Yuji Takahata

Structure-activity relationship studies of carcinogenic activity of polycyclic aromatic hydrocarbons using calculated molecular descriptors with principal component analysis and neural network methods.

Recently a new methodology based on local density of state (LDOS) calculations using topological and semiempirical methods was proposed to identify the carcinogenic activity of polycyclic aromatic hydrocarbons (PAHs). In this work we perform a comparative study of this methodology with principal component analysis (PCA) and neural networks (NN). The PCA and NN results show that LDOS quantum chemical descriptors are relevant descriptors to identify the carcinogenic activity of methylated and non-methylated PAHs. Also, we show that the combination of these distinct methodologies can be an efficient and powerful tool in the structure-activity studies of PAHs compounds. We have studied 81 methylated and non-methylated PAHs, and our study shows that with the use of these methods it is possible to correctly predict the carcinogenic activity of PAHs with accuracy higher than 80%.

Calculation of molecular surface area with numerical factors

Abstract A simple method for the calculation of molecular surface area of small as well as large molecules is presented. The method is based on Higo & Go algorithm [J. Comput. Chem. 10, 376 (1989)] for the supermolecular volume and surface area calculations. We have verified that the average molecular surface area which intercepts the cubes located at the atomic sphere surface is linearly proportional to the side area of the cubes. We determined the proportionality factors and tested them in a system with analytically known surface area and some hydro-carbon and alcohol molecules. Good correlations between calculated surface area and log(S), where S is the aqueous molar solubility of the hydrocarbons and alcohols, have been obtained.

The use of AM1 in structural analyses of primary and secondary enaminones

Abstract The AM1 semiempirical method was employed to study the structural features of some primary and secondary acyclic and cyclic enaminones. This method was shown to be very suitable for these compounds, enabling the correct prediction of the more stable tautomer and geometric forms and other features such as strength of intramolecular hydrogen bonds and rotational barriers. In comparison, its predecessor, the MNDO method, very often gave incorrect or inconsistent results.

SAR ANALYSIS OF SYNTHETIC NEOLIGNANS AND RELATED COMPOUNDS WHICH ARE ANTI-LEISHMANIASIS ACTIVE COMPOUNDS USING PATTERN RECOGNITION METHODS

Abstract Potentially active new neolignan and analogues against leishmaniasis are proposed. Structure-activity relationship (SAR) techniques were employed. Physicochemical properties such as log P , molecular volume, atomic charge and quantum chemical parameters were calculated for a group of synthetic substances for which the biological activities against leishmaniasis are known. Only about half a dozen out of more than twenty parameters were found to be efficient for the classification of the compounds into active and inactive groups.

Inclusion compounds between α-, β- and γ-cyclodextrins: iron II lactate: a theoretical and experimental study using diffusion coefficients and molecular mechanics

Abstract The inclusion compounds between iron II lactate and three different cyclodextrins (CDs) were studied by means of experimental and theoretical data. The importance of iron II in the human metabolism effort the necessity of a minimum concentration to the human life. Malnutrition is one great problem in social politics of many countries on the world. The possibility to the development of novel medicines with the iron II species stable look for an increase on the efficiency for this kind of aid. Kinetics measurements confirm the possibility to stop the oxidation reaction. It was the first indication of efficient molecular encapsulation. Diffusion coefficient measurements were carried out by Taylor–Aris diffusion technique. The decrease of diffusion coefficients measured for iron II lactate when alone and forming the inclusion complexes was obtained for all hosts molecules used. Molecular Mechanics calculations were performed to elucidate the perfect arrange of iron II lactate inside CDs cavity. No great differences were obtained to the binding energy for the different hosts. Using the software HyperChem6.03v MM+, AMBER94 and OPLS Forced Fields for iron atom in two chemical environments (a) vacuum and (b) with addition of 250 water molecules (MM+). The solvent treatment was decisive to the order of stability. This order was β-CD>γ-CD>α-CD, the same order of solubility in water. The results contained in this work confirm the possibility to protect iron II lactate against oxidation.

