A check on rational drug design: crystal structure of a complex of human immunodeficiency virus type 1 protease with a novel gamma-turn mimetic inhibitor

An n→π* Interaction in the Bound Substrate of Aspartic Proteases Replicates the Oxyanion Hole

Aspartic proteases regulate many biological processes and are prominent targets for therapeutic intervention. Structural studies have captured intermediates along the reaction pathway, including the Michaelis complex and tetrahedral intermediate. Using a Ramachandran analysis of these structures, we discovered that residues occupying the P1 and P1' positions (which flank the scissile peptide bond) adopt the dihedral angle of an inverse γ-turn and polyproline type-II helix, respectively. Computational analyses reveal that the polyproline type-II helix engenders an n→π* interaction in which the oxygen of the scissile peptide bond is the donor. This interaction stabilizes the negative charge that develops in the tetrahedral intermediate, much like the oxyanion hole of serine proteases. The inverse γ-turn serves to twist the scissile peptide bond, vacating the carbonyl π* orbital and facilitating its hydration. These previously unappreciated interactions entail a form of substrate-assisted catalysis and offer opportunities for drug design.

Inclusion of multiple fragment types in the site identification by ligand competitive saturation (SILCS) approach

The site identification by ligand competitive saturation (SILCS) method identifies the location and approximate affinities of small molecular fragments on a target macromolecular surface by performing molecular dynamics (MD) simulations of the target in an aqueous solution of small molecules representative of different chemical functional groups. In this study, we introduce a set of small molecules to map potential interactions made by neutral hydrogen bond donors and acceptors and charged donor and acceptor fragments in addition to nonpolar fragments. The affinity pattern is obtained in the form of discretized probability or, equivalently, free energy maps, called FragMaps, which can be visualized with the target surface. We performed SILCS simulations for four proteins for which structural and thermodynamic data is available for multiple diverse ligands. Good overlap is shown between high affinity regions identified by the FragMaps and the crystallographic positions of ligand functional groups with similar chemical functionality, thus demonstrating the validity of the qualitative information obtained from the simulations. To test the ability of FragMaps in providing quantitative predictions, we calculate the previously introduced ligand grid free energy (LGFE) metric and observe its correspondence with experimentally measured binding affinity. LGFE is computed for different conformational ensembles and improvement in prediction is shown with increasing ligand conformational sampling. Ensemble generation includes a Monte Carlo sampling approach that uses the GFE FragMaps directly as the energy function. The results show that some but not all experimental trends are predicted and warrant improvements in the scoring methodology. In addition, the potential utility of atom-based free energy contributions to the LGFE scores and the use of multiple ligands in SILCS to identify displaceable water molecules during ligand design are discussed.

Secondary structure, a missing component of sequence-based minimotif definitions

Minimotifs are short contiguous segments of proteins that have a known biological function. The hundreds of thousands of minimotifs discovered thus far are an important part of the theoretical understanding of the specificity of protein-protein interactions, posttranslational modifications, and signal transduction that occur in cells. However, a longstanding problem is that the different abstractions of the sequence definitions do not accurately capture the specificity, despite decades of effort by many labs. We present evidence that structure is an essential component of minimotif specificity, yet is not used in minimotif definitions. Our analysis of several known minimotifs as case studies, analysis of occurrences of minimotifs in structured and disordered regions of proteins, and review of the literature support a new model for minimotif definitions that includes sequence, structure, and function.

Virtual drug screen schema based on multiview similarity integration and ranking aggregation

The current drug virtual screen (VS) methods mainly include two categories. i.e., ligand/target structure-based virtual screen and that, utilizing protein-ligand interaction fingerprint information based on the large number of complex structures. Since the former one focuses on the one-side information while the later one focuses on the whole complex structure, they are thus complementary and can be boosted by each other. However, a common problem faced here is how to present a comprehensive understanding and evaluation of the various virtual screen results derived from various VS methods. Furthermore, there is still an urgent need for developing an efficient approach to fully integrate various VS methods from a comprehensive multiview perspective. In this study, our virtual screen schema based on multiview similarity integration and ranking aggregation was tested comprehensively with statistical evaluations, providing several novel and useful clues on how to perform drug VS from multiple heterogeneous data sources. (1) 18 complex structures of HIV-1 protease with ligands from the PDB were curated as a test data set and the VS was performed with five different drug representations. Ritonavir ( 1HXW ) was selected as the query in VS and the weighted ranks of the query results were aggregated from multiple views through four similarity integration approaches. (2) Further, one of the ranking aggregation methods was used to integrate the similarity ranks calculated by gene ontology (GO) fingerprint and structural fingerprint on the data set from connectivity map, and two typical HDAC and HSP90 inhibitors were chosen as the queries. The results show that rank aggregation can enhance the result of similarity searching in VS when two or more descriptions are involved and provide a more reasonable similarity rank result. Our study shows that integrated VS based on multiple data fusion can achieve a remarkable better performance compared to that from individual ones and, thus, serves as a promising way for efficient drug screening, taking advantages of the rapidly accumulated molecule representations and heterogeneous data in the pharmacological area.

