Recent applications of density functional theory calculations to biomolecules

The purpose of this overview is to highlight the broad scope and utility of current applications of density functional theory (DFT) methods for the study of the properties and reactions of biomolecules. This is illustrated using examples selected from research carried out within our research group and in collaboration with others. The examples include the hyperfine coupling constants of amino acid radicals, the use of an amino acid as a chiral catalyst for the formation of carbon–carbon bonds in the aldol reaction, hydrogen-bond mediated catalysis of an aminolysis reaction, radiation-induced protein–DNA cross-links, and the mechanism by which an antitumor drug cleaves DNA. We demonstrate that DFT-based methods can be applied successfully to a broad range of problems that remain beyond the scope of conventional electron-correlation methods. Furthermore, we show that contemporary computational quantum chemistry complements experiment in the study of biological systems. Received: 19 December 2001 / Accepted: 8 April 2002 / Published online: 4 July 2002     

Action of HCl on 3-hydroxypyrazolo(iso)quinolines to give 1-chloropyrazoles: evidence for an addition–elimination mechanism by ab initio calculations in gas phase and water

 Addition–elimination reactions involving a nucleophile and a remote leaving group [S^H _N(AE)^tele] are well-known under basic conditions, especially amongst electron-poor six-membered heterocycles, but are less commonly encountered for five-membered heterocycles and are rare under acidic conditions. Concentrated HCl converts 1-hydroxy-1H-pyrazolo[3,4-c] isoquinoline and 1-hydroxy-1H-pyrazolo[3,4-c]quinoline into 3-chloro-1H-pyrazolo[3,4-c]isoquinoline and 3-chloro-1H-pyrazolo[3,4-c]quinoline, respectively. However, apparently neither the isomeric 1-hydroxy-1H-pyrazolo[4,3-c](iso)-quinolines nor the parent 1-hydroxypyrazole undergo this reaction. Additionally, all these systems are refractory under basic conditions. We present a plausible mechanism for the reaction, involving the 3-addition of Cl^- to the diprotonated heterocycle, followed by the elimination of water. Calculations of the initial transition states and intermediates, using optimisation at B3LYP/6-311+G(d,p), including thermochemistry [HF/6-31+G(d)], and single-point Poisson–Boltzmann self-consistent reaction field determination of the free energy of solvation (Jaguar Poisson–Boltzmann self-consistent reaction field), support this mechanism and reproduce the observed order of reactivity, the addition step being 2–4 kcal less favourable for the isomeric 1-hydroxy-1H-pyrazolo[4,3-c](iso)quinolines and provide a rationalisation for the role of strong acid. Received: 27 June 2002 / Accepted: 6 September 2002 / Published online: 14 February 2003     

A density functional theory study of the conformational properties of 1,2-ethanediamine: protonation and solvent effects

 The structures and the conformational energies of nonprotonated, monoprotonated and diprotonated 1,2-ethanediamine have been investigated through density functional theory. The relative performance of local and gradient-corrected functionals is discussed. The existence of hydrogen-bond formation has been determined with electron localisation function calculations. Proton affinities for nonprotonated and monoprotonated 1,2-ethanediamine have been calculated and are in agreement with experimental data. The influence of solvation has been accounted for through the self-consistent isodensity polarisable continuum model. The results for the nonprotonated conformers show that solvation stabilises those conformers which have the lone pair in an antiperiplanar conformation. Solvation of the monoprotonated conformer stabilises significantly the “anti” conformation, which is unstable in the gas phase. For the di-protonated species, solvation stabilises slightly the gauche conformer, which is unstable in the gas phase. Received: 28 September 1999 / Accepted: 2 May 2000 / Published online: 27 September 2000     

A density functional theory study of a concerted mechanism for proton exchange between amino acid side chains and water

 A concerted mechanism for proton exchange between water and the amino acid side chains of cysteine, serine, arginine and glutamic acid has been investigated with hybrid density functional theory. The models used include, besides the amino acid side chain, a number of water molecules ranging from one to five in some cases. The modeling of the amino acids without their backbones is shown to be an excellent approximation. Long-range polarization effects were incorporated through a dielectric cavity method allowing a better comparison to existing measurements for free amino acids in water. The barriers converge rather fast with the number of water molecules for all the present amino acids and the converged values are in reasonable agreement with experiments with discrepancies in the range 2–6 kcal/mol. The dielectric effects were found to be small for all systems except cysteine, where there is a lowering of the barrier by 3–5 kcal/mol. The transition states for these concerted pathways form rings in which the separated charges can be stabilized. Received: 25 October 1999 / Accepted: 5 April 2000 / Published online: 21 June 2000     

