1546 J. Phys. Chem. 1987, 91, zyxwvu 1546-1553 Theoretfcal Models of Diastereomeric Noncovalent Electron-Transfer Complexes. A Thermodynamic and Conformational Investigation Basilio Pispisa,* Antonio Palleschi, and Gaio Paradossi Dipartimento di Chimica, Universitci di Napoli, 801 34 Napoli, Italy, and Dipartimento di Chimica, Universitii di Roma, 00185 Roma, Italy (Received: July 28, 1986) Some of the parameters that were used in the computational studies of noncovalent diastereomericelectron-transfercomplexes between L-adrenaline or L-dopa and [Fe(tetpy)(OH)$ ions anchored to enantiomeric ordered polypeptides have been revised (tetpy zyxwvutsrqp = 2,2’,2’’,2”’-tetrapyridyl). The changes were based on a more realistic approach in evaluating electrostatic energies by using a distance-dependent dielectric constant and partial atomic charges for each atom of reactants. The changes removed some discrepancies previously observed in computed conformations of the diastereomeric adducts with L-dopa and led to a satisfactory comparison between calculated and experimental “differential” thermodynamic functions of binding for both substrates. Furthermore, there is now no dominant steric effect that discriminates between zyxw DL and LL pairs. Instead, binding stereoselectivity results from a complex interplay of ionic and nonbonding forces. Both these interactions enable the redox centers in the diastereomers to experience a different mutual orientation and separation distance which account for kinetic stereoselectivity that predominates over thermodynamic stereoselectivity in the reactions investigated. Solvent reorganization energy changes associated with the diastereomerically related charge-transfer steps were calculated by the ellipsoidal cavity model for short-range electron transfer. These changes reflect the stereochemical control exerted by the ordered polypeptide matrices in the formation of diastereomeric pairs and are chiefly responsible for the observed topochemical phenomena. Introduction Quantitative models for the forces governing molecular asso- ciation are a challenging goal in theoretical physical chemistry and biochemistry. At variance with covalent association, non- covalent binding is rather easy to treat because reliable models of noncovalent forces have been successfully developed over the past years.14 Such forces are generally considered to be van der Waals, hydrophobic, hydrogen-bonding, and Coulombic inter- actions. In treating noncovalent association a major problem has been that of accounting for solvent effect^,^ but this problem may be partly circumvented when dealing with the formation of dia- stereomeric adducts. In such a case, only “differential” energetic contributions are relevant and most of solvent effects may be expected to cancel out upon binding if the modes of binding are similar. In the past few years we were interested in electron-transfer reactions between chiral species. On the basis of previous un- successful investigations6 we felt that, in order to observe ste- reoselectivity in the reaction, the redox couple must form a stable association complex in which the diastereomers have to experience definitely different steric constraints. We then thought to use transition-metal derivatives bound to asymmetric polymers as one of the reactants because the macromolecular ensemble can offer a more efficient discriminating environment for the redox chiral partner than that of a simple asymmetric molecule. The ultimate goal was, in fact, that of obtaining a system that could mimic the stereospecific activity of enzymic materials. With this line of reasoning, we employed [Fe(tetpy)(OH),]+ ions (FeT; tetpy = 2,2’,2’’,2’”-tetrapyridyl) anchored to sodium poly(L-glutamate) (FeTL) or poly(D-glutamate) (FeTD) as en- antiomeric oxidant systems and L-ascorbic acid, L-catecholamines, and L-thiols as reductants, according to eq 1. While ascorbic P-Fe”’T+ + A- F= P-Fe”T + A‘ (1) acid and catechol derivatives were found to undergo, under suitable a stereoselective charge-transfer process with iron(II1) ion, very surprisingly the thiols (such as L-cysteine and reduced L-glutathione) did not. These latter results shall be reported in a separate paper, but they clearly indicate that the basic factors that produce stereochemical control are not yet easily foreseeable. Stereoselectivity apparently involves a delicate balance between opposing forces and must result from a complex interplay of both steric effects and ionic interactions. The knowledge of *Address correspondence to this author at the Universitl di Napoli. 0022-3654/87/209 1 - 1546%0 1 .50/0 the structural features of the diastereomeric encounter complexes is, therefore, of paramount importance for a better understanding of the phenomenon, and conformational energy calculations may be a valuable tool for the purpose. It is the aim of this paper to present hypothetical noncovalent models of FeTD-L-catecholamines and FeTL-L-catecholamines diastereomeric electron-transfer complexes undergoing stereose- lective reaction. Models of this kind were already reported by us9 but those presented here are based on new computational results where electrostatic energies were estimated by using a distance-dependent dielectric constant, corrected for the ionic strength of the medium through the inverse of Debye-Huckel screening length. Using the calculated interaction energies and the molecular parameters of the models, we were able to reproduce the exper- imentally determined differential thermodynamic functions for the formation of these complexes and offer a structural and en- ergetic model to account for, at least partially, the observed to- pochemical effects. For the sake of comparison with computed values, we also summarize here some of the experimental material previously reported,9atogether with new results on the structural features of the oxidant systems and on the thermodynamics of formation of the diastereomeric complexes. Experimental Section Materials. Quaterpyridineiron(II1) complex ions and poly- peptides were obtained as already reported.’3l0 L-Adrenaline (BDH) and L-dopa (Merck) were analytical-grade reagents and used as such. Concentrations were determined by UV absorption? (1) Platzer, K. E. B.; Momany, F. A.; Scheraga, H. A. Int. J. Peptide Protein Res. 1972, 4, 187. Pincus, M. R.; Scheraga, H. A. Arc. Chem. Res. 1981, 14, 299. (2) Levitt, M. J. Mol. Biol. 1974, 2, 393. (3) (a) Case, D.; Karplus, M. J. Mol. Biol. 1978, 123, 697. (b) McCam- (4) De Tar, D. F. J. Am. Chem. SOC. 1981, 103, 107. (5) Gilson, M. K.; Rashin, A.; Fine, R.; Honig, B. J. Mol. Biol. 1985, 183, 503. Zauhar, R. J.; Morgan, R. zyxwv S. Ibid. 1985, zyxw 186, 815. (6) Grossman, B.; Wilkins, R. G. J. Am. Chem. Sor. 1967, 89, 4230. Sutter, J. H.; Hunt, J. B. Ibid. 1969, 91, 3107. Kane-Maguire, N. A. P.; Tollison, R. M.; Richardson, D. E. Inorg. Chem. 1976, 15, 499. (7) (a) Barteri, M.; Pispisa, B. J. Chem. SOC., Faraday Trans. I 1982, 78, 2073. (b) Ibid. 1982, 78,2085. Makromol. Chem., Rapid Commun. 1982, 3, 715. (8) (a) Pispisa, B.; Barteri, M.; Farinella, M. Inorg. Chem. 1983,22, 3166. (b) Pispisa, B.; Farinella, M. Biopolymers 1984, 23, 1465. (9) (a) Pispisa, B.; Palleschi, A.; Barteri, M.; Nardini, S. J. Phys. Chem. 1985, 89, 1767. (b) Pispisa, B.; Palleschi, A. Macromolecules 1986, 19,904. (10) Branca, M.; Pispisa, B.; Aurisicchio, C. J. Chem. Soc., Dalton Trans. 1976, 1543. Cerdonio, M.; Mogno, F.; Pispisa, B.; Vitale, S. Inorg. Chem. 1977, 16, 400. mon, J. A,; Wolynes, P. G.; Karplus, M. Biochemistry 1979, 18, 927. 0 1987 American Chemical Society