Outliers in SAR and QSAR: is unusual binding mode a possible source of outliers?

Outliers in SAR and QSAR: 4. effects of allosteric protein-ligand interactions on the classical quantitative structure-activity relationships

Effects of allosteric interactions on the classical structure-activity relationship (SAR) and quantitative SAR (QSAR) have been investigated. Apprehending the outliers in SAR and QSAR studies can improve the quality, predictability, and use of QSAR in designing unknown compounds in drug discovery research. We explored allosteric protein-ligand interactions as a possible source of outliers in SAR/QSAR. We used glycogen phosphorylase as an example of a protein that has an allosteric site. Examination of the ligand-bound x-ray crystal structures of glycogen phosphorylase revealed that many inhibitors bound at more than one binding site. The results of QSAR analyses of the inhibitors included a QSAR that recognized an outlier bound at a distinctive allosteric binding site. The case provided an example of constructive use of QSAR identifying outliers with alternative binding modes. Other allosteric QSARs that captured our attention were the inverted parabola/bilinear QSARs. The x-ray crystal structures and the QSAR analyses indicated that the inverted parabola QSARs could be associated with the conformational changes in the allosteric interactions. Our results showed that the normal parabola, as well as the inverted parabola QSARs, can describe the allosteric interactions. Examination of the ligand-bound X-ray crystal structures of glycogen phosphorylase revealed that many inhibitors bound at more than one binding site. The results of QSAR analyses of the inhibitors included a QSAR that recognized an outlier bound at a distinctive allosteric binding site.

Determination of Ligand Binding Modes in Hydrated Viral Ion Channels to Foster Drug Design and Repositioning

Target-based design and repositioning are mainstream strategies of drug discovery. Numerous drug design and repositioning projects have been launched to fight the ongoing COVID-19 pandemic. The resulting drug candidates have often failed due to the misprediction of their target-bound structures. The determination of water positions of such structures is particularly challenging due to the large number of possible drugs and the diversity of their hydration patterns. To answer this challenge and help correct predictions, we introduce a new protocol HydroDock, which can build hydrated drug-target complexes from scratch. HydroDock requires only the dry target and drug structures and produces their complexes with appropriately positioned water molecules. As a test application of the protocol, we built the structures of amantadine derivatives in complex with the influenza M2 transmembrane ion channel. The repositioning of amantadine derivatives from this influenza target to the SARS-CoV-2 envelope protein was also investigated. Excellent agreement was observed between experiments and the structures determined by HydroDock. The atomic resolution complex structures showed that water plays a similar role in the binding of amphipathic amantadine derivatives to transmembrane ion channels of both influenza A and SARS-CoV-2. While the hydrophobic regions of the channels capture the bulky hydrocarbon group of the ligand, the surrounding waters direct its orientation parallel with the axes of the channels via bridging interactions with the ionic ligand head. As HydroDock supplied otherwise undetermined structural details, it can be recommended to improve the reliability of future design and repositioning of antiviral drug candidates and many other ligands with an influence of water structure on their mechanism of action.

Outliers in SAR and QSAR: 3. Importance of considering the role of water molecules in protein-ligand interactions and quantitative structure-activity relationship studies

It is frequently mentioned that QSARs have not generally lived up to expectations, especially in cases where high predictability is expected yet failed to deliver satisfactory results. Even though outliers can provide an increased opportunity in drug discovery research, outliers in SAR and QSAR can contort predictions and affect the accuracy if proper attention is not given. The percentages of outliers in QSARs have not changed appreciably over the last decade. In our previous studies, we suggested two possible sources of outliers in SAR and QSAR. In this paper, we suggest an additional possible source of outliers in QSAR. We presented several literature examples that show one or more water molecules that play a critical role in protein-ligand binding interactions as observed in their crystal structures. These examples illustrate that failing to account for the effects of water molecules in protein-ligand interactions could mislead interpretation and possibly yield outliers in SAR and QSAR. Examples include cases where QSAR, considering the role of water molecules in protein-ligand crystal structures, provided deeper insight into the understanding and interpretation of the developed QSAR.

