The influence of short saturated barriers on electron transfer between metal ions

Abstract

Although the metal-ion-containing prosthetic groups of many electron-transport proteins lie at or very near the protein surface, some appear to be relatively inaccessible. Also, it is difficult to image how metalloproteins that are intermediate in the reaction sequence for mitochondrial oxidative phosphorylation could be attached to both sequence neighbors to allow direct-contact electron transfer. Hence, it has been reasonably assumed that electron transport metalloproteins are separated from each other by some distance along the membrane surface [1]. It is of interest to determine how the rate of electron transfer varies with separation distance of redox partners. In this presentation, we will discuss complexes formed from two bridging ligands that separate the metal atoms (Ru(II) and Co(III)) so that direct contact is not possible. One of these ligands, 1,4-dicyano-(2,2.2)-bicyclooctane, is rigid and provides a shortest-distance-beyond-the-first-coordination-sphere of 4 Å, while the other, trans-1,4-dicyanocyclohexane, is flexible and provides a shortest distance of 2 Å (Fig. 1). The general scheme for the electron transfer reaction sequence we have used, involving initial reduction of Ru(III) to Ru(II) by a rapid reaction followed by the slow and irreversible reduction of Co(III) to Co(II), has been developed by Taube and coworkers [2]. The synthesis of the complexes, I and II, was accomplished by use of a labile Ru(III)-trifluoromethanesulfonate intermediate [3] using a method described by Sargeson and coworkers for Co(III) complexes [4] and by the in situ removal of Cl − from Co(III) using (CF 3SO 2) 2O. Estimates for the expected rate of intramolecular electron transfer were made using Marcus' theory as described for ruthenium complexes by Brown and Sutin [5], with a statistical factor to account for the effective concentration of an intramolecular reaction [6]. The results are simple, but interesting nonetheless. Upon reduction of I or II by a stoichiometric or substoichiometric amount of Ru(NH 3) 2+ 6 (in 0.10 M aqueous trifluoromethanesulfonic acid at 25 °C) the Ru(III) nitrile site is reduced to Ru(II) but not reduction of Co(III) to Co(II) is evident even after 20 hours (with concentrations of I or II less than 10 −4 M, so that intermolecular reactions are slow). Addition of excess Ru(NH 3) 2+ 6 leads to formation of Co(H 2O) 2+ 6 at expected rates. Based on a rate reduction ▪ of exp{−0.72(Δr)} [7], one would predict rate reductions of 18 and 4, for I and II, respectively, rather than the observed factors of 10 3 and 10 4 (the E 1 2 values of I and II are different, leading to different predicted rates). The barrier due to a saturated bridge, whether rigid with the shortest distance through the organic structure, as in I, or flexible with the shortest possible distance through the solvent, as in II, is much higher than might have been expected.