Poster 30: Predicting pKa Values from EEM Atomic ChargesStanislav Geidl1, Radka Svobodová Varekova1, Crina-Maria Ionescu1, Ondrej Skrehota1, Tomas Bouchal1, David Sehnal1, Jaroslav Koca1
|1NCBR and CEITEC, Masaryk University, Kamenice 5, 625 00, Brno, Czech Republic|
|The acid dissociation constant pKa is an important molecular property, and its values are of interest in pharmaceutical, chemical, biological and environmental research. The pKa values have found application in many areas, such as the evaluation and optimization of candidate drug molecules, ADME profiling, pharmacokinetics, understanding protein-ligand interactions, etc.. Moreover, the key physicochemical properties like lipophilicity, solubility, and permeability are all pKa dependent. For these reasons, pKa values are important for virtual screening. Therefore, both the research community and pharmaceutical companies are highly interested in the development of reliable and above all very fast methods for pKa prediction. |
Several methods for pKa calculation have been developed . However, pKa values remain one of the most challenging physicochemical properties to predict. A very successful approach for pKa prediction is to use QSPR (Quantitative Structure Property Relationship) models which employ partial atomic charges as descriptors . A drawback of this approach is, that the quantum mechanical (QM) calculation of atomic charges is time consuming. On the other hand, QSPR models which employ conformationally independent empirical charges are not very accurate . One promising solution can be to use empirical atomic charges calculated by EEM (Electronegativity Equalization Method) [4,5] as QSPR descriptors. EEM charges are conformationally dependent and they are able to precisely mimic QM charges.
In our work, we have evaluated the pKa prediction capabilities of QSPR models based on EEM charges. Specifically, we collected 18 EEM parameter sets created for 10 different QM charge calculation schemes. Afterwards, we prepared a training set of 74 substituted phenols. Additionally, we generated for each molecule also its dissociated form by removing the phenolic H. We created QSPR models for each type of QM and EEM charges, and calculated the quality criteria of these models. We found that QSPR models employing EEM charges proved as a good approach for the prediction of pKa. As expected, QM QSPR models provided more accurate pKa predictions than EEM QSPR models. Nevertheless, these differences were not significant. The EEM QSPR models can be created also for other types of molecules to cover large sets of organic compounds. Accordingly, EEM QSPR models constitute a pKa prediction approach which can be applicable for virtual screening.
 A. C. Lee; G. M. Crippen J.: Predicting pKa. Chem. Inf. Model.49 (2009) 2013–2033.
 R. Svobodova Vareková ; S. Geidl; C. Ionescu; O. Skrrehota; M. Kudera; D. Sehnal; T. Bouchal; R. Abagyan; H. J. Huber; J. Koca J.: Predicting pKa values of substituted phenols from atomic charges: Comparison of different quantum mechanical methods and charge distribution schemes. Chem. Inf. Model. 51 (2011) 1795–1806.
 K. C. Gross; P. G. Seybold; C. M. Hadad: Comparison of different atomic charge schemes for predicting pKa variations in substituted anilines and phenols. Int. J. Quantum Chem. 90 (2002) 445–458.
 W. J. Mortier; S. K. Ghosh; S. Shankar J.: Electronegativity equalization method for the calculation of atomic charges in molecules. Am. Chem. SOC. 108 (1986) 4315-4320.
 R. Svobodova Varekova; Z. Jirouskova; J. Vanek; S. Suchomel; J. Koca: Electronegativity Equalization Method: Parameterization and validation for large sets of organic, organohalogene and organometal molecule. Int. J. Mol. Sci. 8 (2007) 572-582.