Representation of Zn(II) complexes in polarizable molecular mechanics. Further refinements of the electrostatic and short-range contributions. Comparisons with parallel ab initio computations

We present refinements of the SIBFA molecular mechanics procedure to represent the intermolecular interaction energies of Zn(II). The two first-order contributions, electrostatic (E(MTP)), and short-range repulsion (E(rep)), are refined following the recent developments due to Piquemal et al. (Piquemal et al. J Phys Chem A 2003, 107, 9800; and Piquemal et al., submitted). Thus, E(MTP) is augmented with a penetration component, E(pen), which accounts for the effects of reduction in electronic density of a given molecular fragment sensed by another interacting fragment upon mutual overlap. E(pen) is fit in a limited number of selected Zn(II)-mono-ligated complexes so that the sum of E(MTP) and E(pen) reproduces the Coulomb contribution E(c) from an ab initio Hartree-Fock energy decomposition procedure. Denoting by S, the overlap matrix between localized orbitals on the interacting monomers, and by R, the distance between their centroids, E(rep) is expressed by a S(2)/R term now augmented with an S(2)/R(2) one. It is calibrated in selected monoligated Zn(II) complexes to fit the corresponding exchange repulsion E(exch) from ab initio energy decomposition, and no longer as previously the difference between (E(c) + E(exch)) and E(MTP). Along with the reformulation of the first-order contributions, a limited recalibration of the second-order contributions was carried out. As in our original formulation (Gresh, J Comput Chem 1995, 16, 856), the Zn(II) parameters for each energy contribution were calibrated to reproduce the radial behavior of its ab initio HF counterpart in monoligated complexes with N, O, and S ligands. The SIBFA procedure was subsequently validated by comparisons with parallel ab initio computations on several Zn(II) polyligated complexes, including binuclear Zn(II) complexes as in models for the Gal4 and beta-lactamase metalloproteins. The largest relative error with respect to the RVS computations is 3%, and the ordering in relative energies of competing structures reproduced even though the absolute numerical values of the ab initio interaction energies can be as large as 1220 kcal/mol. A term-to-term identification of the SIBFA contributions to their ab initio counterparts remained possible even for the largest sized complexes.     

Binding of D- and L-captopril inhibitors to metallo-beta-lactamase studied by polarizable molecular mechanics and quantum mechanics

The bacterial Zn2+ metallo-beta-lactamase from B. fragilis is a zinc-enzyme with two potential metal ion binding sites. It cleaves the lactam ring of antibiotics, thus contributing to the acquired resistance of bacteria against antibiotics. The present study bears on the binuclear form of the enzyme. We compare several possible binding modes of captopril, a mercaptocarboxamide inhibitor of several zinc-metalloenzymes. Two diastereoisomers of captopril were considered, with either a D- or an L-proline residue. We have used the polarizable molecular mechanics procedure SIBFA (Sum of Interactions Between Fragments ab initio computed). Two beta-lactamase models were considered, encompassing 104 and 188 residues, respectively. The energy balances included the inter and intramolecular interaction energies as well as the contribution from solvation computed using a continuum reaction field procedure. The thiolate ion of the inhibitor is binding to both metal ions, expelling the bridging solvent molecule from the uncomplexed enzyme. Different competing binding modes of captopril were considered, either where the inhibitor binds in a monodentate mode to the zinc cations only with its thiolate ion, or in bidentate modes involving additional zinc binding by its carboxylate or ketone carbonyl groups. The additional coordination by the inhibitor's carboxylate or carbonyl group always occurs at the zinc ion, which is bound by a histidine, a cysteine, and an aspartate side chain. For both diastereomers, the energy balances favor monodentate binding of captopril via S-. The preference over bidentate binding is small. The interaction energies were recomputed in model sites restricted to captopril, the Zn2+ cations, and their coordinating end side chains from beta-lactamase (98 atoms). The interaction energies and their ranking among competing arrangements were consistent with those computed by ab initio HF and DFT procedures.     

Interaction of neutral and zwitterionic glycine with Zn2+ in gas phase: ab initio and SIBFA molecular mechanics calculations

