Comparative analysis of QSAR models for predicting pK(a) of organic oxygen acids and nitrogen bases from molecular structure

For 1143 organic compounds comprising 580 oxygen acids and 563 nitrogen bases that cover more than 17 orders of experimental pK(a) (from -5.00 to 12.23), the pK(a) prediction performances of ACD, SPARC, and two calibrations of a semiempirical quantum chemical (QC) AM1 approach have been analyzed. The overall root-mean-square errors (rms) for the acids are 0.41, 0.58 (0.42 without ortho-substituted phenols with intramolecular H-bonding), and 0.55 and for the bases are 0.65, 0.70, 1.17, and 1.27 for ACD, SPARC, and both QC methods, respectively. Method-specific performances are discussed in detail for six acid subsets (phenols and aromatic and aliphatic carboxylic acids with different substitution patterns) and nine base subsets (anilines, primary, secondary and tertiary amines, meta/para-substituted and ortho-substituted pyridines, pyrimidines, imidazoles, and quinolines). The results demonstrate an overall better performance for acids than for bases but also a substantial variation across subsets. For the overall best-performing ACD, rms ranges from 0.12 to 1.11 and 0.40 to 1.21 pK(a) units for the acid and base subsets, respectively. With regard to the squared correlation coefficient r2, the results are 0.86 to 0.96 (acids) and 0.79 to 0.95 (bases) for ACD, 0.77 to 0.95 (acids) and 0.85 to 0.97 (bases) for SPARC, and 0.64 to 0.87 (acids) and 0.43 to 0.83 (bases) for the QC methods, respectively. Attention is paid to structural and method-specific causes for observed pitfalls. The significant subset dependence of the prediction performances suggests a consensus modeling approach.     

The non-grid technique for modeling 3D QSAR using self-organizing neural network (SOM) and PLS analysis: application to steroids and colchicinoids

A novel method for modeling 3D QSAR has been developed. The method involves a multiple training of a series of self-organizing networks (SOM). The obtained networks have been used for processing the data of one reference molecule. A scheme for the analysis of such data with the PLS analysis has been proposed and tested using the steroids data with corticosteroid binding globulin (CBG) affinity. The predictivity of the CBG models measured with the SDEP parameter is among the best one reported. Although 3-D QSAR models for colchicinoid series is far less predictive, it allows for a discussion on the relative influence of the structural motifs of these compounds.     

Determination of pK(a) of acetic and benzoic acids in pyridine

The pK(a)'s of acetic acid and benzoic acid in pyridine as solvent are found to be 10.1 and 9.8, respectively, at 25 degrees . These results are based on measurements of hydrogen ion activities in mixtures of the acids and their tetrabutylammonium salts. Supplementary studies of differential vapour pressure characteristics of solutions of the acids and the acid-salt mixtures, and conductance of tetrabutylammonium benzoate solutions are also incorporated.     

Simple method for determining nucleobase pK(a) values by indirect labeling and demonstration of a pK(a) of neutrality in dsDNA

Phosphorothioates were substituted into double-stranded DNA to study protonated Class I A+.C basepairs by 31P NMR. The method was effective in reporting the A+.C pKa. pKa values near 7.0 were found with optimal nearest-neighbor partners. Such pKa values could expand the catalytic repertoire of nucleic acids.     

A comparative study of diesel analysis by FTIR, FTNIR and FT-Raman spectroscopy using PLS and artificial neural network analysis

Diesel properties determined by ASTM reference methods as cetane index, density, viscosity, distillation temperatures at 50% (T50) and 85% (T85) recovery, and the total sulfur content (%, w/w) were modeled by FTIR-ATR, FTNIR, and FT-Raman spectroscopy using partial last square regression (PLS) and artificial neural network (ANN) spectral analysis. In the PLS models, 45 diesel samples were used in the training group and the other 45 samples were used in the validation. In the ANN analysis a modular feedforward network was used. Sixty diesel samples were used in the neural network training and other 30 samples were used in the validation. Two different ATR configurations were compared in the FTIR, a conventional (ATR1) and an immersion (ATR2) cell. The ATR1 cell presented the best results, with smaller prediction errors (root mean square error of prediction, RMSEP). The comparison of the three PLS models (FTIR-ATR1, FTNIR, and FT-Raman) shows that reasonable values of R² and RMSEP were obtained by the FTIR-ATR1 and FTNIR models in the evaluation of density, viscosity, and T50. The PLS/FT-Raman models presented reasonable results only for the T50 property. None of the techniques was able to generate suitable PLS calibration models for the determination of sulfur content. The ANN/FT-Raman models presented the best performances, with all models presenting R²-values above 85% some of them with RMSEP values significantly smaller than those obtained with FTIR-ATR and FTNIR. The ANN/FT-Raman and ANN/FTIR-ATR1 models were able to estimate the total sulfur content of diesel with 0.01% (w/w) accuracy.     