The electron affinities of transition metal atoms at the CCSD(T) and density functional levels of theory

 The electron affinities of Ti, V, Cr, Fe, Co, Ni, and Cu are computed using the density function theory and CCSD(T) approaches. Overall the CCSD(T) approach yields the best results. For this property, the B3LYP, BLYP, and BP86 functionals perform better than the BPW91, PBEPBE, and PBE1PBE ones. The accuracy of all the methods is higher if the number of 3delectrons is the same in the neutral atom and the anion. This is especially true for the density functional theory methods. Received: 23 January 2002 / Accepted: 1 April 2002 / Published online: 24 June 2002     

Direct ab initio dynamics calculations on the rate constants for the hydrogen-abstraction reaction of C2H5F with O (3P)

The hydrogen-abstraction reaction C₂H₅F+O → C₂H₄F+OH has been studied by a dual-level direct dynamics method. For the reaction, three reaction channels, one for α-abstraction and two for β-abstraction, have been identified. The potential-energy surface information is obtained at the MP2(full)/6-311G(d,p) and PMP2(full)/6-311G(3df,3pd) (single-point) levels. By canonical variational transition-state theory, rate constants for each reaction channel are calculated with a small-curvature tunneling correction. The total rate constant is calculated from the sum of the individual rate constants and the temperature dependence of the branching ratios is obtained over a wide range of temperatures from 300 to 5,000 K. The agreement of the rate constants with experiment is good in the experimental temperature range from 1,000 to 1,250 K. The calculated results indicate that at low temperatures α-abstraction is most likely to be the major reaction channel, while β-abstraction channels will significantly contribute to the whole reaction rate as the temperature increases. Received: 23 January 2002 / Accepted: 23 June 2002 / Published online: 20 September 2002     

Prediction of the Raman spectrum of the aqueous formate anion by a combined density functional theory and self-consistent-reaction-field study

Formate, HCOO⁻, is a small ion that can be used as a convenient model to represent general carboxylic acids as well as the terminal side chains of glutamic and aspartic amino acids in biological systems. Several ab initio techniques (Hartree–Fock and density functional theory methods) were used to study the formate anion in aqueous solution, where the solvent was treated as an electrostatic influence by means of a self-consistent-reaction-field (SCRF) method. The comparison of calculated frequencies with Raman experimental data constitutes a good test for the solvation model employed in this work. The results showed that the application of a SCRF method not only produced smaller errors but also improved the linear correlation between predicted and experimental values. This latter characteristic enables a more efficient application of a linear scaling factor, λ, which corrects the calculated frequencies. In fact, the application of λ to the values calculated for the continuum, when compared to those obtained for a vacuum, resulted in smaller root-mean-square errors of the deviations from the experimental data. Received: 15 May 2000 / Accepted: 19 June 2000 / Published online: 11 September 2000     

The calculation of electron localization and delocalization indices at the Hartree–Fock, density functional and post-Hartree–Fock levels of theory

 Localization, λ(A), and delocalization indices, δ(A,B), as defined in the atoms in molecules theory, are a convenient tool for the analysis of molecular electronic structure from an electron-pair perspective. These indices can be calculated at any level of theory, provided that first- and second-order electron densities are available. In particular, calculations at the Hartree–Fock (HF) and configuration interaction (CI) levels have been previously reported for many molecules. However, λ(A) and δ(A,B) cannot be calculated exactly in the framework of Kohn–Sham (KS) density functional theory (DFT), where the electron-pair density is not defined. As a practical workaround, one can derive a HF-like electron-pair density from the KS orbitals and calculate approximate localization and delocalization indices at the DFT level. Recently, several calculations using this approach have been reported. Here we present HF, CI and approximate DFT calculations of λ(A) and δ(A,B) values for a number of molecules. Furthermore, we also perform approximate CI calculations using the HF formalism to obtain the electron-pair density. In general, the approximate DFT and CI results are closer to the HF results than to the CI ones. Indeed, the approximate calculations take into account Coulomb electron correlation effects on the first-order electron density but not on the electron-pair density. In summary, approximate DFT and CI localization and delocalization indices are easy to calculate and can be useful in the analysis of molecular electronic structure; however, one should take into account that this approximation increases systematically the delocalization between covalently bonded atoms, with respect to the exact CI results. Received: 13 February 2002 / Accepted: 24 April 2002 / Published online: 18 June 2002     