The interaction of Zn²⁺ with glycine (Gly) in the gas phase is studied by a combination of ab initio and molecular mechanics techniques. The structures and energetics of the various isomers of the Gly–Zn²⁺ complex are first established via high‐level ab initio calculations. Two low‐energy isomers are characterized: one in which the metal ion interacts with the carboxylate end of zwitterionic glycine, and another in which it chelates the amino nitrogen and the carbonyl oygen of neutral glycine. These calculations lead to the first accurate value of the gas‐phase affinity of glycine for Zn²⁺. Ab initio calculations were also used to evaluate the performance of various implementations of the SIBFA force field. To assess the extent of transferability of the distributed multipoles and polarizabilities used in the SIBFA computations, two approaches are followed. In the first, approach (a), these quantities are extracted from the ab initio Hartree–Fock wave functions of glycine or its zwitterion in its entirety, and for each individual Zn²⁺‐binding conformation. In the second, approach (b), they are assembled from the appropriate constitutive fragments, namely methylamine and formic acid for neutral glycine, and protonated methylamine and formate for the zwitterion; they undergo the appropriate vector or matrix rotation to be assembled in the conformation studied. The values of the Zn²⁺–glycine interaction energies are compared to those resulting from ab initio SCF and MP2 computations using both the all‐electron 6‐311+G(2d,2p) basis set and an effective core potential together with the valence CEP 4‐31G(2d) basis set. Approach (a) values closely reproduce the ab initio ones, both in terms of the total interaction energies and of the individual components. Approach (b) can provide a similar match to ab initio interaction energies as does approach (a), provided that the two constitutive Gly building blocks are considered as separate entities having mutual interactions that are computed simultaneously with those occurring with Zn²⁺. Thus, the supermolecule is treated as a three‐body rather than a two‐body system. These results indicate that the current implementation of the SIBFA force field should be adequate to undertake accurate studies on zinc metallopeptides. © 2000 John Wiley & Sons, Inc. J Comput Chem 21: 963–973, 2000     

Interaction energies for the purine inhibitor roscovitine with cyclin-dependent kinase 2: correlated ab initio quantum-chemical, DFT and empirical calculations

The interaction between roscovitine and cyclin-dependent kinase 2 (cdk2) was investigated by performing correlated ab initio quantum-chemical calculations. The whole protein was fragmented into smaller systems consisting of one or a few amino acids, and the interaction energies of these fragments with roscovitine were determined by using the MP2 method with the extended aug-cc-pVDZ basis set. For selected complexes, the complete basis set limit MP2 interaction energies, as well as the coupled-cluster corrections with inclusion of single, double and noninteractive triples contributions [CCSD(T)], were also evaluated. The energies of interaction between roscovitine and small fragments and between roscovitine and substantial sections of protein (722 atoms) were also computed by using density-functional tight-binding methods covering dispersion energy (DFTB-D) and the Cornell empirical potential. Total stabilisation energy originates predominantly from dispersion energy and methods that do not account for the dispersion energy cannot, therefore, be recommended for the study of protein-inhibitor interactions. The Cornell empirical potential describes reasonably well the interaction between roscovitine and protein; therefore, this method can be applied in future thermodynamic calculations. A limited number of amino acid residues contribute significantly to the binding of roscovitine and cdk2, whereas a rather large number of amino acids make a negligible contribution.     

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.     

Conformation-dependent intermolecular interaction energies of the triphosphate anion with divalent metal cations. Application to the ATP-binding site of a binuclear bacterial enzyme. A parallel quantum chemical and polarizable molecular mechanics investigation

We have explored the conformation-dependent interaction energy of the triphosphate moiety, a key constituent of ATP and GTP, with a closed-shell divalent cation, Zn2+, used as a probe. This was done using the SIBFA polarizable molecular mechanics procedure. We have resorted to a previously developed approach in which triphosphate is built out from its elementary constitutive fragments, and the intramolecular, interfragment, interaction energies are computed simultaneously with their intermolecular interactions with the divalent cation. This approach has enabled reproduction of the values of the intermolecular interaction energies from ab initio quantum-chemistry with relative errors <3%. It was extended to the complex of a nonhydrolyzable analog of ATP with the active site of a bacterial enzyme having two Mg2+ cations as cofactors. We obtained following energy-minimization a very close overlap of the ATP analog over its position from X-ray crystallography. For models of the ATP analog-enzyme complex encompassing up to 169 atoms, the values of the SIBFA interaction energies were found to match their DFT counterparts with relative errors of <2%.     

Theoretical study on the hydration of hydrogen peroxide in terms of ab initio method and atom-bond electronegativity equalization method fused into molecular mechanics

In this paper, the interaction between hydrogen peroxide (HP) and water were systemically studied by atom-bond electronegativity equalization method fused into molecular mechanics (ABEEM/MM) and ab initio method. The results show that the optimized geometries, interaction energies and dipole moments of hydrated HP clusters HP(H₂O) _n (n = 1–6) calculated by ABEEM/MM model are fairly consistent with the MP2/aug-cc-pVTZ//MP2/aug-cc-pVDZ results. The ABEEM/MM results indicate that n = 4 is the transition state structure from 2D planar structure to 3D network structure. The variations of the average hydrogen bond length with the increasing number of water molecules given by ABEEM/MM model agree well with those of ab initio studies. Moreover, the radial distribution functions (RDFs) of water molecule around HP in HP aqueous solution have been analyzed in detail. It can be confirmed that HP is a good proton donor and poor proton acceptor in aqueous solution by analysis of the RDFs.     

