Comparison of Tetrahymena pyriformis toxicity based on hydrophobicity, polarity, ionization and reactivity of class-based compounds.

A toxicity data set containing the toxicities of 970 hydrophobic, polar and ionizable, nitro substituted and α,ß-unsaturated compounds to Tetrahymena pyriformis was classified into different groups based on the structure and substituted functional groups. Polar, ionizable and reactive compounds exhibit greater toxicity as compared with the non-polar hydrophobic compounds. Step-by-step analysis was carried out between the toxicity and descriptors representing hydrophobicity, polarity/polarizability, ionization and reactivity of compounds. Significant relationships were developed between the toxicity and these descriptors for the compounds. The models developed are simple, interpretable and transparent, using a small number of descriptors that may reflect the interactions of chemicals with the biological macromolecules at the target sites. Hydrophobic parameter log P reflects bio-uptake process compounds. Polarity/polarizability descriptor S reflects the interaction of hydrophilic residues of polar chemicals with biological macromolecules. The fractions of ionized (F (i)) and neutral (F (0)) forms calculated from pK (a) reflect the interactions of ionizable compounds with the macromolecules and effect of ionization of ionizable compounds on the bio-uptake process, respectively. A successful single model was developed by using the descriptors log P, S, F (i) and log F (0) for non-polar, polar and ionizable compounds.

Toxicity of organic pollutants to seven aquatic organisms: effect of polarity and ionization.

The toxicity of organic chemicals to Vibrio fischeri, river bacteria, algae, Daphnia magna and fishes were analysed. The results showed that the toxicity of chemicals to narcotics was dependent on hydrophobicity. A single model for both polar and non-polar narcotics was developed by inclusion of a polarity descriptor as well as the hydrophobic parameter. The highly hydrophobic polar narcotics could be treated as non-polar narcotics because their polar functional group(s) make(s) a relatively small contribution to polarity as compared with their hydrophobicity. In order to investigate the toxic mechanism of action for reactive compounds, the response-surface approach was used to develop models derived from easily calculated descriptors. The stepwise analysis selected the octanol/water partition coefficient and a polarity descriptor to parameterize bio-uptake and reactivity, respectively, for seven species. Benzoic acids can be easily absorbed into the unicellular bacteria, but this is not the case for multicellular D. magna and fish. Their toxicity to V. fischeri is much higher than that to D. magna and carp. Regression analysis was performed based on the model that we developed for ionizable compounds. Good correlations were observed by introducing the correction factor for ionizable compounds. The toxic mechanisms are discussed.

Prediction of biomagnification factors for some organochlorine compounds using linear free energy relationship parameters and artificial neural networks.

Multiple linear regression and artificial neural networks (ANNs) as feature mapping techniques were used for the prediction of the biomagnification factor (BMF) of some organochlorine pollutants. As independent variables, or compound descriptors, the Abraham descriptors often employed in linear free energy relationships were used. Much better results were obtained from the nonlinear ANN model than from multiple linear regression. The average absolute error, average relative error and root mean square error in the calculation of log (BMF) by the ANN model were 0.055, 0.051 and 0.097 for the training set and 0.11, 0.086 and 0.175 for the internal validation set, respectively. The degree of importance of each descriptor was evaluated by carrying out a sensitivity analysis approach for the nonlinear model. The results obtained reveal that the order of importance is the pollutant volume, the pollutant dipolarity/polarizability and the pollutant excess molar refraction. In order to examine the credibility of the obtained ANN model the leave-many-out cross-validation test was applied which gave Q(2)= 0.827 and SPRESS = 0.15. Also the Y-scrambling procedure was applied to the ANN model in order to examine the effect of chance correlations. The results obtained reveal that it is possible to predict the BMFs of organochlorine pollutants using a nonlinear ANN model with Abraham descriptors as inputs.

Olfactory detectability of homologous n-alkylbenzenes as reflected by concentration-detection functions in humans.