A density functional theory study of bridging and terminal oxotitanium(IV) oligomeric and polymeric linear titanoxanes

 This work deals with theoretical investigations on the oxygen–transition metal bond in systems containing linear chains of Ti–O units. From an experimental point of view, in the recent past a number of systems containing linearly arranged Ti–O units were synthesized, in which the Ti atom is complexed with Schiff bases such as acacen and salen. The theoretical study presented here has been carried out applying the density functional theory to model compounds of these systems, in order to shed light on the interactions between the transition metal and oxygen. Calculations have been performed on Ti–O oligomers (dimers, trimers and tetramers) by means of density functional theory at the gradient-corrected level of theory, optimizing the molecular geometries. Calculations have also been performed on linear polymers of the same systems, applying periodic boundary conditions, in order to compare the results with those of oligomeric analogues. Received: 12 January 2002 / Accepted: 16 April 2002 / Published online: 5 July 2002     

Molecular structure and vibrational spectra of indole and 5-aminoindole by density functional theory and ab initio Hartree–Fock calculations

The molecular geometry and vibrational frequencies of indole and 5-aminoindole in the ground state have been calculated by using the Hartree–Fock and density functional method with 6-311++G(d,p) basis set. The optimized geometrical parameters obtained by DFT calculations are in good agreement with the experimental values. Comparison of the observed fundamental vibrational frequencies of indole and 5-aminoindole with the calculated results by density functional and Hartree–Fock methods indicates that B3LYP is superior to the scaled Hartree–Fock approach for molecular vibrational problems. The theoretical spectrograms for FT-IR spectrum of 5-aminoindole have been constructed.     

Analysis of vibrational spectra of 2 and 3-methylpiperidine based on density functional theory calculations

The experimental and theoretical vibrational spectra of 2 and 3-methylpiperidine (abbreviated as 2-MP and 3-MP) were studied. The FT-Infrared spectra of 2-MP and 3-MP molecules were recorded in the liquid phase. The structural and spectroscopic analysis of the title molecules were made by using density functional harmonic calculations. For the title molecules, only one form was found most stable structure by using B3LYP level with the 6-311G (d,p) basis set. Selected experimental bands were assigned and characterized based on the scaled theoretical wave numbers by their total energy distribution (TED).     

Theoretical molecular structures for partially bonded complexes of trimethylamine with SO2 and SO3: ab initio and density functional theory calculations

The self-consistent reaction field (SCRF) method based on Onsager's reaction field theory is applied to investigate the effect of polar media on molecular structures of complexes of trimethylamime (TMA) with SO_x (x=2,3). The calculated SCRF N–S bond lengths at the MPW1PW91/6-311+G(3df) level are in satisfactory agreement with the experimental N–S bond lengths for the TMA–SO_x upon crystallization. The results are enough to demonstrate the usefulness of the reaction field theory in providing qualitative understanding of the medium effect on the partially bonded system such as TMA–SO_x.     

Role of twisting and sliding on the solvation of a stacked cytosine dimer: an ab initio study

Ab initio calculations with inclusion of correlation effects at the MP2/6-31G* level have been used to predict the interaction energy of stacked cytosine dimer (C/C) as a function of twisting and sliding in the gas phase. Systematic calculations have also been carried out on the solvation free energies of various rotated and translated C/C dimers using a polarized continuum model approach at the HF/6-31G* level with a view to probe the role of various degrees of freedom on the free energy of solvation of the C/C dimer. The interaction energy of the C/C dimer decreases upon changing from a parallel to an antiparallel conformation in the gas phase. The 180°-rotated conformation has been found to be the most stable arrangement when compared to other rotated positions. The rotated and translated dimers exhibit lower solvation free energy than the parallel conformation. The decrease in the dipole moment upon rotation from the parallel to the antiparallel conformation indicates the cancellation of charge distribution upon rotation in the z direction of one cytosine base with respect to the other. The calculation reveals that the present approach could not yield association energy, ΔΔG _Asso, in a solvent medium. This may be due to the fact that in the case of floppy molecules the contribution from translational, rotational and vibrational free energies plays a significant role in the calculation of ΔΔG _Asso. Received: 13 December 2001 / Accepted: 25 March 2002 / Published online: 13 June 2002     

