A knowledge-based approach in designing combinatorial or medicinal chemistry libraries for drug discovery. 1. A qualitative and quantitative characterization of known drug databases.

The discovery of various protein/receptor targets from genomic research is expanding rapidly. Along with the automation of organic synthesis and biochemical screening, this is bringing a major change in the whole field of drug discovery research. In the traditional drug discovery process, the industry tests compounds in the thousands. With automated synthesis, the number of compounds to be tested could be in the millions. This two-dimensional expansion will lead to a major demand for resources, unless the chemical libraries are made wisely. The objective of this work is to provide both quantitative and qualitative characterization of known drugs which will help to generate "drug-like" libraries. In this work we analyzed the Comprehensive Medicinal Chemistry (CMC) database and seven different subsets belonging to different classes of drug molecules. These include some central nervous system active drugs and cardiovascular, cancer, inflammation, and infection disease states. A quantitative characterization based on computed physicochemical property profiles such as log P, molar refractivity, molecular weight, and number of atoms as well as a qualitative characterization based on the occurrence of functional groups and important substructures are developed here. For the CMC database, the qualifying range (covering more than 80% of the compounds) of the calculated log P is between -0.4 and 5.6, with an average value of 2.52. For molecular weight, the qualifying range is between 160 and 480, with an average value of 357. For molar refractivity, the qualifying range is between 40 and 130, with an average value of 97. For the total number of atoms, the qualifying range is between 20 and 70, with an average value of 48. Benzene is by far the most abundant substructure in this drug database, slightly more abundant than all the heterocyclic rings combined. Nonaromatic heterocyclic rings are twice as abundant as the aromatic heterocycles. Tertiary aliphatic amines, alcoholic OH and carboxamides are the most abundant functional groups in the drug database. The effective range of physicochemical properties presented here can be used in the design of drug-like combinatorial libraries as well as in developing a more efficient corporate medicinal chemistry library.

Efficiency of signalling through cytokine receptors depends critically on receptor orientation.

Human erythropoietin is a haematopoietic cytokine required for the differentiation and proliferation of precursor cells into red blood cells. It activates cells by binding and orientating two cell-surface erythropoietin receptors (EPORs) which trigger an intracellular phosphorylation cascade. The half-maximal response in a cellular proliferation assay is evoked at an erythropoietin concentration of 10 pM, 10(-2) of its Kd value for erythropoietin-EPOR binding site 1 (Kd approximately equal to nM), and 10(-5) of the Kd for erythropoietin-EPOR binding site 2 (Kd approximately equal to 1 microM). Overall half-maximal binding (IC50) of cell-surface receptors is produced with approximately 0.18 nM erythropoietin, indicating that only approximately 6% of the receptors would be bound in the presence of 10 pM erythropoietin. Other effective erythropoietin-mimetic ligands that dimerize receptors can evoke the same cellular responses but much less efficiently, requiring concentrations close to their Kd values (approximately 0.1 microM). The crystal structure of erythropoietin complexed to the extracellular ligand-binding domains of the erythropoietin receptor, determined at 1.9 A from two crystal forms, shows that erythropoietin imposes a unique 120 degrees angular relationship and orientation that is responsible for optimal signalling through intracellular kinase pathways.

Biophysical tools for structure-based drug design.

The emerging understanding of the structural basis of biological function has made possible a new technology of rational drug design. This technology uses a combination of modeling and intensive computational tools, but also relies critically on direct structural methods for verification and to provide direction for new leads. The present work illustrates applications of this technology to ligand design for several model systems, including iterative modeling and crystallographic design of streptavidin ligands, ranking of streptavidin ligands using comparative molecular field analysis and ranking of thermolysin inhibitors using Poisson-Boltzman methods.

Structural studies of the retroviral proteinase from avian myeloblastosis associated virus.

The structure of the retroviral proteinase from avian myeloblastosis associated virus (MAV) has been determined and refined at 2.2 A resolution. This structure is compared with those of homologous proteinases from Rous sarcoma virus (RSV) and human immunodeficiency type 1 virus (HIV). Through comparison with the structure of a proteinase-inhibitor complex from HIV, a model of a complex between MAV proteinase and a peptide substrate has been generated. Examination of this model suggests structural basis for the diverse specifications of viral proteinases.

PROBIT: a statistical approach to modeling proteins from partial coordinate data using substructure libraries.

A program (PROBIT) has been developed that allows the reconstruction of a complete set of three-dimensional protein coordinates from alpha-carbon coordinates. The program generates a statistical measure of polypeptide conformational behavior for substructures in a defined structural context from a library of highly refined protein structures. These statistics provide a prescription for substructure substitution from the database to allow regeneration of the complete protein structure.

