The theoretical modelling of chemically active transition metal (TM) centres is a notoriously difficult task. The metal-ligand interactions in these complexes are often highly directional and the concoction of suitable analytic interaction potentials can be far from trivial. The situation is rendered even more difficult by the fact that at finite temperature, the system might switch dynamically between different bonding situations or exhibit several energetically close-lying spin states which are all characterized by different coordination numbers and geometries. In this article, we describe the structural, dynamical and reactive properties of complex TM-containing systems with the help of a mixed quantum mechanical/molecular mechanical (QM/MM) molecular dynamics approach, in which the TM centre is described with generalized gradient corrected density functional theory embedded in a classical force field description. The power of such a combined Car-Parrinello/molecular mechanics approach is illustrated with a number of representative examples ranging from enantioselective TM catalysts to radiopharmaceuticals and metalloenzymes.
Three different 3D QSAR methods have been applied for a common pharmacophore model of 45 calcium antagonistically active 1,4-dihydropyridines (DHP) in order to find best correlation of interaction fields and biological activity. Analysis for the entire data set yielded r2/q values in a range starting from 0.821/0.620 (GRID/GOLPE) over 0.872/0.600 (CoMFA) to 0.908/0.744 (CoMSIA). The robustness of these models was tested not only via leave-one-out but also by leave-9-out crossvalidations. Furthermore, models were constructed using a subset of 37 DHPs (training set) allowing the prediction of activity for the residual 8 DHPs (test set). The training set yielded r2/q values starting from 0.826/0.672 (GRID/GOLPE) over 0.872/0.540 (CoMFA) to 0.899/0.662 (CoMSIA). For the test set r values from 0.677 (GRID/GOLPE) over 0.639 (CoMFA) to 0.470 (CoMSIA) were calculated. Besides the statistics, each 3D QSAR model yields further information by analysis of the generated contour maps. Consideration of the CoMFA and CoMSIA fields indicates unfavourable steric interactions for bulky moieties in 4′-position. On the other hand, sterical demanding 2′- and 3′-substituents are favourable and the biological activity of DHPs is further increased if these moieties produce a negative electrostatic potential. In contrast, high π-electron density on top of and parallel to the 4-phenyl ring beside the 2′-position is associated with decreasing activity. This could point to repulsive electronic interactions with binding site residues or to the potential of electron-deficient 4-aryl moieties to behave as electron acceptors in a charge transfer (CT) mechanism.
Phenols are widely used in agriculture and various industries. Many quantitative structure-activity relationships (QSARs) have been developed for phenols. The most toxicologically meaningful QSARs have been established by separating compounds by their mechanisms of action (MOAs). However, correctly determining the MOA of a compound is not easy. Discriminant analysis was employed in this study to separate selected phenols by three MOAs (polar narcosis, weak acid respiratory uncoupling, and soft electrophilicity) using molecular descriptors as discriminating variables. Results showed that quadratic terms of several molecular descriptors were needed in addition to their linear terms as discriminating variables. Cross-validation of the linear discriminant functions showed that a small total error rate for mechanism classification was achieved.
With the advances in genomics, proteomics and functional genomics new therapeutic targets to be tackled by medicinal chemistry are expected. This article reviews applications of first principle methods to address medicinal chemistry issue related to drug/target interactions. Two selected representative case studies involving therapeutically interesting targets are presented. The first case study presents how DFT can contribute to ameliorate scoring functions for drug screening in particular by enabling the discrimination between inhibitors and substrates. The second example shows the use of DFT within the framework of a QM/MM mixed approach for elucidating mechanisms of reaction. This approach allows defining the electronic state and structure of the reaction transition state whose knowledge is essential for designing potent and specific transition state analogs inhibitors. Finally, addressing the issue of other medicinal chemistry related application of DFT we suggest that DFT has indeed the potentiality of becoming very important for challenging now issues presented to medicinal chemistry in the post genomic era.
