Cerius²·Ludi

1       Introduction

Ludi is a method for the de novo design of ligands for proteins (e.g., enzyme inhibitors). It can also suggest modifications of known ligands that may enhance their binding to the target protein.

Ludi allows you to:

• Design de novo candidate protein ligands for the active sites of proteins.

• Suggest modifications of known ligands.

• Manage Ludi libraries.

• Score ligand-inhibitor complexes.

ANALOGS BASED DESIGN -- wherein ligand design is based solely on analogs (i.e., known ligands).

RECEPTOR BASED DESIGN -- wherein ligand design is based on the receptor active site structure.

The Ludi deck of cards:

ACTIVE SITE VIEWER

LUDI LIBRARY

ANALOGS BASED DESIGN

RECEPTOR BASED DESIGN

Background

The design of new and selective ligands for proteins (e.g., enzyme inhibitors) is one of the most important applications in contemporary rational drug design. When the structure of an enzyme is known, it is possible to display it in a modeling environment, such as Cerius², to select potential binding sites by inspection and to design an inhibitor that targets those sites. If, in addition, the structure of one or more protein-inhibitor complexes is known, the design may be aided by a study that identifies essential ligand-protein interactions.

Visually inspecting a protein's active site is of great help to chemists in generating ideas for new potent inhibitors. However, the efficacy of computer-aided drug design is greatly improved by tools that propose potential ligands by an automatic or semiautomatic procedure (Cohen et al. 1990). Indeed, many different approaches towards the de novo design of drugs have been proposed (DesJarlais et al. 1988, Lewis 1990, Goodford 1985, Tomioka 1987).

Two approaches

The known structure approach

One possibility is to search through databases of known structures such as the Cambridge Structural Database (Allen et al. 1983, 1979) and to identify those entries that fit into the active site. The advantages of this approach are that the molecules retrieved from the database do exist and the structures represent low-energy conformations.

However, this approach does not address the issue of conformational flexibility. The crystal structure is not the only low-energy conformation and there are well known cases in which the conformation of an inhibitor bound to the enzyme is different from the crystal structure conformation (Hambley et al. 1986). Moreover, the number and variety of structures is limited by the size of the database used. Both limitations could, at least in principle, be overcome by performing a conformational analysis of the compounds in the active site. It is clear, however, that such a procedure would be very demanding both in CPU time and in disk space usage.

The fragment approach

Another approach uses a library of fragments. The idea is to position molecular fragments into the active site in such a way that hydrogen bonds can be formed with the enzyme and hydrophobic pockets filled with hydrophobic groups. These fragments are then connected by suitable spacer fragments to form a single molecule.

There are several advantages to this approach. It is very fast and due to the large number of possible fragment combinations the variety of molecules that can be generated is enormous.

The Ludi method

Ludi is based on the fragment approach. It suggests how suitable small fragments can be positioned into clefts of protein structures (e.g., an active site of an enzyme) in such a way that hydrogen bonds can be formed with the enzyme and hydrophobic pockets are filled with hydrophobic groups. This positioning is the strength of Ludi because it immediately provides you with ideas about how putative binding sites on the protein might be saturated by fragments and how those fragments might be linked together.