Chemical Graph Theory for Property Modeling in QSAR and QSPR—Charming QSAR & QSPR

Quantitative structure-activity relationship (QSAR) and Quantitative structure-property relationship (QSPR) are mathematical models for the prediction of the chemical, physical or biological properties of chemical compounds. Usually, they are based on structural (grounded on fragment contribution) or calculated (centered on QSAR three-dimensional (QSAR-3D) or chemical descriptors) parameters. Hereby, we describe a Graph Theory approach for generating and mining molecular fragments to be used in QSAR or QSPR modeling based exclusively on fragment contributions. Merging of Molecular Graph Theory, Simplified Molecular Input Line Entry Specification (SMILES) notation, and the connection table data allows a precise way to differentiate and count the molecular fragments. Machine learning strategies generated models with outstanding root mean square error (RMSE) and R2 values. We also present the software Charming QSAR & QSPR, written in Python, for the property prediction of chemical compounds while using this approach.

How medicinal chemists learned about log P

Although log P is now recognized to be a key factor that determines the bioactivity of a molecule, the focus of medicinal chemists on hydrophobicity and log P started with the quantitative structure-activity relationships (QSAR) publications of Hansch and Fujita. Their original publication represents a dramatic change of focus to incorporate consideration of log P after a decade of work unsuccessfully attempting to use the Hammett equation to explain the structure-activity relationships of plant growth regulators. QSAR allows one to explore the quantitative relationship between log P and biological activity even when other factors also influence potency. In particular, Hansch's publications of thousands of QSAR equations demonstrate that a relationship of biological activity with log P is indeed a general phenomenon. Hansch's group also provided data and tools that enable others to explore the relationship between log P and the biological activity of compounds of interest.

Predictive Power of Different Types of Experimental Restraints in Small Molecule Docking: A Review

Incorporating experimental restraints is a powerful method of increasing accuracy in computational protein small molecule docking simulations. Different algorithms integrate distinct forms of biochemical data during the docking and/or scoring stages. These so-called hybrid methods make use of receptor-based information such as nuclear magnetic resonance (NMR) restraints or small molecule-based information such as structure-activity relationships (SARs). A third class of methods directly interrogates contacts between the protein receptor and the small molecule. This work reviews the current state of using such restraints in docking simulations, evaluates their feasibility across broad systems, and identifies potential areas of algorithm development.

Exploring Structure-Activity Relationships with Three-Dimensional Matched Molecular Pairs-A Review

A matched molecular pair (MMP) consists of two small molecules that differ by a few atoms only. The minor structural difference between the molecules allows a detailed analysis of changes in properties. Three-dimensional (3D) MMPs extend the concept of chemical similarity by spatial similarity. Conformations must be generated, and superimpositions have to be calculated. The additional complexity and uncertainty as well as the smaller amount of available experimental data substantially complicates the derivation of models. Nonetheless, there are some benefits that make the transition worthwhile. The 3D concept gives detailed insight into mechanisms behind several methods classically used by the 2D MMP approach. It can help to analyze disrupted series of structure-activity relationships or extend the 2D MMP concept with scaffold hopping. One of the most powerful features is the high confidence structure-activity relationship transfer between series of analogues. Several research groups have approached the problem from different directions. The models vary especially in the 3D similarity measure used and complexity of the applied descriptor selected or designed. Nonetheless, all approaches have increased the amount of information available by incorporating 3D structural information.

Towards cheminformatics-based estimation of drug therapeutic index: Predicting the protective index of anticonvulsants using a new quantitative structure-index relationship approach

The overall efficacy and safety profile of a new drug is partially evaluated by the therapeutic index in clinical studies and by the protective index (PI) in preclinical studies. In-silico predictive methods may facilitate the assessment of these indicators. Although QSAR and QSTR models can be used for predicting PI, their predictive capability has not been evaluated. To test this capability, we developed QSAR and QSTR models for predicting the activity and toxicity of anticonvulsants at accuracy levels above the literature-reported threshold (LT) of good QSAR models as tested by both the internal 5-fold cross validation and external validation method. These models showed significantly compromised PI predictive capability due to the cumulative errors of the QSAR and QSTR models. Therefore, in this investigation a new quantitative structure-index relationship (QSIR) model was devised and it showed improved PI predictive capability that superseded the LT of good QSAR models. The QSAR, QSTR and QSIR models were developed using support vector regression (SVR) method with the parameters optimized by using the greedy search method. The molecular descriptors relevant to the prediction of anticonvulsant activities, toxicities and PIs were analyzed by a recursive feature elimination method. The selected molecular descriptors are primarily associated with the drug-like, pharmacological and toxicological features and those used in the published anticonvulsant QSAR and QSTR models. This study suggested that QSIR is useful for estimating the therapeutic index of drug candidates.