α- and β-cyclodextrin complexes with n-alkyl carboxylic acids and n-alkyl p-hydroxy benzoates. A molecular mechanics study of 1:1 and 1:2 associations

Abstract Molecular mechanics (MM) methods were employed to evaluate stabilization upon formation of inclusion compounds between two different guest molecules and α- and β-cyclodextrins (CDs) for two different stoichiometries 1:1 and 1:2. The two guest molecules studied were n-alkyl carboxylic acids and n-alkyl p-hydroxy benzoates with variety of chain lengths. The computed stability for the inclusion compounds between α-CDs and n-alkyl carboxylic acids reproduced experimental data reported in the literature. The transition between 1:1/1:2 complexes occurred at an alkyl chain length of nC=9. It was previously demonstrated by diffusion coefficients measures that a stable 1:2 stoichiometry inclusion compound could be formed between n-alkyl p-hydroxy benzoates and α-CD for the chain length nC>4. The computed results reproduced the experimental ones. The combination between OPLS and GB/SA resulted in better agreements with experiments than those obtained with MM2 and MM3.

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.

Conformational analysis of the 1,4-dihydropyridines linking the structural aspects to the biological binding event: a study of the receptor-site conformation

Abstract A conformational analysis involving a set of 45 2,6-dimethyl-3,5-dicarbomethoxy-4-X-phenyl-1,4-dihydropyridine derivatives was performed. The potential energy surfaces of the rotation of the phenyl ring around the pyridine ring were constructed for each compound. These were arranged in groups of five in descending order of biological activity. The energy level below 99% of the molecular population for each molecule was superimposed on those groups. The conformationally stable regions permitted to this molecular population were found and compared. The analysis of each group of potential energy surfaces thus constructed revealed that in the biologically active conformation the phenyl ring bisects the pyridine ring and the ortho/meta phenyl substituents are spatially away from the pyridine ring.

A simultaneous study of photoelectron spectra, UV spectra and electron affinities of adenine and guanine with the semiempirical HAM/3 method

Abstract Ionization energies, excitation energies and electron affinities (EAs) of the title molecules were calculated with HAM/3. The photoelectron spectra (PES) and UV spectra of the molecules were interpreted with the aid of the calculated ionization and excitation energies respectively. The accuracy of the calculated EAs was estimated on the basis of the information obtained from the PES and UV spectra. It can be shown that the calculations give reasonable agreement between theory and experiment in all three cases.

2D and 3D QSAR studies of the receptor binding affinity of progestins

Foi realizada uma analise de QSAR-2D com tres descritores sobre a afinidade de ligacao a receptor no citosolo humano. Um conjunto de vinte e tres progesteronas foi dividido em um conjunto de treinamento de dezesseis compostos e em um conjunto de teste de sete compostos. O metodo quântico semi-empirico RM1 foi usado para calcular a geometria e algumas propriedades moleculares. O software DRAGON tambem foi usado para produzir descritores. O software MobyDigs foi usado para selecionar descritores e construir modelos QSAR. O melhor modelo de QSAR foi construido para o conjunto de treinamento usando regressao linear multipla com tres descritores, PW2, Mor15m e GAP-10, resultando em r2 = 0,886, q2 = 0,805, q2boot = 0,723, e q2ext = 0,666. Um conjunto de nove progesteronas adicionais que nao foram usadas para construir o modelo QSAR foi empregado para validacao externa resultando em q2ext = 0,403. A validacao tambem foi feita com funcoes de aptidao RQK. Foi mostrado que o modelo QSAR satisfaz todos os criterios requeridos para validacao, indicando que o modelo de regressao e aceitavel. Dois modelos QSAR-3D foram construidos: o primeiro para avaliar o poder preditivo e o segundo para ser analisado. O poder preditivo obtido com o conjunto de nove compostos para avaliacao externa foi q2ext = 0,476. Usando a representacao grafica dos coeficientes de regressao de PLS, correspondendo as interacoes espacial e eletrostatica, foi possivel obter uma interpretacao mecânica. Foi mostrado que QSAR-2D e 3D juntos satisfazem todos os seis requerimentos do Principio de Setubal (Principio de OECD). A partir das informacoes obtidas pelo QSAR-3D foram construidas quatro novas progesteronas. As atividades de afinidade de ligacao ao receptor destes novos compostos foram varias vezes maiores que qualquer uma do conjunto de vinte e tres progesteronas ja estudado.