Anderson Coser Gaudio

Proposição, validação e análise dos modelos que correlacionam estrutura química e atividade biológica

The present paper aims to bring under discussion some theoretical and practical aspects about the proposition, validation and analysis of QSAR models based on multiple linear regression. A comprehensive approach for the derivation of extrathermodynamic equations is reviewed. Some examples of QSAR models published in the literature are analyzed and criticized.

Seleção de variáveis em QSAR

The process of building mathematical models in quantitative structure-activity relationship (QSAR) studies is generally limited by the size of the dataset used to select variables from. For huge datasets, the task of selecting a given number of variables that produces the best linear model can be enormous, if not unfeasible. In this case, some methods can be used to separate good parameter combinations from the bad ones. In this paper three methodologies are analyzed: systematic search, genetic algorithm and chemometric methods. These methods have been exposed and discussed through practical examples.

HEPT derivatives as non-nucleoside inhibitors of HIV-1 reverse transcriptase: QSAR studies agree with the crystal structures

The interest in the non-nucleoside inhibitors (NNIs) to the reverse transcriptase (RT) as anti-AIDS agents has grown in the last ten years. The compound 1-[(2-hydroxyethoxy)methyl]-6-(phenylthio)thymine (HEPT) is the precursor of the most studied class of NNIs, from which hundreds of derivatives have been synthesized and tested. There are at least twelve QSAR studies about the HEPT derivatives as RT inhibitors. Most of the predictions derived by these studies are related to the nature of the active site near the substituents at positions N-1 and C-5, and at the C-6 phenyl ring. The validity of these models has been checked against the 3-D structure of HIV 1 RT-HEPT complexes available. Most of these predictions were confirmed at the molecular level.

A comparative study of principal component and linear multiple regression analysis in SAR and QSAR applied to 1,4-dihydropyridine calcium channel antagonists (nifedipine analogues)

Abstract The method of principal component analysis was shown to be capable of classifying 1,4-dihydropyridine derivatives into high active and low active groups for various different sets of compounds. Mainly quantum chemical parameters were used with this method. The kind of parameters employed by the method were found to be different to those employed by linear multiple regression analysis for identical sets of compounds.

Validação lateral em relações quantitativas entre estrutura e atividade farmacológica, QSAR

The comparative QSAR is a tool for validating any statistical model that seems to be reasonable in describing an interaction between a bioactive new chemical entity, BIONCE, and the biological system. In order to deeper the understanding of the relationships and the meaning of parameters within the model it is necessary some kind of lateral validation. This validation can be accomplished by chemical procedures using physicochemical organic reactions and by means of biological systems. In this paper we review some of such comparisons and also present a lateral validation between the same set of antimicrobial hydrazides acting against Saccharomyces cerevisiae yeast and Escherichia coli bacterium cells. QSARs are presented to shed light in this important way of stating that the QSAR model is not the endpoint, but the beginning.

Prediction of the binding mode of N2-phenylguanine derivative inhibitors to herpes simplex virus type 1 thymidine kinase

The probable binding mode of the herpes simplex virus thymidine kinase (HSV1 TK) N2-[substituted]-phenylguanine inhibitors is proposed. A computational experiment was designed to check some qualitative binding parameters and to calculate the interaction binding energies of alternative binding modes of N2-phenylguanines. The known binding modes of the HSV1 TK natural substrate deoxythymidine and one of its competitive inhibitors ganciclovir were used as templates. Both the qualitative and quantitative parts of the computational experiment indicated that the N2-phenylguanine derivatives bind to the HSV1 TK active site in the deoxythymidine-like binding mode. An experimental observation that N2-phenylguanosine derivatives are not phosphorylated during the interaction with the HSV1 TK gives support to the proposed binding mode.

Extension of the frontier reactivity indices to groups of atoms and application to quantitative structure-activity relationship studies

Abstract The concept of frontier reactivity index, proposed by Fukui et al. (J. Chem. Phys., 20 (1952) 722), was extended to groups of atoms. The purpose of this procedure is to simulate frontier indices capable of predicting the susceptibility of a chemical group to an external electronic interaction, such as nucleophilic, electrophilic or radical. The group frontier indices can be useful in quantitative structure-activity relationship (QSAR) studies, where substituent constants for groups of atoms are widely used. The group frontier indices are obtained through the sum of the corresponding atomic frontier indices. The atomic frontier indices were computed from AM1 wavefunction coefficients. The frontier indices considered in this article, i.e. frontier electron, orbital and radical densities, are shown to be relatively independent of the molecular geometry. We applied the group frontier index concept to a QSAR study of triazene derivatives. The group frontier indices that appear to be significant in the correlation analysis are consistent with the known hydroxylation mechanism oftriazene induced mutagenesis.