A Multi-Objective Genetic Algorithm for Outlier Removal

Quantitative structure activity relationship (QSAR) or quantitative structure property relationship (QSPR) models are developed to correlate activities for sets of compounds with their structure-derived descriptors by means of mathematical models. The presence of outliers, namely, compounds that differ in some respect from the rest of the data set, compromise the ability of statistical methods to derive QSAR models with good prediction statistics. Hence, outliers should be removed from data sets prior to model derivation. Here we present a new multi-objective genetic algorithm for the identification and removal of outliers based on the k nearest neighbors (kNN) method. The algorithm was used to remove outliers from three different data sets of pharmaceutical interest (logBBB, factor 7 inhibitors, and dihydrofolate reductase inhibitors), and its performances were compared with those of five other methods for outlier removal. The results suggest that the new algorithm provides filtered data sets that (1) better maintain the internal diversity of the parent data sets and (2) give rise to QSAR models with much better prediction statistics. Equally good filtered data sets in terms of these metrics were obtained when another objective function was added to the algorithm (termed "preservation"), forcing it to remove certain compounds with low probability only. This option is highly useful when specific compounds should be preferably kept in the final data set either because they have favorable activities or because they represent interesting molecular scaffolds. We expect this new algorithm to be useful in future QSAR applications.

k-Nearest neighbors optimization-based outlier removal

Datasets of molecular compounds often contain outliers, that is, compounds which are different from the rest of the dataset. Outliers, while often interesting may affect data interpretation, model generation, and decisions making, and therefore, should be removed from the dataset prior to modeling efforts. Here, we describe a new method for the iterative identification and removal of outliers based on a k-nearest neighbors optimization algorithm. We demonstrate for three different datasets that the removal of outliers using the new algorithm provides filtered datasets which are better than those provided by four alternative outlier removal procedures as well as by random compound removal in two important aspects: (1) they better maintain the diversity of the parent datasets; (2) they give rise to quantitative structure activity relationship (QSAR) models with much better prediction statistics. The new algorithm is, therefore, suitable for the pretreatment of datasets prior to QSAR modeling.

VAMMPIRE: a matched molecular pairs database for structure-based drug design and optimization

Structure-based optimization to improve the affinity of a lead compound is an established approach in drug discovery. Knowledge-based databases holding molecular replacements can be supportive in the optimization process. We introduce a strategy to relate the substitution effect within matched molecular pairs (MMPs) to the atom environment within the cocrystallized protein-ligand complex. Virtually Aligned Matched Molecular Pairs Including Receptor Environment (VAMMPIRE) database and the supplementary web interface ( http://vammpire.pharmchem.uni-frankfurt.de ) provide valuable information for structure-based lead optimization.

Perspective: Alchemical free energy calculations for drug discovery

Computational techniques see widespread use in pharmaceutical drug discovery, but typically prove unreliable in predicting trends in protein-ligand binding. Alchemical free energy calculations seek to change that by providing rigorous binding free energies from molecular simulations. Given adequate sampling and an accurate enough force field, these techniques yield accurate free energy estimates. Recent innovations in alchemical techniques have sparked a resurgence of interest in these calculations. Still, many obstacles stand in the way of their routine application in a drug discovery context, including the one we focus on here, sampling. Sampling of binding modes poses a particular challenge as binding modes are often separated by large energy barriers, leading to slow transitions. Binding modes are difficult to predict, and in some cases multiple binding modes may contribute to binding. In view of these hurdles, we present a framework for dealing carefully with uncertainty in binding mode or conformation in the context of free energy calculations. With careful sampling, free energy techniques show considerable promise for aiding drug discovery.

