Method to select chemical molecules having a drug activity by binding to a predetermined target prot

Medicilon's structural biology department offers services supporting structure-based drug discovery from determination of novel targets to final structures. Our platform is one of the earliest established structural biology platforms in China and has been certified by the Shanghai Government. Email:marketing@medicilon.com.cn Web:www.medicilon.com

Method to select chemical molecules having a drug activity by binding to a predetermined target protein, said method comprising: a) identifying (S12) at least a ligand with physical and chemical properties suited to bind with an active site of said target protein; b) fragmenting (S14) said ligand into a set of chemical fragments such that each said chemical fragment (including combination of fragments) has an affinity with at least one site of said target protein; and, in parallel, c) fragmenting (S14) said ligand into scaffold fragment (s), i.e. the remaining of the ligand after removing of the substituent (s) and molecular part fragment (s), i.e. the remaining of the ligand after removing of the connector (s); d) searching (S18) in database of existing molecules, molecules having at least a chemical sub-structure similar to said chemical compound obtained by b) and c); e) validating said found molecules by keeping only molecules showing chemical activity and by simulating their capabilities to combine with said target protein.

The invention relates to the field of computer- aided drug design and, more particularly, de novo drug design. It relates particularly to a method to select a chemical molecule having a drug activity by binding to a predetermined target protein. The invention relates also to a computer program product to perform the method.

Many approaches have been used to discover new chemicals which are suitable for particular purposes, particularly chemicals having a drug activity.

Traditionally, chemical discovery method is based on the screening of very large number of compounds with the goal of finding one or more "hit" compounds for further testing and refinements.

With the advance in computer science and particularly the ability to describe and simulate on a computer a given molecule, new methods to discover chemicals have been developed. This new field is often called computer-aided drug design. It can form a valuable partnership with experiments by providing estimates when experimental approaches are difficult or expensive and by coordinating the experimental data available.

These in silico methods encompass pharmacophoric model identification or quantitative structure activity relationship (QSAR) analysis. But these methods can only be applied to the "forward" problem, which requires the computation of physical, chemical and/or biological properties from the molecular structure of known active molecules .

Another problem, also known as the Λbackward' problem, is the identification of the appropriate molecular structure given some desired physico-chemical or biological properties. Programs to solve, at least partially, this problem are called de novo drug design programs . These programs generate novel active structures fitting the active site of the target protein by searching the whole chemistry space. But they have to solve a huge combinatorial and non linear structure- property correlation problem for which an exhaustive search is impractical.

Some of these programs use genetic algorithms to the design of new molecular structures. For instance, the patent application WO 95/01606 discloses a method for evolving successive populations of molecular structures and evaluating each evolved structure of each population with desired physical and/or theoretical properties.

The document EP 0 818 744 discloses a method for drug candidate identification from a computerized representation of the 3D structure of a binding site on the surface of a biological macromolecule . From the 3D structure, it deduces a functional structure of the binding site to identify favourable and unfavourable interactions between the binding site and a drug candidate molecule. It identifies a molecular fragment capable of placement within the binding site and carrying substituent group. It generates a set of reagent compound that may be reacted to produce a candidate compound comprising the molecular fragment. The method then screens the set of compounds by simulating the interaction with the binding site to retain only those compounds which have a good relationship. Then, the method has an experimental step where, after synthesis of selected candidate compounds, they are evaluated for drug efficacy. This method, as many de novo methods, has a drawback related to the experimental phase. Indeed, the synthesis of a new compound generated only by simulation may be very difficult and even sometimes, impossible. At the opposite of a simulator which is able to construct a virtual molecules atom by atom, the chemist needs to find new chemical reactions to create a new molecule which is a research per se.

Therefore, it would be advantageous to achieve a de novo method which generates already existing or easy to synthesize molecules.

To better address one a more concerns, in a first aspect of the invention, a method to select chemical molecules having a drug activity by binding to a predetermined target protein, said method comprising: a) identifying at least a ligand with physical and chemical properties suited to bind with an active site of said target protein; b) fragmenting said ligand into a set of chemical fragments and combination of fragment such that each said chemical fragment and combination of fragments has an affinity with at least one site of said target protein; and, in parallel, c) fragmenting said ligand into scaffold fragment (s), i.e. the remaining of the ligand after removing of the substituent (s) and molecular part fragment (s), i.e. the remaining of the ligand after removing of the connector (sj ; d) searching in database of existing molecules, molecules having at least a chemical sub-structure similar to said chemical compound obtained by b) and c) ; e) validating said found molecules by keeping only molecules showing chemical activity and by simulating their capabilities to combine with said target protein.

The 3D structure being known, a potential ligand is searched in chemical databases, step S25, based on the properties of the active site. For instance, ligands of protein homologs may be searched. If no ligands are known, step S27, a potential ligand is constructed by using some known de novo drug design programs, step S29, or by virtual, i.e. computer simulated, screening of compounds, step S31.

The de novo drug design program and the screening program use molecule fragment database FD as input .

A molecule database is a computer based database in which each molecule is described by using some molecular representation adapted to computer storage and manipulation, representation well known from the man skilled in the art.

A database of existing molecules means a database of molecules already synthesized. Example of such database are the Comprehensive Medicinal Chemistry database or the free database of commercially available compounds ZINC.