The Non-Grid Technique for Modeling 3D QSAR Using Self-Organizing Neural Network (SOM) and PLS Analysis: Application to Steroids and Colchicinoids

Abstract

A novel method for modeling 3D QSAR has been developed. The method involves a multiple training of a series of self-organizing networks (SOM). The obtained networks have been used for processing the data of one reference molecule. A scheme for the analysis of such data with the PLS analysis has been proposed and tested using the steroids data with corticosteroid binding globulin (CBG) affinity. The predictivity of the CBG models measured with the SDEP parameter is among the best one reported. Although 3-D QSAR models for colchicinoid series is far less predictive, it allows for a discussion on the relative influence of the structural motifs of these compounds.     

Determination of intrinsic bacterial surface acidity constants using a donnan shell model and a continuous pK(a) distribution method

Intrinsic acidity constants (pK(a)(int)) for Bacillus subtilis (Gram+) and Escherichia coli (Gram-) cells were calculated from potentiometric titration data at different salt concentrations. Master curves were generated by replotting charge excess data as a function of pH(S) (pH at the location of surface reactive sites) where pH(S) was determined as a function of Donnan potential, Psi(DON). This potential decreased in magnitude with increasing ionic strength, from -48.5+/-0.2 to -3.5+/-0.0 mV for B. subtilis and -47.9+/-0.3 to -3.5+/-0.0 mV for E. coli at 0.01 and 0.5 M K(+), respectively, indicating an efficient surface charge neutralization by counterions. A fully optimized continuous (FOCUS) pK(a) distribution method revealed four binding sites on B. subtilis and E. coli surfaces from the master curves with pK(a)(int) values of 3.59+/-0.38, 4.33+/-0.57, 5.94+/-0.66, and 8.64+/-0.57 for B. subtilis and 3.73+/-0.44, 4.85+/-0.71, 6.56+/-0.64, and 8.79+/-0.62 for E. coli. These were assigned to functional groups according to reported pK(a) ranges of 2.0-6.0 (carboxylic acid), 3.2-3.5 (phosphodiesters), 5.6-7.2 (phosphoric acid), and 9.0-11.0 (amine groups). Average points of zero salt effect (pH(pzse)) for B. subtilis experiments were 6.63+/-0.21 and 6.42+/-0.08 as a function of pH(bulk) and pH(S), respectively. Under the same criteria, E. coli calculations yielded 5.73+/-0.23 and 5.45+/-0.05. An understanding of metal and proton reactivity on bacterial cell surfaces can be addressed quantitatively through the use of electrostatic and chemical equilibrium modeling techniques proposed in this study. The results are consistent with those of electrical force microscopy studies used to document the intrinsic electrochemical heterogeneity of bacterial cell surfaces.     

pK(a)-directed host-guest assemblies: rational analysis of molecular adducts of 2,4-diamino-6-methyl-1,3,5-triazine with various aliphatic dicarboxylic acids

Molecular adducts of 2,4-diamino-6-methyl-1,3,5-triazine (1) have been prepared with various aliphatic dicarboxylic acids. The molecular complexes (1 a-1 i) thus formed by co-crystallizing 1 with oxalic, malonic, succinic, fumaric, acetylene dicarboxylic, glutaric, thiodiglycolic, diglycolic, and adipic acids have been found to give two types of host-guest assemblies that have voids or channels in a three-dimensional arrangement. The different types of host-guest arrangement appear to result from differences in the acidity of the dicarboxylic acids, that is, acids with pK(a)<3.0 give host networks that consist of 1 and the corresponding acid with water or solvent molecules of crystallization present as guests, whereas acids with pK(a)>3.0 exist as guests in voids in a host network formed by 1. The former arrangement is observed in adducts 1 a, 1 b, 1 e, and 1 h and the latter arrangement is found in adducts 1 c, 1 d, 1 f, 1 g, and 1 i.     