QSAR and molecular graphics analysis of N2-phenylguanines as inhibitors of herpes simplex virus thymidine kinases.

A quantitative structure-activity relationship study of N2-(substituted)-phenylguanines (PHG) as inhibitors of herpes simplex virus thymidine kinase (HSV TK) was performed. The activity of a set of PHG derivatives were analyzed against the thymidine kinase of herpes simplex virus types 1 (HSV1 TK) and 2 (HSV2 TK). Classic and calculated physicochemical parameters were included in the analysis. The results showed that there is an important difference in the activity of the meta substituted PHG derivatives against HSV1 TK and HSV2 TK. The activity of the meta derivatives against HSV2 TK is influenced by a steric effect, which is not observed against HSV1 TK. The superposition of the three-dimensional structures of the active sites of HSV1 TK (crystal structure) and HSV2 TK (homology model) revealed that the amino acid Ile97 is located near the meta position in the HSV1 TK active site, whereas the amino acid Leu97 is located near the meta position in the HSV2 TK active site. This single difference in the active sites of both enzymes can explain the source of the steric effect and serves as an indication that our previously proposed binding mode for the PHG derivatives is plausible. However, another observed mutation in the active site region, Ala168 by Ser168, suggests that an alternative binding mode, similar to that of ganciclovir, could be possible.

MoCalc: A new graphical user interface for molecular calculations

A new computer program called MoCalc (Molecular Calculations) has been designed to help the computational chemistry practitioner in the task of performing and analyzing molecular calculations. MoCalc is a graphical user interface for the MO calculation programs Gamess and Mopac, and uses Rasmol and Babel for molecule display and file conversion, respectively. In its initial version, MoCalc can execute the following operations: (a) create and handle Gamess and Mopac input files; (b) import any kind of molecular geometry supported by Babel and paste it as Cartesian, internal, or Gaussian‐type coordinates on the input file; (c) convert Gamess and Mopac output files to inputs of both programs; (d) edit and validate the keywords that control the Gamess and Mopac calculation procedure; (e) display the input (Mopac) and output (Gamess and Mopac) molecular geometries; (f) run single or multiple (batch) calculations, either interactively or in background; (g) automatically open the output files as soon as the calculation finishes; (h) extract results from the output files, such as energy, charges, dipole, population analysis, wave function, bond orders, and valence analysis, and display them in spreadsheets; (i) calculate reactivity indices derived from the frontier orbital theory and the root‐mean‐square (rms) deviation of input and output geometries. All the results generated by MoCalc can be promptly transferred to text editors and electronic spreadsheets, which facilitate a detailed subsequent analysis and the publication of the results. MoCalc can also perform graphical and numerical comparative analysis of the some results when more than one output file is loaded. The program was coded in Visual Basic and runs in Windows 95/98/NT4/ME/2000/XP environments.

Conformational analysis and the active conformation of N2-phenylguanine derivatives

Abstract A conformational analysis of four N 2 -phenyl-(substituted)-guanine (PHG) derivatives, which are herpes simplex virus thymidine kinase inhibitors (HSV TK), was performed mainly with the semiempirical AM1 method. Nonempirical (ab initio) HF and MP2 calculations were employed to refine some of the results obtained with the semiempirical method. The two dihedral angles that connect the phenyl and the guanine rings were varied for systematic conformational search. Four stable conformations were found: C1 ( θ 1 =184°; θ 2 =58°); C2 ( θ 1 =174°; θ 2 =127°); C3 ( θ 1 =186°; θ 2 =233°) and C4 ( θ 1 =176°; θ 2 =303°), in which θ 1 =N1−C2−N11−C1′, and θ 2 =C2−N11−C1′−C2′. The conformation C3 was found to be equivalent to the bioactive conformation proposed in a previous paper.