Application of Chemometric Techniques for Validation of Computational Methods Applied in Molecular Modeling of 6α-Hidroxyvouacapan-7β, 17β-Lactone with Antiproliferative Activity in Leukemia Cells

Hidroxyvouacapan-6α-7β, 17β-lactone (HVL) is a synthetic derivative of 6α, 7β-dihydroxyvouacapan-17β-oic acid (ADV) which was isolated from the fruits of Pterodon polygalaeflorus Benth. Both compounds showed anti-inflammatory and analgesic activities. According to studies conducted by Euzébio et al. in 2009 and 2010, ADV, HVL and some derivatives showed antiproliferative activity in different line of cancer cells, including the K562 leukemia. In this paper, we propose a combination of quantum-chemical methods, multivariate analysis and statistical to assess semi-empirical (AM1, PM3 and ZINDO), Hartree-Fock (HF) and DFT (B3LYP) methods in different basis set for determining which method better describes in sign and magnitude the geometric parameters of HVL compared to crystallographic data. PCA results were based on three main components, explaining 88.79% of the total variance for the geometrical parameters O2C6, O4C17C14C, C16O1C12, C6C7O3C17 and O3C7C8C14. DFT (B3LYP) method corresponded well with experimental data in hierarchical cluster analysis (HCA). The experimental and theoretical angles were analyzed by linear regression and statistical parameter (correlation coefficients, significance and predictability) were evaluated.