Cooperative ligand reorientations in cytochrome c3: a molecular dynamics simulation.

Molecular dynamics simulations of a tetraheme cytochrome c3 were performed to investigate dynamic aspects of the motion of the axial heme iron ligands. It was found that persistent transitions between alternate axial imidazole orientations of the histidine incorporated in the CXXCH heme binding sequence occurred via correlated motions. The correlated motions involved virtually all of the atoms comprising the polypeptide backbone of the heme binding sequence as well as the histidine imidazole side-chain.

Electrostatic field around cytochrome c: theory and energy transfer experiment.

Energy transfer in the "rapid diffusion" limit from electronically excited terbium(III) chelates in three different charge states to horse heart ferricytochrome c was measured as a function of ionic strength. Theoretical rate constants calculated by numerical integration of the Forster integral (containing the Poisson-Boltzmann-generated protein electrostatic potential) were compared with the experimental data to evaluate the accuracy of protein electrostatic field calculations at the protein/solvent interface. Two dielectric formalisms were used: a simple coulombic/Debye-Hückel procedure and a finite difference method [Warwicker, J. & Watson, H. C. (1982) J. Mol. Biol. 157, 671-679] that accounts for the low-dielectric protein interior and the irregular protein/solvent boundary. Good agreement with experiment was obtained and the ionic-strength dependence of the reaction was successfully reproduced. The sensitivity of theoretical rate constants to the choices of effective donor sphere size and the energy transfer distance criterion was analyzed. Electrostatic potential and rate-constant calculations were carried out on sets of structures collected along two molecular dynamics trajectories of cytochrome c. Protein conformational fluctuations were shown to produce large variations in the calculated energy transfer rate constant. We conclude that protein fluctuations and the resulting transient structures can play significant roles in biological or catalytic activities that are not apparent from examination of a static structure. For calculating protein electrostatics, large-scale low-frequency conformational fluctuations, such as charged side-chain reorientation, are established to be as important as the computational method for incorporating dielectric boundary effects.

Molecular dynamics in ordered structures: computer simulation and experimental results for nylon 66 crystals.

A detailed comparison between molecular dynamics computer simulations and the experimental characterization of molecular motion through deuterium nuclear magnetic resonance (NMR) spectroscopic methods has been carried out for the crystalline phase of nylon 66 (polyhexamethyleneadipamide) at room temperature and just below the melting point. The computer simulations agree quantitatively with the experimental results at room temperature and qualitatively near the crystalline melting point. Both methods demonstrate that individual methylene groups within the crystals exhibit librational motion, which becomes very large in amplitude near the melting point, rather than undergoing discrete conformational jumps; furthermore, the hydrogen-bonded amides are relatively immobile at all temperatures below 230 degrees Celsius. The simulations are shown to be particularly useful for exaning the cooperativity of motion and for providing insight into structural-dynamical correlations. These aspects of the simulations are exemplified by the observation of concerted counterrotation of odd-numbered bonds within the methylene segments and the entropic stabilization of the crystal structure.

X-ray structural studies of the cytokine interleukin 1-beta.

The structure of the human cytokine, interleukin 1-beta (IL1-beta) is composed almost wholly of anti-parallel beta-sheet, organized in a three-fold repeating motif. The beta strands comprising the protein core are interconnected by 11 surface loops that form prominent features on the surface of the molecule. Comparisons of the amino acid sequences of different species of IL1-beta and the related cytokine IL1-alpha show that the majority of conserved residues form the structural core of the molecule, and that most variability occurs in the termini and surface loops that presumably bind the IL1 receptor. These results suggest that there may be some degree of flexibility in interactions made between IL1 and its cell surface receptor.

Molecular dynamics simulation of a phospholipid micelle.

The dynamic character of phospholipid aggregates limits conventional structural studies to the determination of average molecular features. In order to develop more detailed descriptions of phospholipid structure for comparison with experiment, the molecular dynamics of a hydrated lysophosphatidylethanolamine (LPE) micelle, incorporating 85 LPE and 1591 water molecules, have been simulated. Comparison of the initial and equilibrated micelles shows substantial differences both in LPE hydrocarbon chain conformation and polar head-group-solvent interactions. Although these changes produce only subtle effects on the averaged structural properties of the system, the alterations in hydrocarbon chain packing and head-group solvation appear to mimic a polymorphic pretransition from a spherical toward a cylindrical micelle structure.

Molecular dynamics effects on protein electrostatics.

