Pharmacoprint: A Combination of a Pharmacophore Fingerprint and Artificial Intelligence as a Tool for Computer-Aided Drug Design

Structural fingerprints and pharmacophore modeling are methodologies that have been used for at least 2 decades in various fields of cheminformatics, from similarity searching to machine learning (ML). Advances in in silico techniques consequently led to combining both these methodologies into a new approach known as the pharmacophore fingerprint. Herein, we propose a high-resolution, pharmacophore fingerprint called Pharmacoprint that encodes the presence, types, and relationships between pharmacophore features of a molecule. Pharmacoprint was evaluated in classification experiments by using ML algorithms (logistic regression, support vector machines, linear support vector machines, and neural networks) and outperformed other popular molecular fingerprints (i.e., ECFP4, Estate, MACCS, PubChem, Substructure, Klekota-Roth, CDK, Extended, and GraphOnly) and the ChemAxon pharmacophoric features fingerprint. Pharmacoprint consisted of 39 973 bits; several methods were applied for dimensionality reduction, and the best algorithm not only reduced the length of the bit string but also improved the efficiency of the ML tests. Further optimization allowed us to define the best parameter settings for using Pharmacoprint in discrimination tests and for maximizing statistical parameters. Finally, Pharmacoprint generated for three-dimensional (3D) structures with defined hydrogens as input data was applied to neural networks with a supervised autoencoder for selecting the most important bits and allowed us to maximize the Matthews correlation coefficient up to 0.962. The results show the potential of Pharmacoprint as a new, perspective tool for computer-aided drug design.

Rationally designed N-phenylsulfonylindoles as a tool for the analysis of the non-basic 5-HT6R ligands binding mode

Among all of the monoaminergic receptors, the 5-HT₆R has the highest number of non-basic ligands (approximately 5% of compounds stored in 25th version of ChEMBL database have the strongest basic pKa below 5, calculated using the Instant JChem calculator plugin). These compounds, when devoid of a basic nitrogen, exhibit high affinity and remarkable selectivity. Despite a decade of research, no clues have been given for explanation of such an intriguing phenomenon. Here, a series of analogs of four known 5-HT₆R ligands, has been rationally designed to approach this issue. For each of the synthesized 42 compounds, a binding affinity for 5-HT₆R has been measured, together with a selectivity profile against 5-HT_1AR, 5-HT_2AR, 5-HT₇R and D₂R. Performed induced fit docking and molecular dynamics experiments revealed that no particular interaction was responsible for the activity of non-basic compounds. In fact, a plain N-phenylsulfonylindole (1e) was found to possess a moderate (5-HT₆R, K_i = 159 nM) affinity. No other monoaminergic receptor has as simple and selective ligand as this one. Thus, it is stated that it binds to the receptor solely based on its conformation and as such, possesses a minimum amount of features, required for binding. Also, any functional group able to form an additional interaction with the receptor increase the binding affinity, like in the case of two highly active non-basic compounds 3e and 5g (5-HT₆R, K_i = 65 nM and 38 nM, respectively).

The effect of the intramolecular C–H⋯O interactions on the conformational preferences of bis-arylsulfones – 5-HT6 receptor antagonists and beyond

The development of compounds with enhanced activity and selectivity by a conserved spatial orientation of the pharmacophore elements has a long history in medicinal chemistry. Rigidified compounds are an example of this concept. However, the intramolecular interactions were seldom used as a basis for conformational restraints. Here, we show the weak intramolecular interactions that contribute to the relatively well-conserved geometry of N1-arylsulfonyl indole derivatives. The structure analysis along with quantum mechanics calculations revealed a crucial impact of the sulfonyl group on the compound geometry. The weak intramolecular C–H⋯O interaction stabilizes the mutual "facing" orientation of two aromatic fragments. These findings extend the pharmacological interpretation of the sulfonyl group role from the double hydrogen bond acceptor to the conformational scaffold based on intramolecular forces. This feature has, to date, been omitted in in silico drug discovery. Our results should increase the awareness of researchers to consider the conformational preference when designing new compounds or improving computational methods.

The impact of weak intramolecular C–H⋯O interactions on the conformational stability of bis-arylsulfones is discussed, suggesting different role of sulfonyl group in the ligand – 5HT₆ receptor interaction.

Spiro[pyrrolidine-3,3'-oxindoles] and Their Indoline Analogues as New 5-HT6 Receptor Chemotypes

Synthetic derivatives of spiro[pyrrolidinyl-3,3′-oxindole] alkaloids (coerulescine analogues) were investigated as new ligands for aminergic G-protein coupled receptors (GPCRs). The chemical starting point 2′-phenylspiro[indoline-3,3′-pyrrolidin]-2-one scaffold was identified by virtual fragment screening utilizing ligand- and structure based methods. As a part of the hit-to-lead optimization a structure-activity relationship analysis was performed to explore the differently substituted 2′-phenyl-derivatives, introducing the phenylsulphonyl pharmacophore and examining the corresponding reduced spiro[pyrrolidine-3,3′-indoline] scaffold. The optimization process led to ligands with submicromolar affinities towards the 5-HT₆ receptor that might serve as viable leads for further optimization.

Practical application of the Average Information Content Maximization (AIC-MAX) algorithm: selection of the most important structural features for serotonin receptor ligands

The Average Information Content Maximization algorithm (AIC-MAX) based on mutual information maximization was recently introduced to select the most discriminatory features. Here, this methodology was applied to select the most significant bits from the Klekota-Roth fingerprint for serotonin receptors ligands as well as to select the most important features for distinguishing ligands with activity for one receptor versus another. The interpretation of selected bits and machine-learning experiments performed using the reduced interpretations outperformed the raw fingerprints and indicated the most important structural features of the analyzed ligands in terms of activity and selectivity. Moreover, the AIC-MAX methodology applied here for serotonin receptor ligands can also be applied to other target classes.

Average Information Content Maximization--A New Approach for Fingerprint Hybridization and Reduction

Fingerprints, bit representations of compound chemical structure, have been widely used in cheminformatics for many years. Although fingerprints with the highest resolution display satisfactory performance in virtual screening campaigns, the presence of a relatively high number of irrelevant bits introduces noise into data and makes their application more time-consuming. In this study, we present a new method of hybrid reduced fingerprint construction, the Average Information Content Maximization algorithm (AIC-Max algorithm), which selects the most informative bits from a collection of fingerprints. This methodology, applied to the ligands of five cognate serotonin receptors (5-HT2A, 5-HT2B, 5-HT2C, 5-HT5A, 5-HT6), proved that 100 bits selected from four non-hashed fingerprints reflect almost all structural information required for a successful in silico discrimination test. A classification experiment indicated that a reduced representation is able to achieve even slightly better performance than the state-of-the-art 10-times-longer fingerprints and in a significantly shorter time.

Halogen bonding enhances activity in a series of dual 5-HT6/D2 ligands designed in a hybrid bioisostere generation/virtual screening protocol

A hybrid bioisostere generation/virtual screening method combined with narrowing of chemical space through similarity to compounds that are active at the second target was successfully applied for the development of dual 5-HT₆/D₂ receptor ligands.

Expanding the horizons of G protein-coupled receptor structure-based ligand discovery and optimization using homology models

We show the key role of structural homology models in GPCR structure-based lead discovery and optimization, highlighting methodological aspects, recent progress and future directions.