Active site structure of Rieske-type proteins: electron nuclear double resonance studies of isotopically labeled phthalate dioxygenase from Pseudomonas cepacia and Rieske protein from Rhodobacter capsulatus and molecular modeling studies of a Rieske center.

Continuous wave electron nuclear double resonance (CW ENDOR) spectra of [delta-15N,epsilon(-14)N]histidine-labeled phthalate dioxygenase (PDO) from Pseudomonas cepacia were recorded and found to be virtually identical to those previously recorded from [delta,epsilon-15N2]histidine-labeled protein [Gurbiel, R. J., Batie, C. J., Sivaraja, M., True, A. E., Fee, J. A., Hoffman, B. M., & Ballou, D. P. (1989) Biochemistry 28, 4861-4871]. Thus, the two histidine residues, previously shown to ligate one of the irons in the cluster [cf. Gurbiel et al. 1989)], both coordinate the metal at the N(delta) position of their imidazole rings. Pulsed ENDOR studies showed that the "remote", noncoordinating nitrogen of the histidine imidazole ring could be observed from the Rieske protein in a sample of Rhodobacter capsulatus cytochrome bc1 complex uniformly labeled with 15N but not in a sample of PDO labeled with [delta-15N,epsilon-14N]histidine, but this atom was easily observed with a sample of Rh. capsulatus cytochrome bc1 complex that had been uniformly labeled with 15N; this confirmed the conclusion from the CW ENDOR studies that ligation is exclusively via N(delta) for both ligands in the PDO center. Modifications in the algorithms previously used to simulate 14N ENDOR spectra permitted us to compute spectra without any constraints on the relative orientation of hyperfine and quadrupole tensors. This new algorithm was used to analyze current and previously published spectra, and slightly different values for the N-Fe-N angle and imidazole ring rotation angles are presented [cf. Gurbiel et al. (1989) Gurbiel, R. J., Ohnishi, T., Robertson, D. E., Daldal, F., and Hoffman, B. M. (1991) Biochemistry 30, 11579-11584]. This analysis has permitted us to refine the proposed structure of the [2Fe-2S] Rieske-type cluster and rationalize some of the properties of these novel centers. Although the spectra of cytochrome bc1 complex from Rh. capsulatus are of somewhat lower resolution than those obtained with samples of PDO, our analysis nevertheless permits the conclusion that the geometry of the cluster is essentially the same for all Rieske and Rieske-type proteins. Structural constraints inferred from the spectroscopic results permitted us to apply the principles of distance geometry to arrive at possible three-dimensional models of the active site structure of Rieske protein from Rh. capsulatus. Results from this test case indicate that similar procedures should be generally useful in metalloprotein systems. We also recorded the pulsed and CW ENDOR spectra of 57Fe-labeled PDO, and the resulting data were used to derive the full hyperfine tensors for both Fe(III) and Fe(II) ions, including their orientations relative to the g tensor. The A tensor of the ferric ion is nominally isotropic, while the A tensor of the ferrous ion is axial, having A(parallel) > A(perpendicular); both tensors are coincident with the observed g tensor, with A(parallel) of the ferrous ion lying along the maximum g-value, g1. These results were examined using refinements of existing theories of spin-coupling in [2Fe-2S]+ clusters, and it is concluded that current theories are not adequate to fully describe the experimental results.