Three-dimensional quantitative structure activity relationship (QSAR) of cytotoxic active 3,5-diaryl-4,5-dihydropyrazole analogs: a comparative molecular field analysis (CoMFA) revisited study

ABSTRACT: In vitro antitumor evaluation of the synthesized 46 compounds of 3,5-diaryl-4,5-dihydropyrazoles against EAC cell lines and 3D QSAR study using pharmacophore and Comparative Molecular Field Analysis (CoMFA) methods were described. CoMFA derived QSAR model shows a good conventional squared correlation coefficient r2 and cross validated correlation coefficient r2 cv 0.896 and 0.568 respectively. In this analysis steric and electrostatic field contribute to the QSAR equation by 70% and 30% respectively, suggesting that variation in biological activity of the compounds is dominated by differences in steric (van der Waals) interactions. To visualize the CoMFA steric and electrostatic field from partial least squares (PLS) analysis, contour maps are plotted as percentage contribution to the QSAR equation and are associated with the differences in biological activity. BACKGROUND: Pyrazole derivatives exhibit a wide range of biological properties including promising antitumor activity. Furthermore, Aldol condensation assisted organic synthesis has delivered rapid routes to N-containing heterocycles, including pyrazoles. Combining these features, the use of chalconisation-assisted processes will provide rapid access to a targeted dihydropyrazoles library bearing a hydrazino 3D QSAR study using pharmacophore and Comparative Molecular Field Analysis (CoMFA) methods were described for evaluation of antioxidant properties. RESULTS: Chalcones promoted 1 of the 2 steps in a rapid, convergent synthesis of a small library of hydrazinyl pyrazole derivatives, all of which exhibited significant antitumor activity against Ehrlich Ascites Carcinoma (EAC) human tumor cell line comparable to that of the natural anticancer doxorubicin, as a reference standard during this study. In order to understand the observed pharmacological properties, quantitative structure-activity relationship (3D QSAR) study was initiated. CONCLUSIONS: Chalcones heating provides a rapid and expedient route to a series of pyrazoles to investigate their chracterization scavenging properties. Given their favorable properties, in comparison with known anticancer, these pyrazole derivatives are promising leads for further development and optimization.     

A 3D QSAR pharmacophore model and quantum chemical structure--activity analysis of chloroquine(CQ)-resistance reversal

Using CATALYST, a three-dimensional QSAR pharmacophore model for chloroquine(CQ)-resistance reversal was developed from a training set of 17 compounds. These included imipramine (1), desipramine (2), and 15 of their analogues (3-17), some of which fully reversed CQ-resistance, while others were without effect. The generated pharmacophore model indicates that two aromatic hydrophobic interaction sites on the tricyclic ring and a hydrogen bond acceptor (lipid) site at the side chain, preferably on a nitrogen atom, are necessary for potent activity. Stereoelectronic properties calculated by using AM1 semiempirical calculations were consistent with the model, particularly the electrostatic potential profiles characterized by a localized negative potential region by the side chain nitrogen atom and a large region covering the aromatic ring. The calculated data further revealed that aminoalkyl substitution at the N5-position of the heterocycle and a secondary or tertiary aliphatic aminoalkyl nitrogen atom with a two or three carbon bridge to the heteroaromatic nitrogen (N5) are required for potent "resistance reversal activity". Lowest energy conformers for 1-17 were determined and optimized to afford stereoelectronic properties such as molecular orbital energies, electrostatic potentials, atomic charges, proton affinities, octanol-water partition coefficients (log P), and structural parameters. For 1-17, fairly good correlation exists between resistance reversal activity and intrinsic basicity of the nitrogen atom at the tricyclic ring system, frontier orbital energies, and lipophilicity. Significantly, nine out of 11 of a group of structurally diverse CQ-resistance reversal agents mapped very well on the 3D QSAR pharmacophore model.     

The comparative molecular surface analysis (CoMSA) with modified uniformative variable elimination-PLS (UVE-PLS) method: application to the steroids binding the aromatase enzyme

The application of the CoMSA method to analyze 3D QSAR of 50 steroid aromatase inhibitors is described. The 3D QSAR model obtained, reaching a value of cross-validated q(2) = 0.96 (s = 0.31), significantly outperforms those reported in the literature for the CoMFA or CoSA (CoSASA). It is shown that the Uniformative Variable Elimination UVE-PLS or modified iterative UVE procedure (IVE-PLS) can be used for indicating the regions contributing to the binding activity. Thus, after separating the series into two groups of the training and test molecules quite correct external predictions result from the processing of the training set. We proved that the procedure of the data elimination provides stable results, if tested in 50 random runs of the IVE-PLS-CoMSA with different training/test sets. Depending upon the procedure used the quality of the predictions for 25 test molecules is given by SDEP = sum(y(pred)-y(obs))(2)/n)(1/2) = 0.321 - 0.782.     