A quantitative structure-activity relationship (QSAR) of hydroxylated polychlorinated biphenyls (OH-PCBs) potency to inhibit human estrogen sulfotransferase (hEST), was modeled using multivariate methods and calculated semi-empirical descriptors. The QSAR model included 10 selected OH-PCBs, recently reported experimental inhibition potencies of hEST and 17 physico- chemical parameters. The cross-validated explained variance of 0.71 indicates a high predictive capacity of the model. The most important parameters were sub-molecular electronic parameters, such as partial atomic charges and nucleophilic electron densities. The natural substrate estradiol and OH-PCBs have been suggested to inhibit hEST by non-competitive binding to a not yet identified allosteric site. Potential allosteric binding sites in the crystal structure of mouse estrogen sulfotransferase (mEST) and in a homology model of hEST were investigated and discussed using the programs CAST and GRAMM. In general, the most abundant docking sites using GRAMM were in agreement with pockets defined by CAST. These preliminary results should be verified and more detailed studies undertaken when the active dimer of hEST is available. The results indicate that other similar pollutants may well interfere with the sulfonation of estradiol.
Density-Functional-Theory (DFT) provides a general framework to deal with the ground-state energy of the electrons in many-atom systems. Its history dates back to the work of Thomas [1], Fermi [21 and Dirac [3] who devised approximate expressions for the kinetic energy [1.2] and the exchange energy [3] of many-electron systems in terms of simple functionals of the local electron density. These ideas were further elaborated in the Xalpha method of Slater [4], until finally, the foundations of the modern theory were laid down in the mid-sixties by Kohn and collaborators [5, 6]. Since then but particularly in the last two decades the number of applications of DFT to electronic structure problems has grown dramatically. Today DFT is the method of choice for first-principles electronic structure calculations in condensed phase and complex molecular environments. DFT based approaches are used in a variety of disciplines ranging from condensed matter physics, to chemistry, materials science, biochemistry and biophysics. There are several reason for this success: (i) DFT makes the many-body electronic problem tractable at a numerical cost of self-consistent-field single particle calculations, (ii) despite the severe approximations made to the exchange and correlation energy functional, DFT calculations are usually sufficiently accurate to predict materials structures or chemical reactions products; (iii) currently available computational power and modern numerical algorithms make DFT calculations feasible for realistic models of systems like e.g. an interface between two crystalline materials, a carbon nanotube, or the active site of an enzyme.Very often one is interested in modeling atoms that are not static but in motion. This is the case in fluid systems, but even in crystalline materials atoms move at finite temperature. Furthermore structural changes in materials or molecular changes clue to chemical reactions are inevitably associated to atomic motions. A dynamic approach is needed in such situations. Assuming that classical mechanics is adequate to describe atomic motion. atomic trajectories can be calculated numerically by solving the corresponding classical equations of motion. This approach is called molecular dynamics (MD) and was pioneered [7, 8] after the advent of modern digital computers. Today it is a highly successful computational approach widely used in many fields of science to model the dynamics and the statistical properties of classical many-body systems. Usually in MD simulations empirical force fields are used to describe the inter-atomic interactions. Although the accuracy of the force fields is often remarkable, it cannot match the accuracy and the predictive power of quantum mechanical calculations based on DFT, particularly when changes in the electronic structure play a crucial role, like e.g. in chemical reactions. To address this issue a new molecular dynamics scheme was formulated [9] in which the potential energy surface is generated "on the fly" from the instantaneous ground state of the electrons within DFT. This approach, called ab-initio (or first-principles) molecular dynamics (AIMD). is now used with success to model atomistic processes in materials physics, chemistry and biology.Many applications of DFT and AIMD are reviewed in this special issue of QSAR, spanning from studies in physical chemistry to examples in biological and life sciences. It is the purpose of these notes to provide an introduction accessible to the non-specialist to some of the ideas and concepts behind DFT and AIMD.
The electronic properties of small molecules can be calculated quickly and with a reasonable degree of accuracy using semiempirical QM methods. In this study a set of QM properties derived from frontier electron theory have been used to produce a predictive model of the dissociation constants of phenols, benzoic acids and aliphatic carboxylic acids. The pK a values and structures of nearly 500 compounds were extracted from the Physprop database for this purpose. Multiple linear regression was used to search for relationships between pK a and the calculated QM properties. In most cases only a single independent variable, electrophilic superdelocalisability, was needed to produce a good model of pK a . The advantages of our approach are in the speed of calculation and the simplicity of the resultant models. The merits of using semiempirical methods to predict pK a are discussed in relation to previous studies.
