The effective fragment potential (EFP) method is introduced as a way to model the effect of intermolecular hydrogen bonds on molecules described by standard quantum mechanical (QM) methods. The chemical system of interest is divided into two regions: an "active region" (AR) described by QM, and a "spectator region" (SR) that influences the AR via hydrogen bonding. The SR is replaced by an EFP which describes the interaction by three terms: electrostatics, polarization, and exchange repulsion. The potentials are derived from separate ab initio calculations on the prototypical interactions represented by the spectator region. The method is currently being implemented in the quantum chemistry code GAMESS. Some applications involving water in the SR are presented.

Abstract:

Quantum chemistry calculations have been used to study the conversion of spiropyran to merocyanine, both explicitly and with the model compounds pyranand nitrochromene. The minimum energy reaction path for the ring-opening reaction of pyran to form 2,4-pentadienal has been evaluated with Hartree-Fock calculations using the 3-21G and 6-31G(d) basis sets, as well as with semiempirical calculations. The reaction was found to follow a two-step mechanism, where first a conformer of the cis-2,4-pentadienal is formed, particularly a hydrogen-bonded seven-membered ring, and thereafter a rotation about a singlebond to form one of the open-ring conformers of cis-2,4-pentadienal is indicated. Step one in this mechanism is the rate-limiting step, with a barrier of 33kcal/mol at the Hartree-Fock 6-31G(d) level, corrected to 22 kcal/mol with Møller-Plesset second-order perturbation theory. This barrier is significantly higher than that reported previously using semiempirical calculations, and the discrepancy between the two studies is discussed. The energies of reaction for the ring-opening of nitrochromene and spiropyran have also been calculated by ah initio methods, and the effects of microsolvation and bulk solvation have been investigated for these systems. Finite field calculations were carried out to evaluate the polarizabilities and first hyperpolarizabilities and the effects of solvation on these properties.

Abstract:

The transphosphorylation step in the enzyme-catalyzed hydrolysis of phosphate eaters by Ribonuclease A (RNase A) is explored using ab initio quantum chemical methods. For the first time, components found in the RNase A active site are included in the all-electron chemical model, made up of2-hydroxyethyl methyl phosphate monoanion used as the substrate, and small model compounds used to mimic the three important residues, His-12,His-119, and Lys-41, found in the RNase A active site. The remainder of the immediate active site, including ten residues and six bound water molecules, is treated using effective fragment potentials (EFPs) incorporated directly into the Hamiltonian of the quantum system. The EFPs, derived from separate quantum calculations on individual components, are constructed to accurately represent the correct electrostatics and polarization fields of each component. High-resolution X-ray crystallographic data are used to assign the fixed relative positions of each component in the quantum and EFP regions. Characterization of the salient stationary points along the transphosphorylation reaction pathway at the RHF level using a 3-21+G(d) basis set reveals several low-barrier proton transfer steps between the substrate and the active site residues which allow transphosphorylation to occur with modest activation, consistent with the experimental data. Møller-Plesset perturbation theory (MP2) and density functional theory methods utilizing a larger6-31+G(d) basis are also used to explore the effects of electron correlation on the surface energetics. Consistent with expectations, the electrostatic field effects from the EFPs used to represent the non-participating parts of the active site are found to differentially stabilize certain structures along the reaction pathway.

Abstract:

Spectroscopy at a biochemical active site is influenced by local fields and hydrogen-bonds. Quantum calculations of the electronic structure of the entire biomolecule is, of course, impossible, but the chemical system can be modeled by dividing it into an active region (A) described quantum mechanically, and a spectator region (S) that influences A with strong fields and hydrogen-bonds. The all-electron interaction between A and S is replaced by an effective fragment potential (EFP) which represents the interaction as electrostatic, polarization and exchange repulsion terms. The EFP are derived entirely by ab initio model calculations of the S electronic properties and interactions and have been implemented in the quantum chemistry code, GAMESS. Spectroscopic analysis of enzyme active sites using the EFP will examine rhodanese and glutathione bound to glutathione S-transferase. The effect of specific hydrogen-bonds and local helices on spectral shifts is determined.

Abstract:

The exchange repulsion formula proposed by Murrell and co-workers (Proc. Roy. Soc. (Lond.), 1965, A284, 566; J. chem. Phys., 1967, 47, 4916) is considered in detail. Potentially important terms missing in the formalism of Murrell and co-workers are identified and evaluated for the water dimer using several basis sets. Insights into the contributing terms are obtained by using localized molecular orbitals. The results point towards a relatively simple expression for intermolecular exchange repulsion, based on the isolated wavefunctions of the two overlapping species.

