MCSS/HOOK, Molecular Simulations Inc.

Allen B. Richon

Summary of an article by William J. Cromie that appeared in the Harvard Gazette, and reproduced with permission, from MC2, vol 2, no. 1, Molecular Simulations Inc.



http://www.netsci.org/Science/Compchem/feature04.html

Computational drug design is finding more and more applications in today's pharmaceutical and biotech industry. Recently, Professor Martin Karplus and co-workers Drs. Amedeo Caflisch, Diane Joseph-McCarthy, and Leo Caves have developed a computational approach in their search for new drugs to bind to proteins when the crystal structure is known. Their program, Multiple Copy-Simultaneous Search, or MCSS, has been used by Karplus' group to design drug candidates in the fight against the viruses responsible for polio, the common cold, and AIDS. They have recently applied for patents for the design of these anti-viral drugs.

MCSS works by flooding a protein's active site with thousands of copies of organic functional groups; and allowing them to energy-minimize onto the protein surface. The groups that minimize to the same location are inspected, and only those groups that bind most tightly remain while the others are discarded. The end result is a "painting" of the active-site by these functional groups. These functional groups can then be connected to postulate molecules that contain several of these groups. If the molecules then can be synthesized, they can be tested to see how well they actually bind to the protein.

Karplus et al. performed these steps in the design of two molecules to inhibit the polio virus. Their results were sufficiently encouraging, and they are now the subject of the patent filing. A similar approach is underway to develop molecules that would inhibit the AIDS virus. Several candidates are already being synthesized by collaborators at Purdue University. "We have taken the first step," says Professor Karplus. "As the identity and structure of more molecules related to various diseases are determined, rational design methods will be more widely used in finding new drugs." No doubt, if results as promising as those achieved already by the Harvard group continue, MCSS will be at the forefront of this revolution in drug design.

(78 kbytes)

This image shows an MCSSS/Hook result from searching an active site region with 4 functional groups. This functional site map is computed for the active site of hemagglutinin of the influenza A virus with N-methyl acetamide (green), acetic acid (yellow), methyl ammonium (violet), and methanol (white). Lower energy groups are shown, with a surface showing highly negative (blue) and positive (red) regions of electrostatic potential.

MCSS and HOOK provide scientists with a new, systematic approach to ligand design that makes sense. Coupling structural information and computational methods has provided a basis for the rational drug discovery process which has emerged during the last decade. However, techniques for the systematic exploration and screening of new possible ligands have been lacking. And new ligands generated by some exploratory techniques are often difficult or impossible to synthesize in the laboratory.

MCSS and HOOK: Systematic structure-based ligand design

MCSS and HOOK combine to provide a systematic, comprehensive approach to ligand development and de novo ligand design that result in synthetically feasible molecules. Using libraries of functional groups and molecules, MCSS systematically searches for energetically feasible binding sites in a protein. HOOK then systematically searches a database for skeletons which logically connect these binding sites in the presence of the protein. The results are new potential leads, either known compounds from an existing database or new molecules from synthetically feasible skeletons.

MCSS: The MCSS (Multiple Copy Simultaneous Search) method[1,2] exhaustively searches potential active sites in proteins for potential binding points using a unique, computationally efficient approach. In this approach, copies of molecules that correspond to functional groups are randomly placed around a suspected active site. Using energy minimization, new positions for the molecules are simultaneously computed such that each individually interacts with the protein target, but not with each other. MCSS iteratively prunes copies that converge to similar binding locations, reducing the total number of copies during the course of the search. Together, the parallel computation and iterative pruning makes extensive searching possible on computer workstations.

The result is a set of energetically favorable locations and orientations for the input molecules, based on computations with the CHARMm force-field. Using molecules related to functional groups (e.g., methanol, amide, etc.), these sets provide a map of potential binding locations within an active site region for functional groups on ligands. Interaction energies are computed using CHARMm. This technique represents a novel way of characterizing the chemistry of an active site, very different from the traditional grid-probe calculations.

HOOKing a Molecule Together: Given the positions and orientations of these functional sites, HOOK [3] takes the next logical step and attempts to link multiple functional groups with molecular templates taken from a database. For those templates passing this first sieve, a steric overlap with the protein active site is computed, penalizing steric overlap and keeping only those molecules that are topologically feasible. The resulting ligands are, therefore, consistent both with the geometry and chemistry of the binding site. Because the database contains only synthetically feasible templates, this process inherently identifies molecules.

Functions of MCSS and HOOK: These methods permit a variety of structure-based ligand design studies which facilitate functions such as:

  • Determine the classes of molecules that could bind in the active site by computing a functional site map for a variety of functional groups;

  • Screen a database of known molecules against the results of an MCSS functional group search for possible leads among existing molecules;

  • Identify active molecules by searching a database of molecular skeletons that have been developed by your synthetic chemists;
  • Identify potential binding locations and orientations of known leads using MCSS;

  • Use descriptors of functional group position and orientation from other sources, such as NMR, and use HOOK to construct a novel molecule and

  • Locate potential active site by systematic searching of regions of a protein with functional groups.

The QUANTA Interface: A graphical interface to both MCSS and HOOK provides tools for the complete ligand design process. These allow users to:

  • Interactively indentify and view the active site for MCSS searches;

  • Interactively select functional groups, set run-time parameters, and modify ligands;

  • Visualize functional group maps - categorized and colored by interaction energy;

  • Flexible control of HOOK search criteria, including database source and number of functional groups to connect, allowing methylene expansion of skeletons and overlap score;

  • Browse hits from a search of the target protein with display of close contacts and coloring of candidates;

  • Allow further refinement of candidate structures using CHARMm energy minimization and dynamics and

  • Output structures to the Drug Discovery Workbench and Catalyst.

(62 kbytes)

HOOK connects functional groups from MCSS with molecular templates from a database. Shown here is the best hit connecting two of the functional groups while fitting into the steric environment of the active site.

(64 kbytes)

The resulting best hit is superimposed with sialic acid from the crystal structure of hemagglutinin. No alignment of the molecules has been done, indicating that the result of the computational techniques are consistent with the binding observed in the crystal structure.

References:

  1. A. Miranker and M. Karplus, "Functionality maps of binding sites: a multiple copy simultaneous search method", Proteins, Structure, Function and Genetics 11, 29-34 (1991)

  2. A Caflish, A. Miranker, and M.Karplus "Multiple copy simultaneous search and construction of ligands in binding sites: application to inhibitors of HIV-1 aspartic proteinase" J.Med.Chem. 36, 2142-2167 (1993)

  3. M.B. Eisen, D.C. Wiley, M. Karplus and R.E. Hubbard "HOOK: A program for finding novel molecular architectures that satisfy the chemical and steric requirements of a macromolecule binding site." Proteins Structure, Function and Genetics 19, 199-221 (1994).



NetSci, ISSN 1092-7360, is published by Network Science Corporation. Except where expressly stated, content at this site is copyright (© 1995 - 2010) by Network Science Corporation and is for your personal use only. No redistribution is allowed without written permission from Network Science Corporation. This web site is managed by:

Network Science Corporation
4411 Connecticut Avenue NW, STE 514
Washington, DC 20008
Tel: (828) 817-9811
E-mail: TheEditors@netsci.org
Website Hosted by Total Choice