Prorein Science zyxwvutsrqponm (1997), 6:1190-1196. Cambridge University Press. Printed in the USA. Copyright zyxwvutsrqp 0 1997 The Protein Society Optimization of the electrostatic interactions between ionized groups and peptide dipoles in proteins zyx VELIN Z. SPASSOV,'.* RUDOLFLADENSTEIN,' AND ANDREJ D. KARSHIKOFF' 'Centre for Structural Biochemistry, Karolinska Institutet, NOVUM, S-14157 Huddinge, Stockholm, Sweden *Institute of Biophysics, Bulgarian Academy of Sciences, 11 13 Sofia, Bulgaria (RECEIVED January 6, 1997; ACCEPTED March 6, 1997) Abstract The three-dimensional optimization of the electrostatic interactions between the charged amino acid residues and the peptide partial charges was studied by Monte-Carlo simulations on a set of 127 nonhomologous protein structures with known atomic coordinates. It was shown that this type of interaction is very well optimized for all proteins in the data set, which suggests that they are a valuable driving force, at least for the native side-chain conformations. Similar to the optimization of the charge-charge interactions (Spassov VZ, Karshikoff AD, Ladenstein R, 1995, Protein Sci 41516- 1527), the optimization effect was found more pronounced for enzymes than for proteins without enzymatic function. The asymmetry in the interactions of acidic and basic groups with the peptide dipoles was analyzed and a hypothesis was proposed that the properties of peptide dipoles are a factor contributing to the natural selection of the basic amino acids in the chemical composition of proteins. Keywords: electrostatic interactions; Monte-Carlo simulations; peptide dipoles; protein stability The role of the peptide groups in functional properties of proteins has long been in the scope of many studies. Two features of these groups seem to be of dominant importance. The first one is the ability of the peptide amide nitrogens and carbonyl oxygens to form hydrogen bonds and thus to constitute the secondary structure elements of proteins. The second one is their dipolar nature. Taken together, these two properties result in an organized, nonrandom, distribution of the orientation of the peptide dipoles and superposi- tion of their electric field over large portions of the protein mol- ecule. On this basis, the concept of the a-helix macro dipole has been introduced (Wada, 1976). Hol et al. (1978) assumed that the field of an 0-helix macro dipole is equal to the field of a half positive unit charge at the amino end and a half negative unit charge at the carboxyl end. This idea has been applied in analysis of the electrostatic interactions in a number of studies (Friend zyxwvuts & Gurd, 1979; Hol et al., 1981; Warwicker & Watson, 1982; Rogers & Steinberg, 1984). The a-helix macro dipole representation seems, however, to be a rather rough approximation. It has been shown, for example, that the electrostatic effect of an a-helix is not asso- ciated with the macro dipole, but rather with the few dipoles con- fined to the end turns of the helix (Aqvist et al., 1991). Moreover, it has been shown (Aqvist et al., 1991) that models based on this approximation might lead to incorrect conclusions due to neglect- ing the large dielectric effect of the protein/solvent environment. Reprint requests to: A. Karshikoff, Centre for Structural Biochemistry, Karolinska Institutet, NOVUM, S-14157 Huddinge, Stockholm, Sweden; e-mail: aka@csb.ki.se. The alignment of the peptide dipoles in the secondary structure elements was found to be significant for different properties of proteins, such as counterion (Hol et al., 1978) and nucleotide binding (Muegge et al., 1996), enzymatic activity (Warwicker & Watson, 1982; Karshikov et al., 1993), structural stability of pro- teins (Hol et al., 1981; Hol, 1985), but not always as a stabilizing factor (Gilson & Honig, 1989). The fundamental role of the peptide dipoles in protein electro- statics becomes evident in their interactions with the side-chain charges. It has been found that the ionizable groups are situated in regions where favorable interactions with the peptide dipoles occur (Olson & Spassov, 1987). This finding is consistent with the ob- servation that the destabilizing effect of the charge dehydration is compensated by the electrostatic contribution of the peptide di- poles (Bashford & Karplus, 1990; Oberoi et al., 1996). The spatial arrangement of the peptide dipoles and side-chain charges is a result of the delicate balance of different forces sta- bilizing native protein structure, including electrostatic inter- actions. In other words, as far as the orientations and the positions of the peptide dipoles and the positions of the side-chain charges are a result of forces that are not only electrostatic, the magnitude of dipole-charge interactions will implicitly contain the contribu- tion of all other interactions. In a series of papers (Spassov & Atanasov, 1994; Spassov et al., 1994, 1995), we have introduced the concept of optimization of a given type of interactions. It represents the divergence of the value of the energy of selected interactions in the native structure from that calculated for some reference state. The reference state can be defined as a structure where the interactions of interest are set to zero. The divergence, or 1190