Is quantum mechanics necessary for predicting binding free energy?

To take into account polarization effects, the linear interaction energy model with continuum electrostatic solvation (LIECE) is supplemented by the linear-scaling semiempirical quantum mechanical calculation of the intermolecular electrostatic energy (QMLIECE). QMLIECE and LIECE are compared on three enzymes belonging to different classes: the West Nile virus NS3 serine protease (WNV PR), the aspartic protease of the human immunodeficiency virus (HIV-1 PR), and the human cyclin-dependent kinase 2 (CDK2). QMLIECE is superior for 44 peptidic inhibitors of WNV PR because of the different amount of polarization due to the broad range of formal charges of the inhibitors (from 0 to 3). On the other hand, QMLIECE and LIECE show similar accuracy for 24 peptidic inhibitors of HIV-1 PR (20 neutral and 4 with one formal charge) and for 73 CDK2 inhibitors (all neutral). These results indicate that quantum mechanics is essential when the inhibitor/protein complexes have highly variable charge-charge interactions.

A quantitative structure-activity relationship study for structurally diverse HIV-1 protease inhibitors: Contribution of conformational flexibility to inhibitory activity

In this study, we investigated by linear regression model the SAR data of the 15 HIV-1 protease inhibitors possessing structurally diverse scaffolds. First, a regression model was developed only using the enzyme-inhibitor interaction energy as a term of the model, but did not provide a good correlation with the inhibitory activity (R2 = 0.580 and Q2 = 0.500). Then, we focused on the conformational flexibility of the inhibitors which may represent the diversity of the inhibitors, and added two conformational parameters into the model, respectively: the number of rotatable bonds of ligands (deltaSrot) and the distortion energy of ligands (deltaElig). The regression model by adding deltaElig successfully improved the quality of the model (R2 = 0.771 and Q2 = 0.713) while the model with deltaSrot was unsuccessful. The prediction for a training inhibitor by the deltaElig model also showed good agreement with experimental activity. These results suggest that the conformational flexibility of HIV-1 protease inhibitors directly contributes to the enzyme inhibition.

How inaccuracies in protein structure models affect estimates of protein-ligand interactions: computational analysis of HIV-I protease inhibitor binding

The influence of possible inaccuracies that can arise during homology modeling of protein structures used for ligand binding studies were investigated with the molecular mechanics generalized Born surface area (MM-GBSA) method. For this, a family of well-characterized HIV-I protease-inhibitor complexes was used. Validation of MM-GBSA led to a correlation coefficient ranging from 0.72 to 0.93 between calculated and experimental binding free energies DeltaG. All calculated DeltaG values were based on molecular dynamics simulations with explicit solvent. Errors introduced into the protein structure through misplacement of side-chains during rotamer modeling led to a correlation coefficient between DeltaG(calc) and DeltaG(exp) of 0.75 compared with 0.90 for the correctly placed side chains. This is in contrast to homology models for members of the retroviral protease family with template structures ranging in sequence identity between 32% and 51%. For these protein models, the correlation coefficients vary between 0.84 and 0.87, which is considerably closer to the original protein (0.90). It is concluded that HIV-I low sequence identity with the template structure still allows creating sufficiently reliable homology models to be used for ligand-binding studies, although placement of the rotamers is a critical step during the modeling.

Synthesis of amino acid derived seven-membered lactams by RCM and their evaluation against HIV protease

A versatile synthesis of hydroxylated and epoxy 1-azepin 2-ones substituted at N1, C-3 and C-4 or C-7 has been developed. The sequence involves ring-closing metathesis of an amino acid derived diene amide, followed by either epoxidation or dihydroxylation, of the resulting alkene. Assay of the product epoxides (10, 18, 25) and diols (9a, 17, 24) against HIV protease reveals micromolar inhibition.