Abstract:

An effective fragment model is developed to treat solvent effects on chemical properties and reactions. The solvent, which might consist of discrete water molecules, protein, or other material, is treated explicitly using a model potential that incorporates electrostatics, polarization, and exchange repulsion effects. The solute, which one can most generally envision as including some number of solvent molecules as well, is treated in a fully ab initio manner, using an appropriate level of electronic structure theory. In addition to the fragment model itself, formulae are presented that permit the determination of analytic energy gradients and, therefore, numerically determined energy second derivatives (Hessians) for the complete system. Initial tests of the model for the water dimer and water-formamide are in good agreement with fully ab initio calculations.

Abstract:

The internal rotation of formamide with 0-5 water molecules oriented along the N-C bond has been studied by the full ab initio self-consistent field theoryand using the effective fragment (EFP) method. For each case, the EFP geometries, harmonic vibrational frequencies, rotational barriers, and intrinsic reaction coordinates for the internal rotation are found to be in excellent agreement with their ab initio counterparts. The global energy minimum structures for four and five water complexes are predicted to be formamide bonded to two adjacent waters, with all water molecules in a ring. Probably due to the structural constraints, the complexes containing less than four waters have cyclic structures with the two ends of formamide connected by a sequence of water molecules. The internal rotation barrier of formamide-water complexes increases from 15.3 kcal/mol with no water to 19.0 kcal/mol with four waters and seems to saturate at four to five waters. When electron correlation corrections are added, the estimated internal rotation barrier is similar to 20 kcal/mol, in very good agreement with experimental measurements.

Abstract:

Glutamic acid in the gas phase and in aqueous solution was studied with ab initio quantum chemistry calculations. The effective fragment potential (EFP)has been used to study the effects of up to ten water solvent molecules on the neutral and zwitterionic isomers of glutamic acid, as well as to model the reaction path connecting these two isomers. The results with one and two water molecules have been compared to full ab initio ''super molecule'' calculations. This comparison shows that the EFP successfully models the effects of solvent molecules on structure and energetics. In the calculations including zero, one, or two water molecules, the neutral isomer was shown to be more stable than the zwitterion. On the other hand, in the EFP calculations with ten water molecules, where Monte Carlo methods were used to assist in finding the lowest energy conformations, the zwitterion was found to be more stable, which is the expected aqueous phase results.

Abstract:

Excited state geometries of formamide have been explored using the multiconfiguration self-consistent-held method. Optimized equilibrium geometries for the S1 and T1 states are nonplanar with the C-O and C-N bond distances substantially increased from the ground state values. The excitation energies at the ground and excited state geometries are calculated to vary dramatically with non-planar rotation. Raman scattering from the S2 state depends on the transition moment which is shown to vary strongly with geometry. Experimental analyses that project out restricted planar conformations can fit the Raman vibrational pattern but do not inform us about the complicated energy surface for the S2 state which is a resonance embedded in a Rydberg series. Constrained optimizations are used to explore this surface and the variation in the oscillator strength with geometry. Effective fragment potentials (EFP) model the waters in the solvation models. Comparison of the EFP and all-electron structures and energy of binding shows that the EFP adequately replace the all-electron waters. The use of constrained C-2v geometries for the EFP water does not significantly affect either the optimized structure or the energetics of the complex.

Abstract:

The accuracy and efficiency of an approximate formula for the intermolecular Pauli repulsion between closed shell molecules, derived earlier [Mol. Phys.89, 1313 (1996)], is demonstrated for dimers of H₂O, CH₃OH, CH₂Cl₂, CH₃CN, (CH₃)₂CO, and (CH3)₂SO. The energy derivative with respect to a Cartesian coordinate and rigid rotation about the center-of-mass (torques) are presented. The Pauli repulsion energy term is then combined with the Coulomb and classical induction energy terms of the effective fragment potential method [J. Chem. Phys. 105, 1968, 11081 (1996)] to give a general intermolecular interaction potential. This potential is applied to water and methanol clusters.

Abstract:

The recently developed effective fragment potential (EFP) model is applied to the description of a series of small water clusters, (H₂O)_n, n = 3-5. These results are compared with those found in the literature. The model accurately reproduces results obtained at ab initio levels of theory, while the computational cost is comparable to that of models employing empirical potentials. The EFP model thus offers significant promise as an in expensive alternative to the Hartree-Fock methodology in the treatment of small water clusters.

Abstract:

Knowledge of the ionicity of the phosphorane intermediate is important to the analysis of the microscopic mechanism of the hydrolysis of the phosphateester bond by ribonuclease A (RNase A). Five-coordinate uridine vanadate, an analog of the phosphorane, binds to RNase A as the monoanion. The absorption spectra of the vanadate is a probe of the electronic structure of the active site. An in vacuo theoretical model of H₄VO₅^- is calculated to have transitions only in the far ultraviolet (UV). However, H₂VO₅C₂H₄^- has one in the near UV as well as others further into the UV. The transition energy of the monoanion calculated in the field of the protein active site with effective fragment potentials shifts modestly to the red. Broad monoanion absorptions are predicted which would overlap an observed incomplete very broad absorption attributed to the complex of uridine vanadate with RNase A. The absorption bands of neutral ethylene glycol vanadate are predicted to be further to the red but also overlap the experimental absorption.