3D QSAR (COMFA) of a series of potent and highly selective VLA-4 antagonists

The integrin VLA-4 (41) is involved in the migration of white blood cells to sites of inflammation, and is implicated in the pathology of a variety of diseases including asthma and multiple sclerosis. We report the structure-activity relationships of a series of VLA-4 antagonists that were based upon the integrin-binding sequence of the connecting segment peptide of fibronectin (Leu-Asp-Val), and of VCAM-1 (Ile-Asp-Ser), both natural ligands of VLA-4. We explore variation in the ligand derived peptide portion of these antagonists and also in the novel N-terminal cap, which have discovered through chemical optimization, and which confers high affinity and selectivity. Using the X-ray derived conformation of the Ile-Asp-Ser region of VCAM-1, we rationalize the structure-activity relationships of these antagonists using 3D QSAR (COMFA). The COMFA model was found to be highly predictive with a cross-validated R² _CVof 0.7 and a PRESS of 0.49. The robustness of the model was confirmed by testing the influence of various parameters, including grid size, column filtering, as well as the role of orientation of the aligned molecules. Our results suggest that the VCAM-1 structure is useful in generating highly predictive models of our VLA-4 antagonists. The COMFA model coupled with the knowledge that the peptide amides are tolerant to methylation should prove useful in future peptidomimetic design studies.     

An empirical relationship between distribution coefficients of phenols by polyurethane foams and their octanol-water distribution constants and pK(a) values

Sorption of phenol, 3-cresol, 2-, 3-, 4-nitrophenols, 2,4-, 2,6-dinitrophenol, 2,4,6-trinitrophenol, and 1-naphthol by polyether- and polyester-type polyurethane foams (PUF) was investigated. The effects of sorption time, pH, phenol concentration and the structure of tested phenols and PUF were studied. The mechanism of sorption of tested compounds on foams is discussed. It is shown that the hydrophobicity (logP, octanol-water distribution constant) and pK(a) values of the compounds play an important role in the sorption process. A regression equation connecting distribution coefficient of phenols by PUF with their hydrophobicity parameter and pK(a) values were derived. Good correlation between logD and values logP and pK(a) was observed.     

Prediction of LUMO energy and rate constant by comparative molecular field analysis (CoMFA)

We used the comparative molecular field analysis (CoMFA) method to correlate the rate constant (log k) for the S_N2 reaction of benzyl benzenesulfonates and p-methoxybenzylamines. Molecular fields calculated with a C⁺ probe produced a good correlation with a small standard deviation and a high correlation coefficient with cross validation. This study demonstrated that CoMFA is an excellent method in predicting the physicochemical properties of the molecule such as LUMO energy and rate constants. © 1995 by John Wiley & Sons, Inc.     

Measurement of side-chain carboxyl pK(a) values of glutamate and aspartate residues in an unfolded protein by multinuclear NMR spectroscopy

A triple-resonance NMR pulse scheme is presented for measuring aspartic and glutamic acid side-chain pK(a) values in unfolded protein states where chemical shift overlap is limiting. The experiment correlates side-chain carboxyl carbon chemical shifts of these residues with the backbone amide proton chemical shift of the following residue. The methodology is applied to an (15)N, (13)C labeled sample of the N-terminal SH3 domain of the Drosophila protein drk, which exists in equilibrium between folded (F(exch)) and unfolded (U(exch)) states under nondenaturing conditions. Residue-specific pK(a) values of side-chain carboxyl groups are presented for the first time for an unfolded protein (drk U(exch) state), determined from a pH titration. Results indicate that deviations from pK(a) values measured for model compounds are likely due to local effects, while long-range electrostatic interactions appear to be of minor importance for this protein.     

Quantitative structure-toxicity relationships (QSTRs): a comparative study of various non linear methods. General regression neural network, radial basis function neural network and support vector machine in predicting toxicity of nitro- and cyano- aromatics to Tetrahymena pyriformis

Prediction of toxicity of 203 nitro- and cyano-aromatic chemicals to Tetrahymena pyriformis was carried out by radial basis function neural network, general regression neural network and support vector machine, in non-linear response surface methodology. Toxicity was predicted from hydrophobicity parameter (log Kow) and maximum superdelocalizability (Amax). Special attention was drawn to prediction ability and robustness of the models, investigated both in a leave-one-out and 10-fold cross validation (CV) processes. The influence that the corresponding changes in the learning sets during these CV processes could have on a common external test set including 41 compounds was also examined. This allowed us to establish the stability of the models. The non linear results slightly outperform (as expected) multilinear relationships (MLR) and also favourably compete with various other non linear approaches recently proposed by Ren (J. Chem. Inf. Comput. Sci., 43 1679 (2003)).