Conformational Analysis: A New Approach by Means of Chemometrics ALINE THAÍS BRUNI, 1 VITOR B. P. LEITE, 2 MÁRCIA M. C. FERREIRA 1 1 Instituto de Química, Universidade Estadual de Campinas UNICAMP, Campinas, SP, 13083-970 Brazil 2 Departamento de Física, IBILCE, Universidade Estadual Paulista, São José do Rio Preto, SP, 15054-000 Brazil Received 7 August 2000; Accepted 23 July 2001 Abstract: In conformational analysis, the systematic search method completely maps the space but suffers from the combinatorial explosion problem because the number of conformations increases exponentially with the number of free rotation angles. This study introduces a new methodology of conformational analysis that controls the combinatorial explosion. It is based on a dimensional reduction of the system through the use of principal component analysis. The results are exactly the same as those obtained for the complete search but, in this case, the number of conformations increases only quadratically with the number of free rotation angles. The method is applied to a series of three drugs: omeprazole, pantoprazole, lansoprazole—benzimidazoles that suppress gastric-acid secretion by means of H + , K + -ATPase enzyme inhibition. © 2002 John Wiley & Sons, Inc. J Comput Chem 23: 222–236, 2002 Key words: principal component analysis; chemometrics; omeprazole; pantoprazole; lansoprazole; conformational analysis Introduction Experimental techniques are limited and are sometimes insufficient for the study of complex systems. Following recent computa- tional advances, new methods have been applied to the study of compounds and reactions in several fields of science. In medic- inal chemistry and pharmaceutical research, important issues are structure elucidation, conformational analysis, physico-chemical characterization, and biological activity determination. 1 These is- sues are helpful for investigating and elucidating how biological systems evolve and for determining the properties of a given drug. In these areas, methods of theoretical chemistry provide powerful tools for investigating and understanding, at a molecular level, the relationship between chemical structure and biological activity, and also for providing data for the design of new compounds. 2 All chemical information is intimately tied to the three- dimensional atomic arrangement and to the electronic properties of specific sites of a given compound. 3 The natural way to be- gin the theoretical study of a given drug is through structural determination. The main goal of molecular structure determina- tion is to provide a starting point for understanding the physical, chemical, and biological properties of matter. 4 Each different spa- tial arrangement of a molecule, known as a conformation, is defined by the arrangement of its atoms in space, which can be interconverted by rotation about single bonds. 5, 6 There are several ways to find the spatial arrangement of a molecule. Spec- troscopic (microwave, Raman, Infrared, NMR) and diffraction techniques (X-ray, synchrotron, electron, neutron diffraction) are, among others, widely used experimental techniques for struc- tural determination. In this study, we will only focus on the theoretical methods for three-dimensional arrangement determina- tion. Systems with many degrees of freedom have thermodynamic and dynamic properties determined by the nature of their potential energy surfaces. Analysis of molecular conformation space is used for locating stable structures of drug molecules. Potential energy surfaces (PES) can be characterized by their minima, which cor- respond to locally stable configurations, and by the saddle points or transition regions that connect the minima. 7–9 Theoretical cal- culations can be performed in different ways to find minimum energy structures, according to the methodology used. A variety of strategies have been described in recent years. They are capable of locating minimum energy structures on the conformational po- tential energy surfaces. The most common strategies are distance geometry, neural networks, genetic algorithm, simulation methods (Monte Carlo and Molecular Dynamics) and systematic analy- sis. Correspondence to: M. M. C. Ferreira; e-mail: marcia@iqm.unicamp.br Contract/grant sponsors: CNPq (to A.T.B.) and FAPESP (to M.M.C.F. and V.B.P.L.) © 2002 John Wiley & Sons, Inc. DOI 10.1002/jcc.1004