As part of our systematic exploration of chemical determinants for the olfactory potency of vapors towards humans, we measured concentration-detection functions for the odor of the homologous n-alkylbenzenes toluene, ethylbenzene, butylbenzene, hexylbenzene, and octylbenzene. A vapor delivery device based on dynamic olfactometry and calibrated by gas chromatography, served to test groups of 16 to 17 participants. Subjects were young adults from both genders, normosmics, and nonsmokers. Odor functions were tightly modeled by a sigmoid (logistic) function, both at the group and the individual level. Odor detection thresholds (ODTs), defined as the concentration producing a detectability halfway between chance and perfect detection, decreased with alkyl chain length from toluene (79 ppb) to butylbenzene (2.5 ppb), and then increased form butyl to octylbenzene (89 ppb). The "U"-shaped trend of ODTs as a function of alkyl chain length indicated a loss of odor potency beyond a certain molecular size, a phenomenon recently described for chemosensory irritation (chemesthesis) and that will need consideration in structure-activity models of chemosensory potency. Interindividual ODTs' variability for any single odorant amounted to one order of magnitude, in agreement with recent studies of other homologous series but quite smaller than commonly depicted.

Combination of artificial neural network technique and linear free energy relationship parameters in the prediction of gradient retention times in liquid chromatography.

In this work multiple linear regression (MLR) and artificial neural network (ANN) were used to predict the gradient retention times of diverse sets of organic compounds in four separate data sets. Descriptors which were used as inputs of these models are five linear free energy relationship (LFER) solute parameters including E, S, A, B and V. In the first step eight separate multiple linear regression and artificial neural network models were used to predict the gradient retention time for each gradient condition separately. Results obtained in this step reveal that there are significant relations between LFER parameters and gradient retention times of solutes in liquid chromatography. Then MLR and ANN were applied to develop more general models in which several different gradient elution conditions were used. The performances of these models are compared in terms of their standard errors and also correlation analysis. The results obtained reveal that although there are no significant differences between ANN and MLR in separate modeling of the gradient retention times, ANN has a significant superiority over MLR models in developing the general models for various gradient elution conditions. The results of sensitivity analysis on ANN models indicate that the order of importance for input terms in separate ANN models is Vx>B>S>E>A and in the case of combined ANN model is Vx>B>tg>S>E>A, which are in agreement with the order of percentage of significance terms that obtained from the MLR models.

Micellar versus hydro-organic reversed-phase liquid chromatography: a solvation parameter-based perspective.

The performance of the solvation parameter model is examined for micellar liquid chromatography. The results are compared with those offered with hydro-organic eluents, intending to reveal the properties that influence the retention and distinguish the particular behaviour of micellar systems. The retention data of several series of non-ionisable and ionisable compounds (mainly steroids, polyaromatic hydrocarbons, phenols, sulfonamides, beta-blockers, phenethylamines, antihistamines, and diuretics) were used as probe compounds. The micellar mobile phases contained an anionic (sodium dodecyl sulphate), non-ionic (Brij-35), or cationic (cetyltrimethylamonium bromide) surfactant, with or without the addition of an organic solvent (either propanol, butanol, pentanol or acetonitrile). In some instances (steroids, sulfonamides, beta-blockers and diuretics), the processed data were obtained in both micellar and hydro-organic chromatographic modes using the same column. Accuracy in predictions is critically examined and a correction term that takes into account contributions not considered in the original Abraham model, such as electrostatic or steric ones, is suggested to improve the correlations. The proposed correction takes into account differences between the descriptors of ionic and neutral species. The case of compounds showing different degrees of ionisation is discussed. Three solvation parameter approaches that allow the prediction of retention at varying mobile phase composition are proposed, which also revealed differences between the micellar and hydro-organic modes.

Cutoff in detection of eye irritation from vapors of homologous carboxylic acids and aliphatic aldehydes.

Using neat vapors of selected homologous aldehydes (decanal, undecanal, dodecanal) and carboxylic acids (pentanoic, hexanoic, heptanoic, octanoic, nonanoic), we explored the point where a certain homolog (and all larger ones) becomes undetectable by eye irritation (i.e. by ocular chemesthesis). This phenomenon has been observed in other homologous series that also reach a break-point, or cutoff, in chemesthetic detection. Participants (11

Inverse least-squares modeling of vapor descriptors using polymer-coated surface acoustic wave sensor array responses.

