関連イベント情報

Molecular Modeling and Desktop Drug Discovry

Niegel G J chards

BioWeb Coorporation, 12085 Research Drive, Alachua, FL 32615

Background

�@�@The advent of high-throughput screening and combinatorial methods has resulted in a significant paradigm shift in the process of drug discovery during the five past five years. Generation and assay of large libraries of compounds offers the potential to reduce the time for (I) obtaining medicinally important leads, or (ii) optimizing structures with respect to properties such as potency, selectivity and toxicity. Core groups within pharmaceutical companies have also been specifically changed with employing modern techniques in structural biology so as to provide detailed structural information on many target macromolecules. Computational approaches to " rational" drug discovery that were developed from 1982-1990 are being re-evaluated in the light of these scientific developments, especially given the level of investment in hardware and software that was made prior to 1990. In addition, although the cost of computer hardware has been significantly reduced, the licensing and maintenance costs of most software packages remain sufficiently large to raise recurring questions concerning return on investment by managers responsible for groups specializing in computational chemistry. One strategy, "desktop drug discovery", that has proved effective in a number of existing pharmaceutical companies given the low capital costs of providing personal computers, has been the decentralization of many basic modeling studies to medicinal and development chemists throughout the organization. On the other hand, for many of these chemists, including those who have employed modeling methods during their graduate and post-doctoral work, there are usually many barriers to the routine use of computational techniques as part of their research. This workshop seeks to address the educational needs of these individuals at Japanese Pharmaceutical Companies by discussing strategies for integrating basic computational procedures with experimental constrains. In this way, software/hardware investments may be leveraged due to an increase in the number of individuals making use of the computational facilities, with a concomitant tightening of the coupling between theoretical and experimental science.

Goals and Objectives

�@�@When used in combination with experimental methods, the molecular modeling techniques in software packages available from Oxford Molecular and Sony -Tek ( CAChe worksystem, Asp, Tsar, AMBER ) are powerful tools in (I) the generation of structure function hypotheses, (ii) the visualization of small molecule ligands, macromolecules and their complexes, (iii) the a priori evaluation of ligands with potential biological activity, and (iv) understanding chemical reactivity and transition state structure, especially in the design of highly diastereoselective and/or enantioselective reactions. While most modeling approaches can not be used productively in the absence of a large amount of pre-existing experimental information, computational methods and graphics visualization can often be used to prioritize a set of synthetic targets and to develop novel ideas about the relationship between molecular structure and biological function. This introductory workshop seeks to provide a concise overview of current techniques in molecular modeling that are relevant to drug discovery and synthesis. The presentations will be designed for medicinal chemists with a working knowledge of physical organic chemistry, and who have limited experience in using computational methods. the pedagogical approach will be to use " case studies " that can provide a basic recipe for bench chemists to begin to solve modeling problems. At the end of the course, the participants should have an understanding of the applications and limitations of modeling methods, and be able to apply some basic computational methods to their research and synthesis problems. Hence, given a specific drug discovery problem they should be able to decide what methods are applicable ( if any ) and how to integrate them into the usual sequence of experiments that are performed within a project team. They should also acquire an understanding of the interplay between modeling and experiment, especially with regard to strategies for the evaluation of their computer-generated models. the course will also aim to (I) provide chemists with insights into the nature of the primary literature describing molecular modeling methods, and its distribution throughout a wide range of journals, and (ii) facilitate their participation in more advanced workshops on more specialized topics within computational chemistry.

Course Organization

�@�@The proposed course will be held over two consecutive days and will consist of lectures and a number of " hands-on " laboratory exercises. The latter component should provide an opportunity for interested chemists to receive " hands-on " training. Dr. Nigel Richards will also participate in informal discussions after the lecture sessions. BioWeb will organize travel and accommodation for faculty. A draft workshop schedule is outlined at the end of this document.

Curriculum Summary

�@�@Material in the course will be structures, in four sections, to reflect the types of problems that are faced by medicinal and synthetic chemists in many projects of current interest. The workshop lectures will cover as much of the following materials as time allows:

Section 1

Molecular Structure, Conformation and Design

Overview of Computational Methods in Drug Discovery and Chemical Synthesis

Computational Approaches to Peptidomimetic Design and Evaluation

Empirical Force Fields

Energy Minimization Methods

Solvation Models

Conformational Search Procedures

Calibration and Evaluation of Search Results

Section 2

Deriving 3-D Structure-Activity Relationships - Problems and Solutions

Physical Chemistry of Intermolecular Association

The Thermodynamic Origins of Ligand Specificity

Methods for Mapping Potential Ligand Binding Sites in Proteins

Estimating Relative Ligand Affinities

Introduction to Free Energy Calculations

Molecular Similarity - Definitions and Implementation

Superposition Methods - Limitations of the Functional Group Approach

Computing Molecular Electrostatics, Shape and Volume

An Overview of QSAR

Partition Coefficients - Calculation, Measurement and their Biological Relevance

Modeling Biological Target Site on the Basis of Ligand Structure and Selectivity

Computational Methods in the Design of Small Molecular Combinatorial Libraries

Section 3

Transition State Modeling and Prediction of Chemical Reactivity

Overview of Quantum Chemical Techniques

Molecular Orbitals and Atomic Orbitals

The Hartree-Fock Equations

Ab initio Calculations

Semi-Empirical Calculations

AM1 and PM3 Semiempirical Models

Solvation Models

Molecular Electrostatic Potentials

Combining Quantum Mechanics and Empirical Force Field Methods

Visualization of Molecular Properties

Qualitative Methods for Designing Transition State Analogs as Enzyme Inhibitions

Frontier Orbitals, Electrostatics and Steric Interactions in Controlling Reactivity

Modeling Chemical Reactivity - Examples Selected from the Following Topics

Addition of Nitrone Anions to Polarized Olefins

Thiol-catalyzed Amide Hydrolysis

Section 4

Current Issues and Methods in Protein Structure Prediction

Multiple Sequence Alignment - Problems and Solutions

Secondary Structure Prediction Methods

Construction of Tertiary Structure using Homology Modeling

Modeling Transmembrane G Protein-Coupled Receptors and their Interactions with Ligands