Water Effects on Electron Transfer in Azurin Dimers Agostino Migliore,* ,†,‡ Stefano Corni, † Rosa Di Felice, † and Elisa Molinari †,‡ National Center on nanoStructures and bioSystems at Surfaces (S3) of INFM-CNR, Modena, Italy, and Dipartimento di Fisica, UniVersita ` di Modena e Reggio Emilia, Via Campi 213/A, Modena, Italy ReceiVed: July 24, 2006; In Final Form: September 20, 2006 Recent experimental and theoretical analyses indicate that water molecules between or near redox partners can significantly affect their electron-transfer (ET) properties. Here, we study the effects of intervening water molecules on the electron self-exchange reaction of azurin (Az) by using a newly developed ab-initio method to calculate transfer integrals between molecular sites. We show that the insertion of water molecules in the gap between the copper active sites of Az dimers slows down the exponential decay of the ET rates with the copper-to-copper distance. Depending on the distance between the redox sites, water can enhance or suppress the electron-transfer kinetics. We show that this behavior can be ascribed to the simultaneous action of two competing effects: the electrostatic interaction of water with the protein subsystem and its ability to mediate ET coupling pathways. 1. Introduction Protein electron-transfer (ET) reactions represent a major concern in the current nanoscale research for two main reasons: (1) they play a crucial role in vital processes of living cells 1 and (2) modern electronics aims at exploiting the intrinsic functions of biomolecules to implement nanoelectronic devices. 2-4 Long-distance tunneling is the major electron-transfer mecha- nism in proteins, 5,6 and the accurate prediction of the inherent ET rates is a long-standing challenge. Indeed, several factors, subject of experimental 7 and theoretical 8-10 investigations, can concur to determine the rates of biological ET processes, such as the structure and the energies of the donor and acceptor groups, the distance between them, the structure of interposed protein portions, and the thermal atomic motion. Further factors relevant to the intermolecular electron transfer are the docking of the redox partners and the properties of the often intervening solvent. 11 Water is the most important molecular environment for electron transfer. It can affect intermolecular ET rates by means of its electrostatic and quantum interactions with the protein system, which can play a central role in determining the best ET pathway and the activation free energy. 8,12 Many experi- mental 7,13,14 and theoretical 15 studies have been focused on the efficiency of water in mediating electron-transfer reactions, with special attention to its influence on the distance dependence of ET rates. Several questions remain yet debated. 6 Both single- exponential 14 and multiple-exponential 12 decay modes of the ET rates were found. In some circumstances, water appeared to be a poor electron-transfer mediator 14 or appeared not to influence the ET processes significantly, 16 while it has been recently suggested that water molecules in the interface of covalently cross-linked azurin dimers could increase the inherent electron-transfer rate. 13 Even the existence of specific electron- tunneling pathways (not observed directly) is still debated, 17,18 although recent theoretical calculations support the idea of specific ET paths. 19 The present paper is devoted to the ab-initio computation of the electron-transfer matrix elements, or transfer integrals, 20 for the electron self-exchange between Az active sites at different distances, in the presence of two interposed water molecules. The ET system is modeled after the X-ray structure of Az dimers, where such water molecules were observed. 13 The transfer integrals are important factors in controlling the rates of many electron-transfer reactions. Within the context of Marcus’ ET theory, 21 they can be easily combined with estimates of the reorganization energy to evaluate the ET rates, measured in kinetic experiments. Much progress has been recently made in computing transfer integrals 22-27 through several quantum chemical methods, 19,20,22,28,29 and the increasing availability of both electron-transfer kinetic data and powerful computational tools enabled several comparisons between theory and experi- ment. 30 However, electronic couplings between molecules are difficult to calculate accurately, because they are often extremely small, and some inherent problems are still unresolved. 6 In particular, proteins are generally too large systems for exhaustive ab-initio calculations, thus requiring the usage of approximate computational methods, such as semiempirical 31 and protein- fragment 27 approaches. The method exploited in this paper 29 overcomes some usual limitations of such approaches. In particular, it can use a complete multielectron scheme (i.e., does not rely on a single-particle scheme), thus comprising electronic relaxation effects; it does not use empirical parameters; it does not require the knowledge of the transition-state coordinate and of excited-state quantities. We have implemented the method in a density-functional theory (DFT) scheme. 29 DFT is the best compromise between accuracy and computational feasibility for studying large metal-ion complexes, such as the Az sites investigated in our work. DFT is also the best approach for calculating electronic properties of solid crystal systems. Thus, it addresses the desirable purpose of treating biological and inorganic components with the same method, in view of possible applications involving both components together. * Author to whom correspondence should be addressed. Phone: +39- 059-2055315; fax: +39-059-2055651; e-mail: smigliore@unimore.it. † National Center on nanoStructures and bioSystems at Surfaces (S3) of INFM-CNR. ‡ Universita ` di Modena e Reggio Emilia. 23796 J. Phys. Chem. B 2006, 110, 23796-23800 10.1021/jp064690q CCC: $33.50 © 2006 American Chemical Society Published on Web 10/31/2006