In previous work, it was shown that, in principle, vapor descriptors could be derived from the responses of an array of polymer-coated acoustic wave devices. This new chemometric classification approach was based on polymer/vapor interactions following the well-established linear solvation energy relationships (LSERs) and the surface acoustic wave (SAW) transducers being mass sensitive. Mathematical derivations were included and were supported by simulations. In this work, an experimental data set of polymer-coated SAW vapor sensors is investigated. The data set includes 20 diverse polymers tested against 18 diverse organic vapors. It is shown that interfacial adsorption can influence the response behavior of sensors with nonpolar polymers in response to hydrogen-bonding vapors; however, in general, most sensor responses are related to vapor interactions with the polymers. It is also shown that polymer-coated SAW sensor responses can be empirically modeled with LSERs, deriving an LSER for each individual sensor based on its responses to the 18 vapors. Inverse least-squares methods are used to develop models that correlate and predict vapor descriptors from sensor array responses. Successful correlations can be developed by multiple linear regression (MLR), principal components regression (PCR), and partial least-squares (PLS) regression. MLR yields the best fits to the training data, however cross-validation shows that prediction of vapor descriptors for vapors not in the training set is significantly more successful using PCR or PLS. In addition, the optimal dimension of the PCR and PLS models supports the dimensionality of the LSER formulation and SAW response models.

Correlation and prediction of a large blood-brain distribution data set--an LFER study.

We report linear free energy relation (LFER) models of the equilibrium distribution of molecules between blood and brain, as log BB values. This method relates log BB values to fundamental molecular properties, such as hydrogen bonding capability, polarity/polarisability and size. Our best model of this form covers 148 compounds, the largest set of log BB data yet used in such a model, resulting in R(2)=0.745 and e.s.d.=0.343 after inclusion of an indicator variable for carboxylic acids. This represents rather better accuracy than a number of previously reported models based on subsets of our data. The model also reveals the factors that affect log BB: molecular size and dispersion effects increase brain uptake, while polarity/polarisability and hydrogen-bond acidity and basicity decrease it. By splitting the full data set into several randomly selected training and test sets, we conclude that such a model can predict log BB values with an accuracy of less than 0.35 log units. The method is very rapid-log BB can be calculated from structure at a rate of 700 molecules per minute on a silicon graphics O(2).

Ocular and nasal trigeminal detection of butyl acetate and toluene presented singly and in mixtures.

To probe into the rules of trigeminal chemosensory agonism in a binary mixture of chemicals we measured, first, the detectability (i.e., psychometric) function for eye irritation and for nasal pungency of butyl acetate and toluene, singly. (To avoid olfactory biases, nasal pungency was measured in a group of anosmics, i.e., persons lacking a functional sense of smell.) Then, based on the detectability function obtained for the individual chemicals, we prepared mixtures where the 2 components varied in their relative proportions but, if a simple rule of complete sensory agonism (in the sense of dose-additivity) were to hold, the mixtures should be as detectable as the reference concentration of each of the single chemicals. For both trigeminal endpoints (i.e., eye irritation and nasal pungency), the results showed that stimuli of relatively low detectability did show complete sensory agonism, whereas stimuli of relatively high detectability fell short of complete sensory agonism when compared with the detectability of the single substances. Further testing of additional binary and higher order mixtures will confirm whether or not a structure-activity model of trigeminal chemosensory impact of single chemicals, based on selected physicochemical parameters of the stimuli, can also be applied to chemical mixtures.

Solvation descriptors for the polychloronaphthalenes: estimation of some physicochemical properties.

Solvation descriptors for the 75 polychloronaphthalenes have been derived from literature data on various properties. These descriptors (S, the dipolarity/polarizability; B, the hydrogen bond basicity; L, the logarithm of the gas-hexadecane partition coefficient; E, the excess molar refraction; V, the McGowan volume) have been used to estimate properties that may be environmentally relevant. Thus, for all 75 polychloronaphthalenes, we estimate values for the water-octanol partition coefficient, as log POCT, the aqueous solubility, as log S, the gas-water partition coefficient, as log KW, and the gas-dry octanol partition coefficient, as log KOCT. We further show that it is trivial to estimate other properties for all 75 polychloronaphthalenes; these properties include a number of gas-solvent and water-solvent partitions, air-plant and water-plant partitions, and permeation of human skin from water.