Comparative molecular field analysis (CoMFA) and comparative molecular similarity indices analysis (CoMSIA) of some benzimidazole derivatives with trichomonicidal activity

Trichomonosis is a common sexually transmitted infectious disease linked to reproductive health complications. Recently, the benzimidazole nucleus has emerged as a promising scaffold to develop new trichomonicidal agents. Despite the fact that large amounts of experimental data have been accumulated over the past eight years, no quantitative studies have yet been reported on this class of compounds. In our effort to develop new antiparasitic benzimidazole derivatives, we report in this paper CoMFA and CoMSIA studies with an initial set of 70 benzimidazole derivatives with trichomonicidal activity. Four CoMFA models and eight CoMSIA models were generated; ten of these models had values of r(2) > 0.6 and q(2) > 0.5. The best CoMFA model had r(2) = 0.936 and q(2) = 0.634, and the best CoMSIA model had r(2) = 0.858 and q(2) = 0.642. These models were generated by using two conformer selection methodologies (minimum energy conformations and 3D similarity), and three charge types (Mulliken, Gasteiger-Hükel and electrostatic potential atomic charges). The putative active tautomers of 1H-benzimidazole derivatives were selected using 3D-QSAR calculations. All models were validated via an external test set with 13 molecules. The best models satisfied additional validation criteria. The contour maps generated show the most important features that a benzimidazole derivative should have for trichomonicidal activity; they also, suggest that substituents at the 2- and 6-positions are important in the generation of derivatives with strong activity.

Insights into the structure of urea-like compounds as inhibitors of the juvenile hormone epoxide hydrolase (JHEH) of the tobacco hornworm Manduca sexta: analysis of the binding modes and structure-activity relationships of the inhibitors by docking and CoMFA calculations

Substituted urea compounds are well-known as potent inhibitors of juvenile hormone epoxide hydrolase (JHEH) of the tobacco hornworm Manduca sexta. Docking simulations of 47 derivatives inside JHEH were performed to gain insight into the structural characteristics of these complexes. The obtained orientations show a strong similitude with the observed in the known X-ray crystal structures of human soluble epoxide hydrolase (sEH) complexed with dialkylurea inhibitors. In addition, the predicted inhibitor concentration (IC₅₀) of the above-mentioned compounds as JHEH inhibitors were obtained by a quantitative structure-activity relationship (QSAR) method by using comparative molecular field analysis (CoMFA) applied to aligned dataset. The best models included steric and electrostatic fields and had adequate predictive abilities. In addition, these models were used to predict the activity of an external test set of compounds that was not used for building the model. Furthermore, plots of the CoMFA fields allowed conclusions to be drawn for the choice of suitable inhibitors.

Ligand structure controlled allostery in cAMP-dependent protein kinase catalytic subunit

Protein kinase A (cAMP dependent protein kinase catalytic subunit, EC 2.7.11.11) binds simultaneously ATP and a phosphorylatable peptide. These structurally dissimilar allosteric ligands influence the binding effectiveness of each other. The same situation is observed with substrate congeners, which reversibly inhibit the enzyme. In this review these allosteric effects are quantified using the interaction factor, which compares binding effectiveness of ligands with the free enzyme and the pre-loaded enzyme complex containing another ligand. This analysis revealed that the allosteric effect depends upon structure of the interacting ligands, and the principle “better binding: stronger allostery” observed can be formalized in terms of linear free-energy relationships, which point to similar mechanism of the allosteric interaction between the enzyme-bound substrates and/or inhibitor molecules. On the other hand, the type of effect is governed by ligand binding effectiveness and can be inverted from positive allostery to negative allostery if we move from effectively binding ligands to badly binding compounds. Thus the outcome of the allostery in this monomeric enzyme is the same as defined by classical theories for multimeric enzymes: making the enzyme response more efficient if appropriate ligands bind.

