CAMM OF ANTICANCER DRUGS
Karel Huml, Vladimír ubr, Karel Ulbrich
Institute of Macromolecular
Chemistry, Academy of Sciences of the Czech Republic,
Heyrovského nám.2, 162 06 Praha 6, Czech Republic
huml@imc.cas.cz
Keywords : Computer-aided
Molecular Modelling (CAMM), Cancer, Drugs
X-ray and NMR data gathered over the past thirty years has helped us to understand the structure of many complexes at the molecular level providing a detailed description of the interaction between enzymes and their inhibitors. Further study has given us a detailed knowledge concerning the interaction between enzymes and selected substrates, as well. Computer-aided molecular modelling (CAMM) was used as one of the fundamental tools of theoretical chemistry to solve such problems. The binding prediction method as such is an important part of structure-based drug design protocols. Molecular mechanics, the core of CAMM methods, cannot treat the formation/breaking of covalent bonds correctly. However, the noncovalent interaction between molecules can be described in a good agreement with experimental results. Therefore, CAMM methods were utilized parallelly to the experimental methods for a study of lysosotropic anticancer drugs in the Institute [1]. Particularly, interaction of the lysosomal enzyme cathepsin B with polymeric drugs synthetized in our laboratory, leading to the controlled activation of the drug, was under extensive study. A detailed knowledge of the structure and function of the active cleft of the enzyme is a leading information which controls the docking maneuver of the drug approaching the enzyme molecule. The subsequent molecular dynamics, combined with annealing procedure and energy minimization, was used to design a starting model of a noncovalent complex. Calculation of Gibbs free energy was used to estimate quality of competing models of closely related conformations and/or configurations of the substrate under study. A bottleneck of the method is a limited possibility to search the conformational space extensively due to the computing facility available. An intelligent use of additional physico-chemical information is of significant help to focus our attention to reliable models only. Two fragments of polymeric drugs coming into interaction with cathepsin B will be described in detail [2,3], especially, -Gly-Phe-Leu-Gly-NAp and -Gly-Phe-Gly-Lys-NAp, where NAp is a model drug (4-nitroaniline). In any case, the final check of the results is a biochemical experiment, which, unfortunately, is expensive and time consuming.
This work was supported by the Grant Agency
of the Czech Republic: grant 307/96/ K226 and grant 203/96/0111
and by COPERNICUS: grant ERB 3512 PL 941 009.