The overlap multipole expansion procedure is ptilized for the evaluation of the electrostatic molecular potential of poly(dC·dC) and poly(dA·dT). lt is shown that the two sequences of base pairs produce different effects in the significant regions of DNA. In the inner regions of the double helix the potentials are deeper in the major groove than in the minor groove in poly(dG·dC), and the reverse is true in poly(dA·dT). At large distances from the helical axis the regions of the backbone and the minor groove are preferred. The shielding of the phosphate groups by Na+ ions is shown to have an important effect on the discrimination between the two grooves. The region of the minor groove presents always a negative potential for the sequences considered (up to the explored distance of I7 Å from the helical axis), while negative and positive values of the potentials occur, within that distance, in the region of the major groove.

The influence on the molecular electrostatic potential of a B-DNA double helix of the screening of its phosphates by sodium and methylammonium cations is described. The effects of these two counterions are compared with respect to changes in the global potential of DNA and with respect to the potential minima at the nucleophilic sites of the purine and pyrimidine bases.

An exploration is carried out of the possibilities of using Topiol's pseudopotentials in computations of cation-ligand and hydrogen bonding interactions. The trends observed in the results of full SCF computations with different basis sets are reproduced. The larger systematic deviation observed for K+ and Ca++ as compared to Na+ and Mg++ suggests that the total freezing of the electrons in the two large cations is too drastic an approximation in the methodology adopted. A correction is suggested.

Ab initio SCF molecular orbital computations have been performed for a nucleoside, uridine, in a specific conformational arrangement (3'-endo conformation of the ribose, gg orientation of the extracyclic CH2OH group, and anti orientation of the base with respect to the sugar) and the results compared with some of those pertaining to the isolated component fragments, uracil and 3'-endo ribose. While the modifications of the atomic charges or of the energies of the molecular orbitals upon the formation of the nucleoside are small, the perturbations produced in some portions of the corresponding electrostatic molecular potential are more profound and more significant. Those occurring in the vicinity of the CO4 carbonyl group account for the probable increase in the proton affinity of uridine with respect to uracil. Those relevant to the O1′ oxygen indicate that the proton affinity of this atom in the free ribose is appreciably shielded in the nucleoside. The oxygens O2′ and O3′ represent non-negligible sites of attraction for positive ions in the free ribose and in the nucleoside.

The results of ab initio computations including correlation (MP2) performed on the complexes of Na+, monomethylammonium (MMA) and tetramethylammonium (TMA) with a same ligand, benzene, are compared at the same theoretical level (double-zeta basis set, with polarization functions optimized on molecular polarizabilities). The binding energies, corrected for the basis set superposition error, are in satisfactory agreement with the values of the measured gas-phase enthalpies. The nature of the components of the binding are discussed in light of decomposition of the interaction energies in different adducts. The values of the electrostatic, polarization, charge transfer and correlation (dispersion) components, considered at the equilibrium and upon approach to equilibrium, indicate that the order of the Coulomb attraction of these ions for the same ligand is a good indicator of the order of the total energies, but the weight of the other components is far from negligible, and varies from one ion to another.

Ab initio self-consistent-field second-order Møller–Plesset perturbation theory computations including basis set superposition error and zero-point vibrational energy corrections have been performed on the complexation of benzene with the polar head of acetylcholine (ACh). The ACh–benzene complex is about 0.5 kcal/mol less stable than the corresponding tetramethylammonium (TMA)–benzene complex, with a structure a little distorted with respect to the latter. The electronic structure of ACh is little modified by the ligand. Overall, the replacement of one methyl group of TMA by the acetyl tail of ACh does not affect strongly the complexation to benzene, as far as the main interaction is concerned.

Ab initio computations including correlation have been performed in a comparative study of complexes of tetramethylammonium (TMA) with benzene, pyrrole, pyrridine, and imidazole, using polarized Gaussian basis sets of different accuracies. With the best basis (optimized on molecular polarizabilities), the BSSE-corrected binding energies in the most stable complexes of these four Ligands are 9.1, 10.7, 13.3, and 16.3 kcal/mol, respectively, with benzene and pyrrole binding in a plane perpendicular to the TMA axis, and pyridine and imidazole inserting their nitrogen lone pair essentially along the TMA axis. The characteristics of secondary sites of binding of benzene are also determined and the overall results are discussed in connection with the possible role of aromatic amino acids in proteins. (C) 1997 John Wiley & Sons, Inc.

The binding energies of chloride adducts to water, methanol, and isopropanol have been calculated by the molecular orbital method at the self-consistent field (SCF) and Möller-Plesset [MP2, and (partially) MP4] levels. Extended Gaussian basis sets enlarged with both standard valence polarization orbitals and semidiffuse Coulomb polarization orbitals have been used. The best theoretical values obtained by correcting the energy differences between the supermolecule and its fragments for basis-set superposition errors and zero-point vibration energies are in satisfactory agreement with the available reaction enthalpies (i.e., 12.4 kcal/mol computed for water, versus 13.1 experimentally). Equally in agreement with the experimental trend, an increase in the binding energy of Cl− is predicted in going from water to alcohols. The proper introduction of correlation in a sufficiently extended and polarized basis set is essential to the reproduction of this trend. © 1997 John Wiley & Sons, Inc. Int J Quant Chem 63: 567–574, 1997

Theoretical calculations performed on the interactions of acetylcholine with the ‘aromatic gorge’ of acetylcholinesterase indicate the existence of a number of local minima for the substrate. These minima are clustered in four regions of increasing interactions from top to bottom of the gorge, culminating in the region of the ‘active site’. The results allow the delineation of the role of the different aminoacids lining the walls, emphasizing, in particular, that of Trp 279 and Trp 84 while smaller interactions involve tyrosines 70, 121, 130, 334 and Phe 330. The influence of D72 is stressed, as well as the orientating role of A 201 and the strong driving influence of E199.