A novel translationally and rotationally invariant structure descriptor based on the distribution of 3D-atom pairs is described. The new Distance Profiles (DiP) descriptor was applied to two data sets which were previously studied with various 3D-QSAR techniques. DiP compares favorably to the other descriptors for these two data sets and obtains better models in both cases. Since DiP is used in combination with variable selection to achieve interpretability, special emphasize was put on validating the derived models. Avoiding overfilled models was accomplished by constraining the maximum number of variables allowed to select, and by using leave-50%-out cross-validation instead of leave-one-out cross-validation as objective function in variable selection. Furthermore, the derived models were validated with a permutation test where the entire variable selection procedure is repeated each time the response data are scrambled.
The three-dimensional quantitative structure-activity relationship (3D-QSAR) approach using comparative molecular field analysis (CoMFA) was applied to a series of 40 compounds synthesized in our laboratory and evaluated as AANAT inhibitors. The N-bromoacetyltryptamine conformation derived from the X-ray crystal structure of the enzyme bound with a bisubstrate analog, was used to obtain the putative bioactive conformation of these inhibitors. Five statistically significant models were obtained from the randomly constituted training sets (30 compounds) and subsequently validated with the corresponding test sets (10 compounds). The best predictive model (n = 30, q(2) = 0.644, N = 6, r(2) = 0.966, s = 0.145, F = 109.478) can predict inhibitory activity for a wide range of compounds and offers important structural insight into designing novel AANAT inhibitors prior to their synthesis.
This work describes the modeling of human intestinal absorption using one- and two-descriptor nonlinear models. Molecules with known transport mechanisms other than passive transport were removed from the dataset. The human intestinal absorption data were fitted to symmetric and asymmetric sigmoidal curves and surfaces that were based on the arctangent function. The descriptors employed in the models are the sum of hydrogen bond donors and acceptors and the calculated SlogP octanol-water partition coefficient. Despite the simplicity of the model, the quality of fits and predictions is comparable to more complex approaches. As descriptors and predicted values can be calculated rapidly, the method is ideally suited for qualitative predictions for large virtual libraries and the results from de novo design.
4D-QSAR analysis incorporates pharmacophore, conformational and alignment freedom into the development of 3D-QSAR models for training sets of structure-activity data by performing ensemble averaging, the fourth "dimension". The data required to perform 4D-QSAR analysis includes a training set of compounds, usually analogs, and their measured biological activities in a common screen/assay. The 4D-QSAR approach can be applied to both receptor-dependent (RD) and receptor-independent (RI) problems. In the first scheme, the geometry of the receptor (molecular target, usually an enzyme) is available. In contrast, in the second scheme the geometry of the receptor is not part of the data available to perform the analysis. The descriptors in 4D-QSAR analysis are lattice grid cell (spatial) occupancy measures of atoms composing each molecule in the training set realized from the sampling of conformational and alignment spaces. These grid cell occupancy descriptors (GCODs) are generated for a number of different atom
Quantitative Structure-Activity RelationshipsVolume 21, Issue 4 p. 348-356 Review From Narcosis to Hyperspace: The History of QSAR Hugo Kubinyi, Hugo Kubinyi kubinyi@t-online.de BASF AG, Ludwigshafen, Germany Donnersbergstrasse 9, D-67256 Weisenheim am Sand, GermanySearch for more papers by this author Hugo Kubinyi, Hugo Kubinyi kubinyi@t-online.de BASF AG, Ludwigshafen, Germany Donnersbergstrasse 9, D-67256 Weisenheim am Sand, GermanySearch for more papers by this author First published: 04 October 2002 https://doi.org/10.1002/1521-3838(200210)21:4<348::AID-QSAR348>3.0.CO;2-DCitations: 55 AboutPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Share a linkShare onFacebookTwitterLinked InRedditWechat Citing Literature Volume21, Issue4October 2002Pages 348-356 RelatedInformation
Fragmental methods (FMs) have great potential in many practical areas related to the design of new lead compounds. Advanced Algorithm Builder(TM) (AAB) is a new software system which employs FMs in (i) building QSPR, QSAR and SAR models, (ii) converting them to custom (in-house) algorithms and screening filters, and (iii) predicting physical properties and biological activities for new compounds. This review demonstrates how FMs and AAB can be used to substantiate our intuition, interpret observations, validate hypotheses and obtain new algorithms for predicting physical properties and biological activities. Applications for practical and theoretical chemists in the design of new lead compounds are discussed.
