An atomic surface site interaction point description of non-covalent interactions.

Molecular electrostatic potential surfaces (MEPS) calculated using density functional theory have been used to develop a simplified description of the non-covalent interaction properties of organic molecules. The Atomic Interaction Point (AIP) model introduced here represents an evolution of the Surface Site Interaction Point (SSIP) model described previously, in which a molecule is represented by a discrete set of interaction points that define sites of interaction with other molecules. The interaction sites are described by interaction parameters that are equivalent to the experimentally determined H-bond donor and acceptor parameters α and ß. By using high electron density MEPS that lie inside the van der Waals surface, it is possible to obtain accurate interaction parameters and locations for polar sites (s-holes, H-bond donors and acceptors), which are identified as local maxima and minima on the MEPS. For non-polar sites that represent π-systems and halogens, an approach based on molecular orbitals was used to assign the locations of the AIPs, and the interaction parameters were obtained using a lower electron density MEPS that lies close to the van der Waals surface. The AIP descriptions can be implemented directly in the Surface Site Interaction Point Model for Liquids at Equilibrium (SSIMPLE) to calculate solvation free energies, and the free energy of transfer of 1504 compounds from n-hexadecane to water was predicted with a root mean square error of 5 kJ mol-1. AIPs also provide a useful tool for mapping non-covalent interactions in intermolecular complexes, and examples are provided showing how X-ray crystal structures can be converted into AIP interaction maps that allow quantification of the free energy contributions of both polar and non-polar interactions to the stabilities of complexes in solution.

An empirical model for solvation based on surface site interaction points.

Surface site interaction points (SSIP) provide a quantitative description of the non-covalent interactions a molecule makes with the environment based on specific intermolecular contacts, such as H-bonds. Summation of the free energy of interaction of each SSIP across the surface of a molecule allows calculation of solvation energies and partition coefficients. A rule-based approach to the assignment of SSIPs based on chemical structure has been developed, and a combination of experimental data on the formation of 1 : 1 H-bonded complexes in non-polar solvents and partition of solutes between different solvents was used to parameterise the method. The resulting model is simple to implement using just a spreadsheet and accurately describes the transfer of a wide range of different solutes from water to a wide range of different organic solvents (overall rmsd is 1.4 kJ mol-1 for 1713 data points). The hydrophobic effect as well as the properties of perfluorocarbon solvents are described well by the model, and new descriptors have been determined for range of organic solvents that were not accessible by direct investigation of H-bond formation in non-polar solvents.

Application of hydrogen bonding calculations in property based drug design.

A drug can be characterized by "descriptors" that include size (volume) and H-bond acidity and H-bond basicity. These descriptors can be rapidly estimated from structure by a fragment scheme and used to predict physicochemical and transport properties of drug candidates (e.g. logP, solubility, gastrointestinal absorption, permeability and blood-brain distribution). The solvation equations can be interpreted to provide a qualitative chemical insight into biological partition and transport mechanisms. Applications to blood-brain partition and human intestinal absorption (HIA) are discussed.

Fast gradient HPLC method to determine compounds binding to human serum albumin. Relationships with octanol/water and immobilized artificial membrane lipophilicity.

A fast gradient HPLC method (cycle time 15 min) has been developed to determine Human Serum Albumin (HSA) binding of discovery compounds using chemically bonded protein stationary phases. The HSA binding values were derived from the gradient retention times that were converted to the logarithm of the equilibrium constants (logK HSA) using data from a calibration set of molecules. The method has been validated using literature plasma protein binding data of 68 known drug molecules. The method is fully automated, and has been used for lead optimization in more than 20 company projects. The HSA binding data obtained for more than 4000 compounds were suitable to set up global and project specific quantitative structure binding relationships that helped compound design in early drug discovery. The obtained HSA binding of known drug molecules were compared to the Immobilized Artificial Membrane binding data (CHI IAM) obtained by our previously described HPLC-based method. The solvation equation approach has been used to characterize the normal binding ability of HSA, and this relationship shows that compound lipophilicity is a significant factor. It was found that the selectivity of the "baseline" lipophilicity governing HSA binding, membrane interaction, and octanol/water partition are very similar. However, the effect of the presence of positive or negative charges have very different effects. It was found that negatively charged compounds bind more strongly to HSA than it would be expected from the lipophilicity of the ionized species at pH 7.4. Several compounds showed stronger HSA binding than can be expected from their lipophilicity alone, and comparison between predicted and experimental binding affinity allows the identification of compounds that have good complementarities with any of the known binding sites.

On the mechanism of human intestinal absorption.

In order to investigate whether the main step in intestinal absorption in humans is dominated by partition or by diffusion, we have transformed % human intestinal absorption into a first-order rate constant, and have regressed the latter, as logk, against our solvation parameters. The obtained regression coefficients are compared with those for diffusion and partition processes. The coefficients in the logk equation are completely different to those for water/solvent partitions, but are very similar to those for processes (not involving transport through membranes) in which diffusion is the major step. It is suggested that the main step in the absorption process is diffusion through a stagnant mucus layer, together with transfer across the mucusmid R:membrane interface. It is further shown that for strong Bronsted acids and bases, the rate constant for absorption of ionic species is close to that for absorption of the corresponding neutral species, so that to a first approximation the % intestinal absorption can be calculated from properties of the neutral species.

Ionization-specific analysis of human intestinal absorption.

This study presents a mechanistic QSAR analysis of human intestinal absorption of drugs and drug-like compounds using a data set of 567 %HIA values. Experimental data represent passive diffusion across intestinal membranes, and are considered to be reasonably free of carrier-mediated transport or other unwanted effects. A nonlinear model was developed relating %HIA to physicochemical properties of drugs (lipophilicity, ionization, hydrogen bonding, and molecular size). The model describes ion-specific intestinal permeability of drugs by both transcellular and paracellular routes, and also accounts for unstirred water layer effects. The obtained model was validated on two external data sets consisting of in vivo human jejunal permeability coefficients (P(eff)) and absorption rate constants (K(a)). Validation results demonstrate good predictive power of the model (RMSE = 0.35-0.45 log units for log K(a) and log P(eff)). High prediction accuracy together with clear physicochemical interpretation (log P, pK(a)) makes this model particularly suitable for use in property-based drug design.

Predicting penetration across the blood-brain barrier from simple descriptors and fragmentation schemes.

The ability to cross the blood brain barrier (BBB), sometimes expressed as BBB+ and BBB-, is a very important property in drug design. Several computational methods have been employed for the prediction of BBB-penetrating (BBB+) and nonpenetrating (BBB-) compounds with overall accuracies from 75 to 97%. However, most of these models use a large number of descriptors (67-199), and it is not easy to implement the models in order to predict values of BBB+/-. In this work, 19 simple molecular descriptors calculated from Algorithm Builder and fragmentation schemes were used for the analysis of 1593 BBB+/- data. The results show that hydrogen-bonding properties of compounds play a very important role in modeling BBB penetration. Several BBB models based on hydrogen-bonding properties, such as Abraham descriptors, polar surface area (PSA), and number of hydrogen bonding donors and acceptors, have been built using binomial-PLS analysis. The results show that the overall classification accuracy for a training set is over 90%, and overall prediction accuracy for a test set is over 95%.