A new addition to the structural bioinformatics toolbox: 3D models of short bioactive peptides from multiple sequences using feedback restrained molecular dynamics (FRMD).

Modem molecular biology techniques allow the use of new approaches for the 3D mapping of 1D information. Molecular biology techniques are capable of producing large amounts of 1D information (sequences) from a number of different sources (phage displays, ESTs, etc.). In this work, we present a technique that takes advantage of large sets of 1D information and increasingly available computer power to create 3D models. For the purpose of validation the technique is first applied to the modeling of an erythropoietin analog of known 3D structure from 1D information only. The technique is then used to model the immunoreactive region of echinococcus granulosus AgB based on phage raised mimotopes for which there is no previous structural information. The technique here presented is of general application to similar problems where 1D information is available and structure activity relationships (SAR) is needed.

Flexible effective fragment QM/MM method: validation through the challenging tests.

A new version of the QM/MM method, which is based on the effective fragment potential (EFP) methodology [Gordon, M. et al., J Phys Chem A 2001, 105, 293] but allows flexible fragments, is verified through calculations of model molecular systems suggested by different authors as challenging tests for QM/MM approaches. For each example, the results of QM/MM calculations for a partitioned system are compared to the results of an all-electron ab initio quantum chemical study of the entire system. In each case we were able to achieve approximately similar or better accuracy of the QM/MM results compared to those described in original publications. In all calculations we kept the same set of parameters of our QM/MM scheme. A new test example is considered when calculating the potential of internal rotation in the histidine dipeptide around the C(alpha)bond;C(beta) side chain bond.

(Salen)Mn(III)-catalyzed epoxidation reaction as a multichannel process with different spin states. Electronic tuning of asymmetric catalysis: a theoretical study.

The (salen)Mn(III)-catalyzed epoxidation reaction mechanism has been investigated using density functional theory (DFT). There is considerable interest in and controversy over the mechanism of this reaction. The results of experimental studies have offered some support for three different reaction mechanisms: concerted, stepwise radical, and metallooxetane mediated. In this paper, a theoretical examination of the reaction suggests a novel mechanism that describes the reaction as a multichannel process combining both concerted and stepwise radical pathways. The competing channels have different spin states: the singlet, the triplet, and the quintet. The singlet reaction pathway corresponds to a concerted mechanism and leads exclusively to a cis epoxide product. In contrast, the triplet and quintet reactions follow a stepwise mechanism and lead to a product mixture of cis and trans epoxides. We show that the experimentally observed dependence of isomer product ratios on electronic effects connected with the substitution of the catalyst ligands is due to changing the relative position and, hence, the relative activities of the channels with different cis-trans yields. Because the results and conclusions of the present work dramatically differ from the results and conclusion of the recent DFT theoretical investigation (Linde, C.; Akermark, B; Norrby, P.-O.; Svensson, M. J. Am. Chem. Soc. 1999, 121, 5083.), we studied possible sources for the deep contradictions between the two works. The choice of the DFT functional and a model has been shown to be crucial for accurate results. Using high level ab initio calculations (coupled cluster-CCSD(T)), we show that the computational procedure employed in this study generates significantly more reliable numerical results. It is also shown that the smaller cationic model without a chlorine ligand that was used by Linde et al. is too oversimplified with respect to our larger neutral model. For this reason, using the cationic model led to a qualitatively wrong quintet reaction profile that played a key role in theoretical postulates in the earlier work.

alpha- and 3(10)-helix interconversion: a quantum-chemical study on polyalanine systems in the gas phase and in aqueous solvent.

Helices are among the predominant secondary structures in globular proteins. About 90% of the residues in them are found to be in the alpha-helical conformation, and another 10% in the 3(10) conformation. There is a standing controversy between experimental and some theoretical results, and controversy among theoretical results concerning the predominance of each conformation, in particular, helices. We address this controversy by ab initio Hartree-Fock and density functional theory studies of helices with different lengths in a vacuum and in the aqueous phase. Our results show that (1) in a vacuum, all oligo(Ala) helices of 4-10 residues adopt the 3(10) - conformation; (2) in aqueous solution, the 6-10 residue peptides adopt the alpha-helical conformation; (3) there might be two intermediates between these helical conformers allowing for their interconversion. The relevance of these results to the structure and folding of proteins is discussed.