Abstract:

The Effective Fragment Potential (EFP) model of solvation can now be extended beyond aqueous systems due to the development of transferable exchange repulsion potentials. EFPs for methanol and chloroform have been developed, and calculations with these new EFPs agree well with full ab initio calculations. Ab initio calculations have been carried out on zinc tetraphenyl-octobromyl-porphyrin both with and without the EFP solvation model. While the aqueous calculation, which had its geometry optimized, gave good results, the single-point calculations carried out with the two new solvent models indicate the need for geometry optimization.

Abstract:

Porphyrins are a promising class of materials for optical limiting applications, and in the condensed phase solvent effects have been shown to be significant. We report results with a method designed to simulate the effects of discrete solvent molecules, namely the effective fragment potential (EFP) approach which has been implemented for use in ab initio calculations. Further, a simulated annealing (SA) method has been implemented with the EFP solvation model in an attempt to solve the problem of multiple minima in clusters of molecules. The results with this method indicate some success on models of aqueous formamide and aqueous glutamic acid. Ab initio calculations can now be carried out on porphyrins, and the solvation methods are being updated for their use on these systems

Abstract:

The recently developed effective fragment potential (EFP) method is used to study the effect of two, four, six, and eight solvating water molecules on the Menshutkin reaction between ammonia and methyl bromide. The EFP method reproduces all ab initio geometries and energetics (including zero-point energy, thermal, and entropy effects) for the two-water case very accurately. Energetics from all nb initio single-point energies at the EFP geometries for the four, six, and eight water cases are in excellent agreement with corresponding EFP energetics. In the gas phase, the above Menshutkin reaction is kinetically highly unfavorable with a free energy of activation (at 298.15 K) of 40.6 kcal/mol at the RHF level with a double-xi basis set augmented with polarization and diffuse functions. An ion-pair product is found, in agreement with previous work, in which the bromide anion is hydrogen-bonded to an ammonium hydrogen, giving a free energy of reaction of 2.8 kcal/mol. The addition of solvating water molecules has the effect of lowering the barrier and lowering the energy of the ion-pair product relative to the molecule-pair reactant. For eight solvating EFP water molecules, the free energy of activation is 22.8 kcal/mol and the free energy of reaction is -21.9 kcal/mol. Timings indicate that the EFP method allows the inexpensive addition of water molecules to a chemical system, accurately modeling all ab initio calculations with low computational cost.

Abstract:

Calculations on NaCl microsolvated with up to 10 water molecules were performed by treating the NaCl with restricted Hartree-Fock (RHF) 6-31G(d) ab initio wave functions and the waters with the effective fragment potential (EFP) model. Increasing numbers of local minima are found with the addition of successive water molecules. As the number of waters is increased, the Boltzmann-averaged NaCl bond distance lengthens from 2.397 to 3.167 Angstrom, the Mulliken charges increase to +/-0.98, and the Boltzmann-averaged NaCl stretching frequency decreases from 359 to 94 cm(-1). The incremental binding energies (in kcal/mol) as a function of n are 15.4 (n = 1), 15.1 (n = 2), 14.0 (n = 3), 13.9 (n = 4), 12.3 (n = 6), 12.8 (n = 8), 9.1 (n = 10).

Abstract:

An overlap dependent formula for evaluating the charge penetration energy between non-orthogonal molecular orbitals is derived using the Spherical Gaussian Overlap approximation. When combined with an accurate multipole representation of the electrostatic energy, such as in the effective fragment potential method, ab initio electrostatic energies are generally reproduced to within 0.2 kcal/mol for a variety of molecular dimers and basis sets. The only larger error is for the DMSO dimer, where the electrostatic energy is overestimated by 0.7 kcal/mol.

Abstract:

Simulated annealing methods have been used with the effective fragment potential to locate the lowest energy structures for the water clusters (H2O)(n) with n=6, 8, 10, 12, 14, 16, 18, and 20. The most successful method uses a local minimization on each Monte Carlo step. The effective fragment potential method yielded interaction energies in excellent agreement with those calculated at the ab initio Hartree-Fock level and was quite successful at predicting the same energy ordering as the higher-level perturbation theory and coupled cluster methods. Analysis of the molecular interaction energies in terms of its electrostatic, polarization, and exchange-repulsion/charge-transfer components reveals that the electrostatic contribution is the dominant term in determining the energy ordering of the minima on the (H2O)(n) potential energy surfaces, but that differences in the polarization and repulsion components can be important in some cases.