Fourier transformation is suggested as the way to change site and substituent oriented variables (Free-Wilson matrix) into new only latently dependent ones. By their selection according to the statistically important slope values it is possible to find out a model with a low standard error of prediction, SEP%, of activity for not yet prepared compounds. Testing power of the correlation coefficient is␣improved by calculating a probability, PR‰, that the regression is made only by chance. These two indices are used for finding the best model in several investigated cases. Scope and limitations of this transformation are mentioned.
Cyclic nucleotide phosphodiesterases (PDEs) catalyse the hydrolysis of the second messengers adenosine-3',5'-cyclic phosphate cAMP and cGMP At least 11 different PDE types have been described: each of these groups a number of subtypes and splice variants. The PDE types differ in their amino acid sequence, substrate specificity, inhibitor sensitivity and in their organ, tissue and subcellular distribution. The recently solved X-ray structure of PDE4B as well as the results of site-directed mutagenesis experiments on PDE3A, prompted us to further investigate into the molecular mechanism that leads to effective PDE3 inhibition, as a prosecution of our previous studies on characterisation of the catalytic site of PDE family enzymes. On the basis of the experimental data available, a theoretical model of the catalytic site of PDE3A employing homology-modelling techniques was built. On this model thorough docking studies with potent and selective PDE3 inhibitors were performed. The derived inhibition model individuated structural requirements for potent PDE3 inhibition and can now be exploited for rational drug design purposes.
In continuation of a previous QSAR study on the cytotoxicity of 20 sesquiterpene lactones (STLs) of the helenanolide type towards a mouse tumour cell line where a very strong correlation of activity with only two indicator variables encoding the nature of the present alpha,beta-unsaturated carbonyl structure elements (cyclopentenone and alpha-methylcnc-gamma-lactone structure) was found, it was the major goal of this study to establish a QSAR model for a set of STLs with wider structural variability. Cytotoxicity towards the human KB cervix carcinoma cell line was experimentally determined for a set of 40 STLs representing five different structural groups (2 germacranolides, 6 guaianolides, 23 pseudoguaianolides, 8 eudesmanolides and 1 carabranolide (cyclopropane type xanthanolide) and the resulting IC50 values were submitted to a QSAR study using the molecular modelling program MOE. As major result it could be shown that variance in STL cytotoxicity data can be explained to a high degree by electronic and surface properties. QSAR models of considerable internal and external predictivity could be obtained by PCR and PLS analysis of a descriptor set representing fractional accessible molecular surface areas (Q=frASAs). This set of descriptors is calculated by partitioning the molecular surface accessible to a spheric probe of radius 1.4 Angstrom into fractions attributable to atoms within 14 charge intervals from -0.3 to 0.3 e. The applicability of such Q_frASA descriptors is validated by analysis of several sets of literature data, yielding QSAR models of good statistical quality. It is therefore assumed that Q_frASA descriptors may be of wider applicability in QSAR and QSPR.
