Computational structural determination and energy landscape analysis of the hepatic carcinogen 2-(acetylamino)fluorene

 2-(Acetylamino)fluorene (AAF), a potent mutagen and a prototypical example of the mutagenic aromatic amines, forms covalent adducts to DNA after metabolic activation in the liver. A benchmark study of AAF is presented using a number of the most widely used molecular mechanics and semiempirical computational methods and models. The results are compared to higher-level quantum calculations and to experimentally obtained crystal structures. Hydrogen bonding between AAF molecules in the crystal phase complicates the direct comparison of gas-phase theoretical calculations with experiment, so Hartree–Fock (HF) and Becke–Perdew (BP) density functional theory (DFT) calculations are used as benchmarks for the semiempirical and molecular mechanics results. Systematic conformer searches and dihedral energy landscapes were carried out for AAF using the SYBYL and MMFF94 molecular mechanics force fields; the AM1, PM3 and MNDO semiempirical quantum mechanics methods; HF using the 3-21G*and 6-31G* basis sets; and DFT using the nonlocal BP functional and double numerical polarization basis sets. MMFF94, AM1, HF and DFT calculations all predict the same planar structures, whereas SYBYL, MNDO and PM3 all predict various nonplanar geometries. The AM1 energy landscape is in substantial agreement with HF and DFT predictions; MMFF94 is qualitatively similar to HF and DFT; and the MNDO, PM3 and SYBYL results are qualitatively different from the HF and DFT results and from each other. These results are attributed to deficiencies in MNDO, PM3 and SYBYL. The MNDO, PM3 and SYBYL models may be unreliable for compounds in which an amide group is immediately adjacent to an aromatic ring. Received: 26 May 2002 / Accepted: 12 December 2002 / Published online: 14 February 2003     

Theoretical calculation of electrode potentials: Electron-withdrawing compounds

The electrode potential of 2,3-dicyanobenzoquinone in aqueous solution has been calculated relative to parabenzoquinone using a thermodynamic cycle approach that includes accurate gasphase ab initio calculations and calculation of differences in free energies of hydration using the free-energy perturbation method. The discrepancy between the calculated and experimental electrode potential is disappointingly large (99 mV) compared to previous studies using this approach. This, along with the experimental evidence, suggests that the experimental value itself is too large and that theoretical approaches may indeed be as reliable as experimental ones for determining redox properties of molecules such as 2,3-dicyanobenzoquinone. In the light of this discrepancy we have examined the variation of the results with the basis set, inclusion of electron correlation and changes in the parameters used in the molecular dynamics free-energy simulations. The results are shown to be dependent upon the torsional parameters and especially dependent upon the basis set or semiempirical method used to obtain the electrostatic potential-derived charges. The best charge set was determined using the ab initio criteria of completeness—as far as it can be applied to large molecules—and also by studying the effect of hydration on these charges. This was done by allowing the solvent to perturb the wave function prior to the electrostatic potential determination. Thus, 3-21G and 6-31G * basis sets were found to give satisfactory results. Similar results were obtained using semiempirical and ab initio geometries.     

A comparative molecular field analysis (CoMFA) study using semiempirical, density functional, ab initio methods and pharmacophore derivation using DISCOtech on sigma 1 ligands

The Comparative Molecular Field Analysis (CoMFA) was developed to investigate a three-dimensional quantitative structure activity relationship (3D-QSAR) model of ligands for the sigma 1 receptor. The starting geometry of sigma-1 receptor ligands was obtained from the Tripos force field minimizations and conformations were decided from DISCOtech using the SYBYL 6.8. program. The structures of 48 molecules were fully optimized at the ab initio HF/3-21G* and semiempirical AM1 calculations using GAUSSIAN 98. The electrostatic charges were calculated using several methods such as semiempirical AM1, density functional B3LYP/3-21G*, and ab initio HF/3-21G*, MP2/3-21G* calculations within GAUSSIAN 98. Using the optimized geometries, the CoMFA results derived from the HF/3-21G method were better than those from AM1. The best CoMFA was obtained from HF/3-21G* optimized geometry and charges (R2 = 0.977). Using the optimized geometries, the CoMFA results derived from the HF/3-21G methods were better than those from AM1 calculations. The training set of 43 molecules gave higher R2 (0.989-0.977) from HF/3-21G* optimized geometries than R2 (0.966-0.911) values from AM1 optimized geometries. The test set of five molecules also suggested that HF/3-21G* optimized geometries produced good CoMFA models to predict bioactivity of sigma 1 receptor ligands but AM1 optimized geometries failed to predict reasonable bioactivity of sigma 1 receptor ligands using different calculations for atomic charges.     