Experimental determination and calculations of redox potential descriptors of compounds directed against retroviral zinc fingers: Implications for rational drug design.

A diverse set of electrophilic compounds that react with cysteine thiolates in retroviral nucleocapsid (NC) proteins and abolish virus infectivity has been identified. Although different in chemical composition, these compounds are all oxidizing agents that lead to the ejection of Zn(II) ions bound to conserved structural motifs (zinc fingers) present in retroviral NC proteins. The reactivity of a congeneric series of aromatic disulfides toward the NC protein of the human immunodeficiency virus type 1 (HIV-1), NCp7, has been characterized by HPLC separation of starting reagents from reaction products. We calculated the absolute redox potentials of these compounds in the gas phase and in aqueous solvent, using a density functional theory method and a continuum solvation model. Pulsed polarography experiments were performed and showed a direct correlation between calculated and experimentally determined redox propensities. A dependence between protein reactivity and redox potential for a specific compound was shown: Reaction with NCp7 did not take place below a threshold value of redox potential. This relationship permits the distinction between active and nonactive compounds targeted against NCp7, and provides a theoretical basis for a scale of reactivity with retroviral zinc fingers. Our results indicate that electrophilic agents with adequate thiophilicity to react with retroviral NC fingers can now be designed using known or calculated electrochemical properties. This may assist in the design of antiretroviral compounds with greater specificity for NC protein. Such electrophilic agents can be used in retrovirus inactivation with the intent of preparing a whole-killed virus vaccine formulation that exhibits unaffected surface antigenic properties.

Computational studies of the domain movement and the catalytic mechanism of thymidine phosphorylase.

Thymidine phosphorylase (TP) is a dual substrate enzyme with two domains. Each domain binds a substrate. In the crystal structure of Escherichia coli TP, the two domains are arranged so that the two substrate binding sites are too far away for the two substrates to directly react. Molecular dynamics simulations reveal a different structure of the enzyme in which the two domains have moved to place the two substrates in close contact. This structure has a root-mean-square deviation from the crystal structure of 4.1 A. Quantum mechanical calculations using this structure find that the reaction can proceed by a direct nucleophilic attack with a low barrier. This mechanism is not feasible in the crystal structure environment and is consistent with the mechanism observed for other N-glycosidic enzymes. Important catalytic roles are found for the three highly conserved residues His 85, Arg 171, and Lys 190.

Theoretical calculations of glycine and alanine gas-phase acidities.

The gas-phase acidities of glycine and alanine were determined by using a variety of high level theoretical methods to establish which of these would give the best results with accessible computational efforts. MP2, MP4, QCISD, G2 ab initio procedures, hybrid Becke3-LYP (B3LYP) and gradient corrected Becke-Perdew (BP) and Perdew-Wang and Perdew (PWP) nonlocal density functionals were used for the calculations. A maximum deviation of approximately 13 and 18 kJ/mol from experimental data was observed for the computed delta Hacid and delta Gacid values, respectively. The best result was obtained at G2 level, but comparable reliability was reached when the considerably less time consuming B3LYP, BP, and PWP density functional approaches were employed.

Reaction path and free energy calculations of the transition between alternate conformations of HIV-1 protease.

Two different structures of ligand-free HIV protease have been determined by X-ray crystallography. These structures differ in the position of two 12 residue, beta-hairpin regions (or "flaps") which cap the active site. The movements of the flaps must be involved in the binding of substrates since, in either conformation, the flaps block the binding site. One of these structures is similar to structures of the ligand-bound enzyme; however, the importance of both structures to enzyme function is unclear. This transformation takes place on a time scale too long for conventional molecular dynamics simulations, so the process was studied by first identifying a reaction path between the two structures and then calculating the free energy along this path using umbrella sampling. For the ligand-free enzyme, it is found that the two structures are nearly equally stable, with the ligand-bound-type structure being less stable, consistent with X-ray crystallography data. The more stable open structure does not have a lower potential energy, but is stabilized by entropy. The transition occurs through a collapse and reformation of the beta-sheet structure of the conformationally flexible, glycine-rich flap ends. Additionally, some problems in studying conformational changes in proteins through the use of a single reaction path are addressed.

