange' (Figure 15). Upon dimerisation 26% of the total surface area of each monomer is buried to a 1.6 Å radius probe.

The CuZ centre of N2OR comprises four copper ions which adopt the shape of a distorted tetrahedron, and ten ligands (Figure 16): seven histidines residues and three hydroxide ions. Two histidine residues (His270 and His437) of the CuZ centre belong to the loops located on the top of the propeller domain and the remaining five (His79, His80, His128, His325, His376) belong to the inner-most strand of the blades (Figure 16). The seven His ligands are not part of a consensus sequence, the CuZ centre thus differs from other copper centres. As a result of the intermonomer 'domain-exchange', the CuA centre of one monomer is in close proximity (closest distance, 10.2 Å) to the CuZ centre of the second monomer.

The structure of the CuZ centre suggests that there is only one possible site for N2O binding. The CuZ centre could behave as an electron buffer, three copper ions being reduced by the CuA centre prior to substrate processing. The catalytic copper would remain oxidised and, therefore, able to bind the substrate. This electron reservoir could favour a fast electron exchange and prevent the formation of dead-end products.

References
[1] J. Reilly et al., Nature, 401, 549-555 (1999).
[2] J.A. Farrar, A.J. Thomson, M.R. Cheesman, D.M. Dooley, W.G. Zumft, FEBS Lett., 294, 11-15 (1991).

Authors
K. Brown (a), M. Tegoni (a), M. Prudêncio (b), A.S. Pereira (b), S. Besson (b), J.J. Moura (b), I. Moura (b), C. Cambillau (a).

(a) Architecture et Fonction des Macromolécules Biologiques, UPR 9039, CNRS, Marseille (France)
(b) Departamento de Química, Centro de Química Fina e Biotecnologia, Universidade Nova de Lisboa (Portugal)