Theoretical studies on the effects of methods and parameterization on the calculated free energy of hydration for small molecules

Free energies of hydration (FEH) have been computed for 13 neutral and nine ionic species as a difference of theoretically calculated Gibbs free energies in solution and in the gas phase. In‐solution calculations have been performed using both SCIPCM and PCM polarizable continuum models at the density functional theory (DFT)/B3LYP and ab initio Hartree–Fock levels with two basis sets (6‐31G* and 6‐311++G**). Good linear correlation has been obtained for calculated and experimental gas‐phase dipole moments, with an increase by ～30% upon solvation due to solute polarization. The geometry distortion in solution turns out to be small, whereas solute polarization energies are up to 3 kcal/mol for neutral molecules. Calculation of free energies of hydration with PCM provides a balanced set of values with 6‐31G* and 6‐311++G** basis sets for neutral molecules and ionic species, respectively. Explicit solvent calculations within Monte Carlo simulations applying free energy perturbation methods have been considered for 12 neutral molecules. Four different partial atomic charge sets have been studied, obtained by a fit to the gas‐phase and in‐solution molecular electrostatic potentials at in‐solution optimized geometries. Calculated FEH values depend on the charge set and the atom model used. Results indicate a preference for the all‐atom model and partial charges obtained by a fit to the molecular electrostatic potential of the solute computed at the SCIPCM/B3LYP/6‐31G* level. © 2004 Wiley Periodicals, Inc. Int J Quantum Chem, 2004     

The structure of 1-chlorosilatrane: An ab initio molecular orbital and a density functional theory study

The molecular geometries of the 1-chloro-, 1-fluoro-, 1-methyl-, and 1-hydrogenosilatranes were fully optimized by the restricted Hartree-Fock (HF) method supplemented with 3-21G, 3-21G(d), 6-31G(d), and CEP-31G(d) basis sets; by MP2 calculations using 6-31G(d) and CEP-31G(d) basis sets; and by GGA-DFT calculations using 6-31G(d5) basis set with the aim of locating the positions of the local minima on the energy hypersurface. The HF/6-31G(d) calculations predict long (>254 pm) and the MP2/CEP calculations predicted short (∼225 pm) equilibrium Si(SINGLE BOND)N distances. The present GGA-DFT calculations reproduce the available gas phase experimental Si(SINGLE BOND)N distances correctly. The solid phase experimental results predict that the Si(SINGLE BOND)N distance is shorter in 1-chlorosilatrane than in 1-fluorosilatrane. In this respect the HF results show a strong basis set dependence, the MP2/CEP results contradict the experiment, and the GGA-DFT results in electrolytic medium agree with the experiment. The latter calculations predict that 1-chlorosilatrane is more polarizable than 1-fluorosilatrane and also support a general Si(SINGLE BOND)N distance shortening trend for silatranes during the transition from gas phase to polar liquid or solid phase. The calculations predict that the ethoxy links of the silatrane skeleton are flexible. Consequently, it is difficult to measure experimentally the related bond lengths and bond and torsion angles. This is the probable origin of the surprisingly large differences for the experimental structural parameters. On the basis of experimental analogies, ab initio calculations, and density functional theory (DFT) calculations, a gas phase equilibrium (r_e) geometry is predicted for 1-chlorosilatrane. The semiempirical methods predict a so-called exo minimum (at above 310 pm Si(SINGLE BOND)N distance); however, the ab initio and GGA-DFT calculations suggest that this form is nonexistent. The GGA-DFT geometry optima were characterized by frequency analysis. © 1996 by John Wiley & Sons, Inc.     