Modular synthesis of cyclic peptidomimetics inspired by gamma-turns

[reaction: see text] A series of peptidomimetics based on a gamma-turn motif were synthesized using a modular approach, in which N-protected piperidones were reacted with a selection of 2-hydroxyalkyl azides derived from common l-amino acids. Hydrolysis of the initially formed iminium ethers afforded the targeted series of substituted 1,4-diazepin-5-ones.

Protein-ligand binding free energy estimation using molecular mechanics and continuum electrostatics. Application to HIV-1 protease inhibitors

A method is proposed for the estimation of absolute binding free energy of interaction between proteins and ligands. Conformational sampling of the protein-ligand complex is performed by molecular dynamics (MD) in vacuo and the solvent effect is calculated a posteriori by solving the Poisson or the Poisson-Boltzmann equation for selected frames of the trajectory. The binding free energy is written as a linear combination of the buried surface upon complexation, SASbur, the electrostatic interaction energy between the ligand and the protein, Eelec, and the difference of the solvation free energies of the complex and the isolated ligand and protein, deltaGsolv. The method uses the buried surface upon complexation to account for the non-polar contribution to the binding free energy because it is less sensitive to the details of the structure than the van der Waals interaction energy. The parameters of the method are developed for a training set of 16 HIV-1 protease-inhibitor complexes of known 3D structure. A correlation coefficient of 0.91 was obtained with an unsigned mean error of 0.8 kcal/mol. When applied to a set of 25 HIV-1 protease-inhibitor complexes of unknown 3D structures, the method provides a satisfactory correlation between the calculated binding free energy and the experimental pIC5o without reparametrization.

Efficient Evaluation of Binding Free Energy Using Continuum Electrostatics Solvation

The linear interaction energy (LIE) method is combined with energy minimization and finite-difference Poisson calculation of electrostatic solvation for the estimation of the absolute free energy of binding. A predictive accuracy of about 1.0 kcal/mol is obtained for 13 and 29 inhibitors of beta-secretase (BACE) and HIV-1 protease (HIV-1 PR), respectively. The multiplicative coefficients for the van der Waals and electrostatic terms are not transferable between BACE and HIV-1 PR although they are both aspartic proteases. The present approach is about 2 orders of magnitude faster than previous LIE methods and can be used for ranking large libraries of structurally diverse compounds docked by automatic computational tools.

Improved prediction of HIV-1 protease-inhibitor binding energies by molecular dynamics simulations

BACKGROUND

The accurate prediction of enzyme-substrate interaction energies is one of the major challenges in computational biology. This study describes the improvement of protein-ligand binding energy prediction by incorporating protein flexibility through the use of molecular dynamics (MD) simulations.

RESULTS

Docking experiments were undertaken using the program AutoDock for twenty-five HIV-1 protease-inhibitor complexes determined by x-ray crystallography. Protein-rigid docking without any dynamics produced a low correlation of 0.38 between the experimental and calculated binding energies. Correlations improved significantly for all time scales of MD simulations of the receptor-ligand complex. The highest correlation coefficient of 0.87 between the experimental and calculated energies was obtained after 0.1 picoseconds of dynamics simulation.

CONCLUSION

Our results indicate that relaxation of protein complexes by MD simulation is useful and necessary to obtain binding energies that are representative of the experimentally determined values.

Automated docking of highly flexible ligands by genetic algorithms: A critical assessment

An improved version of the fragment-based flexible ligand docking approach SEED-FFLD is tested on inhibitors of human immunodeficiency virus type 1 protease, human alpha-thrombin and the estrogen receptor beta. The docking results indicate that it is possible to correctly reproduce the binding mode of inhibitors with more than ten rotatable bonds if the strain in their covalent geometry upon binding is not large. A high degree of convergence towards a unique binding mode in multiple runs of the genetic algorithm is proposed as a necessary condition for successful docking.

Relation between sequence and structure of HIV-1 protease inhibitor complexes: a model system for the analysis of protein flexibility

The flexibility of different regions of HIV-1 protease was examined by using a database consisting of 73 X-ray structures that differ in terms of sequence, ligands or both. The root-mean-square differences of the backbone for the set of structures were shown to have the same variation with residue number as those obtained from molecular dynamics simulations, normal mode analyses and X-ray B-factors. This supports the idea that observed structural changes provide a measure of the inherent flexibility of the protein, although specific interactions between the protease and the ligand play a secondary role. The results suggest that the potential energy surface of the HIV-1 protease is characterized by many local minima with small energetic differences, some of which are sampled by the different X-ray structures of the HIV-1 protease complexes. Interdomain correlated motions were calculated from the structural fluctuations and the results were also in agreement with molecular dynamics simulations and normal mode analyses. Implications of the results for the drug-resistance engendered by mutations are discussed briefly.