Use of molecular modeling, docking, and 3D-QSAR studies for the determination of the binding mode of benzofuran-3-yl-(indol-3-yl)maleimides as GSK-3beta inhibitors

Molecular modeling and docking studies along with three-dimensional quantitative structure relationships (3D-QSAR) studies have been used to determine the correct binding mode of glycogen synthase kinase 3beta (GSK-3beta) inhibitors. The approaches of comparative molecular field analysis (CoMFA) and comparative molecular similarity index analysis (CoMSIA) are used for the 3D-QSAR of 51 substituted benzofuran-3-yl-(indol-3-yl)maleimides as GSK-3beta inhibitors. Two binding modes of the inhibitors to the binding site of GSK-3beta are investigated. The binding mode 1 yielded better 3D-QSAR correlations using both CoMFA and CoMSIA methodologies. The three-component CoMFA model from the steric and electrostatic fields for the experimentally determined pIC(50) values has the following statistics: R(2)(cv) = 0.386 nd SE(cv) = 0.854 for the cross-validation, and R(2) = 0.811 and SE = 0.474 for the fitted correlation. F (3,47) = 67.034, and probability of R(2) = 0 (3,47) = 0.000. The binding mode suggested by the results of this study is consistent with the preliminary results of X-ray crystal structures of inhibitor-bound GSK-3beta. The 3D-QSAR models were used for the estimation of the inhibitory potency of two additional compounds.

Binding of small-molecule ligands to proteins: "what you see" is not always "what you get"

We review insights from computational studies of affinities of ligands binding to proteins. The power of structural biology is in translating knowledge of protein structures into insights about their forces, binding, and mechanisms. However, the complementary power of computer modeling is in showing "the rest of the story" (i.e., how motions and ensembles and alternative conformers and the entropies and forces that cannot be seen in single molecular structures also contribute to binding affinities). Upon binding to a protein, a ligand can bind in multiple orientations; the protein or ligand can be deformed by the binding event; waters, ions, or cofactors can have unexpected involvement; and conformational or solvation entropies can sometimes play large and otherwise unpredictable roles. Computer modeling is helping to elucidate these factors.

Combined ligand and structure based approaches for narrowing on the essential physicochemical characteristics for CDK4 inhibition

In the absence of an experimentally determined 3D structure of CDK4 (Cyclin-Dependent Kinase 4), QSARs (Quantitative Structure Activity Relationship) have been explored to rationalize binding affinity in terms of physicochemical and structural parameters. Further, docking on a homology model of CDK4 validated the derived QSARs and predicted the binding mode of this series of inhibitors. Relevant parameters and leave-one-out (LOO) cross-validation (q (2)) as well as an external test set validation (r (2) pred) judged the statistical significance and predictive ability of the models. Docking enabled a better understanding of protein-ligand interaction and provided a mechanistic interpretation in terms of physicochemical characteristics. It identified a unique hydrogen bonding between the imidazole of His-95 and the pyridine nitrogen in the ligand. It rationalized the need for R 2 substituents to be bulky and polar, while the substituent at R 8 to be hydrophobic and comparatively less steric. It also explained why at R 6 a variety of substituents are tolerated and how the presence of methyl at R 5 enhances binding affinity.

A ligand's-eye view of protein binding

Docking tools created for structure-based design and virtual screening have also been used to automate ligand alignment for comparative molecular field analysis (CoMFA). Models based on such alignments have been compared with those obtained based solely on shared ligand substructures, but such comparisons have generally failed to distinguish between conformational specification (alignment in the internal coordinate space) and embedding in a shared external frame of reference (Cartesian alignment). Here, large sets of inhibitors were docked into two cyclooxygenase and two reverse transcriptase crystal structures, and the poses generated were evaluated in terms of the CoMFA models they produced. Realigning the conformers obtained by docking by rigid-body rotation and translation to overlay their common substructures improved model statistics and interpretability, provided the protein structure used for docking was reasonably appropriate to the ligands being considered.

Outliers in SAR and QSAR: 2. Is a flexible binding site a possible source of outliers?

Structure-activity relationship (SAR) and/or quantitative structure-activity relationship (QSAR) studies play an important role in a lead optimization of drug discovery research. When there is a lack of ligand-bound protein structural information, one of the assumptions in SAR and QSAR studies is that similar analogs bind to the same binding site in a similar binding mode. In such studies, outliers have often been observed, especially in QSAR. However, most of these studies have focused their attention on the development of QSAR and left outliers unattended. We searched ligand-bound X-ray crystal structures from the protein structure database to find evidences that could indicate a possible source of outliers in SAR or QSAR. Our results showed the possibility of conformational changes in a flexible binding site as one possible source of outliers.