Hydrogen peroxide clusters: the role of open book motif in cage and helical structures

Hartree-Fock (HF) calculations using 6-31G*, 6-311++G(d,p), aug-cc-pVDZ, and aug-cc-pVTZ basis sets show that hydrogen peroxide molecular clusters tend to form hydrogen-bonded cyclic and cage structures along the lines expected of a molecule which can act as a proton donor as well as an acceptor. These results are reiterated by density functional theoretic (DFT) calculations with B3LYP parametrization and also by second-order M?ller-Plesset perturbation (MP2) theory using 6-31G* and 6-311++G(d,p) basis sets. Trends in stabilization energies and geometrical parameters obtained at the HF level using 6-311++G(d,p), aug-cc-pVDZ, and aug-cc-pVTZ basis sets are similar to those obtained from HF/6-31G* calculation. In addition, the HF calculations suggest the formation of stable helical structures for larger clusters, provided the neighbors form an open book structure.     

Benchmark calculations of reaction energies, barrier heights, and transition-state geometries for hydrogen abstraction from methanol by a hydrogen atom

We report benchmark calculations of reaction energies, barrier heights, and transition-state geometries for the reaction of CH(3)OH with H to produce CH(2)OH and H(2). Highly accurate composite methods, such as CBS, G2, G3S, G3X, G3SX, and multi-coefficient correlation methods (MCCMs), are used to calibrate lower-cost methods. We also performed single-level CCSD(T) calculations extrapolated to the infinite-basis limit on the basis of aug-cc-pVXZ (X = 3, 4) correlation consistent basis sets. The benchmark high-level calculations give consensus values of the forward reaction barrier height and the reaction energy of 9.7 kcal/mol and - 6.4 kcal/mol, respectively. To evaluate the accuracy of cost-efficient methods that are potentially useful for dynamics studies of the title reaction, we further include the results obtained by hybrid density functional theory methods and hybrid meta density functional theory methods that have recently been designed for chemical kinetics. Results obtained by popular semiempirical methods are also given for comparison. On the basis of the benchmark gas-phase results, we suggest MC-QCISD/3, MC3BB, and BB1K as reasonably accurate and affordable electronic structure methods for calculating dynamics for the title reaction.     

Ab initio procedure for aqueous-phase pKa calculation: the acidity of nitrous acid

We present an ab initio procedure for accurately calculating aqueous-phase pKa values and apply it to study the acidity of nitrous acid (HNO2, or HONO). The aqueous-phase pK(a) of nitrous acid was obtained from calculated gas-phase acidities and solvation free energies via a thermodynamic cycle and the solvation model chemistry of Barone et al. (J. Chem. Phys. 1997, 107, 3210). Solvation free energies were calculated at the HF/6-31G(d) level using the dielectric-polarizable continuum and the integral equation formalism-polarizable continuum solvent models (D-PCM and IEF-PCM, respectively), with the D-PCM model yielding the most accurate pKa values. For HF free energies of solvation, significant improvements in accuracy could be made by moving to the larger 6-311++G(3df,3pd) and aug-cc-pVQZ basis sets. Solvation free energies were also calculated using the density functional theory (DFT) methods B3LYP, TPSS, PBE0, B1B95, VSXC, B98 and O3LYP, with the most accurate methods being TPSS and VSXC, which provided average errors of less than 0.11 pKa units. Solvation free energies calculated with the different DFT methods were relatively insensitive to the basis set used. Our theoretical calculations are compared with experimental results obtained using stopped flow spectrophotometry. The pKa of nitrous acid was measured as 3.16 at 25 degrees C, and the enthalpy and entropy of nitrous acid dissociation were calculated from measurements as 6.7 kJ mol(-1) and -38.4 J mol(-1) K(-1), respectively, between 25 and 45 degrees C. The UV/visible absorption spectra of the nitrite ion and nitrous acid were also examined, and molar extinction coefficients were obtained for each.     