Molecular mechanisms of resistance: free energy calculations of mutation effects on inhibitor binding to HIV-1 protease.

The changes in the inhibitor binding constants due to the mutation of isoleucine to valine at position 84 of HIV-1 protease are calculated using molecular dynamics simulations. The calculations are done for three potent inhibitors--KNI-272, L-735,524 (indinavir or MK-639), and Ro 31-8959 (saquinavir). The calculations agree with the experimental data both in terms of an overall trend and in the magnitude of the resulting free energy change. HIV-1 protease is a homodimer, so each mutation causes two changes in the enzyme. The decrease in the binding free energy from each mutated side chain differs among the three inhibitors and correlates well with the size of the cavities induced in the protein interior near the mutated residue. The cavities are created as a result of a mutation to a smaller side chain, but the cavities are less than would be predicted from the wild-type structures, indicating that there is significant relaxation to partially fill the cavities.

All-atom models for the non-nucleoside binding site of HIV-1 reverse transcriptase complexed with inhibitors: a 3D QSAR approach.

Several molecular modeling techniques were used to generate an all-atom molecular model of a receptor binding site starting only from Ca atom coordinates. The model consists of 48 noncontiguous residues of the non-nucleoside binding site of HIV-1 reverse transcriptase and was generated using a congeneric series of nevirapine analogs as structural probes. On the basis of the receptor-ligand atom contacts, the program HINT was used to develop a 3D quantitative structure activity relationship that predicted the rank order of binding affinities for the series of inhibitors. Electronic profiles of the ligands in their docked conformations were characterized using electrostatic potential maps and frontier orbital calculations. These results led to the development of a 3D stereoelectronic pharmacophore which was used to construct 3D queries for database searches. A search of the National Cancer Institute's open database identified a lead compound that exhibited moderate antiviral activity.

Design, synthesis, and resistance patterns of MP-134 and MP-167, two novel inhibitors of HIV type 1 protease.

Inhibitors of HIV-1 protease represent a new class of antiretroviral compounds. Here, we report the design and synthesis of two novel C2 symmetry-based inhibitors, MP-134 and MP-167, specifically targeted against HIV-1 variants with reduced sensitivity to another related protease inhibitor, A-77003. In addition, we describe the in vitro selection of viral variants with reduced sensitivity of these two protease inhibitors. An isoleucine-to-valine substitution at residue 84 (I84V) of the HIV-1 protease confers resistance to MP-134, whereas a glycine-to-valine substitution at residue 48 (G48V) confers resistance to MP-167. Testing other protease inhibitors against these variants has revealed specific overlapping patterns of resistance among these agents. These findings have important implications in the design of combination regimens using multiple protease inhibitors and underscore the need to develop non-cross-resistant compounds to be used toward this goal.

Structural mechanisms of HIV drug resistance.

Antiviral therapy for AIDS has focused on the discovery and design of inhibitors for two main enzyme targets of the human immunodeficiency virus type 1 (HIV)--reverse transcriptase (RT) and protease (PR). Despite several classes of promising new anti-HIV agents, the clinical emergence of drug-resistant variants of HIV has severely limited the long-term effectiveness of these drugs. Genetic analysis of resistant virus has identified a number of critical mutations in the RT and PR genes. Structural analysis of inhibitor-enzyme complexes and mutational modeling studies are leading to a better understanding of how these drug-resistance mutations exert their effects at a structural level. These insights have implications of the design of new drugs and therapeutic strategies to combat drug resistance to AIDS.

Flap opening in HIV-1 protease simulated by 'activated' molecular dynamics.

We have used an 'activated' molecular dynamics approach to simulate flap opening in HIV-1 protease. An initial impulse for flap opening was provided by applying harmonic constraints to non-flap residues. After an initial 'melting' phase, the two beta-hairpin structures that constitute the flaps opened to a 25 A gap within 200 ps of simulation. Analysis of backbone torsion angles suggests that flap opening is related to conformational changes at Lys 45, Met 46, Gly 52 and Phe 53. In contrast, similar molecular dynamics simulations on the M46I mutant, which is associated with drug resistance, indicates that this mutation stabilizes the flaps in a closed conformation.