High level ab initio and density functional theory study of bond selective dissociation of CH3SH and CH3CH2SH radical cations

Selective bond dissociation energies for CH₃SH and CH₃CH₂SH radical cations were evaluated with G1, G2, G2MP2, B3LYP, BLYP, and SVWN computational methods. It was determined that both G2 and CBSQ evaluate very accurate bond dissociation energies for thiol radical cations, while gradient-corrected BLYP computes the best energies of three employed DFT methods. For the CH₃CH₂SH radical cation, new, higher than previously estimated selective bond dissociation energies were suggested. Received: 10 September 1997 / Accepted: 9 September 1998 / Published online: 11 November 1998     

Computation of the influence of chemical substitution on the pK a of pyridine using semiempirical and ab initio methods

The physical properties of chemicals are strongly influenced by their protonation state, affecting, for example, solubility or hydrogen-bonding characteristics. The ability to accurately calculate protonation states in the form of pK _as is, therefore, desirable. Calculations of pK _a changes in a series of substituted pyridines are presented. Computations were performed using both ab initio and semiempirical approaches, including free energies of solvation via reaction-field models. The selected methods are readily accessible with respect to both software and computational feasibility. Comparison of calculated and experimental pK _as shows the experimental trends to be reasonably reproduced by the computations with root-mean-square differences ranging from 1.22 to 4.14 pK _a units. Of the theoretical methods applied the best agreement occurred using the second-order M?ller–Plesset/6-31G(d)/isodensity surface polarized continuum solvation model, while the more computationally accessible Austin model 1/Solvent model 2 (SM2) approach yielded results similar to the ab initio methods. Analysis of component contributions to the calculated pK _as indicates the largest source of error to be associated with the free energies of solvation of the protonated species followed by the gas-phase protonation energies; while the latter may be improved via the use of higher levels of theory, enhancements in the former require improvements in the solvation models. The inclusion of alternate minimum in the computation of pK _as is also indicated to contribute to differences between experimental and calculated pK _a values. Received: 27 April 1999 / Accepted: 27 July 1999 / Published online: 2 November 1999     

Orbital-band interactions and the reactivity of molecules on oxide surfaces: from explanations to predictions

 The surface chemistry of oxides is relevant for many technological applications: catalysis, photoelectrolysis, electronic-device fabrication, prevention of corrosion, sensor development, etc. This article reviews recent theoretical works that deal with the surface chemistry of oxides. The account begins with a discussion of results for the adsorption of CO and NO on oxides, systems which have been extensively studied in the literature and constitute an ideal benchmark for testing the quality of different levels of theory. Then, systematic studies concerned with the behavior of adsorbied alkali metals and sulfur-containing molecules are presented. Finally, a correlation between the electronic and chemical properties of mixed-metal oxides is analyzed and basic principles for designing chemically active oxides are introduced. Advances in theoretical methods and computer capabilities have made possible a fundamental understanding of many phenomena associated with the chemistry of molecules on oxide surfaces. Still many problems in this area remain as a challenge, and the approximate nature of most theoretical methods makes necessary a close coupling between theory and experiment. Following this multidisciplinary approach, the importance of band-orbital interactions for the reactivity of oxide surfaces has become clear. Simple models based on band-orbital mixing can explain trends found for the interaction of many adsorbates with oxide surfaces. These simple models provide a conceptual framework for modifying or controlling the chemical activity of pure oxides and for engineering mixed-metal oxides. In this respect, theoretical calculations can be very useful for predicting the best ways for enhancing the reactivity of oxide systems and reducing the waste of time, energy and materials characteristic of an empirical design. Received: 21 June 2001 / Accepted: 8 October 2001 / Published online: 1 February 2002