Theoretical study on infrared vibrational spectra of boric-acid in gas-phase using density functional methods

Density functional theory (DFT) methods with various exchange-correlation functionals such as SVWN, BVWN, BVWN5, BLYP, B1LYP, B3LYP, B3PW91, and BH and H are employed in a theoretical study of molecular boric-acid in gas-phase. In the calculations, the split valence 6-311++G** and 6-31G* basis sets were used. The geometry, zero-point vibrational energies (ZPVEs), and harmonic infrared vibrational (IR) frequencies are predicted. The calculated C_3h-symmetry geometrical parameters are compared with Hartree–Fock (HF) calculation results and experimental data. IR frequencies predicted by the BLYP, B3LYP, and B3PW91 calculations are in good agreement with experimental data. The frequency calculations presented here also suggest that the C_3h-symmetrical structure corresponds to a minimum in the potential energy surface, but neither is D_3h- or C₃-symmetrical structure.     

Ab initio molecular orbital study of 1-fluorosilatrane

The molecular geometry of 1-fluorosilatrane was optimized fully by restricted Hartree–Fock (HF) calculations using the 3-21G, 3-21G(d) and 6-31G(d) basis sets, with the aim of locating the positions of the local minima on the energy hypersurface. The optimized geometries were compared with available experimental (X-ray and ED) and semiempirical data. The ab initio calculations using polarized basis sets are in good agreement with those of previously reported semiempirical calculations, giving a slightly longer equilibrium Si? N distance (～ 256 pm) in the case of the endo minimum. However, the exo minimum predicted by the semiempirical methods is not supported. There was no experimental evidence for the existence of this exo minimum, and the present ab initio calculations suggest that it is highly unstable. There is considerable disagreement among the experimental results in the C? N and C? C bond lengths in various silatranes, their differences being as large as 13 pm. The present calculations predict that these differences may appear because the silatrane skeleton is flexible with low-energy, large-amplitude internal motions which introduce considerable uncertainties into the position of ring carbon atoms. © 1994 by John Wiley & Sons, Inc.     

Comparison of the use of the MNDO and MINDO/3 methods with ab initio methods to study tautomerism in histamine, 2- and 4-methylhistamine

The semiempirical MNDO and MINDO/3 methods are used to study the various tautomeric forms of histamine, 2-methylhistamine, and 4-methylhistamine. Comparisons of the optimized structures and tautomerization energies are made with values obtained from ab initio Hartree-Fock calculations using the 3-21G and STO-3G basis sets. Based on these results and previous comparisons of STO-3G results with x-ray structures, the present results indicate that while there are some differences in the values of the structural parameters, the changes in structure upon tautomerization and/or protonation are very similar. Further analysis of the MNDO and MINDO/3 structures by means of their utilization in 3-21G and STO-3G calculations indicates that either of these semiempirical methods provides reliable values for the structural parameters. Both methods give good qualitative agreement with the ab initio calculations for the relative energies of the various tautomers in the three compounds. In these studies the MNDO method appears to give better quantitative agreement with the 3-21G and STO-3G results than the MINDO/3 method.     

Proton affinities of molecules containing nitrogen and oxygen: Comparing density functional results to experiment

Summary Proton affinities were calculated using density functional theory for 11 small molecules whose primary protonation site is on nitrogen, and eight small molecules that protonate on oxygen. Calculations were performed using both the local spin density approximation and nonlocal gradient corrections to the exchange correlation functional. The results were not sensitive to whether the nonlocal gradient correction was implemented on the final local spin density optimized geometry or whether the correction was included in the self-consistent calculation of the energy at each optimization step. Although negligible basis set dependence was found using the analytic Gaussian basis sets, numerical basis sets required augmentation by a double set of polarization functions to achieve reasonable agreement with experiment. All calculations systematically underestimated oxygen proton affinities.     