Characterization of human immunodeficiency virus type 1 variants with increased resistance to a C2-symmetric protease inhibitor.

Inhibitors of the human immunodeficiency virus type 1 protease represent a promising class of antiviral drugs for the treatment of AIDS, and several are now in clinical trials. Here, we report the in vitro selection of viral variants with decreased sensitivity to a C2-symmetric protease inhibitor (A-77003). We show that a single amino acid substitution (Arg to Gln or Lys) at position 8 of the protease results in a substantial decrease in the inhibitory activity of the drug on the enzyme and a comparable increase in viral resistance. These findings, when analyzed by using the three-dimensional structure of the protease-drug complex, provide a strategic guide for the future development of inhibitors of the human immunodeficiency virus type 1 protease.

Solution influence on biomolecular equilibria: nucleic acid base associations.

This paper consists of two parts. In the first part, the general problem of biomolecular equilibria in solution is considered, stressing that molecular interactions ultimately determine the answer to this problem. It is discussed how computer simulation techniques can reliably treat the problem and several pitfalls of computer simulation to be avoided are pointed out. Other approaches based on modeling and conceptual simplifications such as perturbative methods, long-range interaction approximations, surface thermodynamic approaches, and hydration shell models are discussed. In the second part, the results of Monte Carlo calculations on the associations of nucleic acid bases in water and carbon tetrachloride are presented. Stacked self-associations are found to be preferred in water and hydrogen-bonded complexes are favored in nonpolar solutions, in agreement with experimental data. The influence of the solvent on base associations is explained in terms of solute-solvent and solvent-solvent contributions to the total energy. No enthalpic stabilization of the complexes by the solvent was found. The results are used to examine the validity of various approximations discussed in the first part of the paper.

Monte Carlo simulation of the influence of solvent on nucleic acid base associations.

The results of Monte Carlo calculations of the association between nucleic acid bases in a nonpolar solvent (CCl4) are described. The influence of the solvent on planar and stacked associations of bases was examined by analyzing the total energy of the system, including solute-solute, solute-solvent, and solvent-solvent contributions. Good quantitative agreement with the available experimental data was obtained. Solute-solvent interactions are primarily determined by dispersion forces; consequently, solute-solvent interactions vertical to the solute plane that maximize dispersion interactions are most favored, and a rough proportionally between solute-solvent energy and the surface of the solute was observed. Analysis of solvent-solvent energy is not necessarily reduced when surface area decreases, contrary to the simple cavity concept. "Single molecule probe" calculations were performed to explain the differences in base associations in H2O and CCl4. In CCl4 dispersion forces dominate and planar complexes are stabilized by maximum exposure of molecular planes to the solvent. In H2O electrostatic forces dominate so that the most stable structures are stacked association that allow the maximum number of hydrophilic centers to be exposed to the solvent.

Quantum chemical studies of a model for peptide bond formation. 3. Role of magnesium cation in formation of amide and water from ammonia and glycine.

The SN2 reaction between glycine and ammonia molecules with magnesium cation Mg2+ as a catalyst has been studied as a model reaction for Mg(2+)-catalyzed peptide bond formation using the ab initio Hartree-Fock molecular orbital method. As in previous studies of the uncatalyzed and amine-catalyzed reactions between glycine and ammonia, two reaction mechanisms have been examined, i.e., a two-step and a concerted reaction. The stationary points of each reaction including intermediate and transition states have been identified and free energies calculated for all geometry-optimized reaction species to determine the thermodynamics and kinetics of each reaction. Substantial decreases in free energies of activation were found for both reaction mechanisms in the Mg(2+)-catalyzed amide bond formation compared with those in the uncatalyzed and amine-catalyzed amide bond formation. The catalytic effect of the Mg2+ cation is to stabilize both the transition states and intermediate, and it is attributed to the neutralization of the developing negative charge on the electrophile and formation of a conformationally flexible nonplanar five-membered chelate ring structure.