A Conformational Study on Silacyclohexane. Comparison of ab initio (HF,MP2), DFT,and Molecular Mechanics Calculations. Conformational Energy Surface of Silacyclohexane

The structures and relative energies for the basic conformations of silacyclohexane 1 have been calculated using HF, RI‐MP2, RI‐DFT and MM3 methods. All methods predict the chair form to be the dominant conformation and all of them predict structures which are in good agreement with experimental data. The conformational energy surface of 1 has been calculated using MM3. It is found that there are two symmetric lowest energy pathways for the chair‐to‐chair inversion. Each of them consists of two sofa‐like transition states, two twist forms with C₁ symmetry (twist‐C₁), two boat forms with Si in a gunnel position (C₁ symmetry), and one twist form with C₂ symmetry (twist‐C₂). All methods calculate the relative energy to increase in the order chair < twist‐C₂ < twist‐C₁ < boat. At the MP2 level of theory and using TZVP and TZVPP (Si atoms) basis sets the relative energies are calculated to be 3.76, 4.80, and 5.47 kcal mol^–1 for the twist‐C₂, twist‐C₁, and boat conformations, respectively. The energy barrier from the chair to the twisted conformations of 1 is found to be 6.6 and 5.7 kcal mol^–1 from MM3 and RI‐DFT calculations, respectively. The boat form with Si at the prow (C_s symmetry) does not correspond to a local minimum nor a saddle point on the MM3 energy surface, whereas a RI‐DFT optimization under C_s symmetry constraint resulted in a local minimum. In both cases its energy is above that of the chair‐to‐twist‐C₁ transition state, however, and it is clearly not a part of the chair‐to‐chair inversion.     

Parallel ab initio and molecular mechanics investigation of polycoordinated Zn(II) complexes with model hard and soft ligands: Variations of binding energy and of its components with number and charges of ligands

In this study we compare the binding energies of polycoordinated complexes of Zn²⁺ within cavities composed of model “hard” (H₂O, OH⁻) or “soft” (CH₃SH, CH₃S⁻) ligands. Ab initio supermolecule computations are performed at the HF and MP2 levels using extended basis sets to determine the binding energies and their components as a function of: the number of ligands, ranging from three to six; the net charge of the cavity; and the “hard” versus “soft” character of the ligands. These ab initio computations are used to test the reliability of the SIBFA molecular mechanics procedure, originally formulated and calibrated on the basis of ab initio computations, for such charged systems. The SIBFA intermolecular interaction energies match the corresponding ab initio values using a coreless effective potential split‐valence basis set with a relative error of ≤3%. Extensions to binuclear Zn²⁺ complexes, such as those that occur in the Zn‐binding sites of Gal4 and β‐lactamase proteins, are performed to test the applicability of the methodology for such systems. © 2000 John Wiley & Sons, Inc. J Comput Chem 21: 1011–1039, 2000     

Complexes of thiomandelate and captopril mercaptocarboxylate inhibitors to metallo-beta-lactamase by polarizable molecular mechanics. Validation on model binding sites by quantum chemistry

Using the polarizable molecular mechanics method SIBFA, we have performed a search for the most stable binding modes of D- and L-thiomandelate to a 104-residue model of the metallo-beta-lactamase from B. fragilis, an enzyme involved in the acquired resistance of bacteria to antibiotics. Energy balances taking into account solvation effects computed with a continuum reaction field procedure indicated the D-isomer to be more stably bound than the L-one, conform to the experimental result. The most stably bound complex has the S(-) ligand bridging monodentately the two Zn(II) cations and one carboxylate O(-) H-bonded to the Asn193 side chain. We have validated the SIBFA energy results by performing additional SIBFA as well as quantum chemical (QC) calculations on small (88 atoms) model complexes extracted from the 104-residue complexes, which include the residues involved in inhibitor binding. Computations were done in parallel using uncorrelated (HF) as well as correlated (DFT, LMP2, MP2) computations, and the comparisons extended to corresponding captopril complexes (Antony et al., J Comput Chem 2002, 23, 1281). The magnitudes of the SIBFA intermolecular interaction energies were found to correctly reproduce their QC counterparts and their trends for a total of twenty complexes.     