Density functional calculation of quinone electrode potentials

Here, we have applied density functional methods, in combination with free energy hydration calculations, to calculate two-electron electrode potentials for quinones and naphthoquinones. While we find that the free-energy perturbation method, implemented within a molecular dynamics framework, is superior to the PM 3—SM 3 continuum method for determining free energies of hydration, the computationally less expensive PM 3—SM 3 method does perform well when there is not an internal hydrogen bond. Generally, all the density functional approaches investigated gave good energetics when applied to this problem, but the Beck '88—Vosko—Wilk—Nusair combination of functionals for the exchange-correlation energy gave the best results. The density functional results are marginally better than the Møller-Plesset second-order perturbation results. Moreover, because the results are obtained using a thermodynamic cycle which involves taking differences in total energies, the results are not too dependent on the quadrature scheme used to calculate the exchange-correlation energy. By using semiempirically optimized geometries and the PM 3—SM 3 method for determining free energies of hydration, it has been possible to calculate electrode potentials for a series of large molecules (naphthoquinones) to within about 30 mV of experiment. This result is extremely encouraging and shows that density functional methods offer great promise in the design of redox-active molecules such as bioreductive anticancer agents. © 1995 John Wiley & Sons, Inc.     

Implementation of divide-and-conquer method including Hartree-Fock exchange interaction

The divide-and-conquer (DC) method, which is one of the linear-scaling methods avoiding explicit diagonalization of the Fock matrix, has been applied mainly to pure density functional theory (DFT) or semiempirical molecular orbital calculations so far. The present study applies the DC method to such calculations including the Hartree-Fock (HF) exchange terms as the HF and hybrid HF/DFT. Reliability of the DC-HF and DC-hybrid HF/DFT is found to be strongly dependent on the cut-off radius, which defines the localization region in the DC formalism. This dependence on the cut-off radius is assessed from various points of view: that is, total energy, energy components, local energies, and density of states. Additionally, to accelerate the self-consistent field convergence in DC calculations, a new convergence technique is proposed.     

Performance of the elongation method with larger basis sets

The elongation method proposed by Imamura serves as a theoretical model for polymerization processes. It can now be used together with larger basis sets, Hartree–Fock and density functional methods from the Gaussian 94 package with direct self‐consistent field (SCF). This allows electronic structure calculation of elongating clusters with an efficiency superior to full cluster calculations and a precision superior to previous versions of our elongation method. Performance and accuracy compared with full cluster calculations on a regular polymer using the BLYP/6‐31G(d, p) method. Interaction energies of water and hydrogen fluoride polymers of increasing length are compared between HF, BLYP methods and 4‐31G, 6‐31G(d, p) basis sets: Diffuse and polarization functions have a large influence on the interaction energy on both polymers. Local density of states are calculated for different cluster lengths. They are in good agreement with full cluster calculations. ©1999 John Wiley & Sons, Inc. Int J Quant Chem 74: 35–47, 1999     

The geometry of pyrazole: A test for ab initio calculations

Ab initio calculations on the structure of pyrazole have been carried out at different levels of accuracy. At the Hartree-Fock (HF) level, the performance of several basis sets, namely 3-21G, 6-31G, 6-31G**, and 6–311G** was investigated. The influence of electron correlation effects also was studied by carrying out geometry optimizations at the MP2, MP4, and QCISD levels. The performance of a density functional method also was evaluated. We have also investigated the possible influence of the frozen core approximation on the final optimized geometry. Three different statistical analyses were considered in determining which geometry is closest to the experimental microwave geometry—namely Paul Curtin's diagrams, cluster analysis, and multidimensional scaling. From these analyses, we conclude that there is no asymptotic approach to the experimental geometry by increasing the quality of the theoretical model, although, as expected, the more reliable structures are those obtained at the MP2, MP4, and QCISD levels, as well as those obtained by the B3LYP density functional method. We have also found that the values of the rotational constants are a tight criterion to define the quality of a molecular geometry. © 1995 by John Wiley & Sons, Inc.     

