SELECTION OF NEW TARGET PROTEINS FOR DRUG DESIGN

SELECTION OF NEW TARGET PROTEINS FOR DRUG DESIGN IN GENOME OF MYCOBACTERIUM TUBERCULOSIS Alexander V. Veselovsky V. N. Orechovich Institute of Biomedical Chemistry RAMS, Moscow, Russia e-mail: veselov@ibmh. msk. su

Modern pipeline of new drug development Identify disease Find a drug effective against disease protein (2 -5 years) IN le Ability to decreasing finance and time cost + - Fi le ND A Formulation & Scale-up Human clinical trials (2 -10 years) Fi Preclinical testing (1 -3 years) D Isolate protein involved in disease (2 -5 years) FDA approval (2 -3 years)

Pipeline of target-based and main steps in drug development

Genomics for drug discovery Genome Annotation and classification of genes Drug targets selection

Comparative genomics Human genome Gram(+) bacteria genome Genes-targets of bacteria that differ from human genes D. T. Moir et al. , 1999 Gram(-) bacteria genome

Requirements of “Ideal” Antimicrobial Agent and to Its Target

Target selection (Comparative genomics) favourable similarity Genomes of related species Genomes of other strains of target species Proteins with known spatial structures (PDB) Unfavourable similarity Human genome Target genome Genomes of human symbiont microorganisms

Gene. Mesh – program for protein-targets selection for antimicrobial drug discovery using comparative and functional genomics A. V. Dubanov, A. S. Ivanov, A. I. Archakov (2001) Computer searching of new targets for antimicrobial drugs based on comparative analysis of genomes. Vopr. Med. Khim. 47, 353 -367. (in Russian).

Algorithm of program Gen. Mesh Set of proteins from PDB BLAST Spatial structure ability Genomes of related species Target genome databases Genomes of other strains of target species Human genome BLAST Presence of homologs in genomes of related species Absence of mutations in other strains of target species Absence of homologs in human genome

Target selection in Mycobacterium tuberculosis H 37 Rv using broadened set of genomes for analysis targets for antimycobacterial agents without influencing normal human microflora Common targets for Mycobacteria and fungi

3 D protein structure modelling Approach Homology modelling Limitation Model and template sequence identity must be > 30% Results heavily dependent on human Threading expertise and information from other (Fold recognition) methods for elimination decoy folds Ab initio (De novo) < 150 amino acids * - RMSD of C (A) and residues true positions (%) Accuracy* 1 -3 A 80 -95% 3 -6 A 30 -50% 4 -8 A < 30%

Target selection in genome of Mycobacterium tuberculosis H 37 Rv

Potential Targets Found in Genome of M. tuberculosis H 37 R Freiberg C, Wieland B, Spaltmann F, Ehlert K, Brötz H, Labischinski H. Identification of novel essential Escherichia coli genes conserved among pathogenic bacteria. J Mol Microbiol Biotechnol. 2001 Jul; 3(3): 483 -9. Thanassi JA, Hartman-Neumann SL, Dougherty TJ, Dougherty BA, Pucci MJ. Identification of 113 conserved essential genes using a high-throughput gene disruption system in Streptococcus pneumoniae. Nucleic Acids Res. 2002 Jul 15; 30(14): 3152 -62.

Russian Federal Space Agency Program for protein crystallization in weightlessness International space station (ISC)

Target M. tuberculosis H 37 R

Phosphopantetheine adenylyltransferase of bacteria PPAT 4'-phosphopantetetheine + ATP PPi + 3'dephospho-Co. A + Pi Coenzyme A Penultimate and rate-limited enzyme of bacterial coenzyme A biosynthesis

Comparison of spatial structures of PPAT M. tuberculosis Active site Green – from Russia (1, 6 A) Yellow – 1 TFU. pdb (1, 99 A)

Scheme of virtual screening for new PPAT inhibitors in molecular database Molecular database Experimental testing Database preprocessing Manual selection Docking Compounds selection by scoring functions consensus Calculation of additional scoring function

Discovery ligands from molecular database by docking method

Empirical scoring function The method is fast semi-automated is applicable to 3 -D models does not need extensive training

Accuracy of scoring function

Relationship between scoing functions

Limitation of scoring functions Srt Ligand in solution HLW Receptor HRW Free energy bound water free rotation Sint G = H-T S loosely associated water molecules free water Entropy HLR SW Svib Receptor-Ligand complex Enthalpy

Consensus of scoring functions

The first docking of compounds in PPAT active site 17500 complexes

Active site of phosphopantetheine adenylyltransferase M. tuberculosis

The second docking of compounds in PPAT active site 24000 complexes

Experimental testing of selected ligands

Acknowledgments. This work was supported in part by Russian Federal Space Agency (in frame of ground preparation of space research). Participants: Institute of Bioorganic Chemistry RAS Institute of Crystallography RAS Institute of Biomedical Chemistry RAMS