The active site of HIV-1 protease

The active site of the homodimeric HIV-1 protease includes six amino acids (triads AspThrGly found in each monomer) in amino acid positions 25 to 27 and 25' to 27'. Up to now, the role of Thr26 and Thr26', and Gly27 and Gly27', is unknown. It is hypothesized that strong hydrogen-bonding forces between the Thr26 and Thr26' residues stabilize the conformational state of the active site, and that the function of Gly27 and Gly27' is to accommodate and bind a substrate in a position in which the catalytic Asp25 and Asp25' carboxylate groups can attack the amide moiety of a substrate.

Conformational homogeneity in molecular recognition by proteolytic enzymes

Crystal structures for several hundred protease-inhibitor complexes have been analysed and their superimpositions have been used to demonstrate a universal relationship between inhibitor/substrate conformation and molecular recognition by all aspartic, serine, cysteine and metallo proteases. Proteases universally recognize an extended beta strand conformation in all their peptidic (and non-peptidic) inhibitors and substrate analogues without significant exceptions. This conformational homogeneity is illustrated here for a subset of 180 protease-inhibitor structures which are displayed as (a) structural overlays of multiple inhibitors for each of eight aspartic, eight serine, six metallo and five cysteine proteases; (b) single inhibitors each bound to different proteases; and (c) Ramachandran plots of peptide or pseudo-peptide dihedral angle pairs which demonstrate beta strands (Phi -54 degrees to -173 degrees, Psi 24 degrees to 174 degrees ) like those normally found paired in proteins as beta sheets. However, unlike beta sheets, alpha and 3(10) helices, beta and gamma turns, where the folded main chain amide components are intramolecularly hydrogen bonded and thus unavailable for interaction with proteins, an inhibitor/substrate in an isolated beta strand conformation provides maximum exposure of its hydrogen bonding donors/acceptors and side chain components to a putative protease receptor. This analysis highlights the advantages of a strand conformation over other elements of secondary structure for protease recognition and may lead to generic strategies for inhibitor design.

Predicting structural effects in HIV-1 protease mutant complexes with flexible ligand docking and protein side-chain optimization

We present a computational approach for predicting structures of ligand-protein complexes and analyzing binding energy landscapes that combines Monte Carlo simulated annealing technique to determine the ligand bound conformation with the dead-end elimination algorithm for side-chain optimization of the protein active site residues. Flexible ligand docking and optimization of mobile protein side-chains have been performed to predict structural effects in the V32I/I47V/V82I HIV-1 protease mutant bound with the SB203386 ligand and in the V82A HIV-1 protease mutant bound with the A77003 ligand. The computational structure predictions are consistent with the crystal structures of these ligand-protein complexes. The emerging relationships between ligand docking and side-chain optimization of the active site residues are rationalized based on the analysis of the ligand-protein binding energy landscape.

Solvation studies of DMP323 and A76928 bound to HIV protease: analysis of water sites using grand canonical Monte Carlo simulations

We examine the water solvation of the complex of the inhibitors DMP323 and A76928 bound to HIV-1 protease through grand canonical Monte Carlo simulations, and demonstrate the ability of this method to reproduce crystal waters and effectively predict water positions not seen in the DMP323 or A76928 structures. The simulation method is useful for identifying structurally important waters that may not be resolved in the crystal structures. It can also be used to identify water positions around a putative drug candidate docked into a binding pocket. Knowledge of these water positions may be useful in designing drugs to utilize them as bridging groups or displace them in the binding pocket. In addition, the method should be useful in finding water sites in homology models of enzymes for which crystal structures are unavailable.

Empirical scoring functions: I. The development of a fast empirical scoring function to estimate the binding affinity of ligands in receptor complexes

This paper describes the development of a simple empirical scoring function designed to estimate the free energy of binding for a protein-ligand complex when the 3D structure of the complex is known or can be approximated. The function uses simple contact terms to estimate lipophilic and metal-ligand binding contributions, a simple explicit form for hydrogen bonds and a term which penalises flexibility. The coefficients of each term are obtained using a regression based on 82 ligand-receptor complexes for which the binding affinity is known. The function reproduces the binding affinity of the complexes with a cross-validated error of 8.68 kJ/mol. Tests on internal consistency indicate that the coefficients obtained are stable to changes in the composition of the training set. The function is also tested on two test sets containing a further 20 and 10 complexes, respectively. The deficiencies of this type of function are discussed and it is compared to approaches by other workers.