Comparison of Some Computational Methods for Geometry Optimisation of Phosphorus Acid Derivatives

Several economical methods for geometry optimization, that should be applicable to larger molecules, have been evaluated for 19 phosphorus acid derivatives. MP2/cc-pVDZ geometry optimizations are used as reference points and the geometries obtained from the other methods are evaluated with respect to deviations in bond lengths and angles, from the reference geometries. The geometry optimization methods are also compared to the much used B3LYP/6-31G(d) method. Single point energies obtained by subsequent EDF1/6-31+G(d) or B3LYP/6-31+G(d,p) calculations on the respective equilibrium geometries are also reported relative to the energies obtained from the reference geometries. The geometries from HF/MIDI! optimizations were closer to those of the references than the geometries of the HF/3-21G(d), HF/6-31G(d), and B3LYP/MIDI! optimizations. The EDF1/6-31+G(d) or B3LYP/6-31+G(d,p) single point energies obtained from the HF/3-21G(d), HF/6-31G(d), and B3LYP/MIDI! geometries gave a mean absolute deviation (MAD) from that of the reference geometries of 1.4-3.9 kcal mol m 1 . The HF/MIDI! geometries, however, gave EDF1/6-31+G(d) and B3LYP/6-31+G(d,p) energies with a MAD of only about 0.5 and 0.55 kcal mol m 1 respectively from the energies obtained with the reference geometries. Thus, use of HF/MIDI! for geometry optimization of phosphorus acids is a method that gives geometries of near-MP2 quality, resulting in a fair accuracy of energies in subsequent single point calculations, at a much lower computational cost other methods that give similar accuracies.     

Evaluation of the 1-octanol/water partition coefficient of nucleoside analogs via free energy estimated in quantum chemical calculations

The partition coefficients (logP) of nucleoside analogs determined by the difference in the free energies of hydration and solvation in water-saturated octanol using the thermodynamic integration method are reported. The logP values calculated in this approach are closer to the experimental values compared to other ab initio methods. Solvation free energy in water and octanol, free energy of cavity formation in water and Henry’s constants, and some other parameters are estimated at the density functional theory (DFT) and Hartree-Fock level with 6–31G*, 6–31G, and 6–31+G basis sets. Surface area, mass, refractivity, volume, polarizability, and dipole moment are calculated for some drugs with HF and DFT methods. The results show that log P decreases with the decrease in polarizability and the increase in dipole moment.     

Implementation of the SCC-DFTB method for hybrid QM/MM simulations within the amber molecular dynamics package

Self-consistent charge density functional tight-binding (SCC-DFTB) is a semiempirical method based on density functional theory and has in many cases been shown to provide relative energies and geometries comparable in accuracy to full DFT or ab initio MP2 calculations using large basis sets. This article shows an implementation of the SCC-DFTB method as part of the new QM/MM support in the AMBER 9 molecular dynamics program suite. Details of the implementation and examples of applications are shown.     

Calculation of standard electrode potential of half reaction for benzoquinone and hydroquinone

Geometric parameters, the vibrational frequencies and thermochemical values of benzoquinone and hydroquinone were computed using ab initio molecular orbital calculations (HF) and density function theory (B3LYP) methods with the 6-31G(d) basis set, respectively. The calculated frequencies for benzoquinone and hydroquinone were used for the assignment of the IR frequencies observed in the experimental IR spectrum. Cyclic voltammetry with a glassy carbon electrode of hydroquinone solutions in phosphate buffers at pH 7.0 showed that standard electrode potential of half reaction for benzoquinone and hydroquinone is 0.714V. Standard electrode potential of half reaction for benzoquinone and hydroquinone was calculated using the sum of electronic and thermal free energies, enthalpies of sublimation and energies of solvation for benzoquinone and hydroquinone.