Quantitative Structure-Activity RelationshipsVolume 21, Issue 4 p. 391-396 Review nD QSAR: a Medicinal Chemist's point of view Gerhard Müller, Gerhard Müller gerhard.muller@organon.com N. V. Organon, Lead Discovery Unit, RK4109, Molenstraat 110, PO Box 20, 5340 BH Oss, Netherlands Within the following commentary the author might express his personal opinions which do not necessarily represent those of N.V. Organon.Search for more papers by this author Gerhard Müller, Gerhard Müller gerhard.muller@organon.com N. V. Organon, Lead Discovery Unit, RK4109, Molenstraat 110, PO Box 20, 5340 BH Oss, Netherlands Within the following commentary the author might express his personal opinions which do not necessarily represent those of N.V. Organon.Search for more papers by this author First published: 04 October 2002 https://doi.org/10.1002/1521-3838(200210)21:4<391::AID-QSAR391>3.0.CO;2-LCitations: 2 AboutPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Share a linkShare onFacebookTwitterLinked InRedditWechat Citing Literature Volume21, Issue4October 2002Pages 391-396 RelatedInformation
Due to the importance of the O-H bond dissociation in the antioxidant mechanism of anti-inflammatory phenols, we studied the biradical process Ph-OH --> PhO. + H-. for 25 phenolic compounds using ab initio calculations. Enthalpies of reaction (DeltaH(r)), changes in the electron density at the O-H bond critical point (rho(OH)) and total atomic charges of ortho and para carbon atoms strongly correlate with the in vitro inhibition of cyclooxygenase activity by phenols. The most active compounds have large values of the electron density at the O-H bond (rho(OH)), thus favouring the O-H bond dissociation. In contrast, inactive compounds have small values of the electron density at the O-H bond (rho(OH)), thus reducing the hydrogen donation ability. These results are also supported by the representation of the molecular electrostatic potentials maps. The prediction of the cyclooxygenase inhibitory activity of the proposed QSAR equations is analysed using the multilineal (MLR) method. Finally, the differences in biological activity are examined by analysing the binding interactions of active compounds in the pocket site of human COX-2 enzyme structure derived from crystallographic X-ray data.
Quantitative structure-activity relationships (QSAR) is an area of computational research which correlates structural features and quantities such as the binding affinity, toxic potential, or pharmacokinetic properties of an existing or a hypothetical molecule. This correlation may be obtained by analyzing topology and fields of the molecules of interest or by simulating their spatial interactions with biological receptors or models thereof. While 3D-QSAR simulations allow for a specific interaction scheme with the virtual receptor, they presume the knowledge of the bioactive conformation of the ligand molecules and require a sophisticated guess about manifestation and magnitude of the associated induced fit - the adaptation of the receptor binding pocket to the individual ligand topology.To reduce the human bias - i.e. having to choose the bioactive conformer and to select an induced-fit model - in QSAR studies, we have extended a concept developed at our laboratory (software Quasar) to 4-D (by allowing for a multiple representation of the ligand molecules: conformation, orientation, protonation scheme, stereochemistry) and more recently to 5-D (by allowing for a multiple representation of induced-fit hypotheses). During the simulated evolution the correct entities are identified by the genetic algorithm. The current concept allows for up to 64 representations per ligand and up to six different induced-fit protocols: a subtle, individual, ligand-dependent mechanism, simulated by means of existing fields (steric, electrostatic, hydrogen bond), energy minimization, the lipophilicity potential or, in the simplest case, a linear adaptation. The genetic algorithm cannot only select among the provided six scenarios, but also build any linear or anisotropic combination thereof.
Without a crystallized structure for a human cytochrome P450, the use of computational molecular modeling is one approach to examine the active site requirements of substrate and inhibitor specificity. CYP2D6 is undoubtedly the most studied polymorphic human CYP and it is therefore desirable for drug companies to limit its role in the metabolism of drug-candidate molecules due to the need for therapeutic drug monitoring. The purpose of this study was to use a three-dimensional quantitative structure activity relationship (3D-QSAR) approach for understanding the CYP2D6 substrate requirements. Using literature K-m values (n = 24) derived solely from recombinant sources we were able to build and test one such pharmacophore. This was able to significantly rank-order using the (Spearman's rho coefficient 0.55, p = 0.0022) predicted against observed literature K-m values (n = 28) also derived from recombinant sources. The pharmacophore generated in this study was then fitted into the homology model of the human CYP2D6 based on an alignment of bacterial CYPs and the mammalian CYP2C5 to further validate these modeling approaches. Such models as these represent important tools for quantitative prediction of the level of interaction between a molecule and CYP2D6.