In connection with the current discussions (1,2) on the role of tryptophan, tyrosine and phenylalanine residues in the binding of acetylcholine to the enzyme acetylcholine esterase, to the recognition site for agonists in the acetylcholine receptor and, more generally, in the interactions of choline derivatives with aromatics (see for instance, (3), (4), (5)) it may be useful to report the results of a computation of the intrinsic affinities of the above three side chains for the tetramethylammonium ion (TMA).

In distinction to the classical strategy of molecular modeling minimizing the energy of a small number of molecules, a procedure (FLATER) is developed to treat an aggregate of molecules organized in an array. The molecular energy of a unit cell in the array is minimized relative to a set of general coordinates defining the conformations and arrangement of molecules and the global geometry of the lattice, while allowing complete flexibility of the system. A test performed on the 1,2-dilauroyl-DL-phosphatidylethanolamine:acetic acid crystal of known structure shows that the molecular structures obtained by the procedure are reliable. An application of the procedure to the study of the copacking of lipids and hydrophobic polypetide α-helices in membranes is reported. Its results indicate that the interplay of the interactions that exist in the hydrocarbon region of a membrane concur to favor the assemblage of the helices rather than their dispersion in the lipid phase. This situation springs from the dominance of the helix–helix interactions, insured by the nature and number of the hydrophobic amino acid involved. This program is particularly well suited to studies of the interactions and architecture of molecules in multicomponent membranes.

A theoretical study of the conformation and interaction properties of glycerylmonooleate molecules was carried out as a preliminary step towards understanding their organization into layers. The evolution of the structural and energy characteristics of systems comprising one to seven molecules is described, putting into evidence the respective contributions of the polar heads and hydrocarbon chains. The dependence of the interaction energies on the geometry of the molecular packing shows the flexible and polymorphic properties. A disposition in layer was found, where the packing of the heads controlled by a network of hydrogen bonds is easily accommodated with the packing of the chains, due to the easy rotations around the bonds linking heads and chains.

In order to elucidate the role of the aromatic side-chains in the mechanism of transduction of monovalent cations through the channel of linear gramicidin, two series of analogues containing non-coded aromatic amino acids were synthesized. In the first series, the four tryptophans were replaced by either four L-3-(8-quinolyl)alanyl or four L-3-(4-quinolyl)alanyl residues and single channel conductance measurements showed that these substitutions led to a strong lowering of the channel conductance, which is attributed to a modification of the orientation of the aromatic side-chains due to an increase of their hydrophobicity. In the second series, the analogues contained both tryptophyl and naphthylalanyl residues in various amounts and positions. The single channel conductance data indicated that the conductance was mainly governed by the number of polar residues (Trp) and not by their positions. The conformational consequences of these results are discussed together with their influence on the energy profile of the gramicidin channel.

The constitution of an annulus of glyceryl-monooleate molecules around a gramicidin A monomer has been studied using successive energy optimizations. The average interactions between gramicidin-lipid in the complex are smaller than those between adjacent lipids. The conformations of the lipids remain similar to those optimized in a pure lipid system. The characteristics of the complex are essentially determined by the “roughness” of the monomer surface which results from the nature and succession of its side-chains. As a consequence, the space surrounding the molecule contains a density of lipid atoms smaller than a corresponding space in a pure lipid layer. The implications of these results are discussed in the light of experimental observations.

Energy profiles are calculated, using energy optimization computations, for a sodium cation in the AChR channel and four of its mutants, alpha E241D, beta E247Q, delta E255Q and alpha E241Q, using the model developed previously. The relative energy location of the calculated profiles confirms and specifies the role of each of the Glu residues found in the anionic ring at the bottom of the MII helices. The structural analysis of the results allows the understanding of the differences observed in the conductances for the wild-type and mutant alpha E241D, or for the mutants beta E247Q and delta E255Q in spite of the identity of the global charge of both channels in each couple. The striking correlation observed between the average relative energy location of the profiles and the conductance data appears to provide confirmation of the essential structural features adopted in the model, in particular the inclusion of the Glu(Gln in gamma)-Lys residues in the alpha-helical stretch of the MII helices and the overall location of the internal residues.

A recently developed computational technique, utilized to study different possible lipid/helix aggregates in a membrane layer, shows that hydrophobic helices have an appreciable preference for packing together rather than for staying dispersed in the lipid phase, preference stemming from the strength of the helix-helix interactions due to the hydrophobic nature of the helices.

It is shown by explicit calculation that the distribution of the atomic charges in the constituent molecules of a lipid monolayer or bilayer of glyceryl monooleate creates an intrinsic potential difference between the head region and the hydrocarbon region which tends to repel positive charges towards the exterior and attract negative charges to the interior. The analogies and differences between a bilayer and a monolayer are analyzed. The possible consequences of the intrinsic potential gradient in a lipid layer on the preferred orientation and conformation of a polar neutral molecule are illustrated on the case of a gramicidin A monomer.