Electrostatic calculations have been carried out on a number of structural conformers of tuna cytochrome c. Conformers were generated using molecular dynamics simulations with a range of solvent simulating, macroscopic dielectric formalisms, and one solvent model that explicitly included solvent water molecules. Structures generated using the lowest dielectric models were relatively tight, with side chains collapsed on the surface, while those from the higher dielectric models had more internal and external fluidity, with surface side chains exploring a fuller range of conformational space. The average structure generated with the explicitly solvated model corresponded most closely with the crystal structure. Individual pK values, overall titration curves, and electrostatic potential surfaces were calculated for average structures and structures along each simulation. Differences between structural conformers within each simulation give rise to substantial changes in calculated local electrostatic interactions, resulting in pK value fluctuations for individual sites in the protein that vary by 0.3-2.0 pK units from the calculated time average. These variations are due to the thermal side chain reorientations that produce fluctuations in charge site separations. Properties like overall titration curves and pH dependent stability are not as sensitive to side chain fluctuations within a simulation, but there are substantial effects between simulations due to marked differences in average side chain behavior. These findings underscore the importance of proper dielectric formalism in molecular dynamics simulations when used to generate alternate solution structures from a crystal structure, and suggest that conformers significantly removed from the average structure have altered electrostatic properties that may prove important in episodic protein properties such as catalysis.

Structural origins of high-affinity biotin binding to streptavidin.

The high affinity of the noncovalent interaction between biotin and streptavidin forms the basis for many diagnostic assays that require the formation of an irreversible and specific linkage between biological macromolecules. Comparison of the refined crystal structures of apo and a streptavidin:biotin complex shows that the high affinity results from several factors. These factors include the formation of multiple hydrogen bonds and van der Waals interactions between biotin and the protein, together with the ordering of surface polypeptide loops that bury the biotin in the protein interior. Structural alterations at the biotin binding site produce quaternary changes in the streptavidin tetramer. These changes apparently propagate through cooperative deformations in the twisted beta sheets that link tetramer subunits.

Computer simulation of biological interactions and reactivity.

Computer simulations of molecular motion provide a useful tool for analyzing dynamic aspects of macromolecular structure and function. In many cases, simulations can be compared to experimental results that provide an average estimate of molecular flexibility. For example, variations in computed molecular motions in different regions of a protein structure can be compared to refined B-values obtained from X-ray crystallographic refinement. Such comparisons both provide a detailed view of the motions responsible for crystalline disorder, and allow an evaluation of how crystal packing affects mobility of groups on the protein surface. In these applications, dynamics simulations provide a means of regenerating the temporal dimension of a structure whose average behavior is experimentally well defined in the crystal lattice. An additional benefit of the detailed and instantaneous view of molecular flexibility offered by simulation methods lies in its potential for exploring infrequent structural fluctuations or dynamic states of molecular association that cannot be examined in detail by X-ray methods, but are suggested on the basis of alternative structural information. For example, studies of the effects of surface chemical modification on interacting proteins can produce information concerning the sites, if not the exact details, of the intermolecular interactions. The present work describes some applications of molecular dynamics methods to the study of large molecular aggregates whose dynamic properties thus far have precluded detailed structural descriptions. These include simulations of an electrostatically associated electron transfer complex between cytochromes c and b5, some model systems for trans-membrane ion channels, and a phospholipid micelle.

Molecular dynamics of a cytochrome c-cytochrome b5 electron transfer complex.

Cytochrome c and cytochrome b5 form an electrostatically associated electron transfer complex. Computer models of this and related complexes that were generated by docking the x-ray structures of the individual proteins have provided insight into the specificity and mechanism of electron transfer reactions. Previous static modeling studies were extended by molecular dynamics simulations of a cytochrome c-cytochrome b5 intermolecular complex. The simulations indicate that electrostatic interactions at the molecular interface results in a flexible association complex that samples alternative interheme geometries and molecular conformations. Many of these transient geometries appear to be more favorable for electron transfer than those formed in the initial model complex. Of particular interest is a conformational change that occurred in phenylalanine 82 of cytochrome c that allowed the phenyl side chain to bridge the two cytochrome heme groups.

Effect of basis set on electron populations calculated by using Bader's criterion for partitioning electron density between atoms.

The effect of basis set on Bader's criterion for partitioning electron density between atoms is examined. The major effect is on the position of minimum electron density along the bond of interest. The 6-31G(**) basis leads to the more satisfactory results. The effect of including electron correlation was examined via the use of generalized valence bond wavefunctions. The change in the electron populations was small. The partitioning of electron density is useful in examining the way in which substituents interact with hydrocarbon groups.