Intramolecular chelation of Zn2+ by α‐ and β‐mercaptocarboxamides. A parallel ab initio and polarizable molecular mechanics investigation. Assessment of the role of multipole transferability

α‐ and β‐mercaptocarboxamides constitute the Zn²⁺‐ligating entity of several highly potent metalloenzyme inhibitors. We have studied their interaction energies with Zn²⁺ using the polarizable molecular mechanics procedure SIBFA, and compared them to the corresponding ab initio supermolecule ones. Such validations are necessary to subsequently undertake simulations on complexes of Zn²⁺–metalloenzymes with inhibitors. If the distributed multipoles and polarizabilities are those derived for each ligand in its appropriate Zn²⁺‐binding conformation, a close reproduction of the ab initio binding energies is afforded. However, this representation is not tractable upon increasing the size of the ligands and/or to explore a continuum of binding conformations. This makes it necessary to construct the ligands by resorting to a library of constitutive fragments, namely in this case methanethiolate, formamide, and methane covalently connected together. A close reproduction of the ab initio interaction energies is enabled, but only if the ligand–ligand interactions are computed simultaneously with those occurring with Zn²⁺. This representation accounts for the nonadditivity occurring in the Zn²⁺–methanethiolate–formamide complex, and justifies the use of the distributed multipoles on the fragments for the construction of larger and flexible molecules. © 2001 John Wiley & Sons, Inc. J Comput Chem 22: 1038–1047, 2001     

Joint quantum chemical and polarizable molecular mechanics investigation of formate complexes with penta- and hexahydrated Zn2+: Comparison between energetics of model bidentate,monodentate, and through-water Zn2+ binding modes and evaluation of nonadditivity effects

In order to gain an understanding of the energetics of polycoordinated Zn²⁺ binding to the formate anion (the end side chain of the Asp and Glu residues of proteins), we compare three competing binding modes in the presence of five and six water molecules: a, bidentate binding of Zn²⁺ to both formate oxygens; b, monodentate binding of Zn²⁺ to one formate oxygen; and c, through-water binding of Zn²⁺ to formate, in which the cation remains bound to its first-hydration shell waters and interacts with both formate oxygens through three water molecules. We also investigate a complex d, which is similar to c, in which formate is protonated into formic acid and one water molecule is deprotonated. The computations are carried out using the ab initio self-consistent field/MP2 with three basis sets of increasing size density functional theory, semiempirical AM1 and PM3, and the sum of interactions between fragments ab initio computed (SIBFA) molecular mechanics procedures. The summed energies of the isolated molecules making up the complexes disfavor tautomer d compared to a–c. On the other hand, the ab initio computations give the ordering of intermolecular interaction energies as d formic acid tautomer >b monodentate >a bidentate >c through-water. Whereas the first-order energy E₁ favors both inner-shell Zn²⁺ complexes with formate over the outer-shell complex, the polarization and the charge-transfer components of the second-order energy E₂ both favor the outer-shell complex over the inner-shell one, despite the increased separation between the cation and the highly polarizable formate ion. Energy balances including continuum solvation enthalpies produce an equilibration of complexes a–d. The preference favoring the monodentate complex over the bidentate one is consistent with other ab initio results for formate binding by a fully coordinated Zn²⁺ cation and with structural results from X-ray crystallography. The SIBFA results are consistent with the ab initio results, and the computed interaction energy values match the ab initio ones to within 3%. The effects of nonadditivity are analyzed in the ab initio, SIBFA, and semiempirical computations. ©1999 John Wiley & Sons, Inc. J Comput Chem 20: 1379–1390, 1999     

Comparison of calculated DFT/B3LYP and experimental C and O NMR chemical shifts, ab initio HF/6-31G* optimised structures, and single crystal X-ray structures of some substituted methyl 5β-cholan-24-oates

Single crystal X-ray structures (monoclinic space group P2₁) for methyl 3-oxo-5β-cholan-24-oate and methyl 3,12-dioxo-5β-cholan-24-oate have been solved and compared with HF/6-31G* optimised structures. In the crystalline packings the side chains are connected with weak OC(sp³)HO-type of interactions between C25–H and C24–O–C25 and the keto ends with weak C(sp³)HO=C-type of interactions between C4–H and O=C3. The orientations of the side chains, which steric configurations are of great importance to the biological activity of the molecules, are compared with the experimental structure of methyl 3-hydroxy-5β-cholan-24-oate. Probable reasons for the observed differences are discussed. In addition, ¹³C and ¹⁷O NMR chemical shifts of methyl 3-oxo-5β-cholan-24-oate and methyl 3,12-dioxo-5β-cholan-24-oate as well as the epimeric methyl 3-hydroxy-5β-cholan-24-oate and methyl 3β-hydroxy-5β-cholan-24-oate have been calculated (DFT/B3LYP/6-311G*) and compared with the experimental values by linear regression analyses. In general, the correspondence between the theoretical and experimental parameters is good or excellent.