The solubility of gases and vapours in dry octan-1-ol at 298 K.

Ostwald solubility coefficients of 74 compounds in dry octan-1-ol at 298 K have been determined, and have been combined with literature values and additional values we have calculated from solubilities in dry octan-1-ol and vapour pressures to yield a total of 161 log L(OctOH) values at 298 K. These L(OctOH) values are identical to gas-to-dry octan-1-ol partition coefficients, often denoted as K(OA). Application of the solvation equation of Abraham to 124 values as a training set yielded a correlation equation with n = 124, S.D. = 0.125, r2 = 0.9970 and F = 7731. This equation was then used to predict 32 values of log L(OctOH) as a test set, giving a standard deviation, S.D. of 0.131, an average absolute deviation of 0.085 and an average deviation of -0.009 log units. The solvation equation for the combined 156 log L(OctOH) values was log L(OctOH) = -0.120 - 0.203R2 + 0.560pi2(H) + 3.560 sum(alpha2(H)) + 0.702 sum(beta2(H)) + 0.939 logL16, n =156, r2 = 0.9972, S.D. = 0.125, F = 10573, where, n is the number of data points (solutes), r the correlation coefficient, S.D. the standard deviation and F is the F-statistic. The independent variables are solute descriptors as follows: R2 is an excess molar refraction, pi2(H) the dipolarity/polarisability, sum(alpha2(H)) the overall or summation hydrogen-bond acidity, sum(beta2(H)) the overall or summation hydrogen-bond basicity and L16 is the Ostwald solubility coefficient on hexadecane at 298 K. The equation is consistent with similar equations for the solubility of gases and vapours into methanol, ethanol and propan-1-ol. It is suggested that the equation can be used to predict further values of log L(OctOH), for which the solute descriptors are known, to within 0.13 log units.

Rapid method for the estimation of octanol/water partition coefficient (log P(oct)) from gradient RP-HPLC retention and a hydrogen bond acidity term (zetaalpha(2)(H)).

We propose a rapid method for the measurement of octanol/water partition coefficients (log P(oct)) via fast gradient reversed phase retention and the calculation of the hydrogen bond acidity of the compounds. The cycle time of the generic gradient HPLC method is 5 minutes. The general solvation equation obtained for the log Poct values and the fast gradient Chromatographic Hydrophobicity Indices with acetonitrile (CHI(ACN)) and methanol

Evaluation of human intestinal absorption data and subsequent derivation of a quantitative structure-activity relationship (QSAR) with the Abraham descriptors.

The human intestinal absorption of 241 drugs was evaluated. Three main methods were used to determine the human intestinal absorption: bioavailability, percentage of urinary excretion of drug-related material following oral administration, and the ratio of cumulative urinary excretion of drug-related material following oral and intravenous administration. The general solvation equation developed by Abraham's group was used to model the human intestinal absorption data of 169 drugs we considered to have reliable data. The model contains five Abraham descriptors calculated by the ABSOLV program. The results show that Abraham descriptors can successfully predict human intestinal absorption if the human absorption data is carefully classified based on solubility and administration dose to humans.

Hydrogen bond structural group constants.

The ability of functional groups to act as hydrogen bond acids and bases can be obtained from either equilibrium constants for 1:1 hydrogen bonding or overall hydrogen bond constants. Either method leads to structural constants for hydrogen bonding that in some way are analogous to substituent constants. Extensive lists of these functional group constants are reported. It is shown that those derived from overall hydrogen bond constants are the more useful in analyses of physicochemical and biochemical properties.

Solute-solvent interactions in micellar electrokinetic chromatography. Selectivity of lithium dodecyl sulfate-lithium perfluorooctanesulfonate mixed-micellar buffers.

The solvation parameter model has been applied to the characterization of micellar electrokinetic chromatographic (MEKC) systems with mixtures of lithium dodecyl sulfate and lithium perfluorooctanesulfonate as surfactant. The variation in MEKC surfactant composition results in changes in the coefficients of the correlation equation, which in turns leads to information on solute-solvent and solute-micelle interactions. Lithium perfluorooctanesulfonate is more dipolar and hydrogen bond acidic but less polarizable and hydrogen bond basic than lithium dodecyl sulfate. Therefore mixtures of lithium dodecyl sulfate and lithium perfluorooctanesulfonate cover a very wide range of polarity and hydrogen bond properties, which in turn results in important selectivity changes for analytes with different solute properties.

Unique selectivity of perfluorinated stationary phases with 2,2,2-trifluoroethanol as organic mobile phase modifier.

The selectivity of Luna C18 Xterra C18 and Fluophase (perfluorinated C6) stationary phases has been investigated with aqueous acetonitrile, methanol and 2,2,2-trifluoroethanol mobile phases using linear solvation equations. The gradient retention times of a set of 60 compounds with known molecular descriptors have been determined. Linear solvation equations have been set up to describe the relationship between the gradient retention times and the molecular properties. The selectivity of the stationary phase/mobile phase systems was characterised by the regression coefficients of the molecular descriptors. The perfluorinated stationary phase showed very different selectivity using 2,2,2-trifluoroethanol (TFE) as co-solvent. Compounds with H-bond donor functionality were retained much less than in the other investigated high-performance liquid chromatography (HPLC) systems. This unique selectivity can be explained by the stronger adsorption of trifluoroethanol on the perfluorinated stationary phase surface, than on the hydrocarbon surface. It suggests the importance of the adsorbed organic modifiers in the separation mechanism during reversed-phase HPLC.

Estimation of molecular linear free energy relationship descriptors. 4. Correlation and prediction of cell permeation.

PURPOSE: The passage of molecules across cell membranes is a crucial step in many physiological processes. We therefore seek physical models of this process, in order to predict permeation for new molecules, and to better understand the important interactions which determine the rate of permeation. METHODS: Several sets of cell permeation data reported by Collander have been correlated against calculated Linear Free Energy Relation (LFER) descriptors. These descriptors, taken as the sum of fragmental contributions, cover the size, polarity, polarizabilty, and hydrogen bonding capacity of each molecule. RESULTS: For 36 values of permeation into Chara ceratophylla cells, a model (sd = 0.24) dominated by hydrogen bond acidity is found, while for 63 rates of permeation values into Nitella cells a very similar model yields sd = 0.46. Comparisons between the two cell types are made directly for 17 compounds in both data sets, indicate differences of a similar magnitude to the standard deviations of the above models. The two data sets can be combined to yield a generic model of rates of permeation into cells, resulting in an sd value of 0.46 for a total of 100 data points. CONCLUSIONS: Models allowing accurate prediction of cell permeation have been constructed using 100 experimental data. We demonstrate that hydrogen bond acidity is the dominating factor in determining cell permeation for two distinct species of algal cell.

Connection between chromatographic data and biological data.

There are no previous references to the direct use of GLC data in the correlation of biological processes, but we show that GLC retention data can be used in the correlation of several such processes involving gaseous solutes. There are a number of reports of RP-HPLC and MEKC data being used in the correlation of biological processes, but they are mostly restricted as to the number and type of solute studied. We show that if chromatographic data are used to obtain solvation descriptors for solutes, and if these descriptors are then used in the correlation of biological processes, that this indirect connection is a much more powerful and generally applicable method than is the direct connection between chromatographic data and biological data.

ElogPoct: a tool for lipophilicity determination in drug discovery.

We present an RP-HPLC method, for the determination of logPoct values for neutral drugs, which combines ease of operation with high accuracy and which has been shown to work for a set of 36 molecules comprised largely of drugs. The general features of the method are as follows: (i) compound sparing (< or = 1 mL of a 30-50 microg/mL solution needed), (ii) rapid determinations (20 min on average), (iii) low sensitivity to impurities, (iv) wide lipophilicity range (6 logPoct units), (v) good accuracy, (vi) excellent reproducibility. A linear free energy relationship (LFER) analysis, based on solvation parameters, shows that the method encodes the same information obtained from a shake-flask logPoct determination. To the best of our knowledge a similar performance, on a set of noncongeneric drugs, has not been previously reported. We refer to the value generated via this method as ElogPoct.