Virtual screening models or cruzain inhibitors: development and Eexperimental validation

Author(s)

Alberto Malvezzi

• Document type: Doctoral Thesis
• Press: São Paulo.
• Institution: Universidade de São Paulo (USP). Conjunto das Químicas (IQ e FCF) (CQ/DBDCQ)
• Defense date: 2008-05-09

Examining board members

Antonia Tavares do Amaral; Omar A. EL Seoud; Chuck Shaker Farah; Elizabeth Igne Ferreira; Richard Charles Garratt

• Advisor: Antonia Tavares do Amaral

Abstract

In order to search and identify new cruzain inhibitor(s) - a cysteine-protease of Trypanosoma cruzi, the etiologic agent of Chagas disease - two virtual screening schemes(Models I and II) were proposed, validated- and applied to the ZINC database (3.294.714 compounds). The proposed virtual screening models, bearing a sequence of different physicalchemical, pharmacophore and docking filters, as well as a visual inspection filter, were built from information taken from 13 cruzain complexes and from 20 complexes of other cysteine proteases, having their structures available in PDB. In a first step, a detailed recognition of the cruzain structural features and characteristics was performed through visual inspection of the enzyme environment; followed by the analysis of GRID generated molecular interaction fields; through the identification of molecular interaction properties exposed at the enzyme cavity surface, generated by the CAVBASE program; and by molecular dynamics simulations. The virtual screening Model I, - generated from the structural characteristics recognized from 13 PDB cruzain complexes - when applied to the ZINC database selected 10 compounds. For the experimental validation ofthe model, six ofthese compounds have been acquired and were tested as cruzain inhibitors. It was observed that three of the tested compounds (ZINC02470662, ZINC02682879 and ZINC03192044, respectively) did not show any significant cruzain inhibition, up to 7 mM. Meanwhile the other three tested compounds (ZINC02663001, ZINC01936854 and ZINC03326243, respectively) showed an unspecific cruzain inhibition, suggesting that an enzyme inhibition by promiscuous mechanism occurred. This mechanism was verified by the addition of 0.1% Triton X-100 on the enzymatic assay with a concomitant loss of cruzain inhibition activity. For these compounds, the confirmation of the promiscuous mechanism was also done, observing the loss of enzyme inhibition, after a ten times increase in the cruzain concentration on the enzymatic assay. The virtual screenmg Model II - generated from the structural characteristics recognized from 13 cruzain complexes and 20 complexes of other cysteine proteases, that have been identified on a search for cavities similar to cruzain - selected 55 compounds, when applied to the ZINC database. In order to experimentally validate the model, nineteen compounds have been acquired and were tested as cruzain inhibitors. It has been observed that one compound, ZINC01794422, showed a specific cruzain inhibition (Ki = 21 µM), while the other eighteen showed no significant inhibition, up to 592 µM concentration. The promiscuous mechanism of enzymatic inhibition was not observed, since 0.1% of Triton X-100 was added in ali assays. Additionally, Model II identified two other cruzain inhibitors (ZINC04899534 and ZINC01547017). However, these compounds have not been acquired or tested, since they are known cruzain inhibitors - already described in the literature and are structurally similar to the inhibitors used in the construction of the mode!. Referring to new inhibitors found, Model II showed a hit rate of 5,3%. This value is in agreement with those found in the literature